Akkermansia muciniphila regulates food allergy in a diet-dependent manner

doi:10.21203/rs.3.rs-1745691/v1

Download PDF

Article

Akkermansia muciniphila regulates food allergy in a diet-dependent manner

https://doi.org/10.21203/rs.3.rs-1745691/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 11 Sep, 2023

Read the published version in Nature Microbiology →

Version 1

posted

You are reading this latest preprint version

Alterations in the gut microbiome, including diet-driven changes, are linked to the rising prevalence of food allergy, yet little is known about how specific gut bacteria incite breakdown of oral tolerance. Here, we show that depriving specific-pathogen-free mice of dietary fiber leads to an increase of the mucolytic bacterium Akkermansia muciniphila, which is associated with a surge in the colonic type 2 immune cells and IgE-coated commensals, and microbiota-mediated gut mucosal barrier dysfunction. These changes manifest into exacerbated sensitization to food allergens, ovalbumin and peanut. To demonstrate the causal role of A. muciniphila, we employed a tractable synthetic human gut microbiota in gnotobiotic mice. The presence of A. muciniphila within the microbiota, combined with fiber deprivation, resulted in stronger anti-commensal IgE coating and type 2 immune responses, which worsened symptoms of food allergy. Our study supports a mechanistic link between diet and a mucolytic gut microbe in regulating food allergy.

As the Western world continues to see rising rates of food allergy^1,2 that cannot be explained solely by genetics^3,4, identifing mechanisms of sensitization that are driven by environmental factors has become increasingly important. In food allergy, the gut microbiota has been shown to play an important role, by inducing protection against sensitization^5–7. This is evident from studies in germ-free mice that develop severe anaphylaxis, which can be dampened with the colonization of single microbes or complex communities of bacteria⁷. Moreover, mouse studies utilizing high-fat⁸ or low-fiber⁹ diets have shown that diet-mediated changes in the gut microbiota are associated with food allergy sensitization. Nevertheless, little is known about the causal links between food allergy sensitization and specific members of the microbiome. Identifying such causal links will facilitate the mechanistic understanding of how diet–gut microbe interactions drive sensitization.

Sensitization in atopic diseases begins at barrier sites and, in the context of food allergy, a functional epithelial barrier has been shown to be crucial in preventing sensitization and promoting oral tolerance³. An important part of the intestinal mucosal barrier is the mucus layer, which serves as a first line of defense in the gut, to prevent the microbial invasion, but allow nutrient absorption^10,11. Although gut exposure to allergens is understood to induce immunotolerance in a healthy gut, tolerance breakdown may occur when the mucus layer is genetically ablated, as shown in Muc2^−/− mice, which lack the gene for O-glycosylated mucin-2 (MUC2), the main constituent of the mucus layer¹². Induction of tolerance in the gut through regulatory T cells is mediated by the sensing of MUC2 by local dendritic cells (DC), demonstrating the importance of the mucus barrier in food allergy^12,13.

Diet is an important environmental factor that affects the mucus barrier through the microbiota because: 1) dietary fiber deprivation increases deterioration of the colonic mucus layer by mucolytic gut bacteria^14,15; and 2) Western-style diet increases the permeability of the colonic mucus layer¹⁶. The fiber-deprived microbiota-mediated mucus barrier dysfunction reflects in enhanced susceptibility to infection by Citrobacter rodentium, a pathogen that must cross the mucus layer to reach the gut epithelium and initiate mucosal infection^14,17. Nevertheless, little is known about how the diet-mediated mucus barrier dysfunction together with increased activities of specific mucolytic microbes might impact oral tolerance.

Our previously published data (as a preprint) show that fiber deprivation in mice induces changes in the microbial metabolome, as well as a microbiota-dependent inflammatory landscape within the colonic lamina propria characterized by increases in type 1 inflammatory cells, such as natural killer (NK), CD8⁺ and helper T (Th)17 cells, and type 2 inflammatory cells, including type 2 innate lymphoid cells (ILC2), mast cells, Th2 and Foxp3⁺Gata3⁺ T cells¹⁸. In the current study, we hypothesized that dietary fiber deprivation, which can drive a phenotype characterized by microbiome-mediated colonic mucus erosion and a type 2 immune landscape, together with an increase of specific mucolytic gut microbes, predisposes the host to enhanced allergen sensitization. Using two experimental allergens, ovalbumin and peanut, in specific-pathogen-free and gnotobiotic mouse models, we establish a causal link between Akkermansia muciniphila and allergic sensitization. We show evidence for how a microbiota containing this mucolytic bacterium promotes type 2 immunity in a diet-dependent manner.

Microbiota-mediated mucus layer thinning is associated with gut barrier dysfunction

We fed specific-pathogen-free (SPF) BALB/c mice for 40 days on a fiber-rich (FR) (standard laboratory chow) or a fiber-free (FF) diet, and analyzed the changes in the gut microbiota composition. In line with our previously published results in gnotobiotic mice with a characterized 14-member synthetic human gut microbiota (14SM)¹⁴, the microbiota of FF-fed mice exhibited increased relative abundances of bacteria belonging to the known mucin-specialist genus Akkermansia (approximately 3.4-fold increase in FF- compared to FR-fed mice; change from average relative abundance of 1.7–5.7%) (Fig. 1a). Additionally, other putative mucin-degrading species (Fig. 1a, shown in bold), such as Mucispirillum¹⁹ and Parabacteroides²⁰, were also increased in the FF-fed mice.

In line with previous findings^14,15,17, the FF-fed mice exhibited a two-fold reduction in the colonic mucus layer thickness (Fig. 1b–d, Extended Data Fig. 1a), using fixed tissue samples (Fig. 1c, d). In addition, using an ex vivo ex situ method to assess the penetrability of the mucus on dissected colonic tissues excluding the luminal contents²¹, we show that the colonic mucus displayed similar normalized penetrability to bacteria-sized fluorescent beads between the two diets (Fig. 1e). Intriguingly, with this method, the overall mucus layer thickness, determined as the distance of the beads from the epithelium (Extended Data Fig. 1b–c), was signficantly higher in the FF-fed mice (Fig. 1f). These results suggest that, in the absence of luminal contents—and consequently, in the absence of mucin-degrading microbial enzymes—more mucus was secreted in the FF-fed mice as compared to the FR-fed mice, supporting a stronger goblet cell activity. On the other hand, given that the FF-fed mice exhibited a reduced mucus layer thickness at a steady-state in vivo (Fig. 1b–d, Extended Data Fig. 1a), these results strongly suggest that the reduction in the mucus thickness in the FF-fed mice is a result of increased microbial mucin foraging rather than reduced mucus secretion. In addition, despite a similar penetrability of the mucus to bacteria-sized beads (Fig. 1e), the overall intestinal permeability to FITC-dextran was increased in the FF-fed mice (Fig. 1g), suggesting a higher permeability of the gut mucosal barrier to luminal antigens during fiber deprivation.

Consistent with a type 2-skewed immune landscape in BALB/c mice²², the FF diet did not increase the level of the neutrophilic inflammatory marker lipocalin (LCN-2) (Extended Data Fig. 1d) and only slightly increased the transcript level of the type 1 cytokine IFNγ in the colon (Fig. 1h). By contrast, in the colon, but not in the ileum nor the cecum, the FF-fed mice displayed increased transcriptional expression of both interleukin (IL)-22 and IL-25 (Fig. 1i, j). IL-22 is an important barrier protective cytokine and its expression can be increased during tissue inflammation²³. On the other hand, IL-25 can stimulate the production of type 2 inflammatory cytokines and initiate allergic reactions²⁴. Taken together, these data suggest a latent mucosal inflammatory state in the colon, resulting from the microbiota-driven mucus and epithelial barrier dysfunction.

Colonic barrier dysfunction is associated with local type 2 inflammation and anti-commensal immunity

Using mass cytometry by time-of-flight (CyTOF)-based broad immunophenotyping, our recently prepublished data showed that FF-fed mice exhibit a biased immune landscape in the colonic lamina propria (cLP), with an increase in type 2 inflammatory cells such as ILC2, mast cells and Th2 cells¹⁸. Targeted FACS analysis revealed a higher frequency of eosinophils in the colon and ileum in FF-fed mice (Fig. 2a, two-way ANOVA, diet effect P = 0.0281; and Extended Data Fig. 2a), along with a higher frequency of IL-5-producing CD4⁺ and/or CD8⁺ T cells in the colon and ileum (Fig. 2b–e and Extended Data Fig. 2b). Consistent with an overall increase of activated CD69⁺ T cells¹⁸, FF-fed mice had also more IFNγ-, IL-13-, and IL-17-producing T cells in the colon (Fig. 2b–c), while in the ileum, only IFNγ-producing T cells were enriched in addition to IL-5-producing cells (Fig. 2d–e). To investigate whether the type 2 immune profile in FF-fed mice had an effect beyond the gut, we analyzed the serum. Titers of IgE and IgG1 were not different between the two dietary groups (Fig. 2f and Extended Data Fig. 2c), although the titers of mast cell protease 1 (MCPT1) were reduced in the FF-fed SPF mice but not in the germ-free controls (Extended Data Fig. 2c). Based on the serum data, these results show that FF-fed mice develop a type 2 immune profile that is restricted to the gut.

To determine whether the increased type 2 inflammation within the colon could be reflected in the anti-commensal antibody response, we analyzed the IgA- and IgE-coating of fecal bacteria over the 40-day feeding period. As expected, bacteria were mainly coated with IgA with a detection of about 30% IgA-coated bacteria at the beginning of the timecourse (Fig. 2g). However the titers remained relatively stable between the two diets over time, except higher titers in FF-fed mice compared to FR-fed mice at day 40 (Fig. 2g). In contrast, IgE-coated bacteria were almost undetectable at the beginning of the experiment and increased over time in FF-fed mice only, reaching up to 2% of fecal bacteria at day 40 (Fig. 2h and Extended Data Fig. 2d). In contrast to fecal bacteria after 40 days of feeding, ileal bacteria remained minimally coated with IgE in both FR- and FF-fed mice (Fig. 2i). In line with the barrier breakdown profiles (Fig. 1), these results: 1) support a colon-specific, anti-commensal IgE response driven by the FF diet; and 2) point to a possible route of allergic sensitization through a reduced colonic mucus layer and a more permeable mucosal barrier.

