Introduction – Unconventional T cells and their functions

Several populations of unconventional T cells have been identified in mammals that possess both innate and adaptive characteristics, including γδT cells (not discussed here), invariant natural killer T (iNKT) cells and mucosa-associated invariant T (MAIT) cells.1 Unlike conventional T cells, which typically respond to peptide antigens presented in the context of major histocompatibility complex (MHC) molecules, iNKT and MAIT cells are specialized in the recognition of glycolipid antigens and microbial vitamin derivatives, respectively, two antigenic reservoirs that are invisible to conventional T cells.1

There are several phenotypic and functional features that are strikingly similar between these two populations. iNKT and MAIT cells express restricted T cell receptor (TCR) repertoires, characterized by a canonical and invariant TCRα chain. Mouse and human iNKT cells express a TRAV11-TRAJ18- and TRAV10-TRAJ18-containing TCRα chain, respectively. Both mouse and human MAIT cells express a TRAV1.2-TRAJ33-containing TCRα chain. In addition, these cells are restricted by non-polymorphic MHC class I-related proteins, namely CD1d for iNKT cells2 and MR1 for MAIT cells.3 Both cell types share similar transcriptomic profiles4,5 and are generally considered tissue-resident. Finally, iNKT and MAIT cells exhibit rapid helper and cytotoxic effector functions that are reminiscent of innate responses and are often considered to be innate T lymphocytes.

iNKT cells vastly outnumber MAIT cells in mice, whereas MAIT cells are more prevalent in humans. The most notable difference between iNKT and MAIT cells lies in the biochemical nature of the antigens they recognize. iNKT cells respond to glycolipid antigens such as α-galactosylceramide (αGC) presented by CD1d,6,7 while MAIT cells respond to riboflavin (vitamin B2) metabolic derivatives, which includes 5-(2-oxopropylideneamino)-6-D-ribitylaminouracil (5-OP-RU), presented in the context of MR1.8

iNKT and MAIT cells develop in the thymus and require expression of the transcription factor promyelocytic leukaemia zinc finger (PLZF).9,10,11,12 They differentiate into discrete effector subsets that correspond to CD4 helper T (TH) cell subsets13,14 and are characterized by the expression of signature transcription factors. iNKT1 cells are T-bet+ PLZFlow and preferentially produce IFN-γ, while iNKT17 are ROR-γt+ PLZFint and produce IL-17. Similarly, MAIT1 and MAIT17 cells are characterized by their respective expression of T-bet and ROR-γt. iNKT2 cells are PLZFhigh T-betneg and ROR-γtneg and produce IL-4. No TH2 functional equivalent of MAIT cells has yet been identified. Multiple factors have been shown to control iNKT and MAIT cell effector differentiation, including TCR signal strength and cytokines.15,16,17,18 It is important to note that in humans, iNKT cell subsets have not yet been identified and MAIT cells do not exhibit the clear MAIT1/17 dichotomy found in mice.13,19

The roles iNKT and MAIT cells play in health and disease have been the focus of intense work and the subject of excellent recent reviews.1,20,21,22 Briefly, these cells actively participate in anti-infectious immunity against bacteria, viruses, and other pathogens, and it is arguable that such specialized T cells have evolved in mammals in response to pathogen exposure from their environment. In addition, these cells exert potent functions in tumor immunity20,22,23,24 as well as autoimmune, allergic and inflammatory diseases such as type 1 diabetes and Crohn’s disease.25,26,27,28,29,30,31,32 In this review, we will focus on the interactions between microbes and unconventional T cells and how these interactions influence their function in disease.

