Opposing diet, microbiome and metabolite mechanisms regulate inflammatory bowel disease in a genetically susceptible host

Summary

Inf lammatory bowel diseases (IBDs) are chronic conditions characterized by periods of spontaneous intestinal inflammation and are increasing in industrialized populations. Combined with host genetics, diet and gut bacteria are thought to contribute prominently to IBDs, but mechanisms are still emerging. In mice lacking the IBD-associated cytokine, Interleukin-10, we show that a fiber-deprived gut microbiota promotes deterioration of colonic mucus, leading to lethal colitis. Inflammation starts with expansion of natural killer cells and altered immunoglobulin-A coating of some bacteria. Lethal colitis is then driven by Th1 immune responses to increased activities of mucin-degrading bacteria that cause inflammation first in regions with thinner mucus. However, a fiber-free exclusive enteral nutrition diet also induces mucus erosion, but inhibits inflammation by simultaneously increasing an anti-inflammatory bacterial metabolite, isobutyrate. Our findings underscore the importance of focusing on microbial functions—not taxa—contributing to IBDs and that some diet-mediated functions can oppose those that promote disease.

Graphical Abstract

eTOC Blurb

Pereira, Boudaud et al. demonstrate that a combination of host genetics, reduced dietary fiber, and colonic mucin-degrading gut bacteria triggers inflammatory bowel disease. Microbial mucus deterioration promotes Th1 immune responses that drive lethal colitis. Fiber-free exclusive enteral nutrition protects from lethal colitis via production of an anti-inflammatory bacterial metabolite, isobutyrate.

Introduction

Inflammatory bowel diseases (IBDs) are characterized by periods of spontaneous inflammation in the gastrointestinal tract and occur in people with underlying genetic variations that cause inappropriate immune responses to intestinal antigens, especially ordinarily non-harmful symbiotic gut microbes¹. Despite identification of hundreds of IBD-associated gene polymorphisms², precise mechanisms through which IBDs develop have not been determined. Even when predisposing genetics exist, IBDs do not always occur, suggesting that additional important triggers beyond host genetics and gut microbes are required³.

The incidence of IBDs is increasing in some industrializing countries⁴ and in immigrant populations that move to industrialized countries⁵. Dietary changes associated with industrialization (decreased fiber, increased processed foods, emulsifiers and sugars) are emerging as potential “triggers” that enhance susceptibility^6–9 and some underlying mechanisms are beginning to emerge^7,9. The physiology of the microbes inhabiting the gut is continuously influenced by diet, especially fiber polysaccharides that elude digestion in the upper gastrointestinal tract and arrive in the colon providing nutrients for microbes^10–12. Several studies using mice fed low fiber, high sugar or other “Westernized” diets have shown alterations in the microbiome that correlate with reduced integrity of the mucus barrier, including increased activity of mucin-degrading bacteria and reduced mucus thickness^{9,13, 14,15,16,17,18} and increased mucus penetrability¹⁹.

We investigated the contributions of dietary fiber and mucin-degrading gut bacteria to the development of inflammation in mice lacking the IBD-associated cytokine Interleukin-10 (IL-10). In humans, loss of IL-10 or either of its receptor subunits results in early onset IBD in infants and children²⁰. Conventional Il10^−/− mice develop spontaneous inflammation that is variable between mouse colonies and worsened by the presence of pathobionts like Enterococcus faecalis or Helicobacter spp.²¹, while deriving mice as germfree reduces or eliminates inflammation²². Thus, gut microbes are required for inflammation in Il10^−/− mice, however, the mechanisms mediating disease progression in the presence of commensal bacteria lacking known pathogenic qualities remain unknown.

Our work highlights the idea that microbial functions should be the focus of positive and negative impacts in the development of IBDs. Some of these functions, like mucus degradation, can be genetically complex and difficult to predict from taxonomic or metagenomic data, necessitating the use of simplified systems in which individual functions contributing to IBDs can be investigated at mechanistic levels prior to research in humans.

Results

A combination of host genetic susceptibility, low dietary fiber and intestinal bacteria exacerbate colitis in Il10^−/− mice

We previously determined that colonizing wild-type (WT) germfree mice with a synthetic microbiota containing 14 species (SM14) of sequenced and metabolically characterized human gut bacteria results in erosion of colonic mucus and increased Citrobacter rodentium susceptibility in mice fed a fiber-free (FF) diet¹³. Fiber-deprived WT mice do not develop histologically evident inflammation. Notably, mice colonized with the same SM14 and fed a fiber-rich diet do not experience mucus erosion because metabolism of fiber by gut bacteria decreases the abundance and activity of mucin-degrading species¹³. To determine if diet- and microbiota-driven mucus erosion—and the correspondingly increased proximity of bacteria to host tissue—promotes disease in mice that are genetically susceptible to IBD, we introduced the SM14 into germfree Il10^−/− mice, choosing the C57BL/6J background that has been reported to be most resistant to inflammation²¹. We colonized adult 7–10-week-old Il10^−/− mice fed a fiber-rich (FR) diet and switched a subset of the mice to the FF diet after 14 days of colonization, monitoring body weight up to 60d (Figure 1A). Mice kept on the FR diet (n=14) maintained or gained weight. In contrast, mice switched to the FF diet began losing weight 1–2 weeks after the diet switch and experienced 89% lethality by 60d (n=27). Histology revealed neutrophil infiltration, mucosal erosion, ulceration and edema in the ceca of FF-fed mice, but not those fed FR (Figure 1B). These markers were lower in the ileum and colon (Figure 1C). Cecal inflammation did not develop in SM14-colonized WT mice fed either diet (Figure 1D). Measurements of Lipocalin-2 (LCN-2)—an indicator of inflammation derived from neutrophils, inflamed epithelial cells and macrophages²³—in the cecal lumen provided additional data supporting severe inflammation in SM14-colonized Il10^−/− mice fed the FF diet (Figure 1E). Experiments in which individual variables for diet (FR, FF), colonization (SM14, germfree) and host genotype (WT, Il10^−/−) were individually manipulated support the conclusion that severe inflammation and weight loss only develop in the context of three conditions: IL-10 deficiency, SM14 colonization and FF diet (Figures 1D,E, S1A–D).

Figure 1. Low fiber driven inflammation in Il10^−/− mice.

A. Weights of adult Il10^−/− mice colonized with the SM14 and maintained on a fiber rich (FR) diet or switched to a fiber free (FF) diet at 14d. Curves represent polynomial (quadratic) equations fit to all of the weights gathered at various days. Animals that were euthanized were excluded from the curve at later points. Two FF-fed animals from an early experiment were found dead and were not included in subsequent analyses. Right axis shows survival over time. B. Representative cecal histology of FF (left) and FR (right) fed SM14-colonized Il10^−/− mice (middle, higher magnification insets). Block arrows points to a large ulcer, line arrow to an area of edema. C. Quantitative, blinded histological scoring of ileal, cecal, and colonic tissue from colonized Il10^−/− mice fed the FF diet (n=15–20, one-way ANOVA and post hoc test with Holm-Šídák's multiple comparison test) D. Histological scoring of cecal tissue taken from colonized Il10^−/− mice fed the FR and FF diets, along with additional treatments to manipulate diet, colonization and host genotype variables (n=5–20). E. Cecal Lipocalin-2 (LCN-2) measurements in the same treatments shown in D (n=5–23). F. Mucus penetrability by 1 um-sized beads in the distal colon of Il10^−/− and WT mice fed the FR and FF diets. G. Distance of 1 um-sized beads from the epithelium in the same mice in F. H. Weights of Il10^−/− mice born to either SM14 or SPF parents and weaned to FR or FF. By 79d post weaning, only the SM14/FF group (red) displays weight loss. Curves are fit to weight data as in A. I. Survival curves for 4 separate groups of WT C57bl/6 or Il10^−/− mice colonized by parental transfer of SM14 bacteria at birth. J. Fecal LCN-2 measurements over time in SM14 colonized Il10^−/− pups weaned to FR and FF. K. Endpoint cecal LCN-2 in SM14 and SPF colonized mice (n=4–7). L. Mucus measurements in SM14 and SPF colonized mice. Sample size indicated below each treatment group (n=4–5). M. Cecal LCN-2 in mice fed versions of the FF diet with glucose replaced by dietary fiber from apple, wheat, oat or soluble starch. Concentrations of each supplement are noted below the ingredient (n=5–23). N. Endpoint histology of the cecal tissue from SM14 colonized adult mice shifted to FF at 14d post colonization and then shifted back to FR at either 30d or 40d. (n=5–20). O. Endpoint cecal LCN-2 on mice shown in N. P. Fecal LCN-2 measurements over time on mice shown in N and O. For each time point/treatment the mean is shown along with individual points and error bars represent S.E.M. Experiments in panels F. and G. were done at the University of Luxembourg; all others were done at the University of Michigan. In panels C, D, E, J, K, L, M, bold horizontal bar represents the mean and lighter error bars the S.E.M. P values: * ≤ 0.05; ** ≤ 0.01; ** ≤ 0.001; *** ≤ 0.001; **** ≤ 0.0001; ns = not significant. In panels D, E, K, L, M, N, O, two-way ANOVA and post hoc test with Original FDR method of Benjamini and Hochberg was used for statistics.

Switching SM14-colonized Il10^−/− mice to the FF diet rapidly induced a shift in bacterial composition characterized by decreased fiber-degraders (B. ovatus and E. rectale) and increased abundance of two of the four known mucin-degraders, A. muciniphila and B. caccae (Figure S1E). A similar trend was observed in wild-type mice fed the FF diet, albeit with significantly higher levels of Escherichia coli and Bacteroides thetaiotaomicron in Il10^−/− mice (Figure S1F,G). As expected, feeding the FF diet to Il10^−/− mice resulted in reduced mucus thickness (Figure S1H–K). We hypothesize this erosion of protective mucus is critical for IBD development by increasing bacterial contact with the epithelium and the dysregulated Il10^−/− immune system. Supporting this idea, using bacteria-sized florescent beads in an ex situ mucus penetrability assay²⁴, we determined that FF-fed Il10^−/− mice have higher colonic mucus penetrability and closer proximity of 1μm-sized beads to the host epithelium compared to their FR-fed counterparts (Figures 1F,G, S1L,M). This increase in mucus penetrability was not as pronounced in wild-type mice with SM14 (Figure 1F,G).

