Variation in Microbiome Composition and Faecal Metabolites Are Associated With Differential Susceptibility to DSS‐Induced Colitis
Jessica M. Till, Orion D. Brock, Elyza A. Do, Morgan J. Engelhart, Robert W. P. Glowacki, Shaomin Hu, Ansel Hsiao, Ina Nemet, Philip P. Ahern
TL;DR
Mice from different vendors have different gut microbes and immune responses, making some more likely to develop severe colitis.
Contribution
The study shows that natural microbiome variation influences IBD susceptibility and identifies metabolites linked to disease severity.
Findings
CR mice have higher Th17 levels and faecal IgA compared to JAX mice.
CR mice are more susceptible to DSS-induced colitis than JAX mice.
Co-housing transfers colitis susceptibility from CR to JAX mice.
Abstract
Variation in microbiome composition is linked to differences in intestinal immune phenotypes and can be leveraged to identify microbiome‐driven contributions to phenotypes of interest. Furthermore, such variation has been associated with differing inter‐individual susceptibility to the development of inflammatory bowel disease (IBD), a chronic inflammatory disease of the gastrointestinal tract that is driven by dysfunctional immune‐microbiome interactions. Here, we identified that differences in microbiome composition in C57BL/6 mice from two commonly used commercial vendors, Charles River (CR) and Jackson (JAX) Laboratories, were associated with variation in the intestinal immune phenotype, with CR mice having greater Th17 levels and faecal IgA. In turn, CR mice demonstrated enhanced susceptibility to the dextran sulfate sodium (DSS)‐induced model of colitis compared to JAX mice.…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7| Antibody | Fluorophore | Isotype | Clone | Staining concentration (μg/mL) | Supplier | Catalogue # |
|---|---|---|---|---|---|---|
| Anti‐mouse CD16.2 (FcγRIV) | N/A | Armenian hamster IgG | 9e9 | 5.0 | Biolegend | 149502 |
| Anti‐mouse CD16/32 | N/A | Rat IgG2a, λ | 93 | 5.0 | Biolegend | 101302 |
| Fixable viability dye | Aqua | N/A | N/A | 1 in 300 (following reconstitution as per manufacturer's instructions) | Biolegend | 423102 |
| Anti‐mouse CD4 | APC | Rat IgG2b, κ | RM4‐4 | 0.67 | Biolegend | 116013 |
| Anti‐mouse CD4 | BV605 | Rat IgG2a, κ | RM4‐5 | 0.67 | Biolegend | 100548 |
| Anti‐mouse CD4 | e450 | Rat IgG2a, κ | RM4‐5 | 0.67 | Invitrogen | 48‐0042‐82 |
| Anti‐mouse CD4 | A700 | Rat IgG2a, κ | RM4‐5 | 1.67 | Biolegend | 100536 |
| Anti‐mouse CD4 | PE/Cy7 | Rat IgG2a, κ | RM4‐5 | 0.67 | Biolegend | 100527 |
| Anti‐mouse CD4 | PE | Rat IgG2a, κ | RM4‐5 | 0.67 | Biolegend | 100512 |
| Anti‐mouse TCR‐β chain | A488 | Armenian Hamster IgG | H57‐597 | 1.67 | Biolegend | 109215 |
| Anti‐mouse TCR‐β chain | PerCp/Cy5.5 | Armenian Hamster IgG | H57‐597 | 0.67 | Biolegend | 109227 |
| Anti‐mouse TCR‐β chain | APC | Armenian Hamster IgG | H57‐597 | 0.67 | Biolegend | 109212 |
| Anti‐mouse CD8‐β | PerCp/Cy5.5 | Rat IgG2b, κ | YTS156.7.7 | 0.67 | Biolegend | 126609 |
| Anti‐mouse CD45.2 | A700 | Mouse (SJL) IgG2a, κ | 104 | 1.67 | Biolegend | 109822 |
| Anti‐mouse CD8a | APC | Rat IgG2a, κ | 53–6.7 | 0.67 | Biolegend | 100711 |
| Anti‐mouse CD8a | PE/Cy7 | Rat IgG2a, κ | 53–6.7 | 0.67 | Biolegend | 100722 |
| Anti‐mouse RORγt | APC | Rat IgG1, κ | B2D | 2.0 | Invitrogen | 17‐6981‐82 |
| Anti‐mouse IL‐17A | PE | Rat IgG1, κ | TC11‐18H10.1 | 1.0 | Biolegend | 506903 |
| Anti‐mouse/rat FoxP3 | e450 | Rat IgG2a, κ | FJK‐16s | 1.0 | Invitrogen | 48‐5773‐82 |
- —National Institute of Diabetes and Digestive and Kidney Diseases10.13039/100000062
- —Division of Loan Repayment10.13039/100012893
- —National Heart, Lung, and Blood Institute10.13039/100000050
- —Cleveland Clinic Foundation
- —National Institute of Allergy and Infectious Diseases10.13039/100000060
- —National Institute of General Medical Sciences10.13039/100000057
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGut microbiota and health · Metabolomics and Mass Spectrometry Studies · Inflammatory Bowel Disease
Introduction
1
Inflammatory bowel disease (IBD), encompassing Crohn's disease and ulcerative colitis, is a chronic, debilitating, inflammatory disease of the gastrointestinal (GI) tract estimated to affect ~2.39 million Americans [1]. While the precise aetiology of IBD remains enigmatic, disease pathogenesis arises due to chronic inflammatory responses directed at the resident gut microbiome [2, 3], and microbiome composition has been directly shown to impact IBD susceptibility [4]. Chronic inflammation of the GI tract leads to progressive tissue damage over time, which may lead to scarring and narrowing of the bowel, severe abdominal pain, nutrient malabsorption, fistula formation and colon cancer [5, 6]. Thus, IBD is linked to significant reductions in quality of life for affected individuals. Current IBD therapeutics aim to dampen the aberrant immune responses that mediate pathology, most prominently via neutralising monoclonal antibodies (“biologics”) that target the pro‐inflammatory cytokine responses that lead to tissue inflammation (e.g., anti‐TNF‐α). Despite the efficacy of such approaches, many individuals do not respond to these treatments or become refractory over time [7, 8, 9, 10, 11] and their long‐term use can induce a state of immune suppression and increase individuals' risk for infections and cancer development [12, 13]. While these therapies have revolutionised IBD treatment, these limitations necessitate the need to identify novel therapeutic strategies that limit IBD pathogenesis while sparing protective immunity.
Inter‐individual microbiome variation underlies the pathogenesis of a variety of diseases. Such variation is thought to shape the variable susceptibility to IBD across individuals via the engagement of dysfunctional interactions with host immune or epithelial cells by select microbes. Thus, one promising approach that has the potential to overcome the limitations of current strategies is to target the individual microbial taxa that drive susceptibility to IBD to restore mutualistic host–microbiome interactions. However, the effectiveness of microbiome‐targeted therapies currently remains limited [14, 15, 16, 17]. This is largely due to a lack of defined understanding of how specific microbes modulate disease, making it challenging to design therapeutic strategies that safely and effectively target the microbiome of individuals with IBD. Precision anti‐microbial therapies, where specific microbes or their interactions with the host are targeted to achieve desired therapeutic responses, have proven beneficial in a range of microbe‐driven disorders, including a murine model of IBD [17, 18]. Such approaches have the advantage of avoiding large‐scale disruption of the microbiome but require a detailed understanding of the many host–microbe interactions that promote specific immune responses. While the impact of some specific microbes has been well‐studied [19, 20, 21, 22, 23, 24, 25, 26], the vast majority of resident microbes, and their effects on IBD, remain poorly understood. Furthermore, given the likelihood of large‐scale redundancy within the microbiome, there is an urgent need to uncover the breadth of microbes with inflammatory potential in IBD and the nature of their interactions with the host.
An emerging body of evidence has begun to link changes in the metabolites produced or modified by the resident gut microbiome with IBD [27]. These molecules have the potential to act as therapeutic targets, negating the need to target individual microbes within complex microbiomes. Moreover, they may open new avenues as potential clinical biomarkers, conveying changes in microbial composition or function that may precede or correlate with host inflammatory responses and clinical disease. Therefore, the incorporation of these metabolites into clinical practice may allow for enhanced disease monitoring and proactive changes to therapeutic regimens. The inability to predict long‐term disease severity and the timing of remission periods and relapses remains a major unmet clinical need in the treatment of IBD. Thus, there exists a significant need to identify microbial metabolites, those produced or modified by the microbiome, and to uncover their association with IBD.
To address these gaps in knowledge, we first examined intestinal immune phenotypes and bacterial composition of the faecal microbiomes of groups of C57BL/6 mice from Charles River (CR) and Jackson (JAX) Laboratories. We found that CR mice harboured a distinct microbiome that was associated with elevated colonic Th17 cells and faecal IgA. Next, using the well‐established dextran sulfate sodium (DSS) regulatory T cells‐induced colitis model, we show that CR mice are more susceptible to DSS‐induced colitis than JAX mice and that colitogenic bacteria from CR mice can be transmitted to JAX mice and enhance susceptibility to DSS‐induced colitis during co‐housing. Finally, we uncover select intestinal metabolites that both positively and negatively correlate with colitis severity. Overall, our study identifies microbiome variation in DSS‐induced colitis susceptibility between mice from two common vendors and leverages this variation to identify metabolites whose abundance may have utility in predicting susceptibility to intestinal inflammation.
