A Novel Approach of T Cell Receptor Classification Reveals Dynamic Interactions Amongst Diet, Microbiota, and Host T Cells

Abstract

The intestinal immune system is adapted to maintain constant interactions with environmental stimuli without causing inflammation. The recognition of Ags derived from microbes and diet can induce Treg or effector T cell responses through dynamic regulatory mechanisms, significantly impacting host health and disease. Although several examples of Ag-specific T cell responses to microbial or dietary Ags have been reported, our understanding of the full range of gut T cell responses remains highly limited. In this review, we highlight recent insights into the complexity of gut TCR responses. Different from traditional approaches, such as TCR transgenic mice and peptide MHC tetramers, our novel approach enables comprehensive analysis of entire repertoire of intestinal TCR responses, revealing both aggregated or individual TCR responses to different classes of Ags, which are regulated by bidirectional interactions between diet and microbiota.

INTRODUCTION

The discovery of gene rearrangement in Ag receptors significantly advanced our understanding of immunology. The immense diversity of Ag specificities in T and B cells enables the immune system to recognize individual Ags within a complex mixture of environmental cues. These distinctive features of adaptive immunity are especially prominent in the gut, where the immune system encounters a vast array of Ags derived from diet and microbiota. At this intricate interface, T cells, serving as key coordinators of the adaptive immune system, play a fundamental role in orchestrating the dynamic interplay among gut-residing immune cells.

Depending on the posture on T cells to the given Ags, mucosal immune networks generate distinct immunological outcomes. In response to pathogenic Ags, CD4 T cells are programmed to secrete proinflammatory cytokines such as IFNγ, IL-4, IL-17A, and IL-22, thereby activating innate immune cells (1,2) and promoting the secretion of antimicrobial peptides by intestinal epithelial cells (3). Conversely, in response to innocuous foreign or self-Ags, CD4 T cells act to suppress immune responses by downregulating the expression of costimulatory molecules on dendritic cells (4) or by secreting anti-inflammatory cytokines, TGFβ and IL-10 (5). Under dysbiosis, however, uncontrolled T cell responses to non-pathogenic microbes or dietary Ags lead to development of inflammatory disorders (6,7,8,9,10), implying that CD4 T cells play a central role in maintaining gut immune homeostasis by differentially responding to the distinct classes of Ags. Thus, investigating the underlying mechanisms of how Ag-specific T cell responses are regulated in a highly complex gut environment needs to be addressed in this field. In this review, we spotlight the intricate interaction between diet, microbiota, and host T cells during homeostasis and gut inflammation, which was reported in our recent work of hierarchical TCR classification (11).

REGULATION OF GUT T CELL RESPONSES BY MICROBIOTA

After birth, metazoans are continuously exposed to gut commensal microbial Ags, which display a prolonging effect on host immunity and diseases. One of the most challenging tasks of the gut immune system is to maintain peaceful interactions with harmless foreign Ags while remaining prepared to provide a defensive response against potential pathogenic microbes. Hence, addressing how Treg vs. effector T (Teff) cell responses are regulated by microbial Ags during homeostasis and inflammation has been one of the most important questions to immunologists for decades. In the absence of commensal microbiota, structural changes in the intestines have been observed (6,7,8,9,10). The intestines of the germ-free (GF) mice exhibit significant alterations of nutrition absorption, gut motility and morphology, like those of mice treated with antibiotics. In addition, Peyer’s patches of GF mice are less detected, and the residual ones are less organized compared to those in specific pathogen-free (SPF) mice (12). The alterations of gut-associated lymphoid tissues are shown to be linked to changes in composition of intestinal T cells, such as Th1, Th2, Th17, T follicular helper cells (Tfh cells), or intraepithelial lymphocytes (IELs). Specificities of such intestinal T cells are likely complicated due to the presence of diverse gut luminal Ags although direct examination has not been performed. By using GF mice, multiple immune cells like T and B cells have been reported to be dependent on microbiota for their activation, differentiation, and functionality (13), which likely imply that microbiota contributes to gut T cell responses via stimulating through both innate immune receptors and TCRs.

