Linking Gut Microbial Epitopes & IBD
1st February 2023
The basis of maintaining a healthy gut is the tripartite communication between the intestinal epithelium, the mucosal immune system, and the intestinal microbiome. It is important to maintain conditions that allow appropriate responses by the mucosal immune system to the myriad of antigens to which it is exposed. Induction of inflammatory and humoral responses is obviously important in the response to pathogens; however, for food- and microbiome-derived antigens (and those derived from dead intestinal epithelial cells, which are continuously shed into the gut lumen), it is more appropriate to show immune tolerance. The intestinal immune system is set up to help promote tolerogenic responses to luminal antigens. Tolerance mechanisms to specific antigens may take several forms e.g., the deletion of reactive T and B cells by apoptosis; induction of anergy in these cells (immunological indifference) or, promoting the differentiation of T and B regulatory cells, which actively suppress inflammation through the production of mediators such as IL-10.
Regulating the access of antigens to the mucosal immune system is an important part of facilitating tolerance, with antigen uptake being directed towards specialised epithelial cells called M cells, which are localised over mucosal lymphoid aggregates. Within the lymphoid aggregate, classical dendritic cells (cDCs) can take up and process the antigens and present them on their cell surface, via MHCII molecules. The cDCs migrate, with naïve T and B cells, to the mesenteric lymph nodes (mLNs). Within the lymph nodes, the T and B cells that possess the antigen-appropriate T and B cell receptors can be fully educated to respond to the antigen in the appropriate manner. An alternative uptake route for luminal antigens, via goblet cells in the intestinal mucosa, may also channel antigens towards processing by juxtaposed dendritic cells and subsequent transport to the mLNs.
It has been recognised for a few years in humans, that there is a broad range of T cell receptor-specificity in the T cell population that responds to microbiome-derived antigens. In recent studies, Xavier, Graham and colleagues have investigated which microbiota-derived antigens have a dominant role in driving immune tolerance in the intestine (Pedersen et al., 2022). In these studies, they identified a wide range of bacterial antigens that could stimulate mucosal immune responses. Previously Xavier, Graham and colleagues (Graham et al., 2018) had developed a computation tool that could predict antigenic epitopes presented by MHCII molecules, based on data from studies on mouse DCs. They called this in silico model, BOTA (Bacteria-Originated T cell Antigen predictor), and used it to accurately predict antigenic epitopes from two model microorganisms, Listeria monocytogenes (a pathogen) and Muribaculum intestinale (a commensal).
The researchers used this programme to interrogate data sets (obtained from other in silico analytical platforms) of peptide sequences predicted to bind human HLA class II molecules (based on typical HLA class II variants expressed by a Caucasian US population). They then took these candidate sequences and used them to search for epitopes from which they could be derived, within a representative human microbiome database. Subsequently, it was determined which epitopes were expressed by human subjects in three cohorts: people with Crohn’s disease (67 subjects), people with Ulcerative Colitis (38 subjects) and people without inflammatory bowel disease (27 subjects). The authors focused their analysis on the epitopes that were expressed by more than 85 % subjects in at least one of the cohorts (571 epitopes in total). The results were very striking, with almost half of these most common epitopes associated with just two bacterial phyla, Bacteroidetes (Bacteroidota, at 41 %) and Firmicutes (Bacillota, at 8.5 %), with a similar combined contribution from unknown bacterial species (49.5 %); Proteobacteria, Verrucomicrobia and Actinobacteria accounted for the remainder of epitopes (1 %). They authors calculated that within a phylum, that over 40 % of epitopes were expressed by more than one genus, and that across phyla more than 4 % of epitopes were common to multiple phyla.
Subsequently, work was performed to define how isolated PBMCs responded to the most common epitopes, using a collection of 48 peptides (15 amino acids long) coding for the identified epitopes (called MAPs, for microbiome-associated peptide). From analysis using hBOTA, the top 22 peptides with the highest-predicted HLA class II recognition were selected. The remaining 26 peptides were selected to reflect the diversity of HLA class II variants involved in antigen presentation and the range of bacterial phyla in the human microbiome (including pathogenic species). PBMCs, isolated from 40 healthy volunteers, were cultured for 16 hours in the presence of each peptide; subsequently, supernatants were collected, and cytokine concentrations were quantified using cytometric bead arrays, by flow cytometry. Specifically, levels of IFNγ, IL-17A and IL-10 were measured, as surrogate markers for Th1, Th17 and Treg cell phenotypic responses. Cytokine production was observed in response to each peptide, in the cells isolated from at least one individual. A total of 34 peptides evoked a response in cells from at least 5 individuals, and there were 10 peptides for which a response was seen in samples from at least 10 individuals. Variation in both the overall response profile (in all subjects) to individual peptides and in a subject’s cytokine response to different peptides was also observed.