Results
2
C57BL/6 Mice From Charles River and Jackson Laboratories Harbour Differences in Their Mucosal Immune System and Faecal Microbiome
2.1
Microbiome‐driven immune variation between mice from different vendors or from different housing facilities has previously been effectively leveraged to identify microbes and microbial products that shape immune phenotypes [28, 29, 30, 31, 32, 33]. Specifically, it has recently been demonstrated that C57BL/6 mice from Charles River Laboratories (CR) have an altered mucosal innate immune phenotype relative to C57BL/6 mice from Jackson Laboratories (JAX). It was shown that CR mice possess elevated levels of faecal complement component 3 (C3), which was linked to the presence of specific microbes in the gut [33]. To determine if differences extended to the adaptive immune system, we profiled the colonic CD4+ T cell compartment of mice from CR and JAX (Figure 1A–D, Figure S1A–C). Within the lamina propria, we observed a moderate increase in FoxP3+ regulatory T cells (Tregs) in JAX mice compared to CR mice (Figure S1B,C). Most notably, we observed that CR mice harboured significantly greater colonic T helper (Th) 17 cells (CD4+FoxP3‐RORγt+ and CD4+ FoxP3‐IL‐17A+ T cells) (Figure 1A–D). In addition to colonic T cell populations, we also quantified faecal IgA and observed significantly higher levels in CR than JAX mice (Figure 1E), suggesting greater activation of the adaptive immune system in CR mice. Of note, increased IgA in CR mice was not universal across all batches tested. While six of eight independent batches of CR mice demonstrated significantly higher faecal IgA than JAX mice, two batches of CR mice (indicated by grey data points, Figure 1E) possessed significantly lower faecal IgA levels compared to JAX mice, suggesting that IgA levels vary depending on more factors than simply vendor origin alone.
C57BL/6 mice from Charles River and Jackson Laboratories harbour differences in their mucosal immune system. (A) Frequency of RORγt+ Th17 cells among FoxP3‐CD4+ T cells in the colon of C57BL6NCrl mice (Charles River Laboratories (CR), n = 5) and C57BL6/J mice (Jackson Laboratories (JAX) mice, n = 5). (B) Representative flow cytometry plots showing RORγt staining in live CD4+ T cells in the colon of mice depicted in (A). Isotype control sample is a fluorescence minus one (FMO) control with the appropriate isotype control antibody for RORγt added. (C) Frequency of IL‐17A+IFN‐γ‐ Th17 cells among FoxP3‐CD4+ T cells in the colon of CR (n = 5) and JAX (n = 4) mice following restimulation with PMA and ionomycin. (D) Representative flow cytometry plots showing IL‐17A and IFN‐γ staining in live CD4+ T cells in the colon of mice depicted in (C) following restimulation with PMA and ionomycin. No restimulation control was stained with the same antibody staining mix as the samples, but was not restimulated with PMA and ionomycin. (E) Faecal IgA abundance as assessed by ELISA in CR (n = 60) and JAX (n = 50) mice. Graphs and plots shown are representative of two independent experiments (A–D) or 8 batches of mice (E). Each point represents an individual mouse, and the bar represents the median. Statistical significance was determined using an unpaired Student's t test (A, C, E). Grey data points represent CR IgA‐low batches of mice (E). Flow cytometry plots shown are gated on live CD45.2+CD4+TCRβ+FoxP3‐ cells (B&D).
Given the prominent role of the microbiome in shaping intestinal immune function, we reasoned that differences in microbiome composition may underpin the observed differences in the intestinal immune system between CR and JAX mice. To delineate the differences in the faecal bacterial microbiomes of CR and JAX mice, we performed 16S rRNA sequencing of the V3/V4 region of faecal DNA from CR and JAX mice. Consistent with prior studies [33] we observed several differences between CR and JAX microbiomes (Figure 2A–D). Specifically, JAX mice harboured a greater relative abundance of the genera Lactobacillus and Akkermansia, both of which have been linked to protection against colitis [34, 35, 36, 37, 38]. By contrast, the genus Prevotella was detected in CR mice but not in JAX mice. Prevotella spp. have previously been associated with enhanced host susceptibility to mucosal inflammation [39]. Moreover, we also observed significantly increased relative abundances of the genera Mucispirillum, Bacteroides, Peptococcus, Allobaculum, Parabacteroides, and Clostridium in CR mice compared to JAX (Figure 2D). Notably, we detected the presence of Candidatus Arthromitus in CR mice but not JAX, suggesting the presence of segmented filamentous bacteria (SFB) in the CR microbiome. Given the well‐established role of SFB in promoting the differentiation of non‐pathogenic Th17 cells [28, 40] and IgA in the gut [41, 42, 43], and potential misclassification of SFB as Candidatus Arthromitus [44], we further confirmed the selective presence of SFB in CR faeces in a cohort of mice using SFB‐specific PCR primers [45, 46] (Figure S1D). Thus, the presence of SFB likely explains at least in part the differences observed in mucosal immune phenotypes between CR and JAX mice, although other microbes may also contribute to these differences. While these differences may be attributable to vendor origin and/or genetic differences between the CR and JAX C57BL/6 substrains, they suggest that CR mice harbour a microbiome that is distinct from JAX mice.
*C57BL/6 mice from Charles River and Jackson Laboratories possess distinct faecal microbiome compositions. (A) Stacked bar chart illustrating the relative frequency of the indicated bacterial family as assessed by 16S rRNA gene sequencing (V3/V4 region) of faecal DNA from CR (n = 5 male, n = 5 female) and JAX (n = 3 male, n = 5 female) mice. (B) Heatmap and (C) principal component analysis of weighted UniFrac distance of nonrarefied taxa as assessed by 16S rRNA gene sequencing (V3/V4 region) of faecal DNA in CR male (n = 5), CR female (n = 5), JAX male (n = 3), JAX female (n = 5) mice. (D) Differential relative abundance analysis of select genera in CR (n = 5 male, n = 5 female) and JAX (n = 3 male, n = 5 female) mice. Each stacked bar represents an individual mouse (A). Each column represents an individual mouse (B). Each data point represents an individual mouse. Statistical significance was determined by pairwise PERMANOVA of selected groups, *p ≤ 0.05, **p ≤ 0.01 (C). Each data point represents an individual mouse, and the line represents the median. Statistical significance was determined by Mann–Whitney U test with FDR correction *p ≤ 0.05, *p ≤ 0.01 (D). Group 1 = CR male; Group 2 = CR female; Group 3 = JAX male; Group 4 = JAX female (B, C).
CR Mice Develop More Severe DSS‐Induced Colitis Than JAX Mice
2.2
Next, we sought to determine if the differences in microbiome composition and the adaptive immune compartment were associated with differential susceptibility to intestinal disease in CR and JAX mice. It has previously been demonstrated that DSS‐induced colitis can be driven by pathogenic CD4+ T cells in the context of particular microbiomes [29], and IgA‐targeted microbes have been associated with increased colitogenic potential following DSS treatment [47]. Moreover, it has long been appreciated that there exists significant variability in disease severity in the DSS‐induced mouse model of colitis that is linked to variation in the composition of the microbiome [20]. Thus, due to the differences we observed in the mucosal adaptive immune system and the composition of the gut microbiomes of CR and JAX mice, we hypothesized that CR mice may be more susceptible to developing intestinal inflammation.
To test this hypothesis, cohorts of C57BL/6 mice from JAX and CR were administered DSS in their drinking water (2.5% w/v) for a period of 7 days. After 7 days, DSS was withdrawn and replaced with regular drinking water for a period of 3 or 6 days prior to assessment of colitis severity on Day 10 or Day 13, respectively (Figure 3A). Histological examination revealed that CR mice developed more severe inflammation of the mid and distal colon than JAX mice, and that this was most pronounced on Day 13 (Figure 3B,C). Notably, we observed a high degree of variance in colitis severity at Day 10 in JAX mice, with some mice developing mild disease and others more severe colitis. However, by Day 13, only mild disease was evident, suggesting that JAX mice have a reduced susceptibility to severe colitis and/or an accelerated recovery following withdrawal of DSS. Although the weight loss that is a common hallmark of mice administered DSS does not always correlate with colitis severity [20], we observed that CR mice exhibited more severe weight loss than JAX mice at select timepoints and failed to rebound in weight following the withdrawal of the DSS (Figure 3D). These results are in concurrence with recent findings noting that these two lines of mice demonstrate differential responses to DSS [48]. Similar to observations made at homeostasis, profiling of the intestinal lamina propria revealed that CR mice had increased levels of colonic IL‐17A+ Th17 cells at the experimental endpoint (Day 10) following exposure to DSS, potentially implicating a role for these cells in mediating the enhanced DSS‐induced colitis severity in CR mice (Figure 3E,F). Additionally, JAX mice displayed a trend toward higher percentages of FoxP3+ Tregs (Figure S2A,B). Taken together, these data demonstrate that CR mice are more susceptible to DSS‐induced colitis than JAX mice and harbour elevated Th17 cell populations at both homeostasis and during inflammation. These observations indicate that CR mice may possess factors, immune or non‐immune, which promote more severe inflammation and/or inhibit their ability to resolve intestinal inflammation following DSS administration.
Charles River (CR) mice develop more severe DSS‐induced colitis than Jackson Laboratories (JAX) mice. (A) Experimental outline of the dextran sulfate sodium (DSS)‐induced colitis model. CR and JAX mice were administered DSS in their drinking water (2.5% weight/volume) for 7 days followed by 3–6 days of normal drinking water following which the severity of colitis was assessed histologically. (B) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon of CR and JAX mice at Day 10 (3 days post‐DSS cessation; n = 25 CR, n = 30 JAX) or Day 13 (6 days post‐DSS cessation; n = 8 CR, n = 10 JAX). Scores represent the average of the mid and distal histological scores for each mouse. Two CR mice, from two separate experiments shown for Day 10, were found deceased in the cage prior to the endpoint, and one mouse from a separate independent experiment had to be euthanized early due to pre‐mature meeting of defined endpoint criteria. Two CR mice, from one of the two pooled experiments shown for Day 13, were found deceased in the cage prior to the endpoint. (C) Representative H&E‐stained sections of the colon (40× magnification) of CR and JAX mice from mice in (B). (D) Representative weight loss curve of CR and JAX mice over time, shown as percent of initial body weight measured at Day 1. (E) Frequency of IL‐17A+IFN‐γ‐ Th17 cells among FoxP3‐ CD4+ T cells in the colon of CR (n = 10) and JAX (n = 10) mice following restimulation with PMA and ionomycin as assessed at experiment Day 10 after induction of DSS colitis. (F) Representative flow cytometry plots of IL‐17A and IFN‐γ staining in FoxP3‐ CD4+ T cells following restimulation with PMA and ionomycin. The “no restimulation control” was stained with the same antibody staining mix as samples but was not restimulated with PMA and ionomycin. Flow cytometry plots shown were gated on live CD45.2+CD4+TCRβ+FoxP3‐ cells. Data represents the pooled results from seven independent experiments (B; Day 10 histology) or pooled data from two independent experiments (B; Day 13 histology). Each data point represents an individual mouse, and the bar represents the median. Statistical significance was determined by the Mann–Whitney U test (B). Each data point represents the mean percent of initial weight at that time point for the indicated treatment group, and error bars show the standard deviation of the mean. Statistical significance was determined by mixed‐effects analysis with Šídák's multiple comparisons test comparing the means of each group at each time point (D). Data represents the pooled results from two independent experiments. Each data point represents an individual mouse, and the bar represents the mean. Statistical significance was determined by unpaired Student's t test (E).