As a reductionist approach, gnotobiotic system has been used to address an effect of individual microbial species on immune-modulation, which is done by introducing single species of microbes into mice raised under GF condition (14,15,16). Several microbial species are reported to induce particular T cell responses. First, segmented filamentous bacteria (SFB) are a well-characterized Th17 inducer in the small intestine (15,17). SFB had been associated with IEL development in the past (14). More recent studies demonstrate a unique feature of SFB in their role on immune homeostasis and pathogenesis via SFB-specific Th17 cell development (15,18,19). Due to a substantial effect of Th17 cells on immune homeostasis and pathogenesis (15,19,20), identification of SFB has provided a significant advancement for this field. Second, bacterial species that induce intestinal Treg differentiation have been identified. The colonization of clusters IV, XIVa and XVIII of Clostridium species is shown to enrich TGFβ in the colon resulting in induction of colonic Treg cells (21). Furthermore, cell surface polysaccharide components of Bacteroides fragilis and Bifidobacterium bifidum were shown to promote Treg cell development both under mono-colonization condition and normal conditions that ameliorated experimental colitis (22,23), which collectively suggests a clue on how our immune system tolerates against abundant microbial Ags under normal condition and therapeutic strategies for treating inflammatory disorders. Given these examples, T cell responses in gnotobiotic system seem to be consistent with those in normal condition. However, exceptionally, Akkermansia muciniphila-specific T cell responses are found to be context-dependent (16). Hence, the phenotype of A. muciniphila-specific T cells in mono-colonization condition is only limited to Tfh cells but that is not in normal SPF condition, raising the possibility that T cell responses found under gnotobiotic system may not represent all the T cell responses occurring in normal SPF condition. For this reason, it would be required to verify the results from the gnotobiotic system by performing experiments in the presence of normal complex microbiota. Taken together, it reveals that defined microbes are significant T cell modulators in the intestines but the way they contribute to stimulating intestinal T cells under normal situations needs further investigation.

CONTRIBUTION OF MICROBIAL AND DIETARY Ags TO T CELL RESPONSES IN THE GUT

Environmental Ags delivered through the gut are thought to be the major stimuli of T cells at homeostasis. This is because the recognition of self-Ags in the thymus ensures development of mature T cells exhibiting “lower affinity” to autologous Ags. Given this knowledge, it has been speculated that microbial Ags server as an major immunostimulatory factor at homeostasis (24), which makes the immune system to get activated at a non-pathogenic level and work as an immunological barrier. On the other hand, in the absence of Treg cells, gut Ags drive massive T cell proliferation, resulting in lymphoproliferative disorder accompanied by severe inflammation in multiple organs (25). Regarding the roles of environmental Ags on regulating gut T cell responses, Kim and colleagues (26) examined the effect of diet and microbiota under homeostasis by using Ag-free (AF) mice that are offsprings of GF mice raised with elemental AF diet (27). The paper describes that small intestinal Treg and Teff cells are mainly induced by dietary Ags, whereas colonic T cells are by microbial Ags based on the observation of populational changes seen in SPF vs. GF vs. AF mice. The data in the paper support the notion for preferential presentation of dietary or microbial Ags to the small intestine or the colon, respectively. To extend the observation, neuropillin-1 (Nrp-1) was used and the data show the dominance of Nrp1^lo Treg cells in the small intestines of GF or SPF but not that of AF mice, indicating that small intestinal Treg cells are peripherally induced by dietary Ags (28). Nonetheless, it is possible that Nrp-1 expression may be regulated by secondary effect rather than Ag-specific regulation, which is associated with the controversial debate for using Nrp-1 or Helios as a thymic Treg marker (29).

To study T cell responses to dietary Ags, approaches using model Ags have been also used (30). For example, the OTII TCR Tg mice are widely used as donors for a model Ag-specific T cell system by transferring naïve cells into the mice fed with ovalbumin (OVA) in drinking water. Consistent with populational changes in Treg cells, the transferred OVA-specific CD4 T cells undergo peripheral Treg generation and appear to be enriched in the small intestine (31). In addition to TCR Tg mice, Treg cell responses to dietary Ags have been supported by a study using the MHCII tetramers bound with a dietary peptide (32). In this research, Ag-specific CD4 T cells bound to a gliadin-MHCII tetramer are found to be a mixture of Treg cells and anergic Teff cells that would mediate oral tolerance to dietary Ags. These approaches have provided initial insights into the role of dietary Ags in the induction of intestinal T cell responses.

