MAIT cells and microbial immunity

MR1-dependent T cell responses

The first studies to definitively show that MAIT cells were reactive to bacterial and fungal products presented by MR1 were published by Gold et al. and Le Bourhis et al. in 2010^4,5. Le Bourhis et al. purified human TRAV1–2⁺ CD161⁺ T cells and showed that they could be activated by monocytes infected with Escherichia coli (E. coli) or Mycobacterium abscessus in an MR1-dependent fashion. Due to the low frequency of MAIT cells in mice, transgenic mice expressing TRAV1/TRBV19 (Vα19/Vβ6) TCRs were engineered to show that MAIT cells were activated by a wide array of bacterial and fungal species. In these experiments, Candida albicans (C. albicans), Candida glabrata, E. coli, Klebsiella pneumoniae (K. pneumoniae), Lactobacillus acidophilius, Pseudomonas aeruginosa, Saccharomyces cerevisiae, Staphylococcus aureus (S. aureus), and Staphylococcus epidermidis were all capable of activating MAIT cells. However, Enterococcus faecalis (E. faecalis) and Streptococcus pyogenes (S. pyogenes), as well as five unrelated viruses, were unable to elicit a response by MAIT cells⁵. In parallel, Gold et al.⁴ were studying the human immune response to Mycobacterium tuberculosis (Mtb), and specifically seeking to explain the observation that individuals who had never been exposed to Mtb had a substantial population of T cells that could nonetheless recognize infected cells ex vivo. These cells were cytolytic and made pro-inflammatory cytokines like TNF-α and IFN-γ. In an effort to describe these T cells in more detail, T cell clones were isolated from the blood of healthy individuals and patients with TB. Surprisingly, all of these clones were dependent on MR1 for their recognition of Mtb-infected cells. In addition to Mtb, these clones were broadly reactive to pathogenic microbes such as C. albicans, E. coli, Mtb, Mycobacterium smegmatis (M. smegmatis), Mycobacterium bovis BCG, Salmonella enterica serovar typhimurium (S. enterica typhimurium), and S. aureus, but not viruses or Listeria monocytogenes⁴. Importantly, the data from these two studies suggested that while viruses did not activate MAIT cells, a diverse array of bacteria and fungi could (Table 1). Based on the use of a semi-invariant TCR, and highly conserved presentation molecule, these studies suggested that MAIT cells might detect a conserved microbial ligand.

The ensuing discovery that MAIT cells could recognize a class of small molecules generated during microbial riboflavin biosynthesis^8,9 suggested that this pathway was necessary for the generation of activating MR1 ligands (Table 1). In these and other subsequent studies, it was demonstrated that for some microbes including S. enterica typhimurium^9,15, L. lactis⁹, E. coli¹⁶, and Streptococcus pneumoniae (S. pneumoniae)¹⁷, an ability to synthesize riboflavin through functional Rib enzymes was critical to MR1-dependent recognition of these microbes (Table 1). Specifically, for these bacterial species, the RibD enzyme was required for production of 5-amino-6-D-ribitylaminouracil (5-A-RU), a precursor to riboflavin, and a critical component of these ligands. The 5-(2-oxopropylideneamino)-6-D-ribitylaminouracil (5-OP-RU) ligand, formed when 5-A-RU combines with host or bacterial derived methylglyoxal, is the most potent ligand identified to date⁹. Interestingly, there is now evidence that some bacteria produce non-riboflavin pathway-derived MR1 ligands as well. For example, MR1T clones characterized by TRAV1–2-negative TCR usage recognized S. pyogenes in an MR1-dependent manner even though S. pyogenes lacks the enzymatic pathway for riboflavin biosynthesis¹⁴. Additionally, the 6-formyl pterin (6-FP) ligand, that is antagonistic for MAIT cells⁸, can activate a population of TRAV1–2 negative MR1Ts¹³. As 6-FP is derived from folic acid, another vitamin synthesized by some bacteria and fungi, this is another possible source of microbial MR1 ligands. Clearly, diversity in MR1T TCRs is associated with differential ligand recognition. How diverse MR1 ligands contribute to protective immunity in the context of infection is not yet known.