Diet-driven colonic mucosal barrier dysfunction reflects in exacerbated food allergic responses

Given the microbiome-mediated mucus barrier breakdown and the stronger type 2 immune response in FF-fed mice, we hypothesized that the mice may be rendered more susceptible to food allergen sensitization via the oral route. We applied a protocol²⁵ for ovalbumin (OVA) sensitization using cholera toxin (CTX) as an adjuvant to mice after the 40-day feeding period on FR and FF diets (Fig. 3a). The FF diet did not affect the weight gain of the mice during the 40-days feeding period, neither during the sensitization period (Extended Data Fig. 3). Upon allergen challenge, FF-fed sensitized (OVA-CTX) mice exhibited more severe anaphylactic symptom scores (Fig. 3b) and a greater drop in core body temperature (Fig. 3c), compared to FR-fed sensitized (OVA-CTX) mice. Interestingly, FF-fed mice that were exposed solely to the allergen without adjuvant (OVA) also exhibited increased symptom scores compared to their FR-fed controls (Fig. 3b). In line with our results in Fig. 2, IgE-coated bacteria were more abundant in the feces from FF-fed sensitized mice compared to FR-fed sensitized mice (Fig. 3d). However, sensitization did not increase the overall frequency of IgE-coated bacteria compared to FF-fed unsensitized mice (Fig. 2i), signifying that the increased IgE coating occurs independent of allergen sensitization, likely owing to fiber deprivation occuring already in the first 40 days of feeding. Additionally, serum MCPT1 titers were still lower in FF-fed sensitized mice compared to FR-fed sensitized mice (Fig. 3e). On the contrary, analysis of serum immunoglobulin levels showed an increase in total IgE, but not IgG1, in FF-fed mice compared to FR-fed mice regardless of any experimental sensitization (Fig. 3f). Of note, while the total IgE titers between the diets among three experimental groups did not differ between FR- and FF-fed mice after 40 days of feeding (Fig. 2f), unsensitized mice treated with PBS showed higher titers when fed the FF diet compared to the FR diet (Fig. 3f), suggesting that a feeding period longer than 40 days may result in an atopic phenotype. Similarly, fiber deprivation increased OVA-specific IgE levels but not OVA-specific IgG1 in the serum of sensitized mice (Fig. 3g).

To further support our findings with another type of allergen, we investigated the role of fiber deprivation in the sensitization process to peanut (PN) by adapting a previously published protocol based on PN protein extract, using CTX as an adjuvant⁷ (Extended Data Fig. 4a). As seen with the OVA challenge, neither the diet nor the sensitization affected the weight gain of the mice in the course of the experiment timeline (Extended Data Fig. 4b). FF-fed PN-sensitized (PN-CTX) mice showed more severe symptom scores and temperature drops compared to FR-fed sensitized (PN-CTX) mice (Fig. 3h, i). Although PN-specific IgE and IgG1 were similar between the two diet groups (Extended Data Fig. 4c, d), fiber deprivation increased total IgE and IgG1 levels induced by both PN-CTX or PN sensitization (Extended Data Fig. 4e, f). Additionally, among PN-sensitized mice, FF-fed mice had decreased serum MCPT1 compared to FR-fed mice (Extended Data Fig. 4g), which was similar to the findings from OVA sensitization. Altogether, these results support the idea that fiber deprivation increases food allergen sensitization in SPF mice. As the results between the two dietary antigens were largely consistent, we elected to focus our attention on OVA sensitization for further immune characterizations.

To examine the immunological landscape behind this increased allergic phenotype, we performed broad immunophenotyping of the cLP by CyTOF using a 28-marker panel in OVA-sensitized and control mice. We performed FlowSOM²⁶ clustering (Extended Data Fig. 5) and generated a Uniform Manifold Approximation and Projection (UMAP)²⁷ plot to visualize the cell populations contributing to the immune landscape within the cLP (Fig. 4a). The FF-fed control mice (PBS) exhibited an immune profile characterized by increased cell frequencies of helper T cells (Th1, Th2 and Th17) in addition to NK and CD8⁺ T cells (Fig. 4b–h). Similar profiles were found in sensitized mice (OVA-CTX), where Th1 and Th2 cells were also increased compared to PBS-treated mice (Fig. 4b, c). FoxP3⁺ regulatory T cells (Fig. 4e), CD8⁺ T (Fig. 4f), memory Ly6C⁺CD8⁺ T cells (Fig. 4g) and NK cells (Fig. 4h) were also enriched in FF-fed OVA sensitized mice. Although typically described as type 1 inflammatory cells, CD8⁺ T cells have recently been highlighted for their role in food allergy²⁸. Supporting a role in the fiber deprivation-increased sensitization shown here, IL-5-producing CD8⁺ T cells, but not CD4⁺ T cells, were enriched in the cLP of FF-fed OVA-CTX-sensitized mice (Fig. 4i, j).

Allergic sensitization shifts microbiome in a diet-dependent manner

The composition of the gut microbiome has been linked to allergic sensitization, likely owing to its role in maintaining tolerance via the induction of regulatory T cells and priming of the immune system⁵. Moreover, a microbiome signature has been previously linked with pediatric food allergy²⁹, characterized by increased abundance of the mucin-degrader, Ruminococcus gnavus³⁰, and decreased abundance of several fiber-degrading and probiotic (i.e. Bifidobacterium longum) species compared to healthy controls²⁹. Nevertheless, it remains unclear whether these microbial changes occur prior to or as a consequence of allergic sensitization. Here, to understand how the microbiome is affected by the process of allergic sensitization, using Illumina sequencing platform, we performed 16S rRNA gene sequencing of the V4 region from fecal DNA. Microbiome profiles, as summarized by the Bray–Curtis dissimilarity index, were significantly shifted in FR-fed mice upon challenge, but not in FF-fed mice (Fig. 5a). Of note, the microbiome of mice differed mostly depending on the diet (Fig. 1a, also see reference 18), suggesting that fiber deprivation itself already creates a stronger selective pressure on the microbiome composition than the sensitization process. Barplots of the relative abundances on the family levels revealed increases in the Lachnospiraceae family, and decreases in the families Erysipelotrichaceae, Akkermansiaceae, Clostridia UCG 014, and Tannerallaceae due to allergic sensitization (Fig. 5b, c).

Differential analysis confirmed significant shifts for 24 features in FR-fed sensitized mice and 7 features in FF-fed sensitized mice (Fig. 5c). Of these, 6 features belonging either to the genus Faecalibaculum (sp51 and sp49), to the family Muribaculaceae (sp17), to the genus Eubacterium (sp158 and sp169) or to the family Lachnospiraceae (sp139) were significantly changed due to sensitization in the FF-fed mice, but not in the FR-fed mice (Fig. 5c). On the contrary, among the features that changed more specifically in FR-fed mice, A. muciniphila was most reduced due to sensitization (Fig. 5c). A. muciniphila abundance increased after the 40-day feeding period and decreased after sensitization in both FR- and FF-fed mice (Fig. 5c). However, while this mucolytic bacterium completely disappeared after sensitization in FR-fed mice, the shift was not as evident in FF-fed mice, likely due to the mucin-foraging niche created and supported by fiber deprivation (Fig. 5c, d). Consistently, the shift in microbial metabolism under fiber deprivation could be observed after sensitization through reduced cecal concentrations of short-chain fatty acids (SCFAs), butyrate, propionate and acetate, and increased concentrations of valerate and protein-derived branched-chain fatty acids (BCFAs), isovalerate and isobutyrate (Fig. 5e).

In a previous study, A. muciniphila was reported to be decreased in the fecal microbiota of severe asthmatic patients, and oral supplementation of A. muciniphila to the native complex gut microbiota was able to modulate the immunophenotype and protect mice from allergic airway inflammation³¹. Our results similarly show that A. muciniphila is lost in sensitized FR-fed mice (Fig. 5c, d); nevertheless, the FF-fed mice—which displayed exacerbated food allergy symptoms—did not show a similar loss of this bacterium post sensitization (Fig. 5c, d). Together with the previous study³¹, these results support a link between A. muciniphila and food allergy sensitization, and call attention to better understand the potential diet-dependent causal role of this mucolytic bacterium in sensitization.

Akkermansia muciniphila modulates food allergic sensitization in a diet-specific manner

To investigate the causal role of A. muciniphila in modulating food allergen sensitization, we used germ-free (GF) mice colonized with a fully characterized, human 14-member synthetic human gut microbiota (14SM) in which A. muciniphila can be either included (14SM) or excluded (13SM)¹⁴. After the SM gavage, the mice were fed for a period of 7–10 days on FR or FF diets and underwent OVA sensitization and challenge as per the protocol in Fig. 3a. The microbiota composition differed between pre- and post-sensitization groups (Fig. 6a). In contrast to data in SPF mice (Fig. 5d), A. muciniphila was still present in FR-fed mice after sensitization, reaching 5–11% of the 14SM community (Fig. 6a, Extended Data Fig. 6). Nevertheless, its relative abundance was still increased in FF-fed mice compared to FR-fed mice, reaching 13–18% of the community (Extended Data Fig. 6). In the absence of A. muciniphila, the relative abundances of the three other mucin-degrading bacteria, Bacteroides thetaiotaomicron, Bacteroides caccae and Barnesiella intestinihominis increased in either both dietary groups or in FF-fed mice (Fig. 6a, Extended Data Fig. 6). The low-abundant bacterium Marvinbryantia formatexigens and Desulfovibrio piger also substantially expanded in the absence of A. muciniphila, in FR-fed and FF-fed mice, respectively (Extended Data Fig. 6), whereas the relative abundance of Escherichia coli, Clostridium symbiosum, Bacteroides ovatus, Roseburia intestinalis and Faecalibacterium prausnitzii decreased in either both or one of the 13SM dietary groups (Fig. 6a, Extended Data Fig. 6).

