The role of mucosal‐associated invariant T cells in infectious diseases

What are MAIT cells and what is known about their phenotype and function?

Mucosal associated invariant T (MAIT) cells are αβ T cells with a limited T‐cell receptor (TCR) repertoire that uses mostly TRAV1‐2 recombined with a biased set of TRAJ genes (primarily TRAJ 33, TRAJ 12 and TRAJ 20) and TCR‐β chains.1, 2, 3, 4 MAIT cells are uniquely activated by small molecule vitamin B metabolites presented by the ubiquitously expressed and non‐polymorphic MHC‐class 1‐like molecule, MR1.5, 6, 7, 8, 9 Circulating MAIT cells express certain phenotypic markers that have been used in varying combinations to identify them in human specimens: TRAV1‐2, IL‐18Ra, IL‐23/12Ra, and high expression of both CD26 and CD161.10, 11 Use of different phenotypic definitions of MAIT cells has limited comparability of findings in the field. The recent development of ligand‐loaded MR1 tetramers promises to aid in consistent definitions.12 At this point it appears that MR1 tetramers loaded with 5‐(2‐oxopropylideneamino)‐6‐d‐ribitylaminouracil (5‐OP‐RU) stain the vast majority of MR1‐reactive cells in the peripheral blood of healthy humans, though whether additional classes of MAIT cell activating MR1 ligands exist is a topic of ongoing active investigation.13

MAIT cells comprise 1–10% of circulating CD3⁺ T cells in healthy adults.14, 15 Although their distribution in tissues has not been thoroughly investigated, they have been found to be enriched in the gastrointestinal tract, liver and airway.10, 16, 17, 18, 19 MAIT cell activation can result from either TCR‐dependent signalling (triggered by ligand presented on MR1 by antigen‐presenting cells) or from TCR‐independent cytokine signalling.20, 21, 22, 23 Like other ‘innate‐like’ lymphocytes including invariant natural killer T cells and natural killer cells, MAIT cells respond rapidly to activation by producing cytokines and cytolytic products. Upon TCR‐dependent or TCR‐independent activation, MAIT cells produce interferon‐γ (IFN‐γ), tumour necrosis factor‐α (TNF‐α), cytolytic products (perforin, granulysin and granzymes) and degranulate (exposing CD107a to the cell surface).10, 11, 24, 25, 26 Interleukin‐12 (IL‐12) and IL‐18 have been shown to induce MAIT cell production of IFN‐γ in the absence of TCR signalling.22 Both IL‐15 and IL‐7 have each been shown to license MAIT cells for enhanced production of TCR‐activated pro‐inflammatory cytokines and cytolytic products.23, 26, 27 MAIT cells express the transcription factors promyelocytic leukaemia zinc finger protein (PLZF) and RAR‐related orphan receptor (RORγτ) and do not develop in Rorc ^−/− mice.26, 28 Multiple groups have noted that circulating MAIT cells produce IL‐17 upon PMA/ionomycin stimulation but not TCR‐triggered stimulation.10, 17, 26, 29 A recent study found that compared with peripheral blood MAIT cells, MAIT cells from the female genital tract produced significantly more IL‐17 and IL‐22 and significantly less IFN‐γ or TNF‐α when stimulated with either bacteria or PMA/ionomycin.30 The significance of IL‐17 production by MAIT cells in vivo requires further investigation.

The MAIT cell biology field is still in its early stages. Despite recent increased interest and increasing numbers of papers that examine MAIT cells in human diseases and animal models of diseases, many questions remain unanswered about the function of this immune subset in the healthy and perturbed immune system. Here we review the rapidly evolving literature concerning the role of MAIT cells in infectious diseases and highlight areas of evolving consensus and remaining gaps in knowledge.