Human disease associated with IL-10 dysfunction often presents as early or very early onset IBD in children²⁰. With this in mind, we modified our model to allow natural microbiota transfer in the neonatal period, which also enables direct comparison to conventional specific pathogen free (SPF) mice that are also colonized at birth. We colonized germfree Il10^−/− adult parents fed the FR diet, allowing their pups to be exposed to the maternal SM14 beginning at birth. In pre-weaned pups, the SM14 took on a similar composition as adult FF fed mice (Figure S1N) due to the lack of fiber and the fact that milk oligosaccharides share linkages with mucin O-glycans^25,26. Pups weaned to the FF diet began losing weight after ~39 days post weaning (dpw) (Figure 1H) and experienced 100% mortality by 84 dpw (Figure 1I). In most experiments, we harvested FR- and FF-fed groups at 79 dpw (100d old in University of Michigan facility), in which case mortality among the FF mice was ~82% and all mice weaned to the FR diet survived (Figure S1O). A separate group of FR-fed mice showed 100% survival when maintained for 129 dpw (150d total) as did wild-type mice fed either diet (Figure 1I). Weight loss in pups weaned to the FF diet corresponded with increasing fecal LCN-2 that initiated around the same time (~39 dpw) of declining weight (Figure 1J).

Interestingly, Il10^−/− mice with a conventional SPF microbiota fed the FF diet did not lose weight as severely as those colonized with SM14 (Figure 1H). Compared to SM14-colonized mice fed the FF diet, SPF mice showed lower cecal LCN-2 at 79 dpw (Figure 1K), and less histological damage (Figure S1P), revealing that the more complex, murine microbiota does not promote the same level of inflammation. However, SPF mice fed the FF diet did exhibit reduced mucus thickness, albeit significantly less reduction than SM14-colonized mice (Figure 1L), a point addressed in more detail below.

Restoration of dietary fiber inhibits inflammation

The FR and FF diets differ in several aspects of their composition beyond fiber (Table S1). To more directly test fiber’s contribution, we created versions of the FF diet in which 7.5% of the glucose it contains was replaced with an equivalent amount of food grade, pure fiber from oat, wheat or apple. High sugar has been shown to promote inflammation, including in Il10^−/− mice, and this effect is reduced by replacing some sugar with starch that is digested by the host in the small intestine⁹. As a control for reducing the sugar in our fiber-supplemented diets, we created a diet containing 7.5% highly digestible starch, exchanging free glucose for a glucose polymer available to the host via upper GI digestion. The presence of 7.5% fiber from any of the three sources, but not starch, significantly reduced LCN-2 (Figure 1M) as well as weight loss and histopathology (Figure S2A,B). Even when all of the glucose (44%) was replaced with digestible starch, adult mice colonized with the SM14 developed disease (Figures 1M, S2A,B). This high starch diet more closely approximates a fiber poor human diet, which would be expected to be rich in starchy foods, protein and fat, and still promotes inflammation, implying that direct free sugar availability to the gut microbiota is not a major component in this model. B. ovatus, a proficient degrader of the polysaccharides expected to be present in the supplemental fibers²⁷, was one of the major responders, increasing between 2–3 fold in relative abundance and this increase occurred at the expense of mucin-degrading Akkermansia muciniphila and Bacteroides caccae (Figure S2C–G).

To determine if restoring dietary fiber to mice that had already been fed the disease-promoting FF diet is capable of blocking inflammation, we returned colonized adult mice to the FR diet at either 30d or 40d (16 or 26 days after being switched to FF). Both groups of mice that were returned to FR exhibited no lethal weight loss (Figure S2H) and LCN-2 levels and histology at 60d were lower than mice maintained on the FF diet (Figure 1N, O). Interestingly, temporal fecal LCN-2 measurements showed that both groups experienced a peak of inflammation after fiber had been restored, which then began declining, indicating that low fiber induces disruption to the host–microbe homeostasis in which inflammation is a lagging effect that can eventually be reset (Figure 1P).

The status of the mucus layer is a critical determinant in developing inflammation

To more directly evaluate the disease-promoting role of mucin-degrading bacteria, we colonized adult Il10^−/− mice with a simpler synthetic microbiota containing only the 10 species (“SM10”) previously shown to be unable to grow on mucin oligosaccharides¹³. SM10-colonized mice exhibited 100% survival on the FF diet until 60d post colonization (n=7), significantly lower cecal LCN-2 at 60d (Figure 2A) and reduced histological damage (Figure 2B). Eliminating the mucin-degrading bacteria also decreased mucus thinning (Figures 2C, S2I). Notably, SM10 colonized mice still exhibited some inflammation (Figure 2B). Consistent with the idea that the absence of mucin-degrading bacteria slows, but does not fully block, inflammatory activation of the dysregulated Il10^−/− immune system, SM10-colonized mice that were switched to the FF diet for 136d (150d total colonization) exhibited significantly better survival (50%) compared to FF-fed SM14 mice (Figure 2D).

Figure 2. Mucus integrity is central to diet-induced inflammation.

A. Cecal LCN-2 measurements at 60d after colonization in Il10^−/− mice with full or reduced complexity synthetic microbiota as indicated in the “Colonized” line below each graph: Full SM14, just the SM10 species that do not degrade mucin O-glycans, or SM10 plus individual mucin degraders, B. thetaiotaomicron (Bt), B. caccae (Bc), B. intestinhominis (Bi) or A. muciniphila (Am) (n=7–23). B. Cecal histology scores of the same treatment group shown in A (n=5–20). C. Mucus thickness in SM10 colonized mice fed the FF diet (gray) compared to SM14 colonized mice fed either diet (n=5). D. Survival curves of Il10^−/− adult mice colonized with SM14, SM10 and SM10+Bt to 150d. E. Representative histology from SM10, SM10+Bt and SM10+Am at 60d. F. Regional histology of gastrointestinal tissues from SM10+Bt. G. Cecal LCN-2 from SM14 and SM10+Bt colonized mice. H. Survival curves of Il10^−/− and DKO adult mice colonized with the SM14. I.-L. Individual endpoint weight loss (I.), Cecal LCN-2 (J.), colon histology (K.) and cecal histology (L.) for SM14-colonized Il10^−/−Muc2^−/− double knockout (DKO) mice fed the FR or FF diets (n=5–23). M. Representative colon histology in DKO (left) and Il10^−/− (right) mice showing worse disease in DKO mice. N. Survival curves of mono-colonized DKO adult mice. O.-Q. Cecal LCN-2 (O.), colon histology (P.) and cecum histology (Q.) of mice shown in N. In panels A, B, I-K, O bold horizontal bar represents the mean and lighter error bars the S.E.M. In panels A-C, I-K, O two-way ANOVA and post hoc test with Original FDR method of Benjamini and Hochberg was used for statistics.

Adding back single mucin-degrading bacteria to the SM10 did not elicit the same level of LCN-2 observed with the full SM14 at 60d, indicating that multiple species may act synergistically (Figure 2A). Despite a lack of weight loss and lower LCN-2 at 60d, the presence of either B. thetaiotaomicron or A. muciniphila as the sole mucin-degrader did elicit histological inflammation in the cecum at 60d that was statistically identical to the full SM14 (Figure 2B,E), indicating that weight loss and LCN-2 phenotypes can in some cases be separate from histological damage. A longer, 150d experiment in which mucin-degrading B. thetaiotaomicron was added back to the SM10 revealed that it significantly accelerated lethal weight loss (Figure 2D) and caused histological inflammation that was worse in the cecum (Figure 2F). Thus, this commensal acts like a conditional pathogen in the compromised Il10^−/− host. Interestingly, the elevated cecal histology in SM+Bt colonized mice that were colonized for longer than 60d also occurred without elevated LCN-2 (Figure 2G). Finally, consistent with the idea that low dietary fiber promotes the general proliferation of mucin-degrading bacteria, mice colonized with SM10 plus a single mucin degrader all exhibited expansion of that mucin-degrading bacterium (Figure S2J–O). This contrasts with SM14-colonized mice, which only show expansion of Akkermansia muciniphila and B. caccae in response to low fiber (Figure S2J), implying that some mucin degraders out compete others.

The FR diet could suppress inflammation through mechanisms unlinked to its role in blocking mucin-degrading bacteria. As a separate test of the protective role for mucus in the context of the FR diet, we bred Il10^−/−Muc2^−/− double knockout (DKO) germfree mice, colonized them with the SM14 and fed either the FR or FF diets. These mice lost weight quickly and needed to be sacrificed regardless of diet (Figure 2H) and both groups exhibited severe inflammation (Figure 2I–L). Interestingly, with uniform elimination of MUC2, inflammation in DKO mice developed throughout the lower intestine but was more severe in the colon than in the cecum (Figure 2K–M). Since mucus is known to be thinner and patchier in the cecum compared to colon²⁸, we interpret this to mean that inflammation occurs first in the cecum of SM14-colonized Il10^−/− mice (i.e., with MUC2) because this is a site with high bacterial density and patchier mucus causing it to be where the mucus barrier fails first (note that small intestine also has thin mucus, but far fewer bacteria).

We further leveraged the DKO mice to measure how individual bacteria from different phyla (e.g., with different types of ligands for pattern recognition receptors) elicit inflammation in the context of Il10^−/− and deficient mucus barrier. In support of the conclusion that a variety of commensal bacteria can stimulate inflammation in this system, we observed that monocolonization with two different Bacteroidetes (B. thetaiotaomicron and B. uniformis), A. muciniphila (Verrucomicrobia) and Eubacterium rectale (Firmicutes) all result in strong inflammation, albeit once more with variations in LCN-2 (Figure 2N–Q).

Some human gut symbionts possess transferrable, pathogenic qualities between microbiomes

Because SPF mice fed the FF diet did not develop inflammation as severe as SM14-colonized mice (Figure 1H,K), we performed co-housing experiments in which pups born to SPF- and SM14-colonized mothers were mixed at weaning and exposed to each other’s microbiomes. Co-housing SM14-colonized Il10^−/− pups with SPF mice prevented the weight loss phenotype observed in response to FF-feeding (Figure 3A). However, it did not reduce—and even slightly increased—cecal inflammation as measured by LCN-2 (Figure 3B) or histology (Figure 3C), again indicating that weight loss and inflammation can be uncoupled in the presence of different microbiota. Co-housed mice born to SPF mothers showed a more variable response, with some of these mice developing inflammation that was similar in severity to FF-fed, SM14-colonized mice (Figure 3B,C). Time course analysis of fecal LCN-2 revealed a larger difference between co-housed and non-co-housed SPF mice, with co-housed SPF mice behaving nearly identically to FF-fed SM14 (Figure 3D, solid purple).

Figure 3. Co-housing SPF- and SM14-colonized mice after weaning worsens disease.