As previously reported⁷, GF sensitized FR-fed mice exhibited severe anaphylaxis scores at challenge, with little to no temperature decrease (Fig. 6b, c). The GF FF-fed mice showed similar anaphylaxis scores compared to the GF FR-fed mice, suggesting that the FF diet alone, without the microbiota, is incapable of further increasing the allergic severity. By contrast, 14SM-colonized mice reproduced the same pathological profile as seen with SPF mice, with more severe symptoms in FF-fed sensitized (OVA-CTX) mice compared to FR-fed sensitized (OVA-CTX) mice (Fig. 6b, c). In line with the SPF data (Fig. 3b), when 14SM-colonized mice were exposed only to OVA without adjuvant, FF-fed mice had significantly increased symptoms compared to FR-fed mice, suggesting that the FF diet and its implications to barrier integrity may act as an adjuvant (Fig. 6b). Interestingly, when A. muciniphila was excluded from the microbiota (13SM), sensitized FR-fed and FF-fed mice had similar symptom scores at challenge, and these were overall reduced in comparison to 14SM-colonized FF-fed mice (Fig. 6b). Further building on our data from SPF mice (Fig. 3), these gnotobiotic mouse experiments demonstrate a causal role of A. muciniphila in driving the increased susceptibility to allergic sensitization induced by fiber deprivation. As with the SPF mouse data, neither the FF feeding nor the microbiota composition affected the weight changes in mice during the sensitization period (Extended Data Fig. 7a).

In order to pinpoint the underlying immune pathways activated by the presence of A. mucinphila in the microbiota, we performed CyTOF acquisition of the cLP from these mice (Extended Data Figs. 8 and 7b). Strikingly, while sensitized 14SM-colonized mice were characterized by a strong infiltration of colonic Th2 cells, sensitized 13SM-colonized mice had higher frequencies of Th1, suggesting that A. muciniphila within the microbiota promotes the recruitment of Th2 cells during sensitization (Fig. 6d). By contrast, innate type 2 cells (ILC2, M2 macrophages and mast cells) had the tendency to be more abundant in sensitized 13SM-colonized FR-fed mice compared to 14SM-colonized FR-fed mice (Fig. 6e, Extended Data Fig. 7c). The fiber deprivation decreased the abundances of these innate type 2 cells, as well as Gata3⁺Foxp3⁺CD4⁺ T cells and CD69⁺ Th17 in 13SM-colonized mice (Fig. 6e, Extended Data Fig. 7c–e). In addition, the FF diet increased eosinophil infiltration in 14SM-colonized mice, which was decreased in 13SM-colonized mice (Fig. 6f). These results support a role for A. muciniphila in the maintenance of the allergic type 2 response under fiber deprivation, while the 13SM community would have a regulatory effect by supporting the Th1 response. Intriguingly, 14SM-colonized mice exposed to OVA only, without adjuvant, had profiles more similar to OVA-CTX sensitized 13SM-colonized mice than to 14SM-colonized ones, with lower levels of innate type 2 cells, CD69⁺ Th17 and Gata3⁺Foxp3⁺CD4⁺ T cells, and more Ly6C⁺ CD8⁺ T cells under fiber deprivation (Fig. 6e, Extended Data Fig. 7c–f). These data suggest that A. muciniphila and CTX cooperate to modulate these cell populations. Finally, as shown in SPF mice, the FF diet increased the titers of fecal IgE-coated bacteria in OVA-CTX-sensitized mice. In the absence of A. muciniphila (13SM), this diet effect was reduced (Fig. 6g). Taken together, these data show that A. muciniphila exacerbates allergen susceptibility under fiber deprivation by promoting type 2 inflammation and anti-commensal IgE secretion.

Since A. muciniphila is the most abundant mucin-degrading bacterium in the 14SM model, we assessed whether its absence in the 13SM community impaired the overall microbial enzymatic activity of carbohydrate-active enzymes (CAZymes) and sulfatases that are known to be involved in mucin foraging (Extended Data Fig. 7g). As expected, the activity of the plant glycan-targeting β-glucosidase was reduced under fiber deprivation in both 13SM and 14SM FF groups as compared to their FR counterparts. In line with the increased relative abundances of other mucolytic bacteria in 13SM groups, differences in the mucin-targeting enzyme activities (β-N-acetlyglucosaminidase, α-L-fucosidase and sulfatase) between FR and FF groups within 13 or 14SM remained similar.

Our study shows that fiber deprivation in specific-pathogen-free (SPF) mice promotes food allergic sensitization by impairing the colonic mucosal barrier and inducing a type 2 immune response. The SPF mice exhibited an increased relative abundance of A. muciniphila, along with colonic barrier dysfunction and the higher susceptibility to allergen sensitization. To demonstrate causality, we removed A. muciniphila from a tractable synthetic human gut microbiota (14SM), and the absence of this bacterium within the microbiota results in reduced allergy symptoms. We further show that A. muciniphila—within a fiber-deprived microbiota—enhances allergic severity by increased colonic infiltration of type 2 immune cells. Overall, our results support a mechanistic connection between fiber-deprived diet and a mucolytic bacterium in regulating food allergy.

The fiber deprived gut microbiota-mediated immune profile in SPF and gnotobiotic mice (with A. muciniphila) is characterized by the expansion of Th2 cells and eosinophils in the colonic lamina propria, as well as an increase in IgE coating of fecal bacteria. Although IgE-coated bacteria were recently identified in fecal samples from food allergic children³², the mechanism underlying this process is still elusive. Here, we report that fiber deprivation can promote the binding of IgE to gut bacteria, even in the absence of experimental allergic sensitization.

A. muciniphila has been described having beneficial immune modulating properties^31,33−35; in this study, we show that the impact of this bacterium is context-dependent and can be detrimental in the context of allergic sensitisation when the microbiota is deprived of dietary fiber. In support of our findings, Derrien et al. had previously reported an increase in IL-4 signaling at the transcriptional level in the colon upon monocolonization of germ-free mice with A. muciniphila³⁴, suggesting a role for type 2 immune skewing in the host. Although we did not perform similar monocolonization in our study, it is possible that such type 2 skewing by A. muciniphila gets enhanced within a microbiota, as it was recently shown that its cooperation with other strains is essential for foraging colonic mucus³⁶.

Fiber deprivation increased mucus thinning and intestinal permeability in SPF mice and both may play a role in facilitating recognition of food allergen by immune cells (Fig. 1). Nevertheless, since the global mucolytic capacity of the synthetic community was unchanged in the absence of A. muciniphila, these data suggest that the overall increased gut microbial mucin foraging is not a major driver of the allergic type 2 response in this model. However, the enhanced mucolytic activity of A. muciniphila in the context of a fiber-deprived diet may be required as an additional microbial trigger to promote the breakdown of oral tolerance, likely through increased Th2 immunity. Moreover, our results show that in the absence of A. muciniphila, only the relative abundance of other mucolytic bacteria primarily increased, whereas SCFA-producing microbes (C. symbiosum, R. intestinalis and F. prausnitzii) decreased. Although the Th2 response may be dampened by other unknown changes in the 13SM community, such an effect is nonetheless an indirect consequence of A. muciniphila being absent in the community. It is likely that there are similar other allergy-driving “biomarker” commensals in the gut microbiota that promote tolereance breakdown following dietary changes, either alone or in combination with other biomarker species. Identifying such additional biomarker species would be important in order to use the gut microbiota as a therapeutic/preventative target in food allergy.

Dietary antigens were long thought to be absorbed in the small intestine, where they promote oral tolerance³⁷. In the last decades, however, antigen uptake has been described in the colon through goblet cell-associated antigen passages (GAPs)³⁸, suggesting that allergen sensitization may also develop in the colon. Consistently, we have primarily identified results pertaining to the colon. Recently, we showed that the four mucolytic bacteria among the 14SM community—A. muciniphila, B. thetaiotaomicron, B. caccae and B. intestinihominis—were necessary to mediate the fiber deprivation-induced mucin foraging and severe colitis in the colon but not in the cecum of Il10^−/− mice³⁹, a model for inflammatory bowel disease (IBD). This is likely due to the naturally loose and permeable mucus layer in the small intestine and cecum, whereas its uniform and thick structure in the colon is critical for maintaining host homeostasis with the gut microbiota^12,40. Altogether, these data suggest that the colon is a strategic site for diet–microbiota–mucus interactions and food allergy sensitization.

We have identified inflammatory effects of the fiber-deprived diet on the cLP¹⁸. Here, we present evidence that this inflammatory profile may come as a result of barrier breakdown, allowing greater access of luminal antigens to immune cells. Another study identified a link between acute intestinal infection, which caused a barrier break in the colon, and the subsequent onset of a local dietary antigen (OVA) IgE response⁴¹. These data suggest that either an infection- or diet-mediated breakdown in the colonic barrier can drive an immunological response in the form of food allergy. This further supports findings among patients suffering from IBD and irritable bowel syndrome (IBS), who exhibit a high prevalence of food sensitivity^42–44.

An interesting observation in our broad immunophenotyping data in SPF sensitized mice was the immune skewing for both Th2 and Th1/Th17 that is induced by a fiber-deprived diet. In line with this, in atopic dermatitis, acute inflammation transitioning into a chronic stage was characterized by an overall activation of all Th subsets, rather than Th2 to Th1^45,46. Allergic inflammation at the site of sensitization, as shown in this study in the colon, can support this phenotype. Overall, these data support that food allergy encompasses a vast array of endotypes evolving due to new environmental triggers. Future clinical evaluation and research should focus on immune distinctions and the site of inflammation between food allergic endotypes in order to design personalized diet- and microbiota-based therapies.