Insights from in vitro experiments

Upon exposure to bacteria that produce MR1 ligands, MAIT cell clones cultured from human peripheral blood and polyclonal MAIT cell populations isolated from human peripheral blood (TRAV1‐2⁺ CD161⁺⁺) produce pro‐inflammatory cytokines (TNF‐α, IFN‐γ) and cytolytic products (granzymes, granulysin, perforin).11, 20, 21, 24, 25, 26 This is true when MAIT cells are exposed to bacterial preparations directly (fixed Escherichia coli, Mycobacterium tuberculosis lysate) or when MAIT cells are added to cultured cells that have been experimentally infected (with E. coli, Salmonella typhimurium, Mycobacterium smegmatis, M. tuberculosis).13, 20, 33 The production of cytokines and cytolytic products by MAIT cells in these experiments appears to largely depend on TCR signalling triggered by MR1‐presented ligand and can be augmented by pre‐conditioning with a number of cytokines including IL‐12, IL‐23, IL‐15 and IL‐7.22, 23, 26, 27 Adding MAIT cells to infected cells can result in lysis or apoptosis of the infected cells.1, 25, 26 Moreover, MAIT cells have the ability to decrease mycobacterial burden when added to experimentally infected cell cultures. Chua et al.34 showed that the addition of polyclonal MAIT cells to macrophages experimentally infected with Mycobacterium bovis BCG inhibited bacterial growth. Lepore et al.1 demonstrated that MAIT cell clones with typical (TRAV1‐2/TRAJ33) and atypical (TRAV1‐2/TRAJ12) TCRs decreased M. bovis BCG load in infected macrophages. In sum, these in vitro experiments show that bacterially produced MR1 ligand activates MAIT cells and triggers the secretion of pro‐inflammatory cytokines and cytolytic products that can directly lyse and/or cause apoptosis of the infected host cells. In models of mycobacterial infection MAIT cells enhance the ability of macrophages to inhibit the growth of intracellular organisms. Whether the primary mechanism of control is through enhancing the ability of macrophages to control infection or by killing infected cells is unknown. The effector functions of MAIT cells in the context of bacterial and mycobacterial infection appear to rely in part on bacterially produced ligand being presented by MR1, with cytokines produced in the context of infection potentiating MAIT cell effector functions (Figure 1).

Insights from animal experiments

MR1 is highly conserved among mammals, and although mice have considerably lower levels of MAIT cells than humans, experiments using mice have provided some insights into the role that MAIT cells may play in the immune response to infections with MR1‐ligand producing bacteria.3, 4, 35 These experiments make use of transgenic mouse strains that over‐express MAIT cells (Vα19i‐Tg and Vβ6‐Tg), MR1 knock‐outs (MR1‐KO) of the transgenic strains and wild‐type C57BL/6 mice. These models have allowed comparisons of the outcome of experimental infection in mice with enhanced, intact or disrupted MR1–MAIT cell axes. When exposed to the same initial intraperitoneal inoculum, C57BL/6 MR1‐KO mice developed higher burdens of systemic infection with Klebsiella pneumoniae and had decreased survival compared with wild‐type animals.36 This suggests that MR1 (and presumably MAIT cells) play a protective role and aid in the containment of K. pneumoniae infection. Interestingly, these findings were not replicated in similar infections with three other Gram‐negative organisms that are known to produce MR1‐ligands: E. coli, Shigella dysenteriae and Yersinia enterocolitica. Another group making use of the transgenic over‐expressing MAIT cell mice (both Vα19i‐Tg and Vβ6T), found that MR1‐KO mice had higher bacterial burdens of both E. coli and Mycobacterium abscessus than those with intact MR1.21 Sakala et al.37 recently compared outcomes in the Vα19i‐Tg stain (and its MR1‐KO) and C57BL/6 (and its MR1‐KO) after nasal aerosolized M. tuberculosis infection and found that the MAIT over‐expressing mice had the lowest burden of infection and the MR1 KO mice had the highest. This growing body of evidence suggests that an intact MAIT cell–MR1 axis may be protective against some (but not all) bacterial infections with MR1‐ligand producing organisms.