MR1-independent cytokine driven MAIT cell responses

While MR1-dependent recognition of bacterial and fungal antigens by MAIT cells is well established, there is growing evidence for MR1-independent cytokine-driven responses. It is well-known that virus-specific mouse CD8⁺ T cells can be induced to produce IFN-γ in an antigen independent manner by sensing cytokines including type I IFNs, IL-12 and IL-18. However, in humans the responses to IL-12 and IL-18 are more distinctly linked to cell populations expressing killer cell lectin-like receptor subfamily B member 1 (KLRB1) or CD161, such as MAIT cells^18,19. The role of these cytokines in mediating anti-viral MAIT cell responses is covered by the Klenerman group in this issue. The relative contribution of inflammatory cytokines as compared to MR1-dependent MAIT cell activation in the context of bacterial infection is discussed below. Ussher et al. evaluated the role of IL-12 and IL-18 using a co-culture assay¹⁸. To distinguish the role of TCR-dependent and -independent signaling, THP-1 cells were cultured with either fixed E. coli, which contains MR1 ligands, or with fixed E. faecalis, which does not synthesize riboflavin. In this context, E. faecalis stimulated MAIT cells in a TCR-independent and IL-12 and IL-18-dependent manner. In contrast, while E. coli stimulated TCR-dependent MAIT cell responses early in the co-culture, the responses shifted to TCR-independent responses later¹⁸. More recently, Jo et al.²⁰ found that production of IL-12 and IL-18 by cells infected with fixed E. coli was sufficient to induce cytokine production by purified human liver MAIT cells after a longer co-culture. These data contrast with other studies demonstrating that human MAIT cell responses to live bacterial stimuli are MR1-dependent, even after longer co-incubations e.g.^4,5,14. To explain these discordant results, we speculate that the availability of antigen could be limiting in fixed bacteria allowing for cytokine-driven activation to play a more dominant role. In the context of microbial infection in the mucosa, the relative role of cytokine-driven versus TCR-mediated activation of MAIT cells remains to be elucidated.

Other groups have demonstrated MR1-independent activation of MAIT cells in the context of microbial infection in vivo. In an initial study by Chua et al., murine MAIT cells lost their in vivo anti-mycobacterial activity when the IL-12 p40 subunit was deleted or blocked²¹. Subsequently, Meierovics et al.²² showed that MAIT cell control of Francisella tularensis LVS in macrophages was dependent on IL-12, using blocking with an anti-IL-12p40 antibody. Although addition of the IL-12p40 blocking antibody in both of these cases reversed the ability of MAIT cells to control growth of bacteria, the role of IL-12 itself is unclear, as IL-12p40 is also a subunit of IL-23. Furthermore, in the case of Meierovics et al., ex vivo blocking assays with an anti-MR1 antibody also blocked bacterial growth inhibition and cytokine production to the same extent as the IL-12p40 antibody suggesting possibly redundant or complementary roles in this infection model. In a subsequent in vivo study, the same group demonstrated that MAIT cell stimulation of monocyte differentiation in the context of LVS infection was MR1-independent²³. The differentiation required MAIT cell-dependent production of GM-CSF, and the role of IL-12 or IL-18 in MAIT cell activation was not investigated. Shaler et al. have demonstrated another MR1-independent pathway for MAIT cell activation through bacterial superantigens²⁴. Overall, given the localization of MAIT cells at common sites of infection, both MR1-independent and MR1-dependent MAIT cell activity is likely to play a role in defense against certain infections in vivo.