Acknowledgements

M.S.D. was supported by the Luxembourg National Research Fund (FNR) CORE grants (C15/BM/10318186 and C18/BM/12585940); A.P. was supported by an FNR AFR individual PhD fellowship (11602973); E.T.G. was supported by FNR PRIDE (17/11823097) and the Fondation du Pélican de Mie et Pierre Hippert-Faber, under the aegis of the Fondation de Luxembourg; M.B. was supported by a European Commission Horizon 2020 Marie Skłodowska-Curie Actions individual fellowship (897408); S.Z.C. was supported by FNR PRIDE (19/14254520); and M.N. was supported by the FNR AFR bilateral grant (15/11228353). We acknowledge the National Cytometry Platform (NCP) and the Cytométrie Pitié-Salpêtrière (CyPS) for their assistance with generating cytometry data. The NCP is supported by Luxembourg's Ministry of Higher Education and Research (MESR) funding. We also gratefully acknowledge the support of Christian Jäger, Xiangyi Dong and Floriane Gavotto from the Luxembourg Centre for Systems Biomedicine (LCSB) Metabolomics Platform for GC-MS analyses.

1. Gupta, R. S. et al. Prevalence and Severity of Food Allergies Among US Adults. JAMA Netw. open 2, e185630 (2019).

2. Gupta, R. S. et al. The prevalence, severity, and distribution of childhood food allergy in the United States. Pediatrics (2011). doi:10.1542/peds.2011-0204

3. Akdis, C. A. Does the epithelial barrier hypothesis explain the increase in allergy, autoimmunity and other chronic conditions? Nature Reviews Immunology (2021). doi:10.1038/s41577-021-00538-7

4. Renz, H. & Skevaki, C. Early life microbial exposures and allergy risks: opportunities for prevention. Nat. Rev. Immunol. 21, 177–191 (2021).

5. Plunkett, C. H. & Nagler, C. R. The Influence of the Microbiome on Allergic Sensitization to Food. J. Immunol. (2017). doi:10.4049/jimmunol.1601266

6. Feehley, T. et al. Healthy infants harbor intestinal bacteria that protect against food allergy. Nat. Med. 25, 448–453 (2019).

7. Stefka, A. T. et al. Commensal bacteria protect against food allergen sensitization. Proc. Natl. Acad. Sci. 111, 13145–13150 (2014).

8. Hussain, M. et al. High dietary fat intake induces a microbiota signature that promotes food allergy. J. Allergy Clin. Immunol. (2019). doi:10.1016/j.jaci.2019.01.043

9. Tan, J. et al. Dietary Fiber and Bacterial SCFA Enhance Oral Tolerance and Protect against Food Allergy through Diverse Cellular Pathways. Cell Rep. 15, 2809–2824 (2016).

10. Johansson, M. E. V. & Hansson, G. C. Immunological aspects of intestinal mucus and mucins. Nat. Rev. Immunol. 16, 639–649 (2016).

11. Martens, E. C., Neumann, M. & Desai, M. S. Interactions of commensal and pathogenic microorganisms with the intestinal mucosal barrier. Nat. Rev. Microbiol. 16, 457–470 (2018).

12. Shan, M. et al. Mucus enhances gut homeostasis and oral tolerance by delivering immunoregulatory signals. Science 342, 447–453 (2013).

13. Parrish, A., Boudaud, M., Kuehn, A., Ollert, M. & Desai, M. S. Intestinal mucus barrier: a missing piece of the puzzle in food allergy. Trends Mol. Med. 1–15 (2021). doi:10.1016/j.molmed.2021.10.004

14. Desai, M. S. et al. A Dietary Fiber-Deprived Gut Microbiota Degrades the Colonic Mucus Barrier and Enhances Pathogen Susceptibility. Cell 167, 1339-1353.e21 (2016).

15. Earle, K. A. et al. Quantitative Imaging of Gut Microbiota Spatial Organization. Cell Host Microbe (2015). doi:10.1016/j.chom.2015.09.002

16. Schroeder, B. O. et al. Bifidobacteria or Fiber Protects against Diet-Induced Microbiota-Mediated Colonic Mucus Deterioration. Cell Host Microbe 23, 27–40 (2018).

17. Neumann, M. et al. Deprivation of dietary fiber in specific-pathogen-free mice promotes susceptibility to the intestinal mucosal pathogen Citrobacter rodentium. Gut Microbes (2021). doi:10.1080/19490976.2021.1966263

18. Parrish, A. & Grant, E. et al. Dietary fibers boost gut microbiota-produced B vitamin pool and alter host immune landscape. Res. Sq. doi:10.21203/rs.3.rs-1563674/v2

19. Loy, A. et al. Lifestyle and Horizontal Gene Transfer-Mediated Evolution of Mucispirillum schaedleri, a Core Member of the Murine Gut Microbiota. mSystems (2017). doi:10.1128/msystems.00171-16

20. Pudlo, N. A. et al. Phenotypic and Genomic Diversification in Complex Carbohydrate-Degrading Human Gut Bacteria. mSystems (2022). doi:10.1128/msystems.00947-21

21. Gustafsson, J. K. et al. An ex vivo method for studying mucus formation, properties, and thickness in human colonic biopsies and mouse small and large intestinal explants. Am. J. Physiol. - Gastrointest. Liver Physiol. (2012). doi:10.1152/ajpgi.00405.2011

22. Hsieh, C. S., Macatonia, S. E., O’Garra, A. & Murphy, K. M. T cell genetic background determines default t helper phenotype development in vitro. J. Exp. Med. (1995). doi:10.1084/jem.181.2.713

23. Keir, M. E., Yi, T., Lu, T. T. & Ghilardi, N. The role of IL-22 in intestinal health and disease. J. Exp. Med. (2020). doi:10.1084/jem.20192195

24. Fort, M. M. et al. IL-25 Induces IL-4, IL-5, and IL-13 and Th2-associated pathologies in vivo. Immunity (2001). doi:10.1016/S1074-7613(01)00243-6

25. Ganeshan, K. et al. Impairing oral tolerance promotes allergy and anaphylaxis: A new murine food allergy model. J. Allergy Clin. Immunol. 123, 231-238.e4 (2009).

26. Gassen, S. Van et al. FlowSOM: Using Self-Organizing Maps for Visualization and Interpretation of Cytometry Data. doi:10.1002/cyto.a.22625

27. Mcinnes, L., Healy, J. & Melville, J. UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. (2020).

28. Yu, W. et al. Allergen-specific CD8 + T cells in peanut-allergic individuals. J. Allergy Clin. Immunol. (2019). doi:10.1016/j.jaci.2019.01.011

29. De Filippis, F. et al. Specific gut microbiome signatures and the associated pro-inflamatory functions are linked to pediatric allergy and acquisition of immune tolerance. Nat. Commun. (2021). doi:10.1038/s41467-021-26266-z

30. Crost, E. H. et al. Mechanistic insights into the cross-feeding of Ruminococcus gnavus and Ruminococcus bromii on host and dietary carbohydrates. Front. Microbiol. 9, 2558 (2018).

31. Michalovich, D. et al. Obesity and disease severity magnify disturbed microbiome-immune interactions in asthma patients. Nat. Commun. 10, 5711 (2019).

32. Abdel-Gadir, A. et al. Microbiota therapy acts via a regulatory T cell MyD88/RORγt pathway to suppress food allergy. Nat. Med. 25, 1164–1174 (2019).

33. Ansaldo, E. et al. Akkermansia muciniphila induces intestinal adaptive immune responses during homeostasis. Science 364, 1179–1184 (2019).

34. Derrien, M. et al. Modulation of mucosal immune response, tolerance, and proliferation in mice colonized by the mucin-degrader Akkermansia muciniphila. Front. Microbiol. (2011). doi:10.3389/fmicb.2011.00166

35. Everard, A. et al. Cross-talk between Akkermansia muciniphila and intestinal epithelium controls diet-induced obesity. Proc. Natl. Acad. Sci. U. S. A. 110, 9066–9071 (2013).

36. Luis, A. S. et al. A single sulfatase is required to access colonic mucin by a gut bacterium. Nature 598, 332–337 (2021).

37. Hadis, U. et al. Intestinal Tolerance Requires Gut Homing and Expansion of FoxP3+ Regulatory T Cells in the Lamina Propria. Immunity 34, 237–246 (2011).

38. Knoop, K. A., McDonald, K. G., McCrate, S., McDole, J. R. & Newberry, R. D. Microbial sensing by goblet cells controls immune surveillance of luminal antigens in the colon. Mucosal Immunol. (2015). doi:10.1038/mi.2014.58

39. Pereira, G. V., Boudaud, M. & .... Unravelling specific diet and gut microbial contributions to inflammatory bowel disease. bioRxiv (2022). doi:https://doi.org/10.1101/2022.04.03.486886

40. Johansson, M. E. V., Holmén Larsson, J. M. & Hansson, G. C. The two mucus layers of colon are organized by the MUC2 mucin, whereas the outer layer is a legislator of host-microbial interactions. Proc. Natl. Acad. Sci. U. S. A. 108, 4659–4665 (2011).

41. Aguilera-Lizarraga, J. et al. Local immune response to food antigens drives meal-induced abdominal pain. Nature (2021). doi:10.1038/s41586-020-03118-2

42. Wasielewska, Z., Dolińska, A., Wilczyńska, D., Szaflarska-Popławska, A. & Krogulska, A. Prevalence of allergic diseases in children with inflammatory bowel disease. Postep. Dermatologii i Alergol. 36, 286–294 (2019).

43. Cai, C. et al. Serological investigation of food specific immunoglobulin G antibodies in patients with inflammatory bowel diseases. PLoS One 9, e112154 (2014).

44. Guilarte, M. et al. Diarrhoea-predominant IBS patients show mast cell activation and hyperplasia in the jejunum. Gut (2007). doi:10.1136/gut.2006.100594

45. Tsoi, L. C. et al. Progression of acute-to-chronic atopic dermatitis is associated with quantitative rather than qualitative changes in cytokine responses. J. Allergy Clin. Immunol. 145, 1406–1415 (2020).

46. Gittler, J. K. et al. Progressive activation of TH2/TH22 cytokines and selective epidermal proteins characterizes acute and chronic atopic dermatitis. J. Allergy Clin. Immunol. 130, 1344–1354 (2012).