Additional data suggest that the protective role of MAIT cells in these infections probably occurs in the early days after exposure to pathogen. Chua et al.34 demonstrated that MR1 KO mice had higher burdens of pulmonary infection 10 days after aerosolized infection with M. bovis, but that after 30 days there was no difference in the bacterial burden between the two groups. Using intranasal infection with the live vaccine strain of Francisella tularensis (LVS) in BL/6 wild‐type and MR1‐KO mice, Meierovics et al.38 demonstrated that MAIT cells appear to have a direct, early antibacterial effect in the lungs and a sustained impact on the establishment of an effective adaptive mucosal immune response. Specifically, they demonstrated that upon exposure to LVS F. tularensis, MAIT cells, here defined as double‐negative T cells that were enriched for the canonical MAIT cell TCR Va19‐Ja33, accumulated in the lungs and were associated with brisk production of IFN‐γ, TNF‐α, IL‐17A and iNOS as well as significantly faster clearance of the bacteria from the lungs. The authors explored the relative contribution of MAIT cells and classical CD4⁺ and CD8⁺ T cells to bacterial control and investigated the impact of MAIT cells on the classical adaptive immune response. In mice depleted of classical CD4⁺ and CD8⁺ T cells but with intact MR1, MAIT cells were able to control LVS F. tularensis infection and establish a state of chronic non‐lethal infection for up to 2 months. In contrast, mice depleted of both classical T cells and MR1/MAIT cells died rapidly, suggesting that MAIT cells have the ability to control infection for prolonged periods of time even in the absence of conventional T cells. Importantly, in this study the adaptive immune system was incapable of optimal function in the absence of MR1 and MAIT cells. In the MR1‐KO mice, recruitment to the lungs of activated IFN‐γ producing classical CD4⁺ and CD8⁺ T cells was significantly delayed and attenuated with 60–80% fewer conventional T cells recruited to the lungs at Day 8 following infectious challenge. Taken together these findings suggest that MAIT cells are poised in the mucosal surface of the airways and lungs, where they provide early control of bacterial infection through production of pro‐inflammatory cytokines and antimicrobial products (Fig. 2). Additionally, during this period of early infection they are required for the establishment of optimal adaptive immune response. The mechanisms by which MAIT cells lay the ground work for later adaptive immunity is unknown.

Insights from study of human cohorts

In one of the first studies of MAIT cells in human disease, Gold et al.20 demonstrated that M. tuberculosis‐reactive MR1‐restricted T cells were found in the peripheral circulation of all the people that they assessed, including those who had no evidence of previous exposure to M. tuberculosis. Interestingly, these innately M. tuberculosis‐reactive T cells were found at much lower levels in those individuals with active tuberculosis (compared with those with no history of exposure, or those with evidence of latent infection). Other studies have confirmed that MAIT cell frequencies are decreased during active tuberculosis (and non‐tuberculosis mycobacterial lung disease) in diverse geographic settings.11, 21, 33, 39 Peripheral depletion of MAIT cells is not, however, disease specific, with low levels of peripheral MAIT cells found in individuals with cystic fibrosis experiencing pulmonary exacerbations due to Pseudomonas aeruginosa,40 non‐streptococcal sepsis41 and cholera.42 There is evidence that the degree of peripheral MAIT depletion may correlate with disease severity and may be reversed with antibiotic treatment during recovery from illness.11, 39, 40, 41 Interestingly, in a study of peripheral MAIT cells in intensive care unit patients with sepsis, failure of peripheral MAIT cells to rise during the intensive care unit stay was correlated with the subsequent acquisition of secondary infections.41 Depletion of MAIT cells in the peripheral blood is not specific to bacterial infection and is also observed in infections due to non MR1‐ligand producing organisms (i.e. viral infections, discussed below) and non‐infectious states of disease including autoimmunity, asthma, obesity and diabetes (beyond the scope of this article).43, 44, 45, 46

The mechanism for depletion of peripheral MAIT cells during active bacterial infections is unknown. Although MAIT cells are named for their enrichment in the gut mucosa, there are very few human data about mucosal MAIT cells during bacterial infections that would allow assessment of the hypothesis that peripheral depletion is due to redistribution of MAIT cells to the site of disease. Mycobacterium tuberculosis‐reactive MAIT cells were found in healthy uninfected lung tissue by Gold et al.20 and MAIT cells were found to be present in two M. tuberculosis‐infected lungs by LeBourhis et al.21 without control tissue or paired peripheral sample available for comparison. LeBourhis et al.21 reported enrichment of MAIT cells in the ascites fluid from a single patient with tuberculosis peritonitis, but in a larger series, Jiang et al.33 found diminished frequencies of MAIT cells in pleural fluid from patients with tuberculosis pleuritis and non‐enriched levels of MAIT cells in the peritoneal fluid of patients with tuberculosis peritonitis. Hence, it is not clear whether MAIT cell depletions in bacterial infections are due to redistribution to mucosal sites.