MACRO- AND MICRO-IMMUNOLOGIC ANALYSIS OF GUT TCRs VIA HIERARCHICAL CLASSIFICATION

However, it is unclear what the contribution of TCR-specific responses, studied via the TCR Tg or peptide MHC (pMHC) tetramer systems, is to the total intestinal T cell pool directed against microbial or dietary Ags. To address questions regarding specificities of intestinal T cells beyond the limited examples of specificities, we have developed the systemic TCR classification framework of CD4 T cells in the natural gut environment (11). Hierarchical TCR classification was conducted by controlling exposure of mice to commensal microbiota or/and diet like the previous report (26). In this system, self-, diet-, or microbe-dependent TCRs were determined by proportional increase of each TCR under the layered addition of foreign Ags. Hence, TCRs found in AF condition were considered as self-dependent TCRs, those induced in GF compared to AF as diet-dependent TCRs, finally those in SPF compared to AF and GF as microbe-dependent TCRs (Fig. 1A), which listed up ≤20,000 total TCRs classified from Treg/Teff repertoire of intestinal T cells. This enables us to examine the aggregated (macro-immunologic response) size of self-, diet-, or microbe-dependent TCR responses concurrently and track individual TCR responses (micro-immunologic response) during homeostasis and inflammation.

Figure 1. Hierarchical TCR classification reveals macro- and micro-immunologic perspectives on gut CD4 T cell responses. (A) Diagram illustrating the strategy for hierarchical TCR classification. Sorted Treg or Teff cell samples were used to generate TCR sequencing libraries. TCRs present in the AF condition were regarded as self-dependent TCRs, those induced in GF compared to AF as diet-dependent TCRs, and those induced in SPF compared to GF/AF as microbe-dependent TCRs. The total number of TCRs from the Treg and Teff repertoires of the small intestine and colon was approximately 20,000. (B) Macro-immunologic analysis of classified TCRs. The summed percentages of individual TCRs in each antigenic class reveal the antigenic forces that drive gut TCR responses. This strategy was applied to analyze TCR responses under homeostasis (C) and gut inflammation (D). (E) The most abundant D1 and M2 Treg TCRs were predominantly found in the Treg repertoire at homeostasis but equally present in both Treg and Teff repertoires under inflammation. (F) Loss of Foxp3 or dysregulated differentiation may contribute to the overlap between Treg and Teff TCRs directed against dietary or microbial Ags.

D1, diet-dependent; M2, microbe-dependent.

Although the 10X single cell RNA sequencing has been widely used for analyzing gene expression and TCR repertoire of fully polyclonal T cells, this strategy would preclude reliable classification of TCR reactivity due to the vast diversity of TCRs and low sequencing throughput of current technologies. Indeed, the estimated TCR misclassification rate of fully polyclonal colonic TCR repertoire by 10× sequencing (33) exceeded 50%, which was calculated by hypothetical binomial distribution model (11). Due to the limited information on identical TCRs between individuals, fixed TCRβ chain model is still valuable for dissecting intestinal TCR responses despite of its reduced diversity, which is restricted to the TCRα chain (34,35,36,37,38,39,40,41). It appears that the diversity of this model is enough to drive T cell responses to complex Ags in a broad range of T cell biology. Consistent with reports on the TCR responses of fully polyclonal conditions (16,17,42), TCRs specific to various gut microbes, such as helicobacter species, Bacteroides vulgatus, SFB, A. muciniphila, Cryptosporidum tyzzeri have been identified in TCliβ mice (11,34,36,37,40,43). In addition, self-reactive TCRs that drive thymic Treg development (35,37,38,39,44,45,46), or 8 different self-reactive TCRs that display differential affinities to myelin oligodendrocyte glycoprotein peptide IAb (41) have been described. Finally, the TCRs specific to rodent dietary components, such as soy or corn, were discovered from intestinal Treg repertoire at homeostasis (11). Therefore, this would support the optimal window of TCR diversity permitting macro- and micro-immunologic analysis of the CD4 TCR responses.