47. Steimle, A. et al. Constructing a gnotobiotic mouse model with a synthetic human gut microbiome to study host–microbe cross talk. STAR Protoc. (2021). doi:10.1016/j.xpro.2021.100607

48. Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K. & Schloss, P. D. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the miseq illumina sequencing platform. Appl. Environ. Microbiol. (2013). doi:10.1128/AEM.01043-13

49. Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nature Biotechnology (2019). doi:10.1038/s41587-019-0209-9

50. Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids Res. (2013). doi:10.1093/nar/gks1219

51. Oksanen, J. et al. Package ‘vegan’. A Community Ecology Package. Community Ecol. Packag. (2019).

52. Martinez Arbizu, P. pairwiseAdonis: Pairwise multilevel comparison using adonis. R package version 0.4 (2020).

53. McMurdie, P. J. & Holmes, S. Phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. PLoS One (2013). doi:10.1371/journal.pone.0061217

54. Wickham, H. Package `ggplot2`: Elegant Graphics for Data Analysis. Springer-Verlag New York (2016).

55. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. (2014). doi:10.1186/s13059-014-0550-8

56. Nowicka, M. et al. CyTOF workflow: differential discovery in high-throughput high-dimensional cytometry datasets. F1000Research (2019). doi:10.12688/f1000research.11622.3

57. Leonard, C. et al. Comprehensive mapping of immune tolerance yields a regulatory TNF receptor 2 signature in a murine model of successful Fel d 1-specific immunotherapy using high-dose CpG adjuvant. Allergy Eur. J. Allergy Clin. Immunol. (2021). doi:10.1111/all.14716

58. Steimle, A., Grant, E. T. & Desai, M. S. Quantitative assay to detect bacterial glycan-degrading enzyme activities in mouse and human fecal samples. STAR Protoc. (2021). doi:10.1016/j.xpro.2021.100326

Specifc-pathogen-free (SPF) experiments

All SPF animal experiment protocols were approved by the Animal Welfare Service at the Luxembourg Institute of Health, and further approved by the Veterinary Services Administration within the Ministry of Agriculture (national authorization no. LUPA2018/18, LUPA2019/29). Six-week-old, female BALB/c mice were purchased from Charles River Laboratories (Saint Germain Nuelles, France). Mice were housed in groups of five and assigned randomly to a diet and treatment group. The fiber-rich (FR) mice were fed the Standard CRM (P) Rat and Mouse Breeder and Grower Diet (Special Diets Services, catalogue no. 801722). The fiber-free (FF) diet was custom-manufactured by SAFE diets (Augy, France) according to the TD.140343 diet formulation (Envigo, Indianapolis, IN, USA), as previously described¹⁴. Each cage was provided their respective diets ad libitum with autoclaved water. For baseline readouts (before sensitization protocol), mice were fed for 40 days and then sacrificed for tissue/sample processing.

Germ-free and gnotobiotic experiments

All germ-free and gnotobiotic animal experiment protocols were approved by the Animal Experimentation Ethics Committee (AEEC) at the University of Luxembourg (national authorization no. LUPA2019/50). The germ-free status of the mice was verified by culture-based (both aerobic and anaerobic) methods. For all gnotobiotic experiments, 6–8-week-old BALB/c mice were housed in maintained in isocages and were fed the standard diet SAFE R04, irradiated at 40 kGy (SAFE, Augy, France). Mice were colonized with two consecutive daily intragastric gavages of the synthetic microbiota, as described in Steimle et al.⁴⁷. One week after the gavage, fecal samples were collected, DNA was extracted and purified, and the colonization of individual strains in the synthetic microbiota was confirmed using phylotype-specifc qPCR primers⁴⁷. Diet was switched for half of the cages to the FF diet 10 days after colonization.

Allergen sensitization and challenge

The ovalbumin sensitization protocol was adapted from Ganeshan et al.²⁵ and was conducted as per Fig. 3a. All animals started with a feeding period on either the FR diet or the FF diet prior to sensitization. The feeding period lasted for 40 days for SPF mice and at least 7 days for gnotobiotic and germ-free mice. Mice were sensitized by intragastric gavage of 100 μL of the solution related to their treatment group using a reusable, 20G feeding needle (FST, Heildelberg, Germany), once per week for 8 weeks. All solutions were prepared in autoclaved Dulbecco's phosphate-buffered saline (PBS, Catalogue no. LOBE17-515F, Westburg, Leudsen, Netherlands). Treatment groups were as follows: OVA+CTX (100 µg ovalbumin (Catalogue no. A5503, Merck, Overijse, Belgium), 10 µg cholera toxin (Catalogue no. C8052, Merck, Overijse, Belgium), OVA (100 µg ovalbumin), CTX (10 µg cholera toxin) and PBS (PBS alone). One week after the 8^th gavage, the mice were challenged with 5 mg of OVA in 200 ml of PBS. During the one-hour challenge period, two researchers took core body temperature readouts using a rectal probe (BiosebLab, Vitrolles, France), and recorded the clinical symptom score²⁵; the researchers were blinded to the identities of the mouse treatment groups while recording the symptoms scores. Briefly, “0” was assigned if no symptoms were evident; “1” represents mild scratching, rubbing, or both of the nose, head, or feet; “2” and “3” represent intermediate symptoms (e.g. edema around the eyes or mouth, piloerection, and/or labored breathing); “4” represents significantly reduced motility, tremors, and/or significant respiratory distress; and “5” represents death. Mice were sacrificed 24 h after the allergen challenge (end of the experiment) and serum, tissues and fecal samples were collected for analysis.

The peanut sensitization protocol was adapted from Stefka et al.⁷ and was conducted as per Extended Data Fig. 3a. As with the OVA sensitization protocol, all SPF mice started with a 40-day feeding period on either a fiber-rich (FR) diet or a fiber-free (FF) diet. All solutions were made up in filter-sterilized Tris 20mM, pH 7.2. Peanut protein was extracted from defatted peanut flour (Bell Plantation, Tifton, GA, USA)⁷. Mice were sensitized by intragastric gavage of 400 μL of the solution related to their treatment group: PN+CTX (6 mg PN and 10 µg cholera toxin on days 40 and 42, then 6 mg PN and 15 µg cholera toxin), PN (6 mg PN), CTX (10 µg cholera toxin on days 40 and 42, then 15 µg cholera toxin). On day 75, the mice were challenged twice by gavage with 20 mg of PN, thirty minutes apart. During the one-hour challenge period, two researchers took core body temperature readouts using a rectal probe, and recorded clinical symptom score²⁵; the researchers were blinded to the identities of the mouse treatment groups while recording the symptoms scores. Mice were sacrificed 3 h after challenge, and serum and tissues were retreived for analysis.

Sample processing

All animals were euthanized by cervical dislocation. Blood was extracted immediately by cardiac puncture, and collected into 1.5 ml eppendorf tubes. Serum was separated by incubating blood samples at 37°C for 30 min, followed by centrifugation at 3000 rpm for 30 min. Serum was stored in 1.5 ml eppendorf tubes at –20°C for short term storage, or –80°C for long term storage.

Colons were removed and placed in either Methacarn for mucus layer measurements, or collected in HBSS (w/o) (HBSS without Ca²⁺ and Mg²⁺ (Catalogue no. LOBE10-543F, Westburg, Leudsen, Netherlands) supplemented with 10 mM HEPES (Catalogue no. LOBE17-737E, Westburg, Leudsen, Netherlands) for lamina propria (LP) cell isolation. Ileal tissues were collected in HBSS (w/o) for LP cell isolation. Both ilea and colons were processed using the Lamina Propria Dissociation Kit (Catalogue no. 130-097-410, Miltenyi Biotec, Leiden, Netherlands) and gentleMACS™ Dissociator (Miltenyi Biotec, Leiden, Netherlands), according to the manufacturer’s instructions.

Cecal, fecal and ileal contents were collected in 2 ml screw cap tubes, and flash frozen in liquid nitrogen. All contents were stored at -80°C. Cecal, ileal and proximal colonic tissues were cleaned in PBS and stored in 1 ml RNAprotect Tissue Reagent (Catalogue no. 76106, QIAGEN, Venlo, Netherlands) for up to 1 week, followed by long term storage in -80°C.

Mucus layer measurements

After incubation in Methacarnoys fixative¹⁴, colons were stored in methanol until processing. For processing, samples underwent paraffin embedding and thin cross sections were set on glass slides. Slides were stained with Alcian blue or anti-MUC2, and quantification was determined in a blinded fashion, as previously described¹⁴.

For MUC2 staining, the slides were deparafinized in xylene (Catalogue no. 28973.328, VWR International, Leuven, Belgium) 3 times for 5 min at room temperature. The same 3 consecutive 5 min washes were repeated with pure ethanol (Catalogue no. 10437341, Fisher Scientific, Erembodegem-Aalst, Belgium) to dehydrate the sections and then with fresh Milli-Q water to wash the slides. To retrieve the antigens from the tissue sections, Retrievagen A, pH 6.0 (Catalogue no. 550524, BD Biosciences, Erembodegem-Aalst, Belgium) was heated and immediately poured onto the slides and covered to retain the heat as much as possible. After a 10 min incubation, the slides were left to cool in the solution for 20 min. Then the slides were washed in fresh Milli-Q water 3 times for 5 min, incubated in blocking buffer (Tris-buffered saline (TBS), 10% goat serum (Gibco, Catalogue no. 11540526, Fisher Scientific, Erembodegem-Aalst, Belgium) for 1 h at room temperature, and gently blotted dry. For the staining with the primary antibody, the tissue sections on the slides were covered with the Rabbit monoclonal anti-MUC2 antibody (Catalogue no. ab272692, Abcam, Amsterdam, Netherlands) 1/200 in the blocking buffer, in the dark for 2 hours at RT. The excess liquid was blotted away and the slides were washed by dipping them into fresh TBS for 5 min, 3 consecutive times. The tissue sections were covered with the secondary antibody Alexa Fluor-488-conjugated Goat anti-Rabbit IgG (H+L) (10 mg/mL; Catalogue no. 10729174, Fisher Scientific, Erembodegem-Aalst, Belgium) in blocking buffer, in the dark for 1 h at room temperature. The excess liquid was blotted away and the slides were washed by dipping them into TBS, 3 consecutive times for 5 min. The samples were finally mounted by applying 2–3 drops of ProLong Gold Antifade Mountant with DAPI (Invitrogen, Catalogue no. 11549306, Fisher Scientific, Erembodegem-Aalst, Belgium) and a cover slip to the colon section. After incubating for 24 h at room temperature, the slides were sealed using nail varnish and stored in the dark at 4 °C until they were visualized on the Axio Observer Z1 (Zeiss Carl, Zaventem, Belgium). Images were analysed using the MATLAB software BacSpace¹⁵ to determine mucus layer thickness.