The microbiota and the immune system

Mammals are hosts to a multitude of microbes that colonize their epithelial surfaces to form heterogenous, complex and dynamic ecosystems.33,34,35 The composition of the microbiota is highly personalized and acts as an individual microbial ‘fingerprint’.34 Increased microbiota diversity (or richness) and resilience (i.e., the ability to recover from perturbations) are generally associated with better health.36,37,38 Shifts in microbial composition and diversity from the steady-state—a phenomenon known as dysbiosis—have been associated with numerous diseases, such as obesity, inflammatory bowel disease, and cancer.39,40,41,42 In some instances, disease susceptibility can be transferred upon colonization with the dysbiotic microbiota.43,44,45 The microbiota influences host physiological functions, taking part in essential biological processes, such as nutrient absorption and metabolism,46 barrier function47, and immune education.48,49,50,51

The immune system and the microbiota are engaged in an intimate and interdependent relationship. The loss of microbial diversity and the presence of pathobionts have been associated with inflammation and immune activation,51 which can feed back and fuel further dysbiosis.52 Specific pro-inflammatory microbes have been described, including Clostridium difficile,53,54 certain members of Prevotellaceae,55 the murine norovirus56 and protozoa of the Tritrichomonas genus.57,58 Recent studies suggest that individual components of the innate immune system exert little influence on shaping microbiota composition at steady-state,59,60 however ongoing immune or inflammatory responses can impact microbiota communities through the recruitment and activation of immune cells (e.g., neutrophils, macrophages), the production of pro-inflammatory cytokines, antimicrobial peptides, immunoglobulin (Ig) A and reactive oxygen and nitrogen species.52

In addition, the microbiota exerts a tremendous influence on the development, maturation, and function of the immune system, including T cells. A key role that the microbiota exerts on T cells is the local control of regulatory T (TREG) and TH17 CD4 T cell populations, which balances immune tolerance and pro-inflammatory responses. For example, certain members of the microbiota (e.g., SFB) or pathogens (e.g., Citrobacter rodentium) induce the development of intestinal TH17 cells.61,62,63 Colonization of mice by the protozoa Tritrichomonas similarly exacerbates mucosal and systemic TH1 and TH17 responses.57,58 On the other hand, certain commensals such as Clostridium species have been shown to induce the generation of TREG cells.64,65,66 Finally, monocolonization of germ-free (GF) mice by Bacteroides fragilis has been shown to restore the TH1/TH2 balance.67 B. fragilis also reduced Helicobacter hepaticus-induced inflammation by suppressing TH17 cells and inducing interleukin (IL)-10-producing TREG cells.68 Although iNKT and MAIT cells exert potent antimicrobial functions, how these unconventional T cells and the microbiota influence each other is only beginning to emerge.

Unconventional T cells detect microbial signature molecules

iNKT and MAIT cells are a key component of antimicrobial responses as they detect antigens that are invisible to conventional T cells. The microbial realm contains CD1d- and MR1-binding ligands that potently activate iNKT and MAIT cells through their TCRs. Such ligands can, therefore, be viewed as microbial-associated molecular patterns (MAMPs) and drive specific immune outputs from these innate T cells to shape the resulting immune response to pathogens.

Microbial glycolipids

Several CD1d-restricted antigens have been identified in bacteria and other microbes. For example, α-galacturonosylceramides have been isolated from Sphingomonas and Ehrlichia species,69,70 while α-galactosyldiacylglycerols and α-glucosyldiacylglycerols have been isolated from Borrelia burgdorferi71 and Streptococcus pneumoniae,72 respectively. These ligands have been extensively validated through biological, biochemical, and structural approaches. Other lipid antigens capable of activating iNKT cells include glycolipids from Chlamydia muridarum73 and Helicobacter pylori,74 as well as phospholipids from Mycobacterium bovis bacillus Calmette-Guerin75 and the parasite Leishmania donovani.76 Importantly, α-galactosylceramides have been recently isolated from the cecal and colonic contents of mice77 and from the commensal Bacteroides fragilis.78,79 These findings suggest that the intestinal microbiota may represent an unexplored reservoir of ligands capable of modulating iNKT cell activation and function.