A. Weights of SPF or SM14-colonized mice co-housed at weaning (21d) with pups harboring the opposite colonization. B. Cecal LCN-2 levels in co-housed or non-co-housed SPF and SM14 mice (co-housing status is indicated at the bottom of this and other panels; n=5–11, one-tailed Studenťs t-test and Wilcoxon test). C. Cecal histology of the same treatments shown in B (n=5–11, one-tailed Student’s t-test and Wilcoxon test). D. Cecal LCN-2 over time in co-housed and non-co-housed SPF and SM14 mice. E.–G. Relative abundances of select SM14 species invading SPF co-housed mice, E. coli, A. muciniphila, Ba. intestinihominis, respectively. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Based on 16S rRNA gene sequencing, at least 6 members of the SM14 could be detected in co-housed SPF mice, often transiently or near the end of the time course when inflammation developed (Figures 3E–G, S3A–I). The most prominent of these invading bacteria was the human commensal E. coli strain HS, which appeared in co-housed SPF mice around 28 dpw, prior to the onset of inflammation, and gradually increased, eventually reaching levels >10% in most mice (Figure 3E, solid purple). Two of the mucin-degrading SM14 bacteria (A. muciniphila and Ba. intestinihominis) showed a similar trend, albeit reaching lower levels and with variability among individual mice (Figure 3F,G). Two other bacteria (C. aerofaciens and B. uniformis) showed transient increases around the time inflammation was increasing (~48 dpw) and these organisms decreased thereafter (Figure S3D,E). A test of whether weekly gavages of E. coli HS into SPF mice that were otherwise not exposed to SM14 bacteria failed to support the hypothesis that this species is the sole cause of increased inflammation (Figure S3J,K). Taken together, these co-housing results extend the relevance of our diet-driven model to more complex communities, further implying that one or more of the SM14 bacteria is endowed with conditional pathogenic qualities.

Certain gut bacteria and metabolites associated with exclusive enteral nutrition prevent inflammation despite mucus erosion

A notable characteristic of the disease-promoting FF diet is that its macronutrient composition resembles some exclusive enteral nutrition (EEN) diets used clinically to treat some IBDs (Table S1). EEN diets often contain low or no fiber²⁹ and have proven to be effective at inducing remission in humans, although precise mechanisms are still unknown³⁰. To determine if an EEN diet lacking fiber promotes inflammation in our gnotobiotic Il10^−/− model, we weaned SM14-colonized pups onto a commercial EEN diet, which is normally taken as a liquid but in this case was freeze-dried, pelleted and sterilized. The average weight trajectories of 15 mice supported the conclusion that the low fiber EEN diet promotes some weight loss, although not as severe as the FF diet (Figure 4A). Cecal LCN-2 measurements and histology revealed substantial disease variation in individual animals, with some mice resembling healthy FR-fed mice, some resembling diseased FF-fed mice and some intermediate (Figure 4B,C). Despite experiencing less inflammation, the EEN mice still exhibited reduced mucus thickness, which we expected given the lack of fiber (Figure 4D). Measurements of short- and branched-chain fatty acids (SCFAs, BCFAs) revealed that mice fed the EEN diet displayed between 43.2–723.0-fold (mean 133.4) elevated amounts of a single BCFA, isobutyrate (Figure 4E), suggesting that higher production might help suppress inflammation.

Figure 4. EEN diet improves inflammation in part through isobutyrate production.

A. Weights of mice weaned onto FR, FF, and EEN diets. B. Cecal LCN-2 levels in mice weaned onto FF, FR, and EEN diets (n=7–15). C. Cecal histology scores of the mice shown in B. (n=7–15). D. Mucus thickness measurements in mice shown in B.-C. (n=5). E. Short- and branched-chain fatty acid measurements in mice fed FF, FR, or EEN diets (n=5–15). F. Weights of mice fed the FF diet and water supplemented with either 35mM isobutyrate (orange) or butyrate (purple) (n=6–15). G. Cecal LCN-2 levels in the mice from panel F. H. Cecal histology scores of the mice shown in F. (n=6–15). I. SM14 community composition in mice fed the FR, FF and EEN diets. Red asterisks denote FR, FF and EEN diet comparison. J. Short- and branched-chain fatty acid measurements in culture supernatants of indvidual SM14 bacteria. K. Short and branched chain fatty acid measurements in EEN fed mice colonized with either the full SM14 or an SM14 lacking E. rectale (n=6–15, one-tailed Studenťs t-test and Wilcoxon test). L. Survival of SM14 and SM14 minus E. rectale colonized mice on the FF and EEN diets. M. Cecal LCN-2 measurements of the mice shown in L. (n=6–15, one-tailed Studenťs t-test and Wilcoxon test). N. Cecal histology scores of the mice shown in L. (n=10,15). Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. In panels B, C, D, E, G, H, two-way ANOVA and post hoc test with Original FDR method of Benjamini and Hochberg was used for statistics.

Isobutyrate is an isomer of the well-studied, anti-inflammatory SCFA butyrate, which did not increase (Figure 4E), and isobutyrate is derived from L-valine fermentation by certain gut bacteria³¹. Two other BCFAs (2-methyl butyrate and isovalerate) did not increase despite the amino acids they are derived from (L-leucine and L-isoleucine) being present at similar abundance³² in the soy and milk protein used in the EEN formulation (Figure 4E). This suggests that the increase in isobutyrate is not attributable to increased bulk dietary protein fermentation, which should increase all three BCFAs. Providing either isobutyrate or butyrate (35 mM) in the drinking water of mice fed the disease-promoting FF diet decreased weight loss (Figure 4F) as well as inflammation (Figure 4G,H), revealing that either of these molecules can offset the diet- and microbiome-induced damage.

EEN feeding promoted significant changes in the SM14, most notably 211–278-fold increases (mean 255.6) in E. rectale relative abundance (Figure 4I, compare columns with red asterisks). While this Firmicute is known to produce butyrate, measurements of culture supernatants from 13 of the SM species in medium supplemented with L-valine revealed that E. rectale does not produce isobutyurate under the conditions tested. Rather, isobutyrate was produced by all 4 of the Bacteroides, plus M. formatexigens (Figure 4J). Nevertheless, to determine if the large increase in E. rectale abundance was functionally connected with isobutyrate production, we bred neonatal mice born to parents harboring a version of the SM14 that lacks E. rectale (“SM14 minus E. rectale”) and fed them the EEN diet. Consistent with a causal role for E. rectale in isobutyrate production, mice lacking E. rectale failed to produce isobutyrate (Figure 4K). These mice also exhibited 40% lethality at 79dpw (Figure 4L), although there was not a significant increase in cecal LCN-2 levels or histology relative to the EEN/SM14 group (Figure 4M,N). Taken together, these findings reveal that EEN feeding causes reduced, albeit variable, inflammation even though it elicits similar mucus reduction as the more inflammatory FF diet.

Mucin-degrading bacteria mediate fiber deprivation-induced inflammatory immune responses in a time- and location-dependent manner

Development of intestinal inflammation is a complex process involving innate and adaptive immune responses and critical roles for Th1/Th17 cells have been described in the Il10^−/− colitis model in conventional/SPF mice²¹. Nevertheless, how specific microbial triggers influence the underlying immune pathways and how responses develop over time are less clear. We repeated and validated the same SM14 and SM10 (non-mucin-degrading community) colonization experiments described above with a different C57BL/6J line of Il10^−/− mice (mouse facility: University of Luxembourg) and observed similar weight loss in both adult and post-weaning FF-fed mice colonized with SM14, but not SM10 (Figure S3L–N). Given the temporal variation in colitis phenotypes in FF-fed SM14- (faster colitis) and SM10-colonized (slower colitis) Il10^−/− mice, our model provides unique opportunities to investigate the contributions of not only the microbial triggers (i.e., with or without mucin degraders) but the regionality (cecum vs. colon) of inflammatory pathways associated with increased microbial mucin foraging during fiber deficiency.

In FF-fed SM14-colonized mice, increased LCN-2 was detected in the cecum as early as 35 dpw compared to FR and this diet effect was not observed in mice colonized with SM10 or left germfree (Figure 5A). SM14 mice weaned onto the FF diet also exhibited higher fecal (a proxy for colon) LCN-2 than SM10 mice at 79 dpw, supporting a pro-inflammatory role of mucin-degrading bacteria during fiber deprivation (Figure 5A). An expansion of natural killer (NK) cells was detected in the cecum of FF-fed mice as soon as 35 dpw and this expansion was similar in SM10- and SM14-colonized mice (Figure 5B, left panel; see Figure S4A for flow cytometry sorting scheme and S4B–I for germfree Il10^−/− and SM14-colonized WT controls for flow cytometry experiments). In contrast to the pattern observed in cecum, NK cells expanded in the colons of both FR and FF-fed SM14-colonized mice as soon as 35 dpw, but not in the colons of SM10-colonized mice (Figure 5B, right panel). This further suggests that mucin-degrading bacteria are more important to elicit responses in the colon where mucus is thicker and degradation is required to increase contact. We also observed region-specific host responses that are driven by diet–host genotype interactions, since FF-fed GF Il10^−/− mice, but not SM14 or SM10 colonized WT mice, showed increased NK cells in the cecum, but not in the colon, at 79 dpw compared to FR-fed controls (Figure S4B).

Figure 5. Mucin-degrading bacteria mediate low fiber-induced immune responses.

A. LCN-2 levels in the cecal contents and feces of GF and colonized Il10^−/− mice (n=4–7). Numbers on the X axis indicate the number of samples in which LCN-2 was not detected. B. Proportion of NK cells among CD3⁻CD45⁺ cells in the cecum and colon lamina propria of SM10- and SM14-colonized Il10^−/− mice (n=4–11). C.–F. Proportion of CD3⁺CD4⁺ (C.), CD3⁺CD8⁺ (D.), Th1 (E.) and Th17 (F.) cells among CD45⁺ cells in the cecum and colon of SM10- and SM14-colonized Il10^−/− mice (n=4–11). Helper T (Th) cell subsets are defined as CD3⁺ CD4⁺ Foxp3⁻. Cytokine mRNA expression in the Mesenteric Lymph Nodes (G.-I., MLN) Data are represented as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. In panels A-F, two-way ANOVA and post hoc test with Original FDR method of Benjamini and Hochberg was used for statistics.

While NK cells were more abundant at 35 days and later decreased, T cell recruitment increased over time, generally reaching highest levels at 79 dpw, in both the cecum and colon of SM10- and SM14-colonized Il10^−/− mice (Figures S4C, S5A). In the ceca of both SM10- and SM14-colonized Il10^−/− mice, the FF diet increased CD4⁺ and CD8⁺ T cell populations by 35 dpw, with CD4⁺ T cells increasing further by 79 dpw independently of the diet (Figure 5C,D). However, CD4⁺ T cells also accumulated over time in the ceca of FF-fed GF Il10^−/− mice (Figure S4D), while CD8⁺ T cells were induced by FF-feeding in the ceca of GF Il10^−/− mice, as well as in the ceca and colons of SM14-colonized WT mice (Figures S4E). These results suggest that fiber deprivation induces the recruitment of T cell populations through multiple paths and not all of these are regulated by mucin-degrading bacteria, the microbiota and IL-10 deficiency.