Mucus penetrability assay

Penetrability of the colonic mucus was assessed as described by Gustafsson et al²¹ . Briefly, colons were flushed with ice-cold oxygenated KREBS Buffer “Transport” (116 mM NaCl, 1.3 mM CaCl₂, 3.6 mM KCl, 1.4 mM KH₂PO₄, 23 mM NaHCO₃, and 1.2 mM MgSO₄ – Carl Roth) and open 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 0.6µg/ml SYTO9 (Fisher Scientific, 10237582) in oxygenated KREBS “Base” Glucose Buffer (KREBS “Base” Buffer : 136 mM NaCl, 1.5 mM CaCl₂, 4.3 mM KCl, 1.6 mM KH₂PO₄, 27.1 mM NaHCO₃, and 1.4 mM MgSO₄; KREBS Glucose Buffer containing 10mM Glucose, 5.7 mM sodium pyruvate and 5.1 mM sodium-l-glutamate in KREBS “Base” Oxygenated Buffer – Carl Roth), and the apical chamber was filled with oxygenated KREBS Mannitol Buffer (containing 10mM Mannitol, 5.7 mM sodium pyruvate and 5.1 mM sodium-l-glutamate in KREBS “Base” Oxygenated Buffer). After 10 min incubation in the dark at room temperature, the excess of KREBS Mannitol Buffer is removed and 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 with KREBS Mannitol Buffer to remove excess of beads. The chamber was incubated for 10 min in the dark before being visualized with a microscope Axio Examiner KMAT (Carl Zeiss Microscopy GmbH) using the software Zen3.0 (Blue Edition, Carl Zeiss Microscopy GmbH). 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 (Oxford Instruments Imaris). 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 (Extended data Fig. 1b). The same method was used to compute the mucus thickness as the distance between the median position of the beads and the median position of the epithelium (Extended data Fig. 1b).

Intestinal permeability assay

Mice were fasted for 4 h prior to the assay. They were orally gavaged with FITC-dextran (Catalogue no. FD4-250MG, Merck, Overijse, Belgium) solution (100 mg/mL in PBS) so that they received 44 mg FITC-dextran per 100g of body weight. The mice were anaesthetized 4 hours later by intraperitoneal injection of ketamine (12.5mg/100g body weight) and medetomidine (0.025mg/100g body weight) and blood was collected via cardiac puncture in a 1.5 mL tube protected from light. Serum was separated by incubating blood samples at 37°C for 30 min, followed by centrifugation at 3000 rpm for 30 min. 100 ml of serum diluted with an equal volume of PBS was added to a 96-well black/clear flat bottom polystyrene microplate (Catalogue no. 655906, Greiner, Vilvoorde, Belgium). The concentration of FITC in serum was determined by spectrophotofluorometry with an excitation at 485 nm (20 nm band width) and an emission wavelength of 528 nm (20 nm band width), using serially diluted FITC-dextran (0, 125, 250, 500, 1000, 2000, 4000, 6000, 8000 ng/ml) as a standard. Serum from mice not administered with FITC-dextran was used to determine the background.

Bacterial genomic DNA extraction and qPCR analysis

Bacterial DNA was extracted from fecal pellets using phenol-chloroform extraction. Fecal pellets were disrupted in a mix of 500 µl Buffer A (200 mM NaCl, 200 mM Tris, 20 mM EDTA), 210 µl 20% SDS, 500 µl equal parts phenol:chloroform and approximately 250 µl acid-washed glass beads (Catalogue no. G1277, Merck, Overijse, Belgium) per sample. Mechanical lysis was performed on a Mixer Mill MM 400 (Retsch, Fisher Scientific, Erembodegem-Aalst, Belgium) for 5 min at 30 Hz, then samples were centrifuged at 18000 × g for 3 min at 4°C. The aqueous phase was recovered and 500 µl phenol:chloroform was added, mixed by inversion, and then centrifuged at 18000 × g for 3 min at 4°C. The aqueous phase was recovered and 500 µl pure chloroform was added. Samples were mixed by inversion, and then centrifuged at 18000 × g for 3 min at 4°C. The aqueous phase was recovered into a new tube and 1 volume of isopropanol and 1/10 volume 3 M sodium acetate (pH 5.2) were added. Samples were stored at –20°C for 1 h, followed by centrifugation at max speed for 20 min at 4°C. The supernatants were discarded and the DNA pellets were further cleaned with 1 mL 70% ethanol. Pellets were air-dried for approximately 1 h, and then resuspended in 100 µl of ultrapure water (Invitrogen, Catalogue no. 12060346, Fisher Scientific, Erembodegem-Aalst, Belgium). The DNA was purified using QIAGEN DNeasy Blood & Tissue Kit (Catalogue no. 69506, QIAGEN, Venlo, Netherlands) according to the manufacturer’s instructions. Final quantification was determined with a NanoPhotometer® N60 (Implen, Fisher Scientific, Erembodegem-Aalst, Belgium).

20 ng DNA was amplified within 12.5 µl of master mix: buffer (Catalogue no. 10966034, Invitrogen,), 1.5 mM MgCl₂(Catalogue no. 10966034, Invitrogen), 200 µM dNTP (Catalogue no. 10297117, Invitrogen), GelStar Nucleic Acid Gel Stain (Lonza, Catalogue no. LONZ50535, VWR International, Leuven, Belgium), 0.5 U Platinum™ Taq DNA Polymerase (Catalogue no. 10966034, Invitrogen) and 0.2 µM each of forward and reverse primers¹⁴. The qPCR cycle consisted of pre-denaturation at 95°C for 3 min, followed by 40 cycles of 3 sec denaturation at 95°C, 20 sec annealing at 55°C, and 20 sec extension at 68°C. Samples were held at 95°C for 15 sec post-extension and then a melting curve was generated by heating from 60°C to 95°C with 0.3°C interval increases over 15 sec.

16S rRNA gene sequencing and analysis

DNA concentration was measured using a Qubit® dsDNA HS assay kit (Catalogue no. Q32854, Invitrogen) on a Qubit4 fluorometer (Catalogue no. Q33238, Invitrogen). The V4 region of the 16S rRNA gene was amplified using the dual-index primers described by Kozich et al.⁴⁸ and sequenced on an Illumina MiSeq system. Raw sequences were processed using QIIME 2 (version 2020.6)⁴⁹ with DADA2 for sequence quality control. Sequences were aligned and classified using SILVA 138 reference database⁵⁰. PCoA plots were generated in RStudio version 1.3.1093 (R version 4.0.2), using the package ‘vegan’ version 2.5-7⁵¹, with clustering significance testing using the betadisper() function in vegan and pairwise.adonis() package ‘pairwiseAdonis’ version 0.4⁵² to test for heterogeneity of dispersion and difference in centroids, respectively. Relative abundance plots were generated using the package ‘phyloseq’ version 1.34.0⁵³ and ‘ggplot2’ version 3.3.5⁵⁴. The package ‘DESeq2’ version 1.30.1⁵⁵ was used to perform differential abundance analyses.

RNA extraction

Tissues were defrosted on ice in 1 ml of TRIzol Reagent (Catalogue no. 15596026, Life Technologies, Merelbeke, Belgium). A 0.5 mm metal bead (Catalogue no. 69989, QIAGEN, Venlo, Netherlands) was added prior to mechanical lysis on a Mixer Mill MM 400 (Retsch, Fisher Scientific, Erembodegem-Aalst, Belgium) for 8–10 min at 30 Hz. The extraction was performed according to the TRIzol Reagent protocol with a few adaptations. After bead-beating, samples were centrifuged at 12000 × g for 5 min at 4°C. The supernatant was transferred into a fresh 1.5 ml tube and 200 µl of pure chloroform was added. Samples were mixed by shaking for 15 sec and left to incubate for 3 min at room temperature, then centrifuged at 12000 × g for 15 min at 4°C. The aqueous phase was added to 500 µl isopropanol in a new tube, mixed by shaking for 10 sec and left to incubate for 10 min at room temperature. The samples were then centrifuged for 10 min at 12000 × g at 4°C for the precipitation of total RNA. Supernatant was removed and samples were washed with 1 ml cold 75% EtOH. The pellets were left to dry at 37°C for 5–10 min. Pellets were resuspended in 50 µl nuclease-free water and incubated at 56°C for 15 min. A DNase treatment was performed following the Thermo Scientific DNase I, RNase-free (Catalogue no. EN0521, Thermo Scientific™,) protocol, followed by a purification with Qiagen RNeasy Mini Kit (Catalogue no. 74106, QIAGEN, Venlo, Netherlands) according to the manufacturer’s instructions. Final RNA concentrations were quantifed using a NanoPhotometer® N60 (Implen). Reverse transcriptase was performed to obtain cDNA according to Invitrogen Superscript™ IV Reverse Transcriptase protocol (Catalogue no. 18090010, Invitrogen) in combination with RNaseOUT (Catalogue no. 10777019, Invitrogen).