Anti‐IL‐17A Blockade Does Not Attenuate DSS‐Induced Colitis in CR Mice
2.3
Given that CR mice harboured more IL‐17A+ Th17 cells and enhanced susceptibility to DSS‐induced colitis relative to JAX mice, we sought to determine if IL‐17A played a role in this differential response. Previous work in both mouse models of colitis and in studies of human IBD has described contradictory roles for IL‐17A‐secreting Th17 cells in disease. In the context of the DSS colitis mouse model, deletion/neutralisation of IL‐17A has been found to have beneficial [49], pathogenic [50, 51], or neutral [52] effects on intestinal inflammation and disease severity. To ascertain the impact of IL‐17A within our model system, we administered 0.5 mg of a neutralising anti‐IL‐17A monoclonal antibody (clone 17F3) that has been shown to limit the activity of IL‐17A [53] or an isotype control antibody (clone MOPC‐21) to CR mice every 3 days (beginning 3 days prior to DSS administration) and examined the impact on DSS‐induced colitis (Figure 4A). Administration of anti‐IL‐17A did not significantly reduce or exacerbate colitis severity in CR mice as assessed by colon histological inflammation or overall weight loss when compared to isotype‐treated controls (Figure 4B,C). Collectively, these data suggest that IL‐17A does not directly influence disease severity in the microbiome contexts examined in the present study. This corroborates the notion that the impact of IL‐17A neutralisation on DSS‐induced colitis is highly context‐specific and suggests that other factors may be drivers of differences observed in DSS‐induced colitis susceptibility in CR and JAX mice.
Anti‐IL‐17A blockade does not attenuate DSS‐induced colitis in CR mice. (A) Experimental outline of the anti‐IL‐17A blockade experiment. Mice were injected intraperitoneally with a neutralising anti‐IL‐17A (n = 11) or isotype control (n = 9) monoclonal antibody every 3 days beginning 3 days prior to the start of DSS administration and continuing throughout the duration of the experiment. NW = normal water. (B) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon CR mice treated with the anti‐IL‐17A or isotype control antibody. Scores represent the average of the mid and distal histological scores for each mouse. One isotype‐treated mouse was euthanized prior to the endpoint due to pre‐mature meeting of defined endpoint criteria and was excluded. (C) Representative weight loss curves of CR mice shown as percent of initial body weight measured at Day 1. Data represent the pooled results of two independent experiments (B) or are representative of two independent experiments (C). Each data point represents an individual mouse, and the bar represents the median. Statistical significance was determined by the Mann–Whitney U test (B). Data points show the mean percent of initial weight for the indicated group, and error bars show the standard deviation of the mean. Statistical significance was determined by two‐way ANOVA with Šídák's multiple comparisons test comparing the means of each group at each time point (C).
Co‐Housing of CR and JAX Mice Promotes More Severe DSS‐Induced Colitis in JAX Mice
2.4
There are many differences between C57BL/6N (CR) and C57BL/6J (JAX) mice that may cause different responses to commonly used disease model systems [48], including genetic differences and differences in microbiome composition. Given the differences we observed in the bacterial composition of the faecal microbiome between CR and JAX mice, and recent evidence that has highlighted the composition of the resident gut microbiome as a central driver of variability in the DSS model system [20], we next sought to elucidate whether the microbiome was a critical factor mediating the differential susceptibility to severe disease in our system. To determine if the differences in DSS‐induced colitis severity could be attributable to the microbiome rather than genetic differences between C57BL/6 substrains, we performed co‐housing experiments to facilitate CR and JAX microbiome exchange through coprophagy (Figure 5A). Female mice were selected for this experiment over male mice due to the ability to readily co‐house non‐littermates as adults. To test the impact of microbiome exchange on susceptibility to DSS‐induced colitis, we co‐housed CR mice and JAX mice to promote interchange of the two microbiomes, generating CR^JAX^ mice (CR mice co‐housed with JAX mice) and JAX^CR^ mice (JAX mice co‐housed with CR mice). To control for the impact of co‐housing rather than exposure to new microbiome members, we co‐housed CR mice with non‐littermate CR mice (CR^CR^) and JAX mice with non‐littermate JAX mice (JAX^JAX^). Mice were co‐housed with their new cage mates for either 2 or 4 weeks prior to administration of DSS in the drinking water.
*Co‐housing of CR and JAX mice promotes more severe DSS‐induced colitis in JAX mice. (A) Experimental design schematic. JAX mice and CR mice were co‐housed for the indicated periods to promote microbiome sharing, generating JAXCR mice (JAX mice co‐housed with CR mice) and CRJAX mice (CR mice co‐housed with JAX mice). As controls, JAX mice were co‐housed with non‐littermate JAX mice (JAXJAX) and CR mice co‐housed with non‐littermate CR mice (CRCR). Mice were administered DSS in their drinking water for 7 days followed by 3 days on regular drinking water, prior to assessment of inflammation. (B, C) Faecal IgA abundance measured by ELISA after two (B) or four (C) weeks of co‐housing, but prior to DSS initiation. CRCR (2 weeks n = 5; 4 weeks n = 10), CRJAX (2 weeks n = 5; 4 weeks n = 10), JAXJAX (2 weeks n = 5; 4 weeks n = 9), JAXCR (2 weeks n = 5; 4 weeks n = 10). (D, E) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon of CR and JAX mice at Day 10 (3 days post‐DSS cessation) after 2 weeks (D) or 4 weeks (E) of co‐housing. CRCR (2 weeks n = 4; 4 weeks n = 9), CRJAX (2 weeks n = 5; 4 weeks n = 10), JAXJAX (2 weeks n = 5; 4 weeks n = 10), JAXCR (2 weeks n = 5; 4 weeks n = 10) for each experiment. Scores represent the average of the mid and distal histological scores for each mouse. One mouse in the CRCR treatment group from the four‐week co‐housing experiment was found deceased in the cage prior to the endpoint. (F, G) Weight loss curves of mice from (D, E) shown as the mean percent of the initial body weight measured at Day 1. Each data point represents an individual mouse, and the bar represents the mean. Statistical significance was determined by one‐way ANOVA with Šídák's multiple comparisons test comparing CRJAX mice to CRCR mice and comparing JAXCR mice to JAXJAX mice (B, C). Each data point represents an individual mouse, and the bar represents the median. Statistical significance was determined using Kruskal–Wallis with Dunn's multiple comparisons test comparing CRJAX mice to CRCR mice and comparing JAXCR mice to JAXJAX mice (D, E). Each data point represents the mean percent initial weight at that timepoint for the indicated treatment group, and error bars show standard deviation of the mean. Statistical significance was determined by two‐way ANOVA (F) or mixed‐effects analysis (G) with Dunnett's multiple comparisons test comparing the mean percent initial weight between JAXJAX and each of the other treatment groups *p ≤ 0.05, *p ≤ 0.01 (F, G). Data represent a single experiment (B, D, F) or pooled data from two independent experiments (C, E) or are representative of two independent experiments (G).
To measure the impact of co‐housing on the intestinal immune phenotype, we quantified faecal IgA in these mice following the 2‐ or 4‐week co‐housing period, but prior to DSS exposure. Previous work has illuminated faecal IgA as a marker for identifying potentially colitogenic microbes in the gut [47, 54, 55] and thus may provide insight into changes occurring in the composition of the microbiome with co‐housing. We observed that JAX^CR^ mice displayed elevated levels of faecal IgA compared to JAX^JAX^ mice after 4 weeks of co‐housing, but not after 2 weeks. No changes in faecal IgA levels were observed between CR^JAX^ or CR^CR^ mice with either 2 or 4 weeks of co‐housing (Figure 5B,C). JAX^CR^ mice demonstrated more severe histological inflammation in the mid and distal colon compared to JAX^JAX^ control mice after 4 weeks of co‐housing, while CR^JAX^ developed equivalent inflammation to their CR^CR^ controls (Figure 5D,E). In addition, after 4 weeks of co‐housing, CR^JAX^ and JAX^CR^ mice displayed weight loss trends over the course of the experiment similar to CR^CR^ mice. In contrast, JAX^JAX^ mice displayed significantly less severe weight loss compared to the other treatment groups, especially CR^CR^ mice, coinciding with their development of less severe colitis (Figure 5F,G). Changes in histological disease and weight loss in JAX^CR^ compared to JAX^JAX^ mice were only noted after 4 weeks of co‐housing, indicating that 2 weeks may be insufficient for colitogenic microbes to successfully transfer or that 2 weeks may be insufficient for the imprinting of responses that enhance susceptibility to colitis following colonisation. While the colitis severity of JAX^CR^ mice was not as severe as CR^CR^ mice, the increase in severity observed in JAX^CR^ mice compared to JAX^JAX^ mice suggests that CR mice harbour a microbiome that contains colitogenic agents, which can colonise JAX mice during co‐housing and increase their susceptibility to DSS‐induced colitis.