Consistent with previous descriptions of Il10^−/− colitis in conventional mice²¹ both SM10- and SM14-colonized mice generally exhibited infiltration of Th1 and Th17 cells over time (Figure 5E,F) with fewer changes in Th2 cells (Figure S5B). Despite the IL-10 deficiency, Foxp3⁺ regulatory T cell (Treg) recruitment was also higher during FF feeding of SM14 Il10^−/− mice (Figures S5C). Interestingly, Treg subsets expressing Tbet, RORγt or Gata3 did not follow the same trends (Figure S5D–F). While the abundance of Tbet⁺ Treg cells follow the same trends as inflammatory Th1 cells with higher levels in FF-fed, SM14-colonized mice (Figures S5D) and Gata3⁺ Tregs were increased in FF-fed colonized colons (Figure S5F), the highly suppressive RORγt⁺ Treg population was reduced at 35 dpw in the colon and at 79 dpw in the cecum of FF-fed, SM14-colonized Il10^−/− mice (Figure S5E), thus favoring inflammatory responses. For Th17 cells, in both the cecum and the colon, this infiltration was higher in FF-fed mice compared to FR-fed mice, and in SM14-colonized mice compared to SM10-colonized mice (Figure 5F). By contrast, the FF diet increased Th1 levels in the SM14-colonized colons, but not in the SM10-colonized ones (Figure 5E).

Cytokine protein measurements in cecal tissues revealed increases in IBD-associated markers IL-1β, IL-6, IL-17, IL-22, TNF-α and IFN-γ in fiber-deprived Il10^−/− mice that were colonized with SM14 as adults (Figure S6A–F). Moreover, to provide functional support to the immune cell profiles, we compared levels of Th1- and Th17-type cytokine transcripts in ceca and mesenteric lymph nodes at 35 and 79 dpw in mice colonized at birth. In SM14-colonized Il10^−/− mice, the FF diet led to increased expression of Th1 (IFN-γ, IL-6, TNF-α, IL-12) and Th17 cytokines (IL-17F, IL-22, IL-23 and TGF-β), as well as the mucin-inducing cytokine IL-13 (Figures 5G–I, S6G–R). The expansion Foxp3+ regulatory T cells (Treg) during FF feeding of SM14 Il10^−/− mice (Figures S5C) being independent of IL-10 suggests a mechanism driven by other regulatory mediators such as TGF-β, whose transcript levels increased in SM14-colonized cecal tissues as soon as 35 dpw (Figure S6N). Intriguingly, in the ceca of SM14-colonized mice, both Th1 and Th17 cytokines were induced at 35 dpw, while only Th1 cytokines were maintained at 79 dpw (Figure S6G–N). These cytokine responses tended to develop more slowly in SM10-colonized ceca compared to those with SM14 and, both Th1 and Th17 cytokines were increased by 79 dpw in FF-fed SM10-colonized mice (Figure S6G–N). In the colon-draining mesenteric lymph nodes (MLNs) where naïve T cells are activated and polarized, the Th17-related cytokines, IL-17F and IL-22, were increased by FF-feeding compared to FR in both SM10- and SM14-colonized mice (Figures 5G, S6Q), while the Th1-polarizing cytokines IFN-γ, IL-6 and IL-12 were only increased in SM14-colonized mice, consistent with a dependence on mucin-degrading bacteria to develop Th1 responses (Figures 5H,I, S6P). In summary, our results reveal that, in Il10^−/− mice, the deterioration of the colonic mucus by mucin-degrading bacteria facilitates the induction of not only anti-microbial, but pathogenic Th1 responses in a time- and intestinal site-specific manner.

Alterations to IgA–microbiota interactions precede inflammation

IgA production is a common anti-microbial response in the gut and is usually upregulated during colitis in humans³³ and mouse models³⁴. Mirroring the LCN-2 trend observed for early inflammation (Figure 5A), soluble IgA titers were increased in the cecum at 35 dpw and the feces at 79 dpw, in FF-fed SM14-colonized Il10^−/− mice compared to FR, but not in SM10-colonized mice (Figure 6A). However, prolonged FF-feeding (79 dpw) resulted in depletion of plasma and IgA-producing cells in both the cecum and colon and this loss was also observed in FF-fed WT mice that were colonized with SM14 (Figure S7A, B), suggesting a diet-driven effect on IgA-producing cells rather than an effect of IL-10 loss. This is surprising considering the increased levels of secreted IgA (Figure 6A), but consistent with a previous study showing reduced titers of serum IgA and fewer IgA⁺ B cells in the small intestine of wild-type mice fed a zero-fiber diet compared to high-fiber³⁵. In parallel with reduced IgA-producing cells after FF-feeding, the proportion of IgG-producing B cells was increased at 35 and 79 dpw in both the cecum and colon of SM14-colonized Il10^−/− mice (Figure S7C,D). Consistent with the high production of Th1 cytokines in the presence of mucin-degrading bacteria, FF-feeding increased the proportion of IgG-producing cells only in SM14- but not in SM10-colonized colons (Figure S7D)³⁶.

Figure 6. Fiber deprivation alters IgA–bacteria interactions.

A. Concentrations of free IgA in the cecal contents and feces of GF and colonized Il10^−/− mice (n=4–9). B. Percentages of total IgA-coated bacteria in the feces of FR- and FF-fed, SM14-colonized mice at 21, 56 and 79 dpw (n=5–13). C. Percentages of total IgA-coated bacteria in the feces of SM10-, SM14- or SPF-colonized Il10^−/− mice and SM14-colonized WT mice fed the FR or the FF diet for 79 days (n=5–8). D. IgA-coating profiles of fecal bacteria from SM14-colonized mice fed a FR (left) or a FF (right) diet for 56 days showing the gating strategy of populations being low-coated (IgA^low) and high-coated (IgA^high). Total IgA coating consists of the addition of IgA^high and IgA^low coating. E–F. Percentages of IgA^low- (E.) and IgA^high- (F.) coated bacteria in the feces of indicated groups at 79 dpw (n=5–8). G. IgA-coating index (Kau index) of fecal bacteria from SM14-colonized Il10^−/− and WT mice (n=2–5, multiple unpaired t-test and post hoc test with the two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli). Data are represented as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. In panels A., B., C., E., F., two-way ANOVA and post hoc test with the two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli was used for statistics.

Consistent with a previous study³⁴, the amount of total IgA-coated bacteria was higher in FF-fed mice than in FR-fed mice at 79 dpw (Figure 6B). Along with increased luminal IgA (Figure 6A), this FF diet-induced increase in total IgA-coating was not observed in SM10-colonized mice but was observed in SPF Il10^−/− mice and SM14-colonized WT mice, supporting a mechanism driven by mucin-degrading bacteria (Figures 6C, S8A). Intriguingly, analysis of IgA-coated bacteria revealed the presence of two differentially coated populations in FR-fed mice: a large population with low-coating and a smaller population with high coating, and both populations responded differently to FF-feeding (Figure 6D). The proportion of low-coated bacteria increased in FF-fed mice, while the highly-coated bacteria were almost completely lost in FF-fed mice, as early as 21 days of feeding (Figure 6D–F, S8B). Additionally, highly-coated bacteria were diminished in SM10-colonized FF mice as in SM14-colonized FF mice as soon as 21 dpw and low-coated bacteria only slightly increased by 79 dpw (Figure 6E,F, S8C). Together these results suggest two levels of regulation of IgA coating by the FF diet: a rapid loss of the high coating that occurs independent of mucin-degrading bacteria, and a later increase of low and total coating that is partially dependent on the mucin-degrading bacteria and likely linked to the increased secretion of IgA. Interestingly, the EEN diet that dampened colitis but not mucus thinning in SM14-colonized Il10^−/− mice, increased the proportion of IgA-coated bacteria as soon as 21dpw and conserved the high-coated population compared to mice fed FF (Figure S8D), suggesting additional immune-regulatory properties of EEN diet components.

Given the altered pattern of overall IgA coating, we examined IgA coating of individual SM14 bacteria. Similar to the early loss of highly coated bacteria, changes in IgA-coating of SM14 members also appeared as soon as 21 days of FF-feeding (Figures 6G, S8E). Among the SM14 members, A. muciniphila, D. piger, E. coli and C. aerofaciens showed reduced IgA coating index (ICI) values at one or both timepoints in FF-fed Il10^−/− mice (Figures 6G, S8E). In WT mice, this FF diet-induced reduction in IgA coating was less severe for A. muciniphila, but more pronounced for D. piger and C. aerofaciens, revealing that IL-10 deficiency affects IgA-coating of commensals in a species-dependent manner. In contrast, E. rectale showed a low ICI in FR-fed mice and this increased with FF-feeding (Figure 6G), a condition in which it is present at low levels.

Discussion

The pathophysiology of IBDs is complex and variable, in part because of the large number of different genetic contributions that combine with environmental, microbial and dietary triggers known or hypothesized to influence its development. The diet-driven inflammation model investigated here provides both a case study, in which contributing factors to Il10^−/−-associated inflammation can be explored at mechanistic levels, as well as a more general experimental paradigm in which host genetic, microbiota and dietary variables can be manipulated to examine their effects on disease development. As examples of spontaneous, genetically driven intestinal inflammation continue to be identified^38–40, the ability to work with these murine models in germfree/gnotobiotic conditions and with defined diets holds the potential to uncover foundational principles about these complex diseases.

An important concept that is supported by our findings is that opposing bacterial functions influence disease outcomes. Low fiber induced mucus erosion is deleterious, but mostly in the context of some members of the human SM14 community that appear to have pathogenic qualities and can transfer their pro-inflammatory effect to SPF mice. During EEN feeding, mucus erosion still occurs, but is offset by isobutyrate production. Isobutyrate production is both influenced by diet and contingent upon E. rectale. Thus, these data present actionable hypotheses for future work in humans to determine if patients that respond positively to EEN also show increased isobutyrate. Indeed, one weakness of the current study is that isobutyrate production was not investigated in the context of more complex microbiota, such as germfree Il10^−/− mice colonized with human feces. In these future experiments it may be critical to stratify donor samples based on presence of E. rectale to determine if the presence of this species correlates with production or if other bacteria can be surrogates for the as-yet-unknown role that E. rectale plays. Deconstruction of the EEN formula to identify ingredients that promote E. rectale-dependent isobutyrate production may also lead to optimization of EEN formulations.

The regionality of the FF diet-induced pathogenesis, starting in the cecum and propagating in the colon, is likely due to the nature of the mucus that is affected by the fiber-deprived microbiota. The loose mucus in the cecum already allows more contact between the microbiota and the host under fiber-rich conditions. Thus, the effect of a fiber-deprived microbiota is sensed faster in the cecum than in the colon, where the mucin-degrading activity of the microbiota is required to deteriorate the thick mucus layer before increasing such contacts. Intriguingly, although both Th1 and Th17 responses are known to be regulated by the microbiota, Th1 responses were more dependent on mucin-degrading bacteria than Th17 responses in both the cecum and the colon as revealed by the cytokine (Figure S6) and immune cell profiles (Figure 5), respectively. This supports the immunopathogenic role of mucin-degrading bacteria as the Th1 responses lead to cytotoxic tissue damage, while Th17 responses also support epithelial protective functions.