Cytokine gene expression

qPCR reactions were performed using SYBR Green detection method and Platinum™ Taq DNA Polymerase Kit (Invitrogen) on a C1000 Touch Thermal Cycler (Biorad). Prior to qPCR, cDNA samples were diluted 1/5 in 80µl of nuclease-free water. 1 µl of cDNA was amplified with 11,5 µl of master mix containing buffer (Catalogue no. 10966034, Invitrogen), 2.5 mM MgCl₂(Catalogue no. 10966034, Invitrogen,), 400 µM dNTP (Catalogue no. 1029711, Invitrogen), GelStar Nucleic Acid Gel Stain (Lonza, Catalogue no. LONZ50535, VWR International, Leuven, Belgium), 0.5 U Platinum™ Taq DNA Polymerase (Catalogue no. 10966034, Invitrogen) and 0.2 µM each of forward and reverse primers. The qPCR cycle consisted of pre-denaturation at 94°C for 5 min, followed by 40 cycles of 20 sec denaturation at 94°C, 50 sec annealing at 60°C, and 45 sec extension at 72°C. Samples were held at 72°C for 5 min post-extension and then a melting curve was generated by heating from 65°C to 95°C with 0.3°C interval increases over 15 sec. The mRNA cytokine expression cytokines was normalized to expression of GAPDH. The following primers (Kaneka Eurogentec,Ougrée, Belgium) were used: Ifng forward (5’-ATGAACGCTACACACTGCATC-3’), Ifng reverse (5’-CCATCCTTTTGCCAGTTCCTC-3’), Il22 forward (5’-GCAGCCGTACATCGTCAACC-3’), Il22 reverse (5’-TCCCCGATGAGCCGGACA-3’), Il25 forward (5’-ACAGGGACTTGAATCGGGTC-3’), Il25 reverse (5’-TGGTAAAGTGGGACGGAGTTG-3’), Gapdh forward (5’-AATTCAACGGCACAGTCAAGGC-3’), Gapdh reverse (5’-GTGGTTCACACCCATCACAAA-3’).

Mass cytometry time-of-flight (CyTOF)

Single cell suspension of colonic lamina propria cells were stained with a 28-marker panel, as previously described¹⁸ . Briefly, 3 x 10⁶ cells were stained with 5 µM cisplatin solution (Catalogue no. 201064, Fluidigm, Amsterdam Zuidoost, Netherlands) for 5 min. Cells were washed in FACS buffer (PBS; 2% FBS; 2 mM EDTA) and extracellular cell surface staining mix was added for 30 min at room temperature. Cells were washed twice, fixed using the Foxp3 Transcription Factor Staining Buffer Set (eBioscience, catalogue no. 00-5523-00, Life Technologies, Merelbeke, Belgium) for 45 min at 4°C, and washed in the permeabilization buffer. The intracellular cell staining mix was added to the cells for 30 min at room temperature. Samples were washed twice in FACS buffer, resuspended in Cell-ID™ iIntercalator-Ir (Catalogue no. 201192 A, Fluidigm, Amsterdam Zuidoost, Netherlands) in MaxPar fixation solution (Catalogue no. 201192 A, Fluidigm, Amsterdam Zuidoost, Netherlands) and refrigerated overnight, or for up to five days. Prior to acquisition on the Helios™ mass cytometer (Fluidigm, Amsterdam Zuidoost, Netherlands), samples were washed twice in PBS and then twice in deionized water. Samples were resuspended in deionized water at 0.5 x 10⁶ cells per mL, with 10% calibration beads (EQ Four Element Calibration Beads, catalogue no. 201078, Fluidigm, Amsterdam Zuidoost, Netherlands). Normalized FCS files were imported in FlowJo™ (BD Life Sciences). The files were cleaned to exclude calibration beads and doublets, and CD45⁺ cells were exported as new files, which were later imported into RStudio (version 1.0.143, R version 3.4.4) using the R package flowcore (version 1.44.2) for unsupervised analysis, following a pipeline previously described^56,57.

Cell stimulation and flow cytometry

Prior to staining for cytokines, single-cell suspensions from ileal and colonic lamina propria were incubated for 4 h at 37°C, 5% CO₂, with Lymphocyte Activation Cocktail (Catalogue no. 423304, Biolegend Europe BV, Amsterdam, Netherlands) in DMEM (Catalogue no. LOBE12-604F, Westburg, Leudsen, Netherlands) complemented with 10% FBS, 100 U/ml penicillin/streptomycin (Catalogue no. LOBE17-602E, Westburg, Leudsen, Netherlands), glutamine (Catalogue no. LOBE17-605E, Westburg, Leudsen, Netherlands), 10 mM HEPES (Catalogue no. LOBE17-737E, Westburg, Leudsen, Netherlands) and 0.1% β-mercaptoethanol (Catalogue no. M6250, Merck, Overijse, Belgium). For eosinophils and cytokine expression analysis, cells were washed with PBS and incubated with Zombie NIR (Catalogue no. 423105, Biolegend Europe BV, Amsterdam, Netherlands), for 15 min at 4°C, prior to fixation with the Cytofix/Cytoperm solution kit (BD Biosciences, Erembodegem-Aalst, Belgium). Cells were then washed with FACS buffer, incubated with Fc block (1 mg/10⁶ cells, catalogue no. 553142, BD Biosciences, Erembodegem-Aalst, Belgium) for 15 min and stained with the antibody mix (Extended data Table 1) for 30 min at 4°C. Samples were finally washed and resuspended in PBS for acquisition on a NovoCyte Quanteon Flow Cytometer (ACEA Biosciences). FCS files were analyzed in FlowJo™ (BD Biosciences). For each reported population, counts were normalized to the LD^-CD45⁺/Single Cells/Width, SSC-H subset and results are presented as a percentage of CD45⁺ cells.

Lipocalin-2 ELISA

Fecal samples were resuspended in 500 μl ice-cold PBS with 1% Tween 20, followed by agitation at 2000 rpm for 20 min at 4°C on a thermomixer (Eppendorf, Aarschot, Belgium). Samples were then centrifuged for 10 min at 21000 × g at 4°C. Supernatants were stored at –20°C until further analysis. Lipocalin-2 detection was performed using the Mouse Lipocalin-2/NGAL DuoSet Elisa, R&D Systems (Catalogue no. DY1857, Bio-Techne, Abingdon, UK), according to the manufacturer’s instructions.

MCPT1 ELISA

Mouse mast cell protease 1 (MCPT1) was detected in serum samples using the MCPT-1 Mouse Uncoated ELISA Kit (Catalogue no. 88-7503-88, Life Technologies, Merelbeke, Belgium), according to the manufacturer’s instructions, but adapted for 384-well microplates.

Allergen-specific ELISA

Allergen-specific IgE and IgG1 antibodies were quantified from serum samples by sandwich ELISA. For OVA-specific assays, 20 ml of ovalbumin (Catalogue no. A5503, Merck, Overijse, Belgium) at 100 ng/ml in PBS were added to each well of a 384-well microplate (Catalogue no. 781061, Greiner Bio-One, Vilvoorde, Belgium), and incubated overnight at 4°C. Wells were washed four times with 100 μL of wash buffer (1% Tween-20, 154 mM sodium chloride and 10 mM Trisma-base), and blocked with 75 μl of blocking buffer (15 mM Trizma-acetate, 136 mM sodium chloride, 2 mM potassium chloride and 1% (w/v) BSA (bovine serum albumin) for 2 h at room temperature. Undiluted mouse sera were used to determine the OVA-specific IgE antibody concentrations, using the Mouse Anti-Ovalbumin IgE Monoclonal Antibody (Clone E-C1, catalogue no. 7091, Ams Biotechnology, Abingdon, UK) as serially diluted standard (range 0–1000 pg/ml) in dilution buffer (DB; 15 mM Trizma-acetate, 136 mM sodium chloride, 2 mM potassium chloride, 0.1% (w/v) Tween-20, and 1% BSA). To determine the OVA-specific IgG1 antibody concentrations, mouse sera were serially diluted from 1/400 to 1/12800 in DB. The mouse anti-ovalbumin IgG1 monoclonal antibody (Clone L71, catalogue no. 7093, Ams Biotechnology, Abingdon, UK) was used as serially diluted standard (range 0–1000 pg/ml) in DB. After a washing step performed as described above, 20 ml of diluted standards and samples were added to corresponding wells. The plate was then incubated for 90 min at room temperature. The washing step was repeated and 20 μL of detection antibody, either the phosphatase alkaline-conjugated goat anti-mouse IgE (SouthernBiotech, catalogue no. 1110-04, ImTec Diagnostics, Antwerp, Belgium), either the phosphatase alkaline-conjugated goat anti-mouse IgG1 (SouthernBiotech, catalogue no. 1071-04, ImTec Diagnostics, Antwerp, Belgium) diluted 1/500 in DB, were added to each corresponding well. The plate were then incubated for 90 min at room temperature. Following a last wash, 40 μl of substrate (1 x phosphate tablet (Catalogue no. S0942, Merck, Overijse, Belgium) dissolved in 10 mL of substrate buffer (1 mM 2-Amino-2-methyle-1-propanole, 0,1 mM MgCl2×6H2O) were added to each well. After a last incubation step for 60 min at 37°C, the absorbance was measured at 405 nm using an ELISA plate reader (SpectraMax Plus 384 Microplate Reader and SoftMax Pro 7 Software, Molecular Devices, Berks, UK). The OVA-specific IgE or IgG1 antibodies concentrations were determined for each sample using the corresponding formulated standard curve.

For PN specific assays, same protocol was used with adapted antigen. Peanut protein, extracted from defatted peanut flour¹⁰ (Bell Plantation, Tifton, GA, USA), diluted at 12.5 ng/ml in carbonate-bicarbonate buffer (Catalogue no. C3041, Merck, Overijse, Belgium) was used for the coating overnight. Mouse sera were diluted 1/10 in DB to determine PN-specific IgE antibody levels, and 1/100 to determine PN-specific IgG1 antibody levels. Both detection antibodies (goat anti-mouse IgE-phosphatase alkaline conjugated and goat anti-mouse IgG1-phosphatase alkaline conjugated (SouthernBiotech, ImTec Diagnostics, Antwerp, Belgium) were diluted 1/1000 in DB. The absorbance was measured at 405 nm using an ELISA plate reader (SpectraMax Plus 384 Microplate Reader and SoftMax Pro 7 Software, Molecular Devices, Berks, UK).