The TCR repertoire analysis with hierarchical classification provides several insightful results (11). First, macro-immunologic analysis, which combines percentages of TCRs in each antigenic class (Fig. 1B), reveals antigenic force that drives intestinal Treg or Teff TCR responses during homeostasis and intestinal inflammation. Under homeostasis, unexpectedly, the contribution of self- or microbe-dependent TCRs is bigger than that of diet-dependent TCRs in both small intestinal and colonic Treg repertoires. Cumulatively, self-, microbe-, or diet-dependent TCRs, which are reproducibly found in SPF, account for 35%, 5%, or 15% in Treg TCR repertoires, respectively. Intriguingly, intestinal Teff TCR response is predominantly directed against microbes (Fig. 1C). Notably, microbe-dependent TCRs in Treg vs. Teff repertoires are differentially induced, confirming little overlap between Treg and Teff TCR repertoires at homeostasis (37). By contrast, it has shown that colitic Teff TCR responses are directed against both dietary and microbial Ags, suggesting that both contribute to pathogenic T cell responses during intestinal inflammation (Fig. 1D). More interestingly, intestinal inflammation is accompanied by substantial individual Treg/Teff TCR overlaps, which is limited to TCR responses to diet and microbiota but not self. For instance, the most abundant diet-specific TCR D1 and microbe-specific TCR M2 are predominantly found in Treg repertoire, but not in Teff repertoire, under homeostasis. However, during the intestinal inflammation, both are equally present in both Treg and Teff repertoire (Fig. 1E). Hence, it reveals that T cell responses to diet and microbiota during inflammation are associated with uncontrolled T cell activation. Although more investigations are needed to test whether diet and microbe TCR overlap is induced by loss of Foxp3 on Treg cells or by dysregulated T cell differentiation from naïve T cells (Fig. 1F), it collectively indicates that there is homeostatic Ag-presentation niche for Treg or Teff cell responses but is altered during intestinal inflammation (34,42).

COMPLEX INTERPLAY BETWEEN DIET AND MICROBIOTA

Despite the consensus on the substantial contribution of dietary or microbial Ags to intestinal immunity, it remains unclear how T cell responses to individual gut Ags are tightly modulated. Complex interplay between diet and microbiota has been addressed for their roles in health and diseases (47,48,49,50). First, intestinal homeostasis could be disrupted by high-fat diet via shifting the composition of microbiota as well as impairing intestinal barrier and modulating immune cells, and as a result, exacerbating inflammatory bowel disease (IBD) (51). High-fat diet is known to reduce Bacteroides, Bifidobacteria, and Lactobacillus, thereby promoting imbalance between Firmicutes and Bacteroidetes (52). Such an altered composition of microbiota has been reported to trigger obesity, metabolic change or gastrointestinal disorders (47,49). Second, dietary fiber and its metabolites derived from microbiota could regulate Treg and Teff responses (47), providing immunological barrier while maintaining tolerance to innocuous foreign Ags via diet-microbe interplay. For example, a significant reduction in Th17 cells and CD8αβ IELs is accompanied by the decrease in SFB in the small intestine when mice are fed with low-fiber diet (49). The defect of Th17 responses is associated with the increased susceptibility to Citrobacter rodentium infection, showing that dietary fiber could display a protective function by supporting growth of Th17 stimulating gut bacteria. Interestingly, however, another study indicates that both low-fiber and high-fiber diets decrease SFB-specific T cell responses. Thus, added sugar in customized diet enhances expansion of Faecalibaculum rodentium, which is responsible for replacement of SFB (47). These findings highlight a possible mechanism of immune pathogenicity of dietary sugar (47,53) and remind us of the requisite for precise interpretation on experimental data using a customized diet. Third, dietary interventions have been regarded as an effective model to ameliorate intestinal inflammation, which includes calorie restriction (54), ketogenic diet (55), fasting mimicking diet (56), and elemental diet (57,58,59,60,61,62). For instance, a ketogenic diet containing high-fat and low-carbohydrate is reported to alter composition of microbiota and inhibit group 3 innate lymphoid cells so that ameliorate colonic inflammation (55). Elemental diet has been also used for dietary intervention to control gut inflammation. It has treated IBD patients to improve their gut integrity and reduce inflammation (58,61), which is studied in a murine colitis model as well (57). Although the underlying mechanisms are unclear, these effects may be partly due to reduced T cell stimulation against either diet or gut bacteria. Our hierarchical TCR classification framework suggests that the elemental amino-acid diet (AAD) modulates Teff TCR responses via a change in microbial compositions under homeostasis. However, interestingly, both dietary and microbial Ag-dependent Teff TCR responses contribute to intestinal inflammation and are altered by a diet switch to AAD, through the depletion of macro-immunologic diet-dependent TCR responses and micro-immunologic changes in microbe-dependent TCR responses (Fig. 2A) (11). The effect of dietary treatment on microbiota appears to be largely driven by the non-antigenic components of the elemental diet. Micro-immunologic regulation was also observed in diet-dependent TCR responses, as illustrated in Fig. 2B. A subset of diet-dependent TCRs was preferentially enriched in either the small intestine or colon. Additionally, microbiota was found to increase or decrease percentages of individual diet-dependent TCRs, possibly suggesting that microbiota affect the efficiency of individual dietary Ag-presentation. Thus, both biogeography and microbiota collectively contribute to micro-immunologic regulation of diet-dependent TCR responses (Fig. 2C). In summary, the regulation of intestinal T cell responses is regulated by complex interplay between microbiota and diet, and this suggests a mechanistic basis of the elemental diet treatment.