Along with pro-inflammatory cell profiling, we report infiltration of Treg subsets, whose roles in immune responses are less understood. Treg cells expressing Tbet or RORγt have been proposed as counter-regulators of inflammatory Th1 and Th17 cells, respectively^41,42. Consistent with this, the abundance of Tbet⁺ Tregs follows the same trends as inflammatory Th1 cells. While the role of the expanding Gata3⁺ Tregs in the specific regulation of type 2 inflammatory cells is still unclear, they may constitute a reservoir of Tregs required for tissue repair by 79 dpw in colonized colons^43,44. Finally, despite the general expansion of Tregs, early (35 dpw) loss of the highly suppressive RORγt⁺ subset and the IL-10 deficiency are likely to allow the Th1/Th17 responses to flourish in FF-fed, SM14 mice. While Treg cells have been largely overlooked in this colitis model due to its IL-10 deficiency, our results support alternative mechanisms of immune suppression regulated by dietary fibers and the microbiota that require further attention.

IgAs are among the most abundant proteins in mucus⁴⁵. Their secretion is a key anti-microbial mechanism and IgA coating has been proposed to identify bacteria that are potentially more colitogenic³⁷. Our results demonstrate that different combinations of microbial colonization, diet and host immune status can alter both the amount of intestinal IgA and the bacteria it targets. Our results suggest the possibility that fiber deprivation initiates early disruptions in IgA–microbiota interactions along with a loss of IgA-producing plasma cells.

Many questions remain regarding the development of the multifactorial diseases called IBDs. While host genetics are a mostly permanent trait that is difficult to repair, an exception being stem cell transplant in children lacking IL-10 signaling⁴⁶, the microbiome and especially diet are factors that could be manipulated to delay or reverse disease. The rationale for this is that metabolic pathways like mucin degradation can be present in diverse bacteria and some of these organisms may fulfill similar roles in eroding mucus, while some may not. In contrast, strains of the same species can vary in key metabolic capabilities, including mucus degradation, which is known to vary substantially among strains of Bacteroides⁴⁸. If one considers the effects of the microbiome on IBD development to be a cumulative series of positive and negative stimuli caused by the particular behaviors and metabolites exhibited by the microbes, it should be possible to optimize beneficial processes (e.g., butyrate, isobutyrate) while reducing detrimental events like mucus erosion.

STAR METHODS

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Eric Martens (emartens@umich.edu)

Materials Availability

Bacterial strains and data associated with this study are available to interested parties upon request to Lead Contact.

Data and Code Availability

• All available data deported in this work will be shared by the lead contact upon request.

• This paper does not report original code.

• Any additional requests to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animal models, colonization diet and sample processing

Animal experiments at the University of Michigan followed protocols approved by the University of Michigan Institutional Animal Care and Use Committee (IACUC) and were supervised by licensed verterinarians. Germ-free B6.129P2-Il10tm1Cgn/JZtm mice (Institut für Versuchstierkunde und Zentrales Tierlaboratorium, Hannover, Germany) and wild-type C57BL/6J (University of Bern, Switzerland) were bred at the animal facility of the University of Luxembourg and underwent protocols approved by the Animal Experimentation Ethics Committee of the University of Luxembourg and the Ministre de l’Agriculture, de la Viticulture et du Development rural du Grand-Duché du Luxembourg (LUPA 2020/20). Germfree males and females were colonized at 7–10 weeks of age and none of these mice was involved in any previous experiments/treatments. Mice were either housed alone or in groups of appropriate gender, litter, and dietary experiments. Food and autoclaved distilled water were provided ad libitum. Mice were weighed and monitored at least weekly for diarrhea, prolapses, and general health state and this was increased to daily monitoring in animals that began experiencing weight loss. The Muc2^−/− mice, crossed to C57BL/6J background, were obtained from Leonard Augenlicht (Albert Einstein University). For the generation of Il10^−/− Muc2^−/− double knockout (DKO), the single KO mice were crossed to obtain double heterozygous mice until generation of Il10^−/− Muc2^+/−. Full double KO mice developed spontaneous prolapse in germ-free condition during prolonged breeding periods, thus, the Il10^−/− Muc2^+/− mice were used to generate the double KO mice used for all experiments. Genotyping was performed by extracting DNA from ear punches (Transnetyx). Mice were gavaged with the synthetic microbiota at 6–8 weeks and proceeded to diet change as previously described. The synthetic microbiota bacteria were grown in their respective medium¹³ or a modified YCFA medium⁴⁹ prior to community assembly for gavages. The bacteria were cultivated under anaerobic atmosphere maintained with a gas mixture (85% N₂, 10% H₂, 5% CO₂) to an optical density (absorbance 600nm) ranging from 0.5 to 1.0. The communities were assembled by mixing equal volumes of each specific bacterium and aliquoted into sealed screw cap tubes with its own headspace. Each mouse was gavaged with 0.2 mL of its specific community, depending on the experiment, with freshly prepared inocula for two-three consecutive days. A humane endpoint was used for mice that lost ≥20% of their starting weight and these mice were counted as lethalities in the weight loss and survival curves shown. Animals were euthanized using CO₂ asphyxiation for 5 minutes followed by cervical dislocation. The gastrointestinal tracts were retrieved quickly, to prevent autolysis, and the sections separated. Cecal and colon contents of each animal were flash frozen in liquid nitrogen and kept at −80°C until further use.

The FR diet (Lab Diet 5013) and FF diet (Envigo-Teklad TD.130343) have been previously described¹³. The EEN diet employed was Nestle Nutren 1.5, which was lyophilized and sent to Envigo rodent diets to be formed into pellets, bagged and gamma irradiated prior to use. Apple (Vitacel AF401), oat (Vitacel HF600-30) and wheat (Vitacel WF200) fibers were from J. Rettenmaier (Schoolcraft, MI, USA) and added to the FF diet at 7.5% w/w with a corresponding decrease in the amount of glucose in the FF diet. Soluble, highly digestible starch was provided by Cargill (Gel Instant 12412).

METHOD DETAILS

DNA extraction

DNA from fecal and cecal samples were isolated using a bead-beating phenol:chloroform extraction method followed by DNeasy Blood & Tissue Kit (QIAGEN, USA). In short, samples were weighed between 10–50mg and combined with acid-washed glass beads (212–300mm; Sigma-Aldrich, USA), 500uL Buffer A (200 mM NaCl, 200 mM Tris, 20 mM EDTA, 210 uL SDS (20% w/v, filter-sterelized), and 500 uL phenol:chloroform (Thermo Fisher Scientific, USA). The samples were disrupted using a Mini-Beadbeater-16 (Biospec Products, USA) for 3 minutes at room temperature and centrifuged (10,000 rpm, 4°C, 3 minutes). The aqueous phase was recovered and mixed with an equal volume of phenol:chloroform by gentle inversion and centrifuged (10,000 rpm, 4°C, 3 minutes). The remaining aqueous phase was recovered and mixed with 500 μL of chloroform, mixed by gentle inversion, and centrifuged (10,000 rpm, 4°C, 3 minutes). Recovered aqueous phase was mixed with 1 volume of isopropanol and 1/10 volume of 3M sodium acetate, and stored at −80°C for 20 minutes for DNA precipitation. Samples were centrifuged (15000 rpm, 4°C, 20 minutes), supernatant discarded, washed with 70% ethanol, air-dried, and resuspended in nuclease-free water. The sample DNA extracts were further purified with DNeasy Blood & Tissue kit, following manufacturer protocol (QIAGEN).

RNA extraction, reverse transcription and qPCR

Freshly retrieved ileal, cecal and distal colon tissues were transferred into RNAlater^™ (QIAGEN) and kept at 4°C up to a week. Then, RNAlater^™ was removed and tissues were stored at −80°C until further use. Frozen tissues were transferred into 1 mL of TRIzol reagent (Invitrogen^™), homogeneized with a 5 mm metal bead on a bead beater for 8 min at 30Hz and centrifuged for 3 min at 13000 rpm, 4°C. The supernatant was recovered, mixed thoroughly with 200 μl of chloroform and incubated at room temperature for 2–3 min before a centrifigation for 15 min at 13000 rpm, 4°C. The aqueous phase was recovered, mixed again with an equal amount of chloroform and centrifuged for 15 min at 13000 rpm, 4°C. The aqueous phase was recovered, mixed by inversion with 500 μl of isopropanol, incubated for 10 min at room temperature, and centrifuged for 10 min at 13000 rpm, 4°C. The pellet was washed with 1 ml Ethanol 70% and centrifuged for 5 min at 10000 rpm. The supernatant was discarded and the pellet dried for 5–10 min at 37°C, resuspended with 50 μl nuclease-free water and incubated for 15 min at 56°C. Finally, samples were treated with DNase following the Thermo Scientific DNase1, RNase-free Protocol, and RNA were purified with the RNeasy Mini kit (QIAGEN) according to manufacturer instructions. Final RNA concentrations were quantified by Nanodrop.

Lipocalin-2 (LCN-2) measurements in cecal and fecal contents

Frozen cecal and fecal samples were used to quantify the presence of lipocalin-2 protein (LCN-2) by enzyme linked immunosorbent assay (ELISA). Samples previously frozen (cecal, −80°C, and fecal, −20°C) were weighed in new tubes between 5–5 mg. Samples were kept over dry ice during the weighing. Samples were resuspended in 1 mL PBS (pH 7.4), vortexed for 30s, and kept at 4°C overnight to homogenize. Samples were then extensively vortexed to a homogeneous solution. To measure the LCN-2 levels, a DuoSet mouse lipocalin-2/NGAL ELISA kit (R&D Biosystems, USA) was employed using several dilutions of sample homogeneous solution. Quantification was done following manufacturer protocol.

Soluble IgA measurements in cecal and fecal contents

To measure soluble IgA levels, Nunc^® MaxiSorp^™ 384 well plates (Sigma-Aldrich) were coated overnight with 10 ng/well rabbit anti-mouse IgA (Novus Biologicals, Bio-Techne NB7506) in 20 μl/well of carbonate-bicarbonate buffer (Sigma, Ref.: C3041). After four washes with Washing Buffer (1% Tween-20, 154mM Sodium Chloride and 10mM Trisma-base), 75μl of Blocking Buffer (15 mM Trizma-Acetate, 136 mM Sodium Chloride, 2 mM Potassium Chloride and 1% (w/v) BSA (Bovine Serum Albumin)). After 2h at room temperature, wells were washed again. Sample homogeneous solution and standards (mouse IgA Isotype Control UNLB, Southern Biotech, Imtec Diagnostics, Ref: 0106–01) were diluted in Dilution Buffer (15 mM Trizma-Acetate, 136 mM Sodium Chloride, 2 mM Potassium Chloride, 0.1% (w/v) Tween-20, and 1% BSA) and incubated into the plate at 20 μl/well, room temperature for 90 min. After washing, 20 μl/well of a Phosphatase Alkaline-conjugated goat anti-mouse IgA (Southern Biotech, Imtec diagnostics, Ref: 1040-04), diluted 1/1000 in Dilution Buffer, was added and incubated at room temperature for 90 min. After a final wash, 40 μl/well of substrate (1 phosphate tablet (Sigma, ref S0642-200 TAB) dissolved in 10 mL Substrate Buffer (1 mM 2-Amino-2-methyle-1-propanole, 0.1 mM MgCl2.6H2O)) was added. The plate was incubated at 37°C for 60 min before the absorbance was measured at 405 nm using an ELISA plate reader (SpectraMax Plus 384 Microplate Reader from Molecular Devices; Software: SoftMax Pro 7 Software, Molecular Devices). The IgA concentration was determined for each sample using the formulated standard curve.