To assess the impact of co‐housing CR and JAX mice on the composition of the intestinal microbiome, we performed 16S rRNA sequencing of the V3/V4 region on DNA isolated from faecal pellets collected after 4 weeks of co‐housing but before exposure to DSS (Figure 6A–C). We found that several bacterial genera were transferred from CR to JAX mice during co‐housing, including Prevotella, Mucispirillum, Candidatus Arthromitus, Peptococcaceae, Allobaculum, and Parabacteroides, among others (Figure 6D). The infiltration of these genera into the JAX microbiome was determined by the absence of these genera in JAX^JAX^ but their presence in JAX^CR^ mice after co‐housing. Importantly, many of the genera observed to transfer to JAX^CR^ mice post‐co‐housing were also found to be present in CR but not JAX mice in our original 16S rRNA sequencing of CR and JAX mice at the baseline (Figure 2D). Taken together, these data suggest that these genera may have a direct effect on DSS‐induced colitis severity. Overall, these data implicate members of the CR microbiome as primary drivers of the differences in DSS‐induced colitis severity observed between CR and JAX mice.
*Co‐housing promotes the transfer of bacterial taxa from CR to JAX mice. (A) Stacked bar chart illustrating the relative frequency of the indicated bacterial family as assessed by 16S rRNA gene sequencing (V3/V4 region) of faecal DNA from CRCR, CRJAX, JAXJAX and JAXCR co‐housed mice after 4 weeks of co‐housing. (B) Heatmap and (C) principal component analysis of weighted UniFrac distance of nonrarefied taxa as assessed by 16S rRNA gene sequencing (V3/V4 region) of faecal DNA from CRCR, CRJAX, JAXJAX and JAXCR co‐housed mice after 4 weeks of co‐housing. (D) Differential relative abundance analysis of select genera in CRCR, CRJAX, JAXJAX and JAXCR co‐housed mice after 4 weeks of co‐housing. Each stacked bar represents an individual mouse (A). Each column represents an individual mouse (B). Each data point represents an individual mouse. Statistical significance was determined by pairwise PERMANOVA of selected groups, p ≤ 0.05 (C). Each data point represents an individual mouse, and the line represents the median. Statistical significance was determined by the Mann–Whitney U test with FDR correction (D). Group 1 = CRCR; Group 2 = CRJAX; Group 3 = JAXCR; Group 4 = JAXJAX (B, C). CRCR (n = 5), CRJAX (n = 5), JAXJAX (n = 3), JAXCR (n = 5). Samples acquired 4 weeks post‐co‐housing but prior to DSS exposure (A–D).
Faecal Metabolite Levels Differ Between CR and JAX Mice, but Do Not Directly Impact DSS‐Induced Colitis Severity
2.5
It is increasingly clear that microbiome‐derived metabolites (i.e., metabolites produced by or modified by the microbiome) play an integral role in mediating the myriad impacts of the microbiome on host phenotypes. To determine if the variable susceptibility to DSS‐induced colitis was also associated with differences in the faecal metabolome, we performed targeted faecal metabolite quantitation using liquid chromatography mass spectrometry/mass spectrometry (LC–MS/MS) of JAX and CR mice pre‐DSS exposure and at the experimental endpoint (Day 10 or Day 13) following DSS administration as described above. Of the panel of 40 metabolites assessed (see methods), several metabolites were consistently different between male CR and JAX mice across two independent cohorts of mice (Figure 7A), with an impact of both the vendor origin and DSS administration on metabolite levels in the faeces of these mice. Measurement of metabolites in a female cohort of CR and JAX mice highlighted some differences between males and females, as well as several consistent differences between both sexes of CR and JAX mice (Figure 7B), suggesting a sex‐based effect on some metabolites that warrants further investigation.
Select faecal metabolites may serve as predictive markers of intestinal inflammation severity. (A, B) Faecal metabolites measured by targeted LC–MS/MS in male CR and JAX mice pre‐DSS exposure (n = 16 CR, n = 16 JAX) and at the endpoint (Day 13; n = 9 CR, n = 10 JAX) or female mice pre‐DSS exposure (n = 10 CR, n = 10 JAX) and at the endpoint (Day 10; n = 5 CR, n = 5 JAX) (B). (C–F) Simple linear regression plots of selected metabolite concentrations in faecal samples pre‐DSS (C&E) and at the endpoint (Day 13 (D) or Day 10 (F)) versus the corresponding colitis severity score of each mouse. Correlations shown represent male (C, D) and female (E, F) CR and JAX mice shown in (A, B). Histology scores represent the average of the mid and distal histological scores for each mouse as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon. Each data point represents the indicated faecal metabolite level from an individual mouse, and the bar shows the mean. Statistical significance was determined by one‐way ANOVA with Šídák's multiple comparisons test between CR and JAX mice at each timepoint. Graphs show pooled data from two independent experiments (note, one of these experiments is represented by mice that received an isotype control antibody for the experiments described in Figure 3) (A) or show data from one independent experiment (B). The graph represents pooled data from two independent experiments. Each point represents an individual mouse, and the line represents the line of best fit. Statistical significance and strength of correlation (R 2) was determined by simple linear regression (C–F).
Metabolites such as kynurenic acid, tryptamine, quinic acid, and phenylpropionic acid were all enriched in male CR mice compared to male JAX mice either at the pre‐DSS timepoint or at the endpoint, and similar pre‐DSS patterns were also observed for female CR mice versus female JAX mice pre‐DSS (note, quinic acid was not quantified in the female cohorts). While these differences are not perfectly consistent between males and females, these metabolites are associated with the greater severity of DSS‐induced colitis observed in CR mice. Another metabolite, 4‐hydroxybenzoic acid (4‐HBA), was found to be elevated in male CR mice compared to JAX mice at the endpoint and elevated at the baseline in female CR mice relative to JAX females. This suggests that 4‐HBA may negatively impact DSS‐induced colitis severity. Conversely, metabolites such as serotonin and 4‐hydroxyphenylacetic acid (4‐HPA) were significantly more abundant in the faeces of JAX mice compared to CR at the endpoint in male mice but not female mice. 3‐Hydroxyphenylpropionic acid (3‐HPPA) was elevated in JAX male and female mice relative to CR male and female mice following DSS treatment. While these differences between male and female mice may reflect sex‐based microbiome differences and require further investigation, select metabolites such as 3‐HPPA are associated with reduced DSS severity across sexes, implicating 3‐HPPA in the recovery from or protection against severe DSS‐induced colitis.
Intriguingly, both 3‐HPPA and 4‐HPA were found to be further elevated in response to DSS compared to their baseline levels in male JAX mice, potentially implicating these compounds as protective against severe colitis. 3‐HPPA and 4‐HPA have previously been associated with positive health outcomes in a variety of disease settings [56, 57] and 4‐HPA may be effective at preventing injury‐induced inflammation in both the liver [58] and the lung [59]. Although we did not observe differences in 4‐HPA in our female cohort, its role in other injury‐induced inflammation models and the strong differences observed in male mice prompted its further investigation in our system.
To investigate the role of 4‐HBA, 3‐HPPA and 4‐HPA, the metabolites with the strongest associations with protection from or susceptibility to colitis, we administered them exogenously in the drinking water of mice, followed by assessment of their impact on DSS‐induced colitis severity. As 4‐HBA was elevated in CR mice, and thus potentially inducing more severe colitis, we provided 4‐HBA in the drinking water of JAX mice, which are more resistant to colitis, to determine if it enhanced inflammation. As 3‐HPPA and 4‐HPA were enriched in JAX mice, we posited that these metabolites may play a protective role in DSS‐induced colitis and therefore administered them to cohorts of colitis‐susceptible CR mice to determine their ability to prevent severe disease. Overall, these studies failed to elevate the level of many of the target metabolites in the faeces to levels observed in our initial metabolite screening (Figure 7A), in keeping with the general lack of impact of these metabolites on the course of disease (Figures S3–S5). One potential explanation for this may be that these metabolites are absorbed or degraded in the proximal gut, and potential microbial contributions to the levels of these compounds are likely to concentrate them in areas of higher microbial colonisation. Due to these limitations, these data do not conclusively determine if these metabolites directly modulate DSS‐induced colitis severity.
Select Faecal Metabolites May Serve as Predictive Markers of Intestinal Inflammation Severity
2.6
While the metabolites tested in this study were not observed to impact DSS‐induced colitis severity at the concentrations administered, they may still serve as useful markers to predict disease severity if they correlate with inflammation levels. To determine if any of the quantified metabolites correlated with colitis severity, we performed simple linear regression of the colon histological score (average score of the mid and distal regions) versus the concentration of metabolites measured in the faeces of each mouse pre‐DSS (Figure 7C) and at Day 13 (Figure 7D). This analysis revealed several significant correlations between the concentrations of select metabolites and histological inflammation. The pre‐DSS levels of tryptamine, quinic acid, kynurenic acid and phenylpropionic acid all positively correlated with endpoint colitis score, while 3‐HPPA and 4‐HPA negatively correlated with endpoint disease (Figure 7C). We also observed a positive correlation between pre‐DSS tryptamine levels and disease severity in an independent cohort of female CR and JAX mice (Figure 7E). While no metabolite levels at the endpoint were significantly positively correlated with histological disease, quinic acid demonstrated a trend toward a positive correlation with colitis score. Most strikingly, 3‐HPPA, 4‐HPA and serotonin were significantly negatively correlated with colitis severity following DSS exposure (Figure 7C). Endpoint levels of 3‐HPPA also demonstrated a trend toward a negative association with disease severity in our female cohort, albeit non‐significant, suggesting that even at earlier timepoints in the disease course, higher 3‐HPPA levels may be associated with lower disease severity (Figure 7F). Thus, these metabolites, and particularly 3‐HPPA, show strong associations with protection from severe colitis, suggesting they may be useful predictors of colitis susceptibility and disease activity. Overall, our examination of the faecal metabolomes of CR and JAX mice revealed key differences in the intestinal metabolite landscape of these two groups of mice that correlate with observed differences in DSS‐induced colitis severity.