MAIT cell localization in tissues and expansion in response to infection

MAIT cells may be developmentally programmed to reside in mucosal sites early in life. In support of this, Gold et al. observed that MAIT cells in the thymus, cord blood and peripheral blood express significantly less CD62L, a selectin associated with homing to lymphoid tissues⁵¹. Moreover, Leeansyah et al. observed that human 2^nd trimester fetal MAIT cells already express CD45RO and are functionally mature in mucosal tissues. Fetal MAIT cells were enriched in lung, small intestine and liver compared to spleen and thymus^6,53. Furthermore, at these sites, they proliferated and produced cytokines in response to E. coli infection, suggesting the development of an early arm of antibacterial mucosal immunity. Numerous studies in mice, NHP, and humans have extensively mapped MAIT frequencies in tissues. Circulating MAIT cells in mice are less numerous than in humans (<0.1% versus 1–10%, respectively)^10,11. However, the use of an MR1 tetramer identifies enrichments in the murine lung, liver, intestines, and female genital tract at up to 2–4% of T cells^11,15. A pioneering study of NHP observed MR1 tetramer⁺ cells enriched in the liver and bronchial alveolar lavage as compared to the blood⁴⁸. Within humans, studies demonstrate a higher frequency of MAIT cells are present in the intestine and jejunal mucosa (5–60% of CD4-T cells)^6,10, liver parenchyma and vasculature (20–50% of T cells)^{6,24,36,54,55}, placental intervillous blood (2–4% of T cells)⁴⁴ and the airway and lungs (2–4% of T cells)⁴. Interestingly, MAIT cells are present at a lower frequency in the human female genital tract⁵⁶ and lymph nodes³⁶ as compared to the blood. That different tissues may contain clonotypic enrichments of MAIT cells with different TCRs is suggested by the observation that the Vα7.2/Jα12 transcript is enriched relative to the Vα7.2/Jα33 transcript in the human intestine, kidney, and ovary samples³⁶.

In the context of infection, numerous studies have demonstrated that circulating MAIT cell frequencies decrease^{4,5,39–43,46,47,57–60}. The prevailing hypothesis is that MAIT cells have been recruited to infected tissues. For example, intranasal infection of mice with live attenuated F. tularensis, mycobacteria, or S. enterica typhimurium caused the recruitment of proliferating MAIT cells to the lungs^21,22,61. In a unique model of human bacterial infection, volunteers challenged with oral wild-type Salmonella typhi were evaluated for development of typhoid fever⁴⁶. In those who developed disease, MAIT cells became activated, decreased in frequency from the blood, and increased expression of gut homing chemokine receptors. Additionally, Greene et al. evaluated the in vivo response by MAIT cells to mycobacteria in NHP. Here, Greene et al. used an MR1 tetramer to track MAIT cells in NHP that were vaccinated intradermally with BCG and then challenged with pulmonary Mtb⁴⁸. They observed an increase in Ki-67 expression, a cellular marker of proliferation and activation, on MAIT cells at the site of vaccination and systemically during Mtb infection. Together, these studies support the hypothesis that MAIT cells are recruited to and can expand at mucosal sites in response to bacterial infections.

To further explore MAIT cell expansion in response to infection, Chen et al. administered attenuated S. enterica typhimurium intranasally to mice. Here, MAIT cells accumulated locally (up to 50% of T cells) within one-week following challenge in an MR1-dependent manner and these expansions were sustained for at least 50 days¹⁵. Additionally, these investigators and one other found that intranasal delivery of purified antigen (5-OP-RU) was not sufficient to induce MAIT accumulation^15,32. Instead, costimulation with Rib-pathway-deficient S. typhimurium or TLR 2, 3, 6, 9 agonists led to MAIT cell enrichment in the lungs that was sustained for at least 10 weeks and MR1 dependent¹⁵. While TLR-agonists alone are sufficient to activate many cells within the innate arm of the immune system, they were not sufficient for MAIT cells^4,5. These data would suggest that mucosal MR1T cell expansion requires both TCR-dependent signals in conjunction with co-stimulation, such as TLR signaling. This illustrates how MAIT cells cooperate with innate immune stimuli as part of the early local response to infection.