Short- and branched-chain fatty acid quantification

Short-chain and branched-chain fatty acids (SCFAs) standards mixture was obtained from Sigma (CRM46975). 13C-short chain fatty acid stool mixture (Sigma, SBR00035-1mL) was used as the internal standard (IS). Analytical reagent-grade 3-nitrophenylhydrazine (3NPH)·HCl (Cat#N21804), EDAC·HCl (Cat#341006); HPLC grade pyridine (Cat#270407); LC–MS grade acetonitrile (Cat#34851), water (Cat#270733), and formic acid (Cat#5438040450) were also purchased from Sigma–Aldrich. The working standard solutions were created by performing serial dilution from the 10mM stock solution down to nM range using freshly prepared 50% (v/v) aqueous acetonitrile in water. The chemical derivatization protocol was modified from Han et al. ⁵⁰. Briefly, 20μL of the working standard solutions or samples was mixed with 40μL of 200mM 3NPH in 50% aqueous acetonitrile, 120mM EDAC-6% (v/v) pyridine solution in the same solvent and 4μL of the IS in a Verex glass vial. The mixture was reacted at 40°C for 30 min. After reaction, 96μL of 0.1% formic acid in 10% acetonitrile solution was added to the mixture to quench the reaction. 30μL of the reaction solution was then transferred to a new HPLC vial and 2-μL aliquot of each solution was injected into the LC-MS/MS instrument. Each modified SCFA was optimized in Agilent MS for detection through Agilent Optimizer 2.0. All optimized SCFAs information was combined, and a LC-MRM MS method was created. Retention time for each SCFA was determined from two transitions. Then the MRM MS method was transformed into a dynamic MRM MS or dMRM MS method with all the RTs and MS information for the final LC-MS/MS acquisition method.

LC-MS/MS analysis was performed on the Agilent Technologies Triple Quad 6470 LC/MS system consist of 1290 Infinity II LC Flexible Pump (Quaternary Pump), 1290 Infinity II Multisampler, 1290 Infinity II Multicolumn Thermostat with 6 port valve and 6470 triple quad mass spectrometer. Agilent Masshunter Workstation Software LC/MS Data Acquisition for 6400 Series Triple Quadrupole MS with Version B.08.02 is used for calibration, compound optimization and sample data acquisition.

A Waters Acquity UPLC BEH TSS C18 column (2.1 × 100mm, 1.7μm) column was used with mobile phase A) consisting of 0.1% formic acid in water; mobile phase (B) consisting of 0.1% formic acid in acetonitrile. Gradient program: mobile phase (B) was held at 15% for 1 min, increased to 55% in 19 min, then to 99% in 20 min and held for 2 min before going to initial condition and held for 4 min. The column was at 40 °C and 2 μl of sample was injected into the LC-MS with a flow rate of 0.3 ml/min. Calibration of the 6470 MS was achieved through Agilent ESI-Low Concentration Tuning Mix. Source parameters: Gas temp 300 °C, Gas flow 5 l/min, Nebulizer 45 psi, Sheath gas temp 250 °C, Sheath gas flow 11 l/min, Capillary −3500 V, Delta EMV −200 V. Dynamic MRM scan type is used with 0.07 min peak width. dMRM transitions and other parameters for each compound were list in a separate sheet. Delta retention time of plus and minus 1 min, fragmentor of 40 eV and cell accelerator of 5 eV were incorporated in the method. Data analysis was performed by Agilent Mass Hunter Quantitative Analysis for QQQ B.10.00 for integration. Results were exported to CVS file for further analysis.

Lamina propria cell extraction and flow cytometry analysis

Cecal and colonic lamina propria cells were extracted using the Lamina Propria Dissociation Kit and gentleMACS Dissociators (Miltenyi Biotec, Germany) according to the manufacturer’s instruction. After digestion, cells were resuspended in PB buffer (PBS, pH 7.2, 0.5 % BSA) and counted. For analysis of T cells and NK cells with the expression of transcription factors, the FOXP3/Transcription Factor Staining Buffer kit (eBioscences – 00-5523-00) was used along with the following anti-mouse antibodies: BV605-conjugated anti-CD4 (Biolegend, RM4–5; 1/700), BV650-conjugated anti-B220 (BD Biosciences, RA3-6B2; 1/88), BV711-conjugated anti-CD3 (Biolegend, 17A2; 1/88), BV780-conjugated anti-CD45 (BD Biosciences, 30-F11; 1/88), FITC-conjugated anti-CD335/NKp46 (Biolegend, 29A1.4; 1/100), PE-Cy5-conjugated anti-CD8 (Biolegend, 53–6.7; 1/700), eF450-conjugated anti-FoxP3 (eBiosciences, FJK-16s; 1/200), PE-conjugated anti-GATA3 (Biolegend, 16E10A23; 1/44), PE-eF610-conjugated anti-EOMES (eBiosciences, Dan11mag; 1/100), PE-Cy7-conjugated anti-Tbet (Biolegend, 4B10; 1/44), APC-conjugated anti-RORgt (eBiosciences, AFKJS-9; 1/22). For B cells analysis and immunoglobulin expression, the BD Cytofix/Cytoperm^™ Fixation/Permeabilization Solution Kit (BD Biosciences – 554714) was used along with the following anti-mouse antibodies: eF450-conjugated anti-B220/CD45R (eBiosciences, RA3-6B2; 1/700), eF506-conjugated anti-CD19 (eBiosciences, 1D3; 1/88), BV711-conjugated anti-CD3 (Biolegend, 17A2; 1/88), BV780-conjugated anti-CD45 (BD Biosciences, 30-F11; 1/88), APC-conjugated anti-CD138 (Biolegend, 281–2; 1/100), FITC-conjugated anti-IgA (eBiosciences, mA-6E1; 1/700), PerCP-Cy5.5-conjugated anti-IgD (Biolegend, 11–26c.2a; 1/200), PE-conjugated anti-IgE (Biolegend, RME-1; 1/44), PE-Cy5-conjugated anti-IgM (BD Biosciences, R6–60.2; 1/100), PE-Cy7-conjugated anti-IgG (Biolegend, Poly4053; 1/44). Breifly, 1.5 million cells were washed twice with PBS prior to 15 min of staining with the Zombie NIR^™ Fixable Viability dye (BioLegend – 423105), followed by 2 washes with FACS Buffer (PBS, 5% fetal bovine serum) and fixation according to manufacturer instructions. Cells were then incubated at 4°C for 15 min with Fc Block (Rat anti-mouse CD16 and CD32; BD Pharmingen – Cat.553141) and 30 min with the antibody mixes. Both Fc Block and antibodies were diluted in the permeabilization buffer provided with the fixation kits. Finally, cells were washed twice in their respective permeabilization buffer and resuspended in FACS Buffer for acquisition on a NovoCyte Quanteon Flow Cytometer System (Agilent). The data were then analyzed on FlowJo.

Analysis of IgA-coating of bacteria and sorting

Frozen fecal samples were homogeneized in 1 ml ice-cold PBS and centrifuged for 3 min at 100g, 4°C. Supernatant was filtered through a 70 μm straining sieve and centrifuged for 5 min at 10,000g, 4 °C. The pellet was resupended in 1 ml ice-cold PBS, the OD₆₀₀ detected on a Nanodrop and the amount of bacteria computed as follow: 2 OD₆₀₀ = 10⁹ bacteria. Bacteria were pelleted again for 5 min at 10,000g, 4 °C, and resuspended in 500 μl of staining buffer (PBS, 5% goat serum). After 20 min of incubation on ice, 1×10⁹ bacteria were washed with 1 ml ice-cold PBS and stained for 30 min at 4°C with 4 μg of FITC-conjugated anti-mouse IgA antibody (Southern Biotech) in 100 μl of staining buffer. Cells were then washed once and resupended in 1 ml PBS, and 100 μl of bacteria was pelleted and frozen until further analysis. Remaining bacteria where centrifuged and resuspended in 90 μl of staining buffer and 10 μL of anti-FITC MicroBeads. After 15 minutes of incubation at 4°C, bacteria were washed, resuspended in 500 μl of staining buffer and applied onto a LS column for sorting of IgA⁺ and IgA⁻ fractions with a QuadroMACS^™ Separator (Miltenyi Biotec). IgA-coated and IgA-uncoated fractions were centrifuged and dry pellets were stored at −80°C until further analysis. For analysis of IgA coating of bacteria, frozen pellets were defrost on ice and washed with 1 ml of staining buffer. Since all samples were not sorted at the same time, bacteria were stained again for 30 min with 0.5 μg of FITC-conjugated anti-mouse IgA antibody (Southern Biotech) in 100 μl of staining buffer to refresh and harmonize the staining between batches. After a washing step, DNA was stained for 20 minutes with diluted 1:4000 in 200 μl of DNA staining solution (0.1 M HEPES, 0.9 % NaCl, pH 7.2). Finally, bacteria were washed twice with PBS, fixed for 20 min in 4% PFA, washed again and analyzed on a NovoCyte Quanteon Flow Cytometer System (Agilent). Data were then analyzed on FlowJo. For analysis of IgA affinity, bacteria were treated with 200 μl of PBS containing 3M NaSCN for 15 min at 4°C, 800rpm, prior to blocking and staining with FITC-conjugated anti-mouse IgA antibody as mentioned above.

Mucus measurements

The colons were sectioned from the colon-cecum junction to the anus, and immediately fixed in freshly made Carnoy’s fixative (methanol:chloroform:glacial acetic acid, 60:30:10 v/v). The distal part of the small intestine was fixed in freshly made Carnoy’s fixative together with the half blunt end of the cecum. Fixed tissues were kept in Carnoy’s fixative for three hours and exchanged for fresh fixative for another 24 hours. The tissues were then washed in 100% methanol and kept until placed in cassettes for histology preparation. The remaining empty half of the cecum were flash frozen in liquid nitrogen and kept at −80°C until further use.