Discussion
3
The studies presented here describe how variation in microbiome composition between two different commercial vendors of C57BL/6 mice is associated with distinct intestinal immune phenotypes and susceptibility to DSS‐induced colitis. We leverage this variation to identify faecal metabolites whose abundance relates to the severity of colitis. Variation in microbiome composition among different vendors, and even different vivaria within institutions, has proven a useful means to identify microbes that coordinate immune function and disease susceptibility [30, 31, 32, 60]. Our finding that CR mice have elevated IL‐17A+ Th17 cells and faecal IgA levels compared to their JAX counterparts coincides with a recent report, indicating that CR mice have a more activated innate immune system than JAX mice [33]. Furthermore, this was associated with enhanced susceptibility to DSS‐induced colitis in CR mice. Differences in susceptibility to DSS‐induced colitis have been well documented, and this has been attributed primarily to variation in microbiome composition [20]. While our study does not exclude a contribution of genetic differences among the C57BL/6 substrains used here, co‐housing experiments transmitted DSS susceptibility to JAX mice, supporting the notion that the primary driver of this variation is differences in microbiome composition. Microbiome‐driven variation in DSS colitis has proven an obstacle for researchers due to inconsistencies in disease course and severity over time and across labs, limiting the utility of the model for the testing of interventions that promote or reduce colitis severity. The finding that mice from two commonly used mouse vendors show distinct susceptibility may provide an opportunity to leverage these mice as standards for testing interventions that limit colitis (CR mice) or identifying microbes that exacerbate disease (JAX mice).
Despite elevated IL‐17A+ Th17 cells in CR mice, we found that IL‐17A neutralisation had no impact on colitis. The role of IL‐17A in the DSS‐colitis model is nuanced and has been shown to have pathogenic [49], protective [50, 51], and neutral effects [52]. This mirrors human IBD, where neutralisation of IL‐17A has not proved clinically efficacious [61] despite its success in the treatment of other diseases like psoriasis [62]. While the reasons for this variation are not fully understood, it is likely that the impact of IL‐17A is shaped by microbiome composition. Th17 cells are known to have both microbiocidal effects via IL‐17A [28, 60, 63, 64, 65] and immunosuppressive effects via IL‐10 [66] which may lead to divergent outcomes upon IL‐17A neutralisation. Indeed, in the murine T cell transfer colitis model, CD4+ T cell expression of the transcription factor T‐bet is essential for colitis development only in particular microbiome contexts [67]. This reinforces the idea that the immune networks driving IBD, and therefore the efficacy of therapies targeting these responses, vary based on the composition of the microbiome. Increased intestinal Th17 cells and IgA have been linked to SFB colonisation [28, 40, 41, 42, 43]. We found that CR mice harboured SFB, thus likely explaining these observations. Although SFB has been implicated in the pathogenesis of DSS‐induced colitis [68], to our knowledge, this has not been directly tested, likely due to the enormous challenges in the isolation of pure SFB in culture [69]. Thus, its involvement in DSS‐induced colitis remains unclear. Our finding that IL‐17A plays no role in the pathogenesis of DSS‐induced colitis in CR mice further suggests that SFB is likely not the primary cause of differential susceptibility. While colitic microbes are posited to trigger IgA [47, 54, 55], IgA binding itself is likely protective, as luminal IgA deficiency is linked to more severe DSS‐induced colitis [70]. Prevotella has also been implicated in the pathogenesis of DSS‐induced colitis [39] and has been observed as a member of the CR microbiome in our study and others [33]. Further identification of the causative agents using gnotobiotic systems has proven uniquely challenging due to the high susceptibility to systemic disease when a minimal microbiome is present. Euthanasia of mice at a premature timepoint, which is too early to support the development of colitis (data not shown), is often necessary. Therefore, much work remains to identify the microbial agent(s) involved in promoting more severe colitis in CR mice.
Microbial‐derived/modified metabolites play a profound role in modulating host immune and epithelial function [71], and their targeting represents a promising strategy for manipulating the impact of the microbiome on host health [27], yet our understanding of how these factors impact IBD remains in its infancy [72, 73]. Using a targeted metabolomic approach, we identified metabolites whose presence was associated with colitis susceptibility and severity. Among those, the most striking in CR mice was an elevation of 4‐hydroxybenzoic acid (4‐HBA) during colitis in CR mice. In addition, we also observed elevated faecal levels of 3‐hydroxyphenylpropionic acid (3‐HPPA) and 4‐hydroxyphenylacetic acid (4‐HPA) in the faeces of JAX mice both pre‐ and post‐DSS exposure. However, direct testing of the role of these metabolites through exogenous provision in the drinking water did not significantly impact colitis severity, likely due to a failure to elevate their levels to those seen in mice with microbial communities with elevated levels of these metabolites. Moreover, these intervention studies were largely designed to be exploratory and therefore favoured small sample sizes in order to test many potential disease‐modulating compounds that could then be more rigorously tested if biologically interesting results were obtained. Therefore, these studies are significantly underpowered in their ability to determine true impacts on disease. Despite this, we found that pre‐and post‐DSS faecal levels of many of the metabolites screened in this study were associated with disease severity. Most consistently, we observed across three independent cohorts of CR and JAX mice that baseline levels of tryptamine and endpoint levels of 3‐HPPA were positively and negatively associated with endpoint colitis severity, respectively. Thus, these metabolites may have the potential to act as biomarkers of susceptibility to disease and disease progression in IBD. It is possible that these correlations are reliable only in the context of the communities studied here. Therefore, the utility of these metabolites as useful predictors of inflammation susceptibility remains to be rigorously tested across different microbiome types and in humans.
Collectively, our data build upon the current understanding of how microbiome composition drives variation in susceptibility to IBD, and how this variation can be leveraged to uncover metabolites associated with susceptibility or protection from intestinal inflammation.
Materials and Methods
4
Mice
4.1
Conventionally housed mice from Jackson Laboratories (C57BL/6J, strain 00064; mice used were obtained from Room RB16 from the west coast facility) and Charles River Laboratories (C57BL/6NCrl, strain 027; mice used were obtained from room K61) were purchased from the indicated vendors 1–4 weeks prior to use in each experiment and were allowed to acclimatise in a non‐barrier facility for at least 1 week prior to the initiation of experiments. All mice were subject to a 14‐h light 10‐h dark cycle throughout the duration of their housing. Cages contained 1/8‐in. corncob bedding and mice were provided nesting material for enrichment. Mice utilised were 8–12 weeks of age and were age‐matched in each experiment. Male mice were used predominantly due to excessive weight loss observed in female mice following DSS‐colitis. However, female mice were used in several experiments and mice were sex‐matched within each experiment. The sex of mice used is indicated within the figure legends. Mice were fed ad libitum LabDiet, Cat. # 5010—Laboratory Autoclavable Rodent Diet throughout the experiment. Mice were given non‐acidified drinking water ad libitum via water bottles for the duration of their housing. In this facility, the following viral and microbial agents are tolerated: Mouse Norovirus (MNV), Murine Chapparvoviruses (MuCPV/MKPV), Helicobacter spp., Corynebacterium bovis , Rodentibacter spp., Entamoeba muris, and Tritrichomonas spp.
Colon Lamina Propria Cell Isolation
4.2
Whole colons were isolated at necropsy and trimmed of fat, opened longitudinally and washed of their contents in sterile PBS (prepared in‐house from Caisson, Cat. # PBP01) supplemented with 0.1% weight/volume BSA (Sigma, Cat. # A3059; PBS/BSA hereafter). Colons were then divided into 3–5 evenly sized pieces and stored in sterile PBS/BSA in a 50 mL Falcon tube (Falcon, Cat. # 352070) on ice until ready to be processed. To remove the epithelial cell layer, each colon was incubated in ~20 mL PBS supplemented with EDTA (PBS/EDTA, final concentration 5 mM; Corning Cat. # MT‐46034CI) in a 50 mL conical tube, which was placed horizontally at 37°C with orbital shaking at 180 rpm for 20 min. After incubation, tubes were shaken vigorously by hand for ~20 s. After shaking, the supernatant was removed, and the tissue was subject to one additional identical incubation and shaking. Colons were then washed twice by incubating in complete RPMI (RPMI containing 20 mM HEPES (Thomas Scientific Cat. # C748G40), 3% volume/volume FBS (Gibco, Cat. # 10437028)) and 5 mL of penicillin/streptomycin (50 units/mL of each antibiotic; prepared in‐house from Penicillin‐G (Sigma, Cat. # P3032‐100) and streptomycin (Gibco Cat. # 11860‐038)) to remove residual EDTA prior to digestion. The remaining colon tissue was then digested to liberate the lamina propria cells by incubating the tissue in 10 mL of complete RPMI supplemented with 20 μL of collagenase (reconstituted at 50 mg/mL in PBS, Type VIII, Sigma, Cat. # C2139) and 15 μL of dispase (0.075 U/mL; BD Biosciences, Cat. # 354235) per sample. Samples were incubated horizontally with orbital shaking at 180 rpm for 30 min at 37°C. After incubation, samples were vortexed for ~30 s and the media from each digestion was harvested, passed through a 40 μM cell strainer (Fisher Cat. # 7201430) into a new 50 mL collection tube to collect cells digested from the tissue and stored on ice. The remaining tissue was subjected to a second round of digestion using the same method, and the harvested supernatant was pooled with the supernatant from the first digestion. ~10 mL of PBS/EDTA was added to halt the digestion and cells were spun down at 453 × g for 10 min. The supernatant was discarded, and the cell pellets were resuspended by flicking and stored on ice. Sample cell counts were obtained using a haemocytometer using trypan blue (Fisher Cat. # 25‐900‐CI) to distinguish and count live cells.