Total IgE and IgG1 ELISA

Total IgE and IgG1 were quantified from serum samples by sandwich ELISA. The same protocol as for allergen ELISA was used, with adapted reagents. Rat anti-mouse IgE at 60 ng/well or rat anti-mouse IgG1 at 20 ng/well (SouthernBiotech, catalogue no. 1130-01 and 1144-01 respectively, ImTec Diagnostics, Antwerp, Belgium) was diluted in carbonate buffer and used as capture antibody. Mouse sera were serially diluted from 1/10 to 1/320 to determine the total IgE antibodies concentration, and the mouse IgE isotype control (SouthernBiotech, catalogue no. 0114-01, ImTec Diagnostics, Antwerp, Belgium) was used as the serially diluted standard in DB (range 0–250 pg/ml). To determine the total IgG1 antibody concentrations, mouse sera were serially diluted from 1/500 to 1/16000 in DB, and the Mouse IgG1 isotype control (SouthernBiotech, catalogue no. 0102-01, ImTec Diagnostics, Antwerp, Belgium) was used as serially diluted standard in DB (range 0–2000 pg/ml). Each detection antibody, either the phosphatase alkaline-conjugated goat anti-mouse IgE (SouthernBiotech, catalogue no. 1110-04, ImTec Diagnostics, Antwerp, Belgium) or the phosphatase alkaline-conjugated goat anti-mouse IgG1 (SoutherBiotech, catalogue no. 1071-04, ImTec Diagnostics, Antwerp, Belgium), was diluted 1/500 in DB for the assays. The total IgE or IgG1 concentration was calculated the same way as for the OVA-specific IgE/IgG1 concentration.

Flow cytometry for Ig-coated bacteria

Quantification of Ig-coated bacteria was determined from frozen fecal samples or tissue contents by flow cytometry. Samples were homogenized in 500 µl ice-cold PBS by mixing at max speed on a ThermoMixer for 20 min at 4°C. Samples were topped up with 500 µl PBS followed by centrifugation at 100 × g for 5 min at 4°C. The supernatant was passed through a 70 µm strainer (Pluriselect, catalogue no. 43-10070-70, ImTec diagnostics, Antwerp, Belgium) followed by centrifugation at 10000 × g for 5 min at 4°C. The pellet was resuspended with 1 mL PBS for measurement of the optical density at 600 nm and quantification of the bacteria. Approximately 10⁹bacteria were used per staining. Samples were incubated with 500 µl PBS with 5% goat serum (Gibco, catalogue no. 11540526, Fisher Scientific, Erembodegem-Aalst, Belgium) at 4°C for 20 min. After a centrifugation at 10000 × g for 5 min at 4°C, pellets were resuspended in PBS with 5% goat serum and the appropriate antibody: FITC-conjugated anti-mouse IgA (Clone mA-6E1, eBioscience, catalogue no. 11-4204-83, Life Technologies, Merelbeke, Belgium), PE-conjugated anti-mouse IgE (Clone RME-1, catalogue no. 406908, Biolegend Europe BV, Amsterdam, Netherlands), or PE-conjugated IgG1 isotype (Clone RTK2071, catalogue no. 400408, Biolegend Europe BV, Amsterdam, Netherlands). After incubating for 30 min at 4°C, samples were washed in 1 mL PBS. Bacterial pellets were resuspended in 200 µl DNA staining solution (0.9 % NaCl in 0.1 M HEPES, pH 7.2) with 1:4000 SYTO™ 60 Red Fluorescent Nucleic Acid Stain (5 mM Solution in DMSO, cat no. S11342, Invitrogen) and incubated for 20 min at 4°C. Samples were washed in 1 mL PBS and resuspended in 200 µl PBS for aquisition on a NovoCyte Quanteon Flow Cytometer (ACEA Biosciences). FCS files were imported into FlowJo (BD Biosciences) for analysis. IgE-coated bacteria were quantified using a negative gate constructed with the corresponding isotype control sample (Extended Data Fig. 2d).

p-nitrophenyl-based enzyme assays

The enzyme assays were performed as per the previously described protocol⁵⁸.

Statistical analysis

Unless otherwise specifed, all statistical analyses, were performed in GraphPad Prism version 8.0 using either an unpaired t-test, a Mann–Whitney test, or a two-way ANOVA, with P values adjusted using the Benjamini-Hochberg method. For each experiment, n indicates the number of mice per group.

Data availability

The raw fastq files from 16S rRNA gene sequencing have been deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under accession numbers PRJEB53451 (https://www.ebi.ac.uk/ena/browser/view/PRJEB53451) and PRJEB51707 (https://www.ebi.ac.uk/ena/browser/view/PRJEB51707). The mass cytometry datasets for colonic lamina propria have been uploaded to the FlowRepository database under accession number FR-FCM-Z5G2 (https://flowrepository.org/).

Code availability

The analysis pipelines used in this study can be found at https://github.com/DII-LIH-Luxembourg.

Table 1 is available in the supplementary files.

Yes there is potential Competing Interest. Author M.S.D. works as a consultant and an advisory board member at Theralution GmbH, Germany.

Table1Antibodylist.xlsx
Supplemental Table
ExtendedDataFigure1.jpg
Extended Data Fig. 1 a, Alcian blue-stained transverse colonic sections of mice fed a fiber-rich (FR) or a fiber-free (FF) diet for 40 days. Black arrow heads indicate the edges of the mucus layer. b, Cartoon showing how the mucus penetratibility was determined using the images in panel c. c, Representative images of distal colonic tissues showing the 1-mm-sized beads (red dots) laying on top of the epithelium (green dots) using an ex situ ex vivo method. The space between the beads and the epithelium exemplify the unpenetrable mucus layer. d, Fecal lipocalin-2 (LCN-2) measured at the beginning (D1) and at the end (D40) of the 40-day feeding period (n=10). All scatter plots are represented with mean +/- SEM.
ExtendedDataFigure2.jpg
Extended Data Fig. 2 a, Gating strategy for eosinophils, defined as Siglec-F⁺CD11b⁺ singlets. b, Gating strategy for cytokine-expressing CD4⁺ and CD8⁺ T lymphocytes. c, Serum titers of IgG1 and mouse mucosal mast cell protease 1 (MCPT1) (n=2–10, two-way ANOVA, P values adjusted using the Benjamini-Hochberg method). d, Representative plots for IgE-coated bacteria in fiber-rich and fiber-free fed mice. Each sample was stained with an α-IgE antibody (red) or an isotype control (blue).
ExtendedDataFigure3.jpg
Extended Data Fig. 3 Weight curves of the mice over the 40-days feeding period and the sensitization period to OVA-CTX (n=5–10, mixed-effects model with matched values per mice, time effect P<0.0001).
ExtendedDataFigure4.jpg
Extended Data Fig. 4 a, Schematic timeline for peanut protein (PN) sensitization with cholera toxin (CTX) as an adjuvant and control groups. Samples were collected 3 h after the challenge. b, Weight curves of the mice over the 40-days feeding period and the sensitization period to PN-CTX (n=10, mixed-effects model with matched values per mice, time effect P<0.0001). c–g, Serum titers of PN-specific IgE (c, n=9–10), PN-specific IgG1 (d, n=9–10), total IgE (e, n=5–10), total IgG1 (f, n=5–10), and mouse mast cell protease 1 (g, MCPT1, n=10). Two-way ANOVA, P values adjusted using the Benjamini-Hochberg method. Fiber-rich (FR, green), fiber-free (FF, red). All scatter plots are represented with mean +/- SEM.
ExtendedDataFigure5.jpg
Extended Data Fig. 5 Heatmap of the 28 annotated clusters identified by FlowSOM from CyTOF analysis of colonic lamina propria cells of SPF OVA-sensitized mice 24h after challenge. Numbers in the heatmap indicate the relative expression of each marker (x-axis) among cell populations (y-axis). Percentages shown in parantheses represent frequencies of populations among CD45⁺ cells.
ExtendedDataFigure6.jpg
Extended Data Fig. 6 Relative abundances of indicated strains from feces of FR-fed (green) and FF-fed (red) mice, at the end of the experiment (n=3–6, two-way ANOVA, P values adjusted using the Benjamini-Hochberg method). All scatter plots are represented with mean +/- SEM.
ExtendedDataFigure7.jpg
Extended Data Fig. 7 a, Weight curves of the mice over the sensitization period to OVA-CTX (n=3–9, two-way ANOVA with matched values per mice, time effect P<0.0001). b, Uniform manifold approximation and projection (UMAP) of the 28 unique populations identified using FlowSOM among CD45⁺ cells from the colonic lamina propria of OVA-CTX sensitized, 14SM- and 13SM-colonized mice 24h after challenge. Highlighted populations are the ones showing differences between groups. c–f, Frequencies of indicated cell populations among cLP CD45⁺ cells (n=5–10, two-way ANOVA, individual P values adjusted using the Benjamini-Hochberg method). g, Enzymatic activity of select carbohydrate-active enzymes (CAZymes) measured in the feces of OVA-CTX-sensitized, 14SM- and 13SM-colonized mice (n=3–7, two-way ANOVA, P values adjusted using the Benjamini-Hochberg method). Fiber-rich (green), fiber-free (red). All scatter plots are represented with mean +/- SEM.
ExtendedDataFigure8.jpg
Extended Data Fig. 8 Heatmap of the 28 annotated clusters identified by FlowSOM from CyTOF analysis of colonic lamina propria cells of germ-free and gnotobiotic OVA-sensitized mice 24h after challenge. Numbers in the heatmap indicate the relative expression of each marker (x-axis) among cell populations (y-axis). Percentages shown in parantheses represent frequencies of populations among CD45⁺ cells. All scatter plots are represented with mean +/- SEM.

Download PDF

Journal Publication

published 11 Sep, 2023

Read the published version in Nature Microbiology →

Version 1

posted

You are reading this latest preprint version

Akkermansia muciniphila regulates food allergy in a diet-dependent manner

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Microbiota-mediated mucus layer thinning is associated with gut barrier dysfunction

Colonic barrier dysfunction is associated with local type 2 inflammation and anti-commensal immunity

Diet-driven colonic mucosal barrier dysfunction reflects in exacerbated food allergic responses

Allergic sensitization shifts microbiome in a diet-dependent manner

Discussion

Declarations

Acknowledgements

References

Materials And Methods

Table 1

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1