Slides were deparaffinized by submerging in xylene (Sigma-Aldrich, USA) for five minutes, followed by another xylene incubation for five minutes. Afterward, the slides were dehydrated twice in 100% ethanol for 5 minutes. The slides were then quickly washed in Milli-Q water and antigens were retrieved by submerging in antigen retrieving solution (10 mM sodium citrate, pH 6.0). The submerged sections were heated to 90°C for 10 minutes and cooldown in room temperature for 20 minutes. Slides were quickly dipped three times in Milli-Q water and blotted to remove excess liquid. To better hold liquid, a PAP pen was used to draw around the tissue area for the subsequent steps. The sections were blocked by covering the tissue in blocking buffer (1:10 goat serum (Sigma, USA) in Tris-buffered Saline (TBS; 500 mM NaCl, 50 mM Tris, pH 7.4)) and incubated for an hour at room temperature. For the primary antibody staining, the tissue was covered with a 1:200 dilution of Mucin 2 antibody (H-300) (Santa Cruz Biotechnology, USA) in blocking buffer and incubated for two hours at room temperature. Following the incubation, the slides were rinsed three times in TBS, for five minutes each. The secondary antibody staining was performed by covering the tissue with a 1:200 dilution of Alexa Fluor 488 conjugated goat anti-rabbit IgG antibody (Thermo Fisher Scientific, USA) in blocking buffer for one hour at room temperature in dark. The tissue sections were washed twice in TBS for 5 minutes, gently blotted, and covered with ProLong Gold Antifade reagent with DAPI (Invitrogen, USA), covered with cover slips and sealed with nail polish. The slides were kept at room temperature for 24 hours in dark, then kept in 4°C until imaging. The mucus layer in the sections were visualized using a Zeiss Apotome by taking pictures across fecal pellets and stitching the images together to compose a single image. Mucus layer measurements were performed by using BacSpace as described by Earle et al.

Histological examination of intestinal tissue sections

Extent of histologic lesions was scored on a semi-quantitative basis by a trained pathologist (KE) in a blinded fashion, using a modification of the scoring system of Bugni et al. ⁵¹. Inflammation, epithelial damage, epithelial hyperplasia and dysplasia and the presence or absence of submucosa edema were scored on a scale of 1–4 according to the following table. Scores for each category were added to determine a total score. Cecum and colon were scored separately.

Score	0	1	2	3	4
inflammation	Normal	Scattered PMNs or plasma cells in lamina propria or submucosa	Coalescing mucosal and/or submucosal inflammation	Widespread submucosal inflammation	Severe diffuse transmural inflammation
epithelial damage	None	Focally dilated glands and/or attenuated surface epithelium, and/or clusters of sloughed cells in lumen	Focally extensive gland dilation and/or surface epithelial attenuation, many sloughed cells in lumen	Mucosa erosions	Mucosal ulceration
hyperplasia/dysplasia	Normal	Hypertrophy and/or hyperplasia present	Cellular and/or glandular dysplasia present
Submucosal edema	Absent	Present

Mucus penetrability assay

Penetrability of the colonic mucus was assessed as described by Gustafsson et al ²⁴. Briefly, colons were flushed with ice-cold oxygenated KREBS Buffer (116 mM NaCl, 1.3 mM CaCl₂, 3.6 mM KCl, 1.4 mM KH₂PO₄, 23 mM NaHCO₃, and 1.2 MgSO₄ – Carl Roth) and opened along the mesenteric axis. The longitudinal muscle layer was removed by blunt dissection and the distal mucosa was inserted in a perfusion chamber. The basolateral chamber was filled with 0.6 μg/ml SYTO9 (Fisher Scientific - 10237582) in oxygenated KREBS Glucose Buffer (KREBS Buffer containing 10mM Glucose, 5.7 mM sodium pyruvate and 5.1 mM sodium-l-glutamate), and the apical chamber was filled with oxygenated KREBS Mannitol Buffer (KREBS Buffer, containing 10mM Mannitol, 5.7 mM sodium pyruvate and 5.1 mM sodium-l-glutamate). After 10 min incubation in the dark at room temperature, FluoSphere^™ carboxylate beads (1μm, red 580/605 – Invitrogen – F882) were applied on top and let to sediment on the tissue for 5 min in the dark at room temperature. The apical chamber was then gently washed several times to remove excess of beads. The chamber was incubated for 10 min in the dark before being visualized with a microscope. For each tissue, 4–7 confocal images were taken in XY stacks from the epithelium at the bottom to the beads on top, with 5 μm-intervals between sections. Images were then analyzed with Imaris software, and the penetrability was computed by comparing the distance between the outer border of the beads and the epithelium with the distance between the most inner beads and the epithelium.

16S rRNA gene-based community analysis

PCR and library preparation were performed by the University of Michigan Microbiome Core lab as described by Kozich et al. ⁵². The V4 region of the 16S rRNA gene was amplified using the dual-index primers described by Kozich et al, 2013. The normalized and amplicon size evaluated samples were sequenced using an Illumina MiSeq. The raw sequences were analyzed using mothur v1.42.3 ⁵³ with the included controls: PBS control during the sequencing and ZymoBIOMICS Microbial Community DNA Standard (cat# D6306) for error analysis. Sequences were aligned to the reference Silva database version 132 for contamination analysis and SPF experiments. Moreover, gnotobiotic mice sequences were also aligned to a custom reference database containing the 16S v4 region from each of the 14 bacteria members. The R package “vegan” was used to calculate the principal component analysis from the Bray-Curtis dissimilarity index.

QUANTIFICATION AND STATISTICAL ANALYSIS

All the statistical analyses were done using Prism GraphPad (Version 9.4.0). All data were analyzed using two-way ANOVA and post hoc test with Original FDR method of Benjamini and Hochberg, with the exception of Figure 1C (one-way ANOVA and post hoc test with Holm-Šídák's multiple comparison test), Figure 3B (one-tailed Studenťs t-test and Wilcoxon test), Figure 4M–N (one-tailed Studenťs t-test and Wilcoxon test), Figure 6G (multiple unpaired t-test and post hoc test with the two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli), and Figure 6A–C–E–F (two-way ANOVA and post hoc test with the two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli). All Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; whereas ns indicates comparisons that are not significant. Exact numbers of animals (n) used for each individual experiments and details of the statistical tests used are indicated in the respective figure legends.

Supplementary Material

1

Table S1. Nutrition and ingredient comparison between mice diets in this work, related to Star Methods – Experimental model and subject details.