Vitamin B metabolites

Riboflavin or vitamin B2 is essential for energy generation in all organisms.80 Although mammals acquire this metabolite through their diet, bacteria are able to obtain riboflavin through import and/or de novo biosynthesis.81 Bacteria that possess riboflavin biosynthetic pathways potently activate MAIT cells through their TCR. These include S. pneumoniae, Klebsiella pneumoniae, Salmonella enterica serovar Typhimurium, Escherichia coli, Pseudomonas aeruginosa, and Mycobacterium tuberculosis,8,82,83 as well as several bacterial commensal species from the Bacteroidetes and Proteobacteria phyla.84 The genetic ablation or inhibition of the riboflavin metabolism in several of these bacteria abolished MAIT cell responses, confirming that this metabolic pathway produces MAIT cell antigens.8,85,86 Riboflavin biosynthesis, therefore, constitutes a telltale sign of microbial activity that MAIT cells can respond to, with the riboflavin derivatives 5-OP-RU and 5-(2-oxoethylideneamino)-6-D-ribitylaminouracil (5-OE-RU) being potent agonists.85 It is important to note that not all MR1-binding ligands are capable of activating MAIT cells. For example, although the folic acid (vitamin B9) metabolite 6-formylpterin (6-FP), the aspirin analog 3-formylsalicylic acid (3-FSA), and a derivative of the drug methotrexate all bind within the MR1 antigen-binding pocket, they do not stimulate MAIT cells, suggesting that non-agonist and agonist ligands may compete to regulate MAIT cell activation.8,87

Microbes regulate iNKT and MAIT cell activation

Aside from the detection of glycolipid and vitamin metabolites, the activation and response of iNKT and MAIT cells can be controlled by many microbe- and host-derived factors.

Microbe-induced pro-inflammatory cytokines are essential co-signals

iNKT and MAIT cells can indirectly detect the presence of microbes through the recognition of MAMPs. Antigen-presenting cells (APCs) activated by natural or synthetic ligands of the toll-like receptors (TLR) 2, TLR3, TLR4, TLR5, TLR8, and TLR9 release pro-inflammatory cytokines that can activate MAIT cells to produce IFN-γ.88,89,90 Similarly, stimulation of TLR4, TLR7 and TLR9, Dectin-1, Rig-I, and Nod1/2 on APCs leads to IFN-γ production by iNKT cells.70,91,92,93,94,95,96 In these instances, MAIT and iNKT cells are activated by IL-12 and/or IL-18 produced by activated APCs such as dendritic cells and monocytes/macrophages. In addition, iNKT and MAIT cells can respond to IL-1β, IL-7, IL-15, IL-18, IL-21, IL-23, IL-25, and IL-33 alone or in certain combinations.97,98,99,100,101,102,103,104 Whereas some of these cytokines have proliferative effects (e.g IL-7 and IL-15), others induce the production of helper cytokines (e.g., IL-1β, IL-23, IL-25, and IL-33). Pro-inflammatory cytokines can activate innate T cells on their own or act as co-signals to potentiate weak TCR-mediated signals90,105,106,107,108 and, at least for iNKT cells, appear to be the main activation pathway in response to bacteria.105 Pro-inflammatory signals also underpin iNKT and MAIT cell activation during viral infection and sterile inflammation, such as during cancer and autoimmunity.1,20,21,22

Microbes regulate CD1d and MR1 expression

Regulating the expression of CD1d and MR1 molecules by microbes and their by-products constitutes another mechanism to control iNKT and MAIT cell activation and potentially prevent unwanted responses towards commensals. MR1 is essentially sequestered within the endoplasmic reticulum at steady-state109,110 and its expression at the plasma membrane can, in the presence of riboflavin derivatives such as 5-OP-RU, be induced or reinforced by MAMPs and/or type I and II interferons.111,112,113,114 In addition, viruses such as herpes virus can downregulate surface MR1 levels and lead to impaired TCR-mediated activation.115 CD1d is expressed quite ubiquitously and its expression can be increased by type I and II interferons, MAMPs such as LPS, all-trans retinoic acid, the short-chain fatty acid butyrate, and potentially vitamin D.116,117,118,119,120,121,122,123 Viruses such as herpes virus or human immunodeficiency virus-1 are also able to regulate CD1d expression, perhaps as a mechanism for immune evasion.118,124,125

Bile acids and microbial by-products may influence iNKT and MAIT cell responses

Some microbes may regulate iNKT and MAIT cell activation and function through their production of metabolites as well as modulation of bile acids within the intestine (Fig. 1). Primary bile acids are secreted into the small intestine during digestion and metabolized by the gut microbiota into secondary bile acids upon reaching the large intestine. Lithocholic acid (LCA) is a secondary bile acid produced by Clostridium and Eubacterium species that inhibits the function of TH1 CD4 T cells by binding to the vitamin D receptor.126 Since the vitamin D receptor is expressed by both iNKT and MAIT cells and is required for the normal development and function of iNKT cells,4,122,127 it is possible that LCA similarly modulates iNKT and MAIT cell function.