Figure 2. Bidirectional interactions between diet and microbiota regulate gut TCR responses. (A) Antigenic changes induced by a diet switch to AAD. Following diet switch to AAD, macro-immunologic depletion of diet-dependent TCR responses and micro-immunologic changes in microbe-dependent responses may contribute to the amelioration of intestinal inflammation. (B) Micro-immunologic regulation of individual diet-dependent TCRs, influenced by biogeography and microbiota. (C) Diagram illustrating the contribution of gut bacteria to dietary Ag presentation, in the context of regulation by biogeography.

Given the complexity and dynamic changes in the gut bacteria, network analysis amongst TCRs and bacterial amplicon sequence variants (ASVs) may be useful for anticipating TCR responses directed to specific gut bacterial Ags. Hence, network analysis of combined data sets with individual TCRs and 16S ASVs could be valuable for examining which gut bacteria are responsible for inducing T cell responses in colitis. In this perspective, bacterial ASVs connected or non-connected to colitic TCRs were analyzed to identify the composition of T cell-stimulating gut bacteria in intestinal inflammation (11). Consistent with this idea, ASVs connected to microbe-dependent Teff TCRs showed a positive correlation with weight loss during colitis, whereas ASV non-connected did not. In this context, Akkermansiaceae and Muribaculaceae appeared to be poorly immunogenic, despite their expansion during colitis (34). In contrast, it is likely that a fraction of Lachnospiraceae, Prevotellaceae, and Erysipelotrichaceae families, of which are connected to microbe-dependent TCRs in network analysis, contribute to Teff TCR responses during intestinal inflammation. This is in line with hypothesis that subsets of these families are involved with human IBD (63,64). Meanwhile, network analysis with diet-dependent TCR-associated gut bacteria would also enhance our understanding of food allergy in the future, in which T cells responses to diet are regulated by gut bacteria. Accordingly, it reveals the possibility that network analysis with hierarchically classified TCRs and bacterial ASVs would provide a comprehensive insight into interactions amongst diet, microbiota, and host T cells, which are critical for identifying gut bacteria that stimulate T cells in colitis or regulate T cell responses to diet in the context of food allergy.

CONCLUDING MARKS

Due to the biologically complex nature of the gut, understanding Ag-specific T cell responses to commensal microbial and dietary Ags has been a significant challenge. Although mono-colonization experiments have been useful to study Ag-specific T cell responses, T cell responses seen in gnotobiotic setting do not fully recapitulate those in the presence of normal microbiota as shown by Akkermansia-reactive T cell responses (16). Consistent with this idea, individual SFB-specific TCRs were found to be differentially induced in SFB-containing SPF vs. SFB-mono-colonization condition indicating that complex microbial interactions might regulate differential SFB epitope presentation (11). In addition, previous studies have focused on a limited number of Ag-specific responses using TCR transgenic mice or pMHC tetramers. We propose that applying the TCR classification framework to normal TCR repertoire in SPF mice provides a comprehensive tool for exploring T cell responses at both macro- and micro-immunological levels. This approach enables the simultaneous quantification and analysis of aggregated and individual TCR responses to distinct classes of Ags under normal conditions. This approach would be suitable to study the dynamic interactions amongst natural dietary components, microbiota and host T cells.

Abbreviations

AAD

amino-acid diet

AF

Ag-free

ASV

amplicon sequence variant

GF

germ-free

IBD

inflammatory bowel disease

IEL

intraepithelial lymphocyte

Nrp-1

neuropillin-1

OVA

ovalbumin

pMHC

peptide MHC

SFB

segmented filamentous bacteria

SI

small intestine

SPF

specific pathogen-free

Teff

effector T

Tfh

T follicular helper