Key resources table.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
eF450-conjugated anti-mouse B220/CD45R (RA3-6B2)	eBiosciences	Cat#48-0452-82
BV650-conjugated anti-mouse B220/CD45R (RA3-6B2)	BD	Cat#563893
APC-conjugated anti-mouse CD138 (281-2)	Biolegend	Cat#142506
eF506-conjugated anti-mouse CD19 (1D3)	eBiosciences	Cat#69-0193-82
BV711-conjugated anti-mouse CD3 (17A2)	Biolegend	Cat#100241
FITC-conjugated anti-mouse CD335/NKp46 (29A1.4)	Biolegend	Cat#137606
BV605-conjugated anti-mouse CD4 (RM4-5)	Biolegend	Cat#100548
BV780-conjugated anti-mouse CD45 (30-F11)	BD	Cat#564225
PE-Cy5-conjugated anti-mouse CD8 (53-6.7)	Biolegend	Cat#100710
PE-eF610-conjugated anti-mouse EOMES (Dan11mag)	eBiosciences	Cat#61-4875-82
eF450-conjugated anti-mouse FoxP3 (FJK-16s)	eBiosciences	Cat#48-5773-82
PE-conjugated anti-mouse GATA3 (16E10A23)	Biolegend	Cat#653804
FITC-conjugated anti-mouse IgA (mA-6E1)	eBiosciences	Cat#11-4204-83
FITC-conjugated anti-mouse IgA (polyclonal)	Southern Biotech	Cat#1040-02
PerCP-Cy5.5-conjugated anti-mouse IgD (11-26c.2a)	Biolegend	Cat#405710
PE-conjugated anti-mouse IgE (RME-1)	Biolegend	Cat#406908
PE-Cy7-conjugated anti-mouse IgG (Poly4053)	Biolegend	Cat#405315
PE-Cy5-conjugated anti-mouse IgM (R6-60.2)	BD	Cat#553409
APC-conjugated anti-mouse RORgt (AFKJS-9)	eBiosciences	Cat#17-6988-82
PE-Cy7-conjugated anti-mouse Tbet (4B10)	Biolegend	Cat#644824
rabbit anti-mouse IgA	Novus Biologicals, Bio-Techne	Cat#NB7506
mouse IgA Isotype Control UNLB	Southern Biotech	Cat#0106-01
Phosphatase Alkaline-conjugated goat anti-mouse IgA	Southern Biotech	Cat#1040-04
phosphate tablet	Sigma	Cat#S0642-200 TAB
Fc Block (Rat anti-mouse CD16 and CD32)	BD Pharmingen	Cat#553141
anti-FITC MicroBeads	Miltenybiotec	130-048-701
Mucin 2 antibody (H-300)	Santa Cruz Biotechnology	SC-15334
Alexa Fluor 488 goat anti-rabbit IgG	Invitrogen	A11008
Bacterial and virus strains
Akkermansia muciniphila: DMS 22959, type strain	DSMZ	Cat#DMS 22959
Bacteroides caccae: DSM 19024, type strain	DSMZ	Cat#DSM 19024
Bacteroides ovatus: DSM 1896, type strain	DSMZ	Cat#DSM 1896
Bacteroides thetaiotaomicron: DSM 2079, type strain	DSMZ	Cat#DSM 2079
Bacteroides uniformis: ATCC 8492, type strain	ATCC	Cat#ATCC 8492
Barnesiella intestinihominis: YIT11860	DSMZ	Cat#DSM 21032
Clostridium symbiosum: DSM 934, type strain, 2	DSMZ	Cat#DSM 934
Collinsella aerofaciens: DSM 3979, type strain	DSMZ	Cat#DSM 3979
Desulfovibrio piger: ATC 29098, type strain	ATCC	Cat#ATC 29098
Escherichia coli HS	ATCC	N/A
Eubacterium rectale: DSM 17629, A1-86	DSMZ	Cat#DSM 17629
Faecalibacterium prausnitzii: DSM 17677, A2-165	DSMZ	Cat#DSM 17677
Marvinbryantia formatexigens: DSM 14469, type strain, I-52	DSMZ	Cat#DSM 14469
Roseburia intestinalis: DSM 14610 type strain, L1-82	DSMZ	Cat#DSM 14610
Biological samples
Chemicals, peptides, and recombinant proteins
Acetic acid 100% 1l	Carl Roth	6755.1
Chloroform, 99.8+%, Certified AR for Analysis, Stabilised with Amylene, Fisher Chemical	Fisher Scientific	10122190
Méthanol 5 litres	Carl Roth	4627.6
Phenol/Chloroform/Isoamyl Alcohol (125:24:1 Mixture, pH 4.3, Liq.), Fisher BioReagents	Fisher Scientific	10699543
Sodium chloride, 1 kg, ≥99,5 %, p.a., ACS, ISO	Carl Roth	3957.1
TRIS, 1 kg, PUFFERAN® ≥99,9 %, p.a.	Carl Roth	4855.2
Ethylenediamine tetraacetic acid disodium salt dihydrate, 1 kg	Carl roth	X986.2
Sodium Dodecyl-Sulfate	Fisher Scientific	BP166
Isopropanol, Extra Pure, SLR, Fisher Chemical	Fisher Scientific	10477070
Ethanol	VWR	1.08543.0250
RNA Stabilization Reagent (250 ml)	QIAGEN	76106
RNAprotect	QIAGEN	76506
TRIzol™ Reagent	Ambion / Fisher Scientific	15596018
carbonate-bicarbonate buffer	Sigma	C3041
Invitrogen™ Random primers (Invitrogen™ 48190011)	Fisher Scientific	10646313
Thermo Scientific™ DNase I, RNase-free (Thermo Scientific™ EN0521)	Fisher Scientific	10649890
Invitrogen™ SuperScript™ IV Reverse Transcriptase	Fisher Scientific	15317696
Invitrogen™ Platinum™ Taq DNA Polymerase	Fisher Scientific	10358742
Invitrogen™ Eau distillée exempte de DNase / RNase UltraPure™	Fisher Scientific	12060346
Invitrogen™ SYBR™ Green I Nucleic Acid Gel Stain, 10,000X concentrate in DMSO	FisherScientific	10710004
Invitrogen™ dNTP Set (100 mM) Solution	Fisher Scientific	10083252
Invitrogen™ RNaseOUT™ Recombinant Ribonuclease Inhibitor	Fisher Scientific	10154652
Tween-20	Sigma	P1379
Trisma-base	Sigma	T1503
Trizma-Acetate	Sigma	T1258
Potassium chloride, CELLPURE® ≥99 %	Carl Roth	HN02.2
Bovine Serum Albumin	Sigma	A9647
Zombie NIR™ Fixable Viability dye	BioLegend	423105
fetal bovine serum	Gibco	10500 - 064
Magnesium chloride (MgCl2.6H2O)	Sigma	M9272
Normal Goat Serum	Imtec Diagnostics	0060-01
Gibco™ Goat Serum, New Zealand origin	Fisher	11540526
HEPES Buffer, 1M stock in normal saline, 100 ml	Westburg	LO BE17-737E
SYTO™ 60 Red Fluorescent Nucleic Acid Stain	Invitrogen	S11342
paraformaldehyde	Sigma	441244
paraffin	Sigma	18512
Xylenes, ≥98.5%, ACS Reagent, xylenes plus ethylbenzene basis, Honeywell™ Riedel-de Haën™	FisherScientific	15692690
Sodium citrate	Fisher	S279
Invitrogen™ ProLong™ Gold Antifade Mountant with DAPI, P36931	FisherScientific	11549306
Invitrogen™ SYTO™ 9 Green Fluorescent Nucleic Acid Stain	Fisher Scientific	10237582
Sodium chloride, 1 kg, ≥99,5 %, p.a., ACS, ISO	Carl Roth	3957.1
Potassium chloride, CELLPURE® ≥99 %	Carl Roth	HN02.2
Potassium Phosphate, Monobasic	Carl Roth	P018.1
Sodium acetate trihydrate, 500 g	Carl roth	HN05.1
Magnesium sulfate	Carl Roth	P027.1
Calcium chloride dihydrate	Carl Roth	HN04.2
Pyruvic acid sodium salt, 25 g	Carl roth	8793.1
L-Glutamic acid, 25 g	Carl roth	1743.1
Sodium hydrogen carbonate, 1 kg	Carl roth	HN01.2
ethylenediaminetetraacetic acid	Sigma	EDS
Volatile Free Acid Mix	Sigma	CRM46975
3C-Short Chain Fatty Acids Stool Mixture	Sigma	SBR00035-1mL
3-Nitrophenylhydrazine hydrochloride	Sigma	N21804
EDAC Hydrochloride	Sigma	341006
HPLC grade pyridine	Sigma	270407
LC–MS grade acetonitrile	Sigma	34851
LC–MS grade water	Sigma	270733
LC–MS grade formic acid	Sigma	5438040450
Critical commercial assays
DNeasy Blood & Tissue kit	QIAGEN	69506
Buffer ATL (200 ml)	QIAGEN	19076
Qiagen proteinase K (10 ml)	QIAGEN	19133
RNeasy Mini kit	QIAGEN	74106
DuoSet mouse lipocalin-2/NGAL ELISA kit	R&D Biosystems	DY1857
ZymoBIOMICS Microbial Community DNA Standard	Zymo Research	D6306
Lamina Propria Dissociation Kit	Miltenyi Biotec	130-097-410
FOXP3/Transcription Factor Staining Buffer kit	eBioscences	Cat#00-5523-00
BD Cytofix/CytopermTM Fixation/Permeabilization Solution Kit	BD Biosciences	Cat#554714
Deposited data
Experimental models: Cell lines
Experimental models: Organisms/strains
Mice B6.129P2-Il10tm1Cgn/JZtm	Institut für Versuchstierkunde und Zentrales Tierlaboratorium	N/A
Mice wild-type C57BL/6J	Germ-Free Mouse Facility – University of Michigan and University of Bern	N/A
Mice C57BL/6J Muc2−/−	Albert Einstein University	N/A
Mice C57BL/6J Il10−/− Muc2−/−	This study	N/A
Mice C57BL/6J Il10−/−	Germ-Free Mouse Facility – University of Michigan	N/A
Oligonucleotides
Cytokine Il5 Forward	TGAGGCTTCCTGTCCCTACT	Eurogentec
Cytokine Il5 Reverse	CCACACTTCTCTTTTTGGCGG	Eurogentec
Cytokine Il6 Forward	TAGTCCTTCCTACCCCAATTTCC	Eurogentec
Cytokine Il6 Reverse	TTGGTCCTTAGCCACTCCTTC	Eurogentec
Cytokine Il10 Forward	CTGCCTGCTCTTACTGACTGG	Eurogentec
Cytokine Il10 Reverse	TGGGAAGTGGGTGCAGTTATT	Eurogentec
Cytokine Il12p40 Forward	TGTGGAATGGCGTCTCTGTC	Eurogentec
Cytokine Il12p40 Reverse	GCCTTTGCATTGGACTTCGG	Eurogentec
Cytokine Il13 Forward	CCTGGCTCTTGCTTGCCTT	Eurogentec
Cytokine Il13 Reverse	GGTCTTGTGTGATGTTGCTCA	Eurogentec
Cytokine Il17f Forward	GGTAGCAGCTCGGAAGAACC	Eurogentec
Cytokine Il17f Reverse	TGGAATTCACGTGGGACAGA	Eurogentec
Cytokine Il22 Forward	TCGTCAACCGCACCTTTATG	Eurogentec
Cytokine Il22 Reverse	CCCGATGAGCCGGACA	Eurogentec
Cytokine Il23 Forward	TGGAGCAACTTCACACCTCC	Eurogentec
Cytokine Il23 Reverse	GGGCAGCTATGGCCAAAAAG	Eurogentec
Cytokine Il33 Forward	CGTTCTGGCCTCACCATAAGA	Eurogentec
Cytokine Il33 Reverse	CCGTGGATAGGCAGAGAAGT	Eurogentec
Cytokine Ifng Forward	ATGAACGCTACACACTGCATC	Eurogentec
Cytokine Ifng Reverse	CCATCCTTTTGCCAGTTCCTC	Eurogentec
Cytokine Tnfa Forward	AGCCCACGTCGTAGCAAAC	Eurogentec
Cytokine Tnfa Reverse	GATAGCAAATCGGCTGACGG	Eurogentec
Cytokine Tgfb Forward	TGATACGCCTGAGTGGCTGTCT	Eurogentec
Cytokine Tgfb Reverse	CACAAGAGCAGTGAGCGCTGAA	Eurogentec
Hsp90 Forward	CAGAAGGCTGAGGCAGACAA	Eurogentec
Hsp90 Reverse	ATCATGCGGTAGATGCGGTT	Eurogentec
Synthetic microbiota 14 species primers	N/A
Recombinant DNA
Software and algorithms
SoftMax Pro 7 Software	Molecular Devices	N/A
FlowJo	BD Biosciences	N/A
BacSpace	Earle et al., 2015	N/A
Zen 3.0	Carl Zeiss Microscopy GmbH	N/A
mothur v1.42.3	Schloss et al., 2009	N/A
Imaris	Oxford Instruments Imaris	N/A
Agilent Masshunter Workstation Software LC/MS Data Acquisition for 6400 Series Triple Quadrupole MS with Version B.08.02	Agilent Technologies	N/A
Other
Acid-washed glass beads (212–300 mm)	Sigma-Aldrich	Cat#G1277
Anaerobic chamber	Coy manufacturing	Vinyl Type A + Type B
Nunc® MaxiSorp™ 384 well plates	Thermo Scientific	464718
SpectraMax Plus 384 Microplate Reader	Molecular Devices	ABS PLUS
gentleMACS Dissociators	Miltenyi Biotec	130-096-427
NovoCyte Quanteon Flow Cytometer Systems 4 Lasers	ACEA Biosciences / Agilent	2010097
70 μm straining sieve	Clear Line	141379C
QuadroMACS™ Separator	Miltenyi Biotec	130-090-976
LS column	Miltenyi Biotec	130-042-401
Zeiss Apotome	Zeiss	Discontinued
Microscope Axio Examiner KMAT	Carl Zeiss Microscopy GmbH	491406-9880-010
Perfusion chambers for bead assay	Gustafsson et al., 2012	N/A
FluoSphere™ carboxylate beads, 1μm, red 580/605	Invitrogen	F8821
Screw cap tube, 1.5ml, non-skirted, sterile, white cap	Greiner	716261
Microtube, PP, 2ml, screw cap, 5×50pc, S	Greiner	723261
Stainless Steel Beads, 5 mm (200)	Qiagen	69989
White PCR plates (25 pcs)	Bioplastics	AB70659
RETSCH Broyeur mélangeur MM 400 (Tissue Lyser)	Fisher Scientific	10573034
Retsch™ Rack d’adaptateur en PTFE	Fisher Scientific	10122852
CFX96TM RealTime System (C1000 Thermal Cycler)	Biorad	1855195
Implen™ NanoPhotometer™ N60 Micro-Volume UV-VIS Spectrophotometer (N60-Touch)	Fisher Scientific	15442203
Agilent Technologies Triple Quad 6470 LC/MS system consist of 1290 Infinity II LC Flexible Pump (Quaternary Pump)	Agilent Technologies	Reference
1290 Infinity II Multisampler, 1290 Infinity II Multicolumn Thermostat with 6 port valve and 6470 triple quad mass spectrometer	Agilent Technologies	Reference
Waters Acquity UPLC BEH TSS C18 column (2.1 × 100mm, 1.7μm) column	Waters

Dietary fiber-deprived gut microbiota elicits lethal colitis in Il10^−/− mice

Mucin-degrading gut bacteria weaken mucus and drive inflammatory immune responses

Alterations in bacterial IgA coating and increased NK cells precede inflammation

Exclusive enteral nutrition promotes bacterial isobutyrate, reducing inflammation