Fig. 1: Modulation of unconventional T cell function by microbe-derived metabolites and other molecules.
figure 1

Lithocholic acid (LCA), a by-product of bile acid metabolism, is a ligand for the vitamin D receptor (VDR), which is expressed by iNKT and MAIT cells. Binding of LCA to VDR may inhibit their production of TH1 cytokines (e.g., IFN-γ and TNF-α). Indole is a tryptophan catabolite that may regulate IL-17 and IL-22 production by unconventional T cells through interaction with the transcription factor RORγt. The short-chain fatty acid (SCFA) butyrate regulates gene expression through inhibition of histone deacetylases (HDAC). Butyrate may inhibit HDAC3 in iNKT cells, which has been shown to inhibit their cytokine secretion. Other tryptophan catabolites may regulate iNKT cell cytokine response through interaction with the aryl hydrocarbon receptor (AhR). Finally, inhibition of AhR within intestinal epithelial cells (IEC) alters interactions with iNKT cells. This results in the loss of IL-10 production by IECs and exacerbated IL-13/IFN-γ responses in iNKT cells, which leads to intestinal inflammation.

Full size image

Indole is a catabolite of tryptophan produced by Gram-negative and Gram-positive microbes containing tryptophanase, including E. coli, Bacteroides, Clostridium, and Vibrio species.128 Indole inhibits TH17 CD4 T cell differentiation through interactions with RORγt,129 and may similarly inhibit the differentiation and/or function of iNKT17 and MAIT17 cells. Other tryptophan catabolites such as tryptamine, indolelactic acid, and indoleacetic acid are ligands for the aryl hydrocarbon receptor (AhR). AhR is expressed by human iNKT cells and has been shown to regulate their IL-17 and IL-22 production,130 and AhR deficiency results in increased IFN-γ production by iNKT cells and exacerbates ConA-induced liver injury in mice.131

Oxazoles are heterocyclic aromatic organic compounds that are found in certain diets as well as in microcins, which are antimicrobial peptides produced by E. coli. Oxazoles can induce tryptophan metabolism and AhR activation in intestinal epithelial cells, and result in exacerbated pro-inflammatory responses of iNKT cells and colitis.132

Butyrate is a short-chain fatty acid derived from the fermentation of indigestible food by certain bacterial taxa belonging to Firmicutes and Bacteroidetes.133,134,135 Butyrate has gained increasing attention in recent years due to its capacity to regulate diseases such as inflammatory bowel disease and colorectal cancer.136,137 The primary mechanism of action of butyrate is the epigenetic regulation of gene expression via inhibition of histone deacetylase (HDAC),138 among which HDAC3 is required for iNKT cell development139 and has been shown to inhibit cytokine secretion by iNKT cells in a model of antibody-induced arthritis.140 Therefore, butyrate-producing microbes can participate in the control of iNKT cells via the butyrate-HDAC3 axis.

Although relatively little is known compared to conventional T cells,141,142 the emerging data suggest that iNKT and MAIT cells respond to a complex network of microbe-derived molecules and their by-products to fine tune ensuing immune responses. Specific interactions between the microbiota and unconventional T cells are discussed below.