Flow Cytometry
4.3
Staining for flow cytometric‐based assessment was performed in a 96‐well U‐bottom plate (Falcon, Cat. # 353077). PBS supplemented with 0.1% weight/volume BSA (PBS/BSA) was used to dilute all antibodies for surface staining and used for all washes unless otherwise stated. For surface staining, all centrifugation steps were performed at 453 × g for 5 min at 4°C. For wash steps, a multichannel pipette was utilised to pipet 200 μL of PBS/BSA per well and the cells were re‐suspended by pipetting up and down in the well. Cells were plated, centrifuged, and the supernatant was removed by flicking the solution out of the plate. If staining for cytokines, cells in all samples and fluorescence‐minus‐one (FMO) controls were re‐stimulated by incubating them in 200 μL/well of complete RPMI (5% volume/volume FBS) supplemented with PMA (0.05 μg/mL final concentration, Sigma, Cat. # P1585), Ionomycin (0.5 μg/mL final concentration, Sigma, Cat. # I0634) and Brefeldin A (10 000×, Biolegend Cat. # 420601) for 3 h at 37°C in a humidified incubator supplemented with 5% CO_2_. Single‐colour compensation controls and non‐re‐stimulated control samples were incubated in complete RPMI. Following re‐stimulation, cells were washed with PBS/BSA twice prior to blocking Fc receptors to prevent non‐specific antibody staining in downstream steps. Fc blocking was performed using two antibody clones (93 and 9e9) and live/dead staining was performed using a fixable viability dye diluted in PBS/BSA according to the table below. All samples and (FMO) controls were re‐suspended in 50 μL of Fc block/viability solution per well for 30 min in the dark at 4°C. Compensation controls were re‐suspended in PBS/BSA. Cells were then spun down and each well was re‐suspended in 50 μL of surface staining solution. Surface staining solution contained surface antibodies diluted as shown in the table below in PBS/BSA. Cells were stained for 30 min in the dark at 4°C. Cells were then washed twice in PBS/BSA to remove unbound surface stain. Cells were fixed overnight in the dark at 4°C using the ThermoFisher fixation/permeabilization buffer (ThermoFisher, Cat. # 00‐5123‐43 and 00‐5223‐56) at a 1:3 ratio of concentrate to diluent. Cells were fixed by addition of 100 μL of fixative per well with immediate mixing by pipetting up and down. Following overnight fixation, cells were spun down (note: after fixation, all centrifugation steps occurred at 652 × g for 5 min at 4°C) and washed in PBS/BSA twice. Cells were permeabilized by resuspending them in 50 μL of permeabilization buffer (1:10 dilution of ThermoFisher permeabilization buffer ThermoFisher, Cat. # 00‐8333‐56) in Milli‐Q water with 2% volume/volume normal rat serum (Invitrogen, Cat. #10710C or Sigma, Cat. #R9759) per well and incubating for 30 min in the dark at 4°C. Cells were then spun down and re‐suspended in 50 μL of intracellular staining mix, which contained antibodies targeting intracellular targets diluted according to the table below in permeabilization buffer. Cells were incubated in the intracellular staining mix for 30 min in the dark at 4°C. Unbound intracellular stain was washed off by centrifuging and washing cells twice in 200 μL of permeabilization buffer, with resuspension of cells by pipetting up and down during each wash step. Cells were washed a final time in PBS/BSA, centrifuged, resuspended in 200 μL of PBS/BSA and transferred to flow tubes (Sigma, Cat. #CLS4401) for sample acquisition. All samples were acquired using a LSRII Fortessa (BD Biosciences) and BD FACSDiva software. All data was analysed in FlowJo software (versions 10.10.0) (Table 1).
Faecal IgA Quantification
4.4
Faecal samples were snap‐frozen after collection and stored at −80°C prior to processing. Samples were weighed and re‐suspended in sterile PBS with the protease inhibitor (ThermoScientific, Cat. # A32965) at a concentration of 100 mg of faeces/mL in Eppendorf tubes. Samples were homogenised in the PBS by physical dissociation with a sterile wooden stick, followed by vortexing. Samples were centrifuged at 14000 × g for 10 min at 4°C. Supernatants were transferred to new tubes. Samples were diluted 1:500 and IgA was quantified by ELISA using the Invitrogen Uncoated Mouse IgA ELISA kit (Invitrogen, Cat. # 88‐50 450‐88) following manufacturers' specifications. Sample absorbance was determined at 450 nm and 570 nm using an Agilent Biotek Synergy H1 microplate reader with Gen5 Version 3.16 software.
Faecal Pellet Bacterial DNA Isolation and Quantification
4.5
Faecal pellets were weighed and processed using the Qiagen DNeasy PowerSoil Kit (Qiagen, Cat. # 47014) following the manufacturer's specifications. Briefly, faecal pellets were homogenised using the manufacturer's provided bead tubes in a Qiagen TissueLyser II for 10 min at max speed. The resultant homogenate was processed according to the manufacturer's specifications to isolate total DNA. Final elution of DNA was performed in 50 μL of nuclease‐free water (Invitrogen, Cat. # AM9937). DNA concentration was measured by the Qubit dsDNA Broad Range Assay Kit (Thermofisher, Cat. # Q32853) following the manufacturer's specifications.
16S rRNA Sequencing and Analysis
4.6
16S rRNA sequencing of the V3/V4 gene region was performed by SeqCenter (Pittsburgh, PA). Briefly, samples were prepared using a Quick 16S Kit (Zymo Research) with phased primers, which are specific for the V3/V4 region (341F/806R) of the 16S rRNA gene. Samples were sequenced using a P1 600cyc NextSeq2000 Flowcell to create 2 × 301 bp paired‐end reads. Paired‐end 250 nucleotide reads were demultiplexed and analysed using the QIIME 22024.10 software package as described previously [74].
SFB Detection by PCR and Gel Electrophoresis
4.7
Faecal samples from JAX and CR mice were weighed and processed as outlined above using the Qiagen DNeasy PowerSoil Kit following the manufacturer's specifications. Resultant DNA was quantified by Qubit dsDNA Broad Range Assay Kit following the manufacturer's specifications. DNA isolated from faecal samples was used as the template for 50 μL polymerase chain reactions containing the following reagents: 100 ng of DNA, 1 μL of the forward primer (779F, 10 μM), 1 μL of the reverse primer (1008R, 10 μM) [45, 46], 25 μL of the Dreamtaq Green PCR Master Mix (2×) (ThermoScientific, Cat. #K1081) and 21 μL of the nuclease‐free water (Invitrogen, Cat. #AM9937). Reactions were conducted under the following cycling parameters in a BioRad T100 Thermal Cycler: 95°C (5 min), 30 cycles of 95°C (1 min), 60°C (30 s), 72°C (1 min), final extension at 72°C (10 min). Reactions were run on a 1.5% agarose (Fisher, Cat. # BP160‐500) gel cast with 1× Tris‐Acetate‐EDTA (TAE) and 0.12 μL ethidium bromide (Fisher Scientific, Cat. # BP1302‐10) per mL of agarose gel. Gel was run at 100 V for 1 h and imaged on a BioRad Gel Doc system using Image Lab version 5.2.1. The presence of SFB was confirmed by the visualisation of a 229 bp band on the gel, which corresponded with the predicted PCR product length for the primer set used.
Dextran Sulfate Sodium‐Induced Colitis Model
4.8
A 2.5% weight/volume solution of “Colitis Grade” DSS (MP Biomedical Cat. # 160110) was prepared from a sterile stock container in a laminar flow biological safety cabinet. DSS powder was dissolved in sterile non‐acidified water. DSS water was administered to mice via water bottles for 7 days. After cessation of DSS, mice were provided normal drinking water for up to 6 days, with the exact date of euthanasia based on clinical signs of disease. Unless otherwise stated, all mice in each experiment were euthanized on the same day. Body weight was measured at least every other day throughout the experiment. Any mouse that lost greater than 25% of its original body weight was euthanized.
Histological Disease Scoring
4.9
Colons were divided into even thirds to denote the proximal, mid and distal colon. Tissue sections (~3–4 mm) from the middle of each piece were removed and fixed in 10% neutral buffered formalin (StatLab, Cat. # 28600‐5) for at least 36 h. Tissue sections were transferred to 70% ethanol for longer‐term storage. Sections were paraffin‐embedded, cut into 5 μM sections and mounted on slides. Slides were stained by haematoxylin and eosin (H&E). Slides were scored blindly by a trained pathologist using the scoring criteria previously published [75]. Briefly, sections were scored based on the severity and extent of the inflammation (i.e., the layers of the colon affected) from 0 to 3 and the crypt damage and the percent of the area of the tissue involved in the inflammation from 0 to 4. The total score for each tissue section was calculated as follows: (inflammation severity × extent of inflammation) + (crypt damage × percentage of involved area) to yield a final disease score with a range from 0 to 25.
In Vivo Antibody Neutralisation Studies
4.10
Mice were injected intraperitoneally with 0.5 mg in a volume of 0.2 mL of anti‐mouse IL‐17A (BioXCell, Cat. # BE0173; clone 17F3) or anti‐mouse IgG1 isotype (BioXCell, Cat. # BE0083; clone MOPC‐21) every 3 days throughout the duration of the experiment, beginning 3 days pre‐DSS exposure. All stock antibodies were diluted in sterile PBS and stored at −20°C until administration.
LC–MS/MS Analysis
4.11
All targeted LC–MS/MS analyses were performed on a chromatographic system consisting of two Shimadzu LC‐30 ad pumps (Nexera X2), a CTO 20 AC oven operating at 40°C and a SIL‐30 AC‐MP autosampler in tandem with an 8050 triple quadrupole mass spectrometer (Shimadzu Scientific Instruments Inc). The following ion source parameters were applied: nebulizing gas flow, 3 L/min; heating gas flow, 10 L/min; interface temperature, 300°C; desolvation line temperature, 250°C; heat block temperature, 400°C; and drying gas flow, 10 L/min. For data analysis, Lab Solution Version 5.89 (Shimadzu) was used. GC–MS/MS analysis was performed on a Trace 1310 gas chromatograph (ThermoFisher Scientific) in tandem with a Thermo TSQ‐Evo triple quadrupole.