Tissue Resident MAIT Cells Display Unique Functionality

Upon stimulation, MAIT cells can make a variety of cytokines that have the potential to promote an antibacterial or anti-fungal response. It has been widely observed that they produce the pro-inflammatory cytokines IFN-γ and TNF-α^{4–8,10,24,38,62,63}. MAIT cell effector function in response to TCR signaling is triggered more rapidly than conventional T cells. For example, a comparison of human T cell responses to Staphylococcal superantigen found that MAIT cells were activated to secrete IFN-γ more rapidly than conventional T cells²⁴. In this way, the authors concluded that MAIT cells provide an early source of proinflammatory mediators that could be pathogenic or protective depending on the nature of the bacterial stimulus. This may be due to upregulation of cytokine transcripts that poise MAIT cells for robust responsiveness⁶⁴. In the context of bacterial infection, murine MAIT cells produce IL-17, while less than 1–2% of circulating NHP and human MAIT cells produce this cytokine^6,7,24,38,48. However, human MAIT cells from discrete tissue sites, disease states, and inflammatory environments have distinct cytokine profiles. MAIT cells in the spleen, thymus, liver, lungs, placental intervillous blood, and colon, share a similar cytokine profile with those in the circulation^{7,44,50,51,64,65}. In contrast, MAIT cells in the female genital tract⁵⁶, intrasinusoidal liver⁵⁵, spinal fluid of patients with ankylosing spondylitis^62,66, salivary glands of patients with primary Sjogren’s Syndrome⁶⁷, pleural fluid of patients with tuberculosis pleurisy⁵⁷, and the blood of patients with ulcerative colitis⁶⁸, make more IL-17 and less TNF-α and IFN-γ. Specifically, up to 22% of MAIT cells in the female genital tract produced IL-17 in response to E. coli infection ex vivo in contrast to those from the blood⁵⁶. Notably, Tang et al. observed that the pre-treatment of MAIT cells in vitro with IL-7 and IL-23 increased the IL-17 polarization⁵⁵. Like IL-17, IL-22 is another cytokine preferentially expressed by MAIT cells at different tissue sites. IL-22 is a cytokine associated with host defense and repair of mucosal tissue and is generally not produced by circulating MAIT cells⁷. However, MAIT cells from the spleen and small intestine from 2^nd trimester fetuses could make IL-22 in response to fixed-E.coli, while IFN-γ producing cells were found in the small intestine, lung, and spleen⁵³. Collectively, these data suggest that tissue-resident MAIT cells will have a phenotype and function tailored to their anatomic location, proximity to microbes, and inflammatory environment. In this regard, MAIT cells might contribute directly to microbial control, could enable the subsequent development of adaptive immunity, or could contribute to autoimmunity.

MAIT cell cytolytic capacity

Cytolytic properties of immune cells, including MAIT cells, are crucial for eliminating intracellular bacterial infection, either through the induction of apoptosis of the target cells, or via the introduction of anti-bacterial proteins. Cytotoxic T cells, in humans both CD8⁺ and CD4⁺, contain granzyme B, perforin, and granulysin in preformed granules, allowing for rapid delivery to the target cell. MAIT cells also express cytotoxic molecules and are capable of lysing bacterially infected cells ranging from macrophages to epithelial cells^{4,6,7,36,45,55,69}. However, data suggest that resting MAIT cells do not express high levels of granzyme B like conventional T cells, but instead, can be licensed to produce granzyme B by bacterial infection, TCR stimulation, or IL-12/15/18^57,64,69,70. Importantly, this cytolytic ‘licensing’ could be maintained at least 6 days in an in vitro culture⁶⁹. Furthermore, MAIT cells were found to be specifically activated in the context of volunteers challenged with an attenuated strain of Shigella dysenteriae⁴⁵. MAIT cell lysis of epithelial cells infected with S. dysenteriae was dependent upon MR1. Collectively, MAIT cells are cytolytic effectors that have may have licensing mechanisms to prevent inflammatory cytolytic damage that would be more likely given their abundance and proximity at the mucosal interface.