Regulation of the microbiota by unconventional T cells

iNKT cell/CD1d-mediated shaping of the microbiota

Analysis of the gut microbiota through 16S rRNA sequencing using TRAJ18−/− mice (lacking iNKT cells) and CD1d−/− mice (lacking all CD1d-restricted T cells, including iNKT cells) has revealed that CD1d and/or iNKT cells may influence the composition of the microbiota (Table 1).143,144,145,146,147,148 Two studies have reported similar alterations in CD1d−/−143 and TRAJ18−/− mice,144 namely an increased prevalence of the Bacteroidetes, Deferribacteres, Proteobacteria, and TM7 phyla and several genera including Lactobacillus, Bacteroides, Mucispirillum, and Prevotella relative to wild-type (WT) mice. In a third study, bacteria from the phylum Tenericutes and genera Bifidobacterium, Prevotella, and Olsenella were found to be enriched in TRAJ18−/− compared to WT mice, while the Deferribacteres and Actinobacteria phyla were markedly decreased.147 It is important to note that environmental factors, such as diet, housing conditions, and maternal transmission are known to have a major impact on microbiota composition and should, therefore, be carefully controlled in all microbiota studies.33,34,149,150,151,152,153 Although co-housing of mice after weaning has been widely used in efforts to normalize the microbiota in various experimental settings, a recent study demonstrated that this method is ineffective at normalizing mucosa-associated microbes and that only littermate mice from heterozygous breeding should be used to minimize these confounding effects.152 In support of this, we found that adjacently bred non-littermate CD1d−/− and WT mice had major microbiota alterations, whereas CD1d+/+, CD1d+/− and CD1d−/− littermate mice had similar bacterial composition.143 Two of the studies discussed in this section lacked information on the experimental designs used and likely compared non-littermate mice, which puts the relevance of the findings into question.144,147 In addition, the TRAJ18−/− used in these studies were recently shown to have a less diverse T cell population and to completely lack MAIT cells as an artifact of the genetic targeting method employed.154,155 Several new strains of TRAJ18−/− have been generated to alleviate the unintended consequences of the genetic manipulation, which should help clarify the biological functions of iNKT cells in disease and in shaping the intestinal microbiota.156,157,158

Table 1 Summary of 16S rRNA sequencing studies assessing the role of iNKT and MAIT cells in shaping the intestinal microbiota composition.
Full size table

An initial study found that CD1d−/− mice had altered microbial communities and increased bacterial translocation compared to WT littermates, which was attributed to altered Paneth cell function.145 However, this study contained no statistical analyses and it is unclear how much the genotype contributed to community dissimilarity. Additionally, a more recent study reported microbial alterations between CD1d−/− and CD1d+/− littermates, between CD1dfl/fl CD11c-cre (which conditionally lack CD1d only on CD11c-expressing cells) and CD1dfl/fl littermates, and following oral αGC administration in WT mice.146 However, it is unclear whether the genotype or other factors such as maternal transmission, caging or sex primarily explained the observed shifts in bacterial composition. In sum, it is presently unclear whether the presence and/or activation of iNKT cells exert a strong influence in shaping the commensal microbial communities in mice.

Influence of MAIT cells and MR1 on the microbiota

Only a few studies have addressed whether MAIT cells or MR1 shape the microbiota thus far. Recent studies showed that MR1−/− mice exhibited differences in the relative abundance of several bacterial families, including Bacteroidaceae and Lactobacilliaceae, compared to WT mice, which conferred resilience to antibiotic perturbations and was associated with increased production of bile acids and resistance to Clostridium difficile infection.159,160 However, it is important to note that these authors initially compared MR1−/− mice bred in house with commercially obtained WT mice, while additional analyses were performed on mice that were co-housed for four weeks post-weaning (Table 1). As discussed above, co-housing is not a reliable method to normalize environmental differences.152 In addition, the relative contribution of genotype and other factors to the observed variation in microbiota composition was not analyzed. An important message that can be drawn from these studies is that littermate-controlled experiments carefully designed to minimize variability and robust statistical methods are needed in order to draw biologically meaningful conclusions on the immune-mediated shaping of the gut microbiota.59

Modulation of innate T cells by the microbiota

Microbial colonization impacts innate T cell development and homeostasis

The microbiota is important for the education of the adaptive immune system, and early life exposure ensures proper T and B cell maturation, differentiation, and tolerance.49,161,162 This also appears to be true for iNKT and MAIT cells. Although the development and tissue homing capabilities of iNKT cells are largely unaffected by the absence of the microbiota in GF mice, their prevalence is increased in the lungs, small intestine and colon.27,163,164,165 In addition, iNKT cells in GF mice have an altered TCRβ repertoire and are overall less mature and hyporesponsive to αGC stimulation.163 Similar defects are also observed in mice harboring a restricted microbiota characterized by a reduced prevalence of Sphingomonas species, which are known to produce activating iNKT cell ligands.164 However, whether these changes are specific to certain iNKT cell subsets and their tissue distribution is not known. Interestingly, mono-colonization of GF mice with Sphingomonas yanoikuyake, but not E. coli, corrected the defects, suggesting that the commensal-derived antigenic load directly participates in iNKT cell homeostasis and/or function.163 Furthermore, the accumulation of iNKT cells in the lungs and intestines of GF mice can be reversed by microbiota colonization early in life but not after weaning.27