Selected gut microbial metabolites were quantified by a previously published stable‐isotope‐dilution LC–MS/MS method [76] with some modifications. Metabolites quantified: serotonin, 2‐, 3‐, and 4‐hydroxy hippuric acid, tryptamine, hippuric acid, phenylacetylglutamine (PAGln), 5‐hydroxy‐indole‐3‐acetic acid, phenylacetylglycine (PAGly), indole‐3‐acetylglycine (InacGly), indole‐3‐lactic acid, N‐acetyl‐tryptophan, indole‐3‐acetic acid, trans‐indole‐3‐acrylic acid, indole‐3‐propionic acid, phenylpyruvic acid, phenylacetic acid, phenylpropionic acid, phenyllactic acid, phenylacrylic acid, benzoic acid, 4‐hydroxy‐phenylacetic acid, 4‐hydroxy‐phenyllactic acid, 4‐hydroxy‐phenylpyruvic acid, 4‐hydroxy‐phenylpropionic acid, 4‐hydroxy‐phenylacrylic acid, 2‐, 3‐, and 4‐hydroxy‐benzoic acid, indoxyl sulfate, 3‐hydroxy‐phenylpropionic acid, p‐cresol‐sulfate, trimethylamine‐N‐oxide (TMAO), indole‐3‐carboxaldehyde, phenylsulfate (Phe‐SO_4_), 3,4‐benzoic acid, quinic acid, kynurenic acid and amino‐benzoic acid. Briefly, stable‐isotope‐dilution LC–MS/MS was used for the quantification of metabolites in (20 μL) of the faecal extract. Ice‐cold methanolic solution of internal standards (D_4_‐serotonin; D_5_‐tryptophan; D_5_‐hippuric acid; D_2_‐5‐hydroxy‐indole‐3‐acetic acid; D_5_‐indole‐3‐acetic acid; D_2_‐indole‐3‐propionic acid; D_5_‐phenylacetylglutaimine; D_5_‐phenylacetylglycine, D_5_‐phenylacetic acid; D_4_‐4‐hydroxyphenylacetic acid; D_5_‐benzoic acid; D_9_‐phenylpropionic acid; D_7_‐p‐cresol sulfate; D_4_‐indoxyl sulfate; D_3_‐creatinine, and D_9_‐TMAO) was added to faecal extracts (80 μL), followed by vortexing and centrifuging (21 000 × g; 4°C for 15 min). The clear supernatant was then transferred to glass vials with microinserts. A XSelect HSS T3 column (3.0 × 100 mm; 3.5 μm) (Cat. # 186004780, Waters, Ireland) was used for chromatographic separation. A gradient of solvent A (0.1% acetic acid in water) and B (0.1% acetic acid in acetonitrile) was used for chromatographic separation with a flow rate of 0.4 mL/min and a 1 μL injection volume. Electrospray ionisation in positive and negative ion modes with multiple reaction monitoring (MRM) was performed under the following conditions: m/z 177.2: 160.1 for serotonin; m/z 181.2: 154.1 for D_4_‐serotonin; m/z 195.8: 121, for 2‐, 3‐ and 4‐hydroxyhippuric acid; m/z 161.0: 144.1 for tryptamine; m/z 180.0: 105.1 for hippuric acid; m/z 185.0: 110.1 for D_5_‐hippuric acid; m/z 1920.0: 161.1 for 5‐hydroxy‐indole‐3‐acetic acid; m/z 194.2: 148.1 for D_2_‐5‐hydroxy‐indole‐5‐acetic acid; m/z 206.2: 118.1 for indole‐3‐lactic acid; m/z 247.2: 159.1 for N‐acetyl‐tryptophan; m/z 210.0: 150.1 for D_5_‐tryptophan; m/z 176.0: 130.1 for indole‐3‐acetic acid; m/z 181.2: 134.2 for D_5_‐indole‐3‐acetic acid; m/z 188.0: 115.1 for indole‐3‐acrylic acid; m/z 190.0: 130.1 for indole‐3‐propionic acid; m/z 191.8: 130.1 for D_2_‐indole‐3‐propionic acid; m/z 265.2: 130.1 for PAGln; m/z 271: 130.1 for D_5_‐PAGln; m/z 193.8: 76.1 for PAGly; m/z 198.9: 76.1 for D_5_‐PAGly; m/z 76.0: 59.1 for TMAO; m/z 85.0: 66.2 for D_9_‐TMAO; m/z 113.9:44.2 for creatinine and m/z 117.0:47.2 for D_3_‐creatinine; m/z 146.2:118.1 for indole‐3‐carboxaldehyde; m/z 190.0: 144.1 for kynurenic acid in positive ion mode and m/z 135.0: 91.0 for PAA; m/z 140.2: 95.9 for D_5_‐PAA; m/z 212.0: 79.8 for indoxyl sulfate; m/z 216.0: 79.9 for D_4_‐indoxyl sulfate; m/z 187.0: 106.9 for p‐cresol sulfate; m/z 193.9: 114.1 for D_7_‐p‐cresol sulfate; m/z 150.9: 106.9 for 4‐hydroxyphenylacetic acid; m/z 157.1: 112.9 for D_6_‐4‐hydroxyphenylacetic acid; m/z 181.1: 163.1 for 4‐hydroxyphenyllactic acid; m/z 136.9: 93.1 for 2‐, 3‐ and 4‐hydroxybenzoic acid; m/z 165.1: 121.0 for 3‐ and 4‐hydroxyphenylpropionic acid; m/z 165.5: 146.9 for phenyllactic acid; m/z 162.9: 91.0 for phenylpyruvic acid; m/z 163.1: 119.1 for 4‐hydroxyphenylacrylic acid; m/z 121.2: 77.05 for benzoic acid; m/z 126.5:82.1 for D_5_‐benzoic acid; m/z 149.0: 105.1 for phenylpropionic acid; m/z 158.2: 114.1 for D_9_‐phenypropionic acid; m/z 147.0: 103.1 for phenylacrylic acid and m/z 153.2: 108.9 for 3,4‐dihydroxybenzoic acid; m/z 191.1: 84.9 for quinic acid in negative ion mode.
Gut Microbial Metabolite Intervention Studies
4.12
Sterile water solutions were prepared within a biological safety cabinet using autoclaved water and autoclaved glassware. Three equimolar water solutions were prepared for administration to CR mice: 1: 3‐hydroxyphenylpropionic acid (2.5 g/L) (Fisher, Cat. # AAL0127906) + sodium bicarbonate (1.27 g/L) (Fisher, Cat. # S631‐3) in water, 2: 4‐hydroxyphenylacetic acid (2.3 g/L) (Fisher, Cat. # H029025G) + sodium bicarbonate (1.27 g/L) in water, 3: sodium bicarbonate (1.27 g/L) in water. Solutions were mechanically agitated until completely dissolved in the water. Two equimolar solutions were prepared to administer to JAX mice: 1: sodium 4‐hydroxybenzoate (2.5 g/L) (Sigma, Cat. # H3766) in water and 2: sodium bicarbonate (1.31 g/L) in water. All metabolite water and control water solutions were stored at 4°C in the dark for up to 2 weeks. Water solutions were protected from light exposure by wrapping the storage vessels and cage water bottles in aluminium foil. Metabolite and control water solutions, along with a normal drinking water control for both CR and JAX, were administered to mice continuously for 14 days. One day after administration of the metabolite or control water solutions, DSS was added to the drinking water of the mice at a final concentration of 2.5% weight/volume. DSS was administered for 7 days after which mice were switched back to metabolite or control drinking water without DSS for 6 days or until mice had achieved criteria for the experiment endpoint.
Statistical Analysis and Graph Generation
4.13
Statistical analyses were performed using the built‐in analysis within GraphPad Prism version 10.2.1. The specific statistical tests used are indicated within the figure legends. Experimental schematics were created in BioRender; Engelhart, M. (2025) https://BioRender.com/6s4h9wp; https://BioRender.com/30bu2oe; https://BioRender.com/mkhn1do; https://BioRender.com/8o37dpo.
Author Contributions
J.M.T. conceived of, designed and performed all experiments relating to immune phenotyping, microbiome analysis and assessment of colitis susceptibility and played a primary role in writing the manuscript. O.D.B., M.J.E., R.W.P.G. assisted in experiments and provided intellectual input throughout the course of the studies. S.H. performed a histological assessment of the severity of colitis. I.N. performed and analysed metabolite assessments and provided intellectual input. E.A.D. and A.H. performed 16S V3/V4 analysis and graph generation for the respective figures. P.P.A. conceived of and designed experiments, oversaw data analysis and played a primary role in writing the manuscript.