In their analysis of the cytokine response, the authors focused on IL-10 and IL-17A, based on the role of these two cytokines in regulating intestinal immune responses (tolerance and the suppression of inflammation in the case of IL-10; chronic inflammation in the case of IL-17). They ranked the peptides according to summed Z-score for each of these cytokines, with skewed weighing applied to those with a moderate or high response. When cross-referenced to their origin, eight of the top-10 ranked peptides were associated with the Bacteroidetes phylum and of these, five showed strong homology to amino acid domains within TonB-dependent receptors (TBDRs), which are a family of bacterial outer membrane transport proteins (these included the top-3 ranked peptides). This result echoed previous work by same researchers, which had identified a TBDR epitope in mice that had two-thirds sequence homology with the top-ranked and fifth-ranked peptide (these were MAP22 and MAP36, respectively). The epitope identified in mice was found within a tethered, plug-domain, which could fit into and block the pore domain (a β-barrel) in SusC (Starch Utilisiation System), a TBDR involved in oligosaccharide transport. Further in silico analysis found nearly 30,000 TBDR-associated epitopes from Bacteroidetes, with more that half of these having sequency homology to MAP22 and MAP36, and half of these epitopes showing association with the intestinal microbiome. Four peptides were synthesised to represent the TBDR plug domain in Bacteroidetes and used to stimulate cultured PBMCs (from three subjects); one peptide (termed TBDRP3) was immunostimulatory and provoked significant IL-10 secretion, relative to a peptide with scrambled sequence.
Using the larger cohort of 40 healthy individuals, the TBDRP3 was found to elicit an anti-inflammatory IL-10 (Treg-like) response in cultured PBMCs from most individuals, although some responded in a pro-inflammatory (Th1-like) manner, with a dominant IFNγ response. The researchers also studied a cohort of patients with inflammatory bowel disease (IBD), with 36 people with Crohn’s disease (11 with active disease) and 32 people with Ulcerative Colitis (9 with active disease). There was a trend for patients with active IBD to show greater cytokine responses to TBDRP3 than individuals in remission; however, these increases did not attain statistical significance, except for the increase in IL-17A in PBMC samples from patients with active Crohn’s disease. There was also a trend for subjects with active Crohn’s disease to show more elevated cytokine levels than those patients with active Ulcerative Colitis. The authors proposed that these data supported the hypothesis that T cell responses to a dominant commensal-derived epitope were skewed towards a Th17 phenotype (and away from a Treg tolerogenic phenotype) in patients with active CD.
In order to examine the response of different T cell populations of the mucosal immune system, intraepithelial and mesenteric lymph node (mLN) CD4 T cells were isolated from mice. The cells were incubated in the presence of tetramers of the mouse MHCII molecule, H2-I-Ab, labelled with either APC or PE and loaded with a 14-amino acid peptide, corresponding to a conserved epitope of a SusC-like protein from Bacteroidales (previously, Graham, Xavier and co-workers had shown this peptide to be immunostimulatory in mice). The CD4 cells were sorted and the cells positive for both APC and PE were collected and single cell sequencing was performed. Application of Uniform Manifold Approximation and Projection to data obtained from differential gene expression analysis demonstrated that cells could be separated into 17 distinct clusters. The SusC peptide-specific intraepithelial lymphocytes (IELs) and mLN cells localised to different clusters and demonstrated different phenotypes, with the IELs predominantly identified as CD4+ cytotoxic T cells. The SusC-specific mLN cells were associated with a T follicular helper (Tfh) cell phenotype; in addition, there were populations of mLN cells demonstrating evidence of clonal expansion, including resting (naïve) T cells, activated T regulatory cells, and a discrete population of immunosuppressive Tfh cells.
The authors’ findings expand our knowledge of which commensal-derived antigens the intestinal mucosal immune system is responding to. They demonstrate that conserved epitopes can show wide-ranging expression across different phyla of bacteria and within a phylum. The microbiome, the efficiency of antigen processing and the differential expression of MCHII alleles are all subject to individual variation (there are currently more than 9000 different HLA II-coding alleles described in humans); this was clearly illustrated by the results, which showed subject-specific immune responses to a matrix of different peptide epitopes. This work supports the hypothesis that epitope dominance is modulated by abundance of the epitope, accessibility of the epitope within the phagosome (of the antigen presenting cell) and the avidity of the interaction between the peptide epitope and MHCII molecules. Using a dominant epitope corresponding to the bacterial outer membrane, oligosaccharide transport protein SusC, the authors show that an immunosuppressive response was observed in immune cells from healthy individuals, but that cells from patients with active Crohn’s disease showed an inflammatory response to the same molecule. Finally, this new work demonstrates that epitope response is conserved across species. Mouse intestinal T cells also respond to a SusC peptide, resulting in the clonal expansion of T cells with SusC-specific T cell receptors, with these T cells demonstrating a range of phenotypes dependent on their site of origin (i.e. IELs and mLNs), as might be expected.
This research raises the possibility of determining which microbiome-derived epitopes play important roles in inflammatory bowel disease (on an individual basis). In addition, it suggests that monitoring changes in an individual’s immune response to specific epitopes may aid the prediction of relapse in IBD and other inflammatory disorders. Finally, the work serves to highlight the usefulness of mouse models for the investigating the functional dynamic between mucosal immunity and the microbiome.