Tthe development of MAIT cells is profoundly affected by the absence of the microbiota in GF mice3,10,13,82,166 (Fig. 2). Although colonization of GF mice with a complex human microbiota is able to restore MAIT cell numbers to that of specific pathogen-free (SPF) mice, they are unable to reach the higher levels normally found in humans.10 Similarly, mono-association of GF mice with Enterobacter cloacae, Lactobacillus casei, Bacteroides thetaiotaomicron, Bifidobacterium animalis, but not E. faecalis, also restores MAIT cell numbers to SPF levels.82 In addition, early-life exposure to riboflavin-synthesizing bacteria, such as Enterobacteriaceae, can restore MAIT cell development in GF mice, whereas colonization of adult GF mice has no effect.100 Together, these findings suggest that riboflavin-synthesizing commensals can stimulate MAIT cell development and/or homeostasis (Fig. 2). Likewise, topical application of 5-OP-RU to the skin of GF mice early in life, but not in adulthood, is sufficient to restore MAIT cell numbers.100 Another recent study in mice has suggested that microbial metabolites have the potential to translocate to the thymus, upon which they can be captured by MR1-expressing cells and presented to developing thymocytes to induce MAIT cell maturation.166 In humans, MAIT found in the umbilical cord blood are immature, whereas some mature MAIT cells can be found in peripheral blood after birth and accumulate with age.167,168 This suggests that MAIT cell maturation likely involves microbial antigens and appears to be fine-tuned by the TCRβ chain.167,168 Collectively, these findings suggest that early-life colonization by the microbiota has a profound and long-lasting impact on the development and homeostasis of both iNKT and MAIT cells.

Fig. 2: Early-life microbial colonization stimulates MAIT cell development and maturation.
figure 2

MAIT cells are virtually absent in germ-free (GF) mice. Early-life colonization of GF mice with diverse microbiota from mice or humans restores MAIT cell development. Similarly, oral, intraperitoneal, or topical administration of riboflavin-producing bacteria or the MAIT ligand 5-OP-RU restores MAIT cell development and homeostasis. Interestingly, exogenous 5-OP-RU can traffic to the thymus to stimulate MAIT cell ontogeny.

Full size image

The microbiota influences innate T cell function in disease

The gut microbiota modulates iNKT cell function in disease. The accumulation of iNKT cells in the lungs and intestines of GF mice was correlated with exacerbated ovalbumin-driven asthma and oxazolone-induced colitis, which could be improved by administration of neutralizing anti-CD1d antibodies.27 The authors of this study went on to show that GF mice mono-colonized with B. fragilis from birth were protected from chemically induced colitis in adulthood as a result of normalized abundance of colonic iNKT cells79 1. The protective effect of B. fragilis was dependent on the expression of the glycosphingolipid GSL-Bf717, as GF mice colonized with mutant B. fragilis lacking GSL-Bf717 were not protected.143,169,170 Although iNKT cell functional response (e.g., cytokine production) was not directly assessed, these studies demonstrated that early-life microbial colonization is necessary to tune down the pro-inflammatory functions of iNKT cells at mucosal sites.