Funding
This research was supported by: (i) funds provided from the Cleveland Clinic Foundation and R01DK126772 from the National Institute of Diabetes and Digestive and Kidney Diseases issued to the lab of P.P.A., (ii) R01HL160747 from the National Heart, Lung, and Blood Institute, issued to I.N., F31AI179030 from the National Institute of Allergy and Infectious Diseases issued to E.A.D., R35GM158026 from the National Institute of General Medical Sciences and R01AI157106 and R01AI178908 from the National Institute of Allergy and Infectious Diseases issued to A.H., and R01AI181382 from the National Institute of Allergy and Infectious Diseases issued to A.H. and P.P.A. In addition, R.W.P.G. received support from the NIH division of Loan Repayment, grant # LRP0000016021 and LRP0000045724.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: (A) Representative gating strategy for intestinal T cell phenotyping. (B) Frequency of FoxP3+ cells among CD4+ T cells in the colon of CR (n = 15) and JAX mice (n = 13). (C) Representative flow cytometry plots showing FoxP3 staining in live CD4+ T cells in the colon of mice from (B). Isotype control sample is a fluorescence minus one (FMO) control with the appropriate isotype control antibody for FoxP3 added. Flow cytometry plots shown are gated on Live CD45.2 + CD4 + TCRβ+ cells. (D) Gel image of PCR products amplified using SFB‐specific primers. Well 1: 100 base pair ladder; Wells 2–6: male CR faecal DNA; Wells 7–11: male JAX faecal DNA; Well 12: negative control (water) reaction. The graph represents pooled data from three independent experiments (B). Statistical significance was determined by unpaired Student's t test. Each data point represents one individual mouse, and the bar represents the mean (B). Plots are representative of three independent experiments (C). Figure S2: (A) Frequency of FoxP3+ cells among CD4+ T cells in the colon of CR (n = 10) and JAX mice (n = 10) as assessed at experiment Day 10 after induction of DSS colitis. (B) Representative flow cytometry plots showing FoxP3 staining in live CD4+ T cells in the colon of mice from (A). Isotype control sample is a fluorescence minus one (FMO) control with the appropriate isotype control antibody for FoxP3 added. Flow cytometry plots shown are gated on Live CD45.2+CD4+TCRβ+ cells. Statistical significance was determined by unpaired Student's t test. Each data point represents one individual mouse, and the line represents the mean (A). Graph represents pooled data from two independent experiments. Plots are representative of two independent experiments (B). Figure S3: (A) Schematic of host metabolism of administered metabolites outlined in Figure 5. (B, C) Metabolites measured by targeted LC–MS/MS in faecal samples from male CR (B) and JAX mice (C) at the endpoint. CR: NW (n = 5), NaHCO_3_ (n = 3), 3‐HPPA (n = 2), 4‐HPA (n = 5). JAX: NW (n = 5), NaHCO_3_ (n = 5), 4‐HBA (n = 5). (D, E) Metabolites measured by targeted LC–MS/MS in serum samples from male CR (D) and JAX mice (E) at the endpoint. CR: NW (n = 5), NaHCO_3_ (n = 3), 3‐HPPA (n = 3), 4‐HPA (n = 5). JAX: NW (n = 5), NaHCO_3_ (n = 5), 4‐HBA (n = 5). Each data point represents an individual mouse, and the bar represents the mean. Statistical significance was determined by Kruskal–Wallis test with Dunn's multiple comparisons test (B–E) Figure S4: (A) Experimental outline of experiments to test the impact of specific metabolites on DSS‐induced colitis severity in CR and JAX mice. Mice were pre‐treated with normal water, water supplemented with NaHCO_3_ or water supplemented with a given metabolite with or without NaHCO_3_ to promote solubility (see methods for details) for 1 day prior to DSS administration. DSS was administered in the drinking water in combination with the control or treatment waters previously described for 7 days followed by administration of control or metabolite treatment water without DSS for 3–6 days following DSS cessation at which time the severity of inflammation was assessed. Note: CR mice were euthanized on Day 10 prior to the intended endpoint of Day 13 due to the majority of mice pre‐maturely meeting defined endpoint criteria. (B) Weight loss curves of CR mice shown as mean percent initial body weight measured at Day 1. (C) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon of male CR mice at Day 10 (3 days post‐DSS cessation), NW (n = 5), NaHCO_3_ (n = 3), 3‐HPPA (n = 3), 4‐HPA (n = 5). Scores represent the average of the mid and distal histological score for each mouse. One mouse in the NaHCO_3_ group and two mice from the 3‐HPPA group were found deceased in the cage prior to the experiment endpoint at Day 10. One mouse from the NaHCO_3_ group was euthanized prior to the experiment endpoint due to pre‐mature meeting of defined endpoint criteria. (D) Weight loss curves of male JAX mice shown as mean percent of initial body weight measured at Day 1. (E) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon of male JAX mice at Day 13 (6 days post‐DSS cessation), NW (n = 5), NaHCO_3_ (n = 5), 4‐HBA (n = 5). Scores represent the average of the mid and distal histological score for each mouse. Each data point represents the mean percent initial weight at that timepoint for the indicated treatment group; error bars show standard deviation of the mean. Statistical significance was determined by mixed‐effects analysis (B) or two‐way anova (D) with Dunnett's multiple comparisons test at each timepoint comparing the percent initial weight between NaHCO_3_ group and each of the other treatment groups *p ≤ 0.05, **p ≤ 0.01 (B&D). Each data point represents an individual mouse, and the bar represents the median. Statistical significance was determined by Kruskal–Wallis with Dunn's multiple comparisons test comparing NaHCO_3_‐treated mice to each of the other treatment groups (C&E). NW = normal water, NaHCO_3_ = water supplemented with NaHCO_3_ (CR = 1.27 g/L; JAX = 1.31 g/L), 3‐HPPA = water supplemented with 3‐OH‐phenylpropionic acid (2.5 g/L) + NaHCO_3_ (1.27 g/L), 4‐HPA = water supplemented with 4‐OH‐phenylacetic acid (2.3 g/L) + NaHCO_3_ (1.27 g/L), 4‐HBA = water supplemented with sodium benzoate (2.5 g/L). Figure S5: (A) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon of female CR mice treated with NW (n = 4), NaHCO_3_ (n = 4), 3‐HPPA (n = 5), or 4‐HPA (n = 5) at Day 13 (6 days post‐DSS cessation). Scores represent the average of the mid and distal histological score for each mouse. One mouse from the NW group and one mouse from the NaHCO_3_ group were found deceased in the cage prior to the experiment endpoint. (B) Weight loss curves of female CR mice shown in (A) as mean percent of initial body weight measured at Day 1. (C) Metabolites measured by targeted LC–MS/MS in serum samples collected from female CR mice treated with NW (n = 4), NaHCO_3_ (n = 4), 3‐HPPA (n = 5), or 4‐HPA (n = 5) shown in (A) at the endpoint. (D) Colonic inflammation severity as assessed by blinded histological scoring of H&E‐stained sections of the mid and distal colon of female JAX mice treated with NW (n = 5), NaHCO_3_ (n = 5), or 4‐HBA (n = 5) at Day 13 (6 days post‐DSS cessation). Scores represent the average of the mid and distal histological score for each mouse. (E) Weight loss curves of female JAX mice shown in (D) as mean percent of initial body weight measured at Day 1. (F) Metabolites measured by targeted LC–MS/MS in serum samples collected from female JAX mice treated with NW (n = 5), NaHCO_3_ (n = 5), or 4‐HBA (n = 5) shown in (D) at the endpoint. (G) Metabolites measured by targeted LC–MS/MS in faecal samples collected from female CR mice treated with NW (n = 4), NaHCO_3_ (n = 4), 3‐HPPA (n = 5), or 4‐HPA (n = 5) shown in (A) at the endpoint. (H) Metabolites measured by targeted LC–MS/MS in faecal samples collected from female JAX mice treated with NW (n = 5), NaHCO_3_ (n = 5), or 4‐HBA (n = 5) shown in (D) at the endpoint. Each data point represents an individual mouse, and the bar represents the median. Statistical significance was determined by Kruskal Wallis with Dunn's multiple comparisons (A&D). Each data point represents the mean percent initial weight at that timepoint for the indicated treatment group, and error bars show standard deviation of the mean. Statistical significance was determined by mixed‐effects analysis (B) or two‐way ANOVA (E) with Dunnett's multiple comparisons test comparing the mean percent initial weight of the NaHCO_3_ group to each of the other treatment groups ^ns^ P ≥ 0.05, *p ≤ 0.05, **p ≤ 0.01 (B&E). Each data point represents an individual mouse, and the bar represents the mean. Statistical significance was determined by Kruskal Wallis with Dunn's multiple comparisons test (C, F–H). NW = normal water, NaHCO_3_ = water supplemented with NaHCO_3_ (CR = 1.27 g/L; JAX = 1.31 g/L), 3‐HPPA = water supplemented with 3‐OH‐phenylpropionic acid (2.5 g/L) + NaHCO_3_ (1.27 g/L), 4‐HPA = water supplemented with 4‐OH‐phenylacetic acid (2.3 g/L) + NaHCO_3_ (1.27 g/L), 4‐HBA = water supplemented with sodium benzoate (2.5 g/L).
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1J. D. Lewis , L. E. Parlett , M. L. Jonsson Funk , et al., “Incidence, Prevalence, and Racial and Ethnic Distribution of Inflammatory Bowel Disease in the United States,” Gastroenterology 165, no. 5 (2023): 1197–1205.e 2, 10.1053/j.gastro.2023.07.003.37481117 PMC 10592313 · doi ↗ · pubmed ↗
- 2Y. Chen , W. Cui , X. Li , and H. Yang , “Interaction Between Commensal Bacteria, Immune Response and the Intestinal Barrier in Inflammatory Bowel Disease,” Frontiers in Immunology 12 (2021): 761981, 10.3389/fimmu.2021.761981.34858414 PMC 8632219 · doi ↗ · pubmed ↗
- 3R. J. Xavier and D. K. Podolsky , “Unravelling the Pathogenesis of Inflammatory Bowel Disease,” Nature 448, no. 7152 (2007): 427–434, 10.1038/nature 06005.17653185 · doi ↗ · pubmed ↗
- 4G. J. Britton , E. J. Contijoch , I. Mogno , et al., “Microbiotas From Humans With Inflammatory Bowel Disease Alter the Balance of Gut Th 17 and RO Rgammat(+) Regulatory T Cells and Exacerbate Colitis in Mice,” Immunity 50, no. 1 (2019): 212–224.e 4, 10.1016/j.immuni.2018.12.015.30650377 PMC 6512335 · doi ↗ · pubmed ↗
- 5T. Jess , M. Gamborg , P. Munkholm , and T. I. Sørensen , “Overall and Cause‐Specific Mortality in Ulcerative Colitis: Meta‐Analysis of Population‐Based Inception Cohort Studies,” American Journal of Gastroenterology 102, no. 3 (2007): 609–617, 10.1111/j.1572-0241.2006.01000.x.17156150 · doi ↗ · pubmed ↗
- 6F. Marrero , M. A. Qadeer , and B. A. Lashner , “Severe Complications of Inflammatory Bowel Disease,” Medical Clinics of North America 92, no. 3 (2008): 671–686, 10.1016/j.mcna.2007.12.002.18387381 · doi ↗ · pubmed ↗
- 7S. Ben‐Horin , U. Kopylov , and Y. Chowers , “Optimizing Anti‐TNF Treatments in Inflammatory Bowel Disease,” Autoimmunity Reviews 13, no. 1 (2014): 24–30, 10.1016/j.autrev.2013.06.002.23792214 · doi ↗ · pubmed ↗
- 8V. Billioud , W. J. Sandborn , and L. Peyrin‐Biroulet , “Loss of Response and Need for Adalimumab Dose Intensification in Crohn's Disease: A Systematic Review,” American Journal of Gastroenterology 106, no. 4 (2011): 674–684, 10.1038/ajg.2011.60.21407178 · doi ↗ · pubmed ↗