Additional hints on how the gut microbiota functionally impacts iNKT cells and their ability to regulate disease can be gleaned from a few other studies. For example, glycolipid-mediated activation of iNKT cells during the course of dextran sulfate sodium (DSS)-induced colitis has shown inconsistent results between studies.143,169,170 The discrepancies may be attributed to different experimental designs, including the dose, frequency, and route of glycolipid injection, as well as environmental factors such as housing conditions, diet, and microbiota differences between animal facilities. In a previous study, we demonstrated that WT mice born from dams transplanted with two discrete microbial communities have different clinical outcomes in response to the synthetic iNKT cell ligand OCH9, which suggests that iNKT cell responses and/or their ability to modulate immune and inflammatory responses are regulated by the intestinal microbiota.143 Short-term broad spectrum antibiotic treatment in adult mice increased colonic iNKT cell numbers, and subsequent reconstitution with a dysbiotic microbiota (isolated from DSS-treated mice) skewed the cells towards a pro-inflammatory phenotype.171 This microbiota-driven skewing of iNKT cell phenotype was also observed in ulcerative colitis patients.28 Antibiotic treatment also had similar effects on liver iNKT cells. In mice, accumulation of IFN-γ-producing iNKT cells in the liver was driven by the upregulation of CXCL16 in liver sinusoidal endothelial cells as a result of the reduction in microbiota-mediated primary to secondary bile acid conversion in the gut, and in this case was associated with protective function against intrahepatic tumor growth.172 Taken together, these studies suggest that the gut microbiota has the capacity to regulate iNKT cell homeostasis in certain tissues, as well as their functional response during inflammation and tumor development (Fig. 3).

Fig. 3: The microbiota modulates iNKT cell function in disease.
figure 3

iNKT cells in GF mice are pro-inflammatory and fuel allergen-induced airway inflammation and colitis. Early-life colonization with a complex microbiota, or the sphingolipid-producing commensal Bacteroides fragilis restores normal iNKT cell functions and limits inflammation. Bile acids produced by the liver are modified by the microbiota in the gut and recirculate back to the liver where they recruit iNKT cells via the chemokine CXCL16. Accumulated iNKT cells provide antitumor immunity in the liver.

Full size image

MAIT cells have also been shown to play a role in multiple infectious and sterile diseases, however to what extent and how the microbiota regulates their function remains unclear. MAIT cells provide initial immunity against Mycobacterium tuberculosis infection due to their robust granzyme B production.173 Interestingly, their capacity to produce granzyme B correlates with the abundance of specific bacterial species, namely B. ovatus and P. merdae of the class Bacteroidia. Furthermore, antibiotic-induced dysbiosis was found to increase early infection levels of M. tuberculosis in mice, which correlated with reduced MAIT cell numbers in the lungs.174 The inoculation of a complex microbiota into antibiotic-treated mice restored MAIT cell numbers and conferred protection, suggesting that the microbiota modulates MAIT cell proliferation and activity in response to infection. In addition, the loss of intestinal epithelial integrity and bacterial translocation in patients with alcohol-related liver disease led to reduced circulating MAIT cell numbers and impaired IL-17 and granzyme B production.175 Similar MAIT cell defects were found in children with type 1 diabetes, which is associated with a loss of intestinal barrier integrity.25 Furthermore, MR1−/− non-obese diabetic (NOD) mice have greater intestinal permeability, decreased expression of tight-junction proteins, abnormal mucus distribution, and accelerated disease progression compared to MR1+/− NOD mice.25 Therefore, there is mounting evidence that the microbiota can modulate MAIT cell function during infectious and inflammatory diseases.

Concluding remarks

While our understanding of the biology of iNKT and MAIT cells has seen tremendous progress in recent years, it has for the most part focused on how host-derived factors control the development and function of these innate T cells, with the influence of environmental factors, such as the microbiota being mainly overlooked. Whereas it remains unclear whether innate T cells influence the composition of the microbiota in mammals, the influence of the microbiota on iNKT and MAIT cells has now come into clearer focus. These findings raise several fundamental and practical questions: Can the heterogeneity observed in innate T cell compartments between individuals and populations be explained, at least in part, by environmental differences such as microbiota composition? Are microbiota-innate T cell interactions relevant to human disease and therapies targeting these cells? What are the microbial determinants that modulate innate T cell homeostasis and functional properties? And most importantly, can we manipulate the gut microbiota or harness their by-products in order to modulate iNKT and MAIT cell function in disease? Innate T cells constitute an integral component of the immunologist’s toolkit to develop novel targeted immunotherapies,24 and prospects of their selective manipulation learned through their dialogue with the gut microbiota are certainly worthy of careful examination.