The few MAIT cells that remain in circulation during active pulmonary tuberculosis have diminished functional capacity, producing less IFN‐γ upon stimulation and decreased transcription of effector genes (IFN‐γ, TNF‐α, IL‐17F, granulysin and granzyme B).33, 39 These functionally impaired MAIT cells express higher levels of pro‐apoptotic markers and higher levels of the inhibitory co‐stimulating molecule programmed death‐1 (PD‐1) than MAIT cells from non‐tuberculous patients.33, 39, 47 In vitro blockade of PD‐1 can increase IFN‐γ production in circulating MAIT cells from patients with tuberculosis.33

In summary, human data provide little indication about the role that MAIT cells are playing in bacterial infectious disease, but do demonstrate that during active bacterial infections MAIT cells become depleted from the peripheral blood and those that remain display a dysfunctional phenotype. Using the mouse model, Sakala et al.37 found that peripheral MAIT cells up‐regulate integrins and cell trafficking molecules to migrate to the lungs soon after aerosolized mycobacterial infection. Given that the animal data suggest migration of peripheral MAIT cells to the lung upon respiratory exposure and protective activity of MAIT cells in the lungs, the lack of human data about the role of MAIT cells at mucosal sites during bacterial infection is a considerable gap in the field.

HIV

Investigations of MAIT cells from several cohorts of HIV‐infected people from around the world (USA, Europe, South Africa and Australia) have shown that people with chronic HIV infection have significantly lower levels of peripheral MAIT cells than comparable healthy controls.49, 50, 51, 52, 53, 54 Interestingly, the degree of MAIT cell depletion does not correlate to the usual markers of HIV disease progression, including the degree of peripheral CD4⁺ T‐cell depletion or the level of viral replication in the blood (HIV viral load).51, 54 Even individuals who appear to have a natural ability to control HIV infection (long‐term non‐progressors and elite controllers) have been found to experience similar levels of MAIT cell depletion to those observed in people with chronic progressive HIV infection.50 Leeansyah et al.26, 53 have found that the MAIT cells that remain in the peripheral circulation of patients with HIV are functionally impaired compared with MAIT cells from healthy people. Upon bacterial stimulation, they have found that MAIT cells from people with untreated chronic HIV produce lower levels of effector cytokines (IFN‐γ, TNF‐α, IL‐17) and cytolytic products (granzyme A, granzyme B, CD107a and perforin) upon stimulation with bacterial MR1 ligand.26, 53 Interestingly, these authors have found that in vitro treatment with IL‐7 has the capacity to restore the cytolytic response to bacterial exposure in MAIT cells from HIV‐infected people.26 In contrast, Fernandez et al.55found that MAIT cells from individuals with chronic untreated HIV retained comparable levels of IFN‐γ and TNF‐α production upon stimulation with purified MR1 ligand. The contrasting findings about the functionality of residual MAIT cell populations may be due to differences in the functional assays used by each group. Fernandez et al. stimulated MAIT cells from HIV‐positive patients with an hMR1 over‐expressing cell line that had been pulsed with purified synthetic MAIT cell activating antigen. In contrast, Leeansyah et al.26, 53 stimulated peripheral blood mononuclear cells with paraformaldehyde‐fixed whole E. coli. Differences in the strength of stimulation and cytokine co‐stimulation or inhibition may explain the discrepant findings. At this point, therefore, there is consensus in the finding of MAIT cell depletion but lack of consensus on the functional capacity of those MAIT cells that remain in circulation in patients with chronic HIV (Figure 2).

MAIT cell depletion appears to occur very early (within the first few weeks) after HIV infection.49, 54, 55 To date there have been no animal or human studies following individuals longitudinally before and after HIV infection to address the possibility that individuals with pre‐existing low levels of peripheral MAIT cells are more likely to become HIV infected. Nonetheless, the leading hypothesis in the field is that events surrounding acute HIV infection result in peripheral MAIT cell depletion. Circulating MAIT cells express high levels of CCR5, which is the co‐receptor that facilitates HIV entry into CD4⁺ T cells. However, Cosgrove et al.49 performed in vitro infections with a CCR5‐trophic HIV strain and found that peripheral MAIT cells are no more likely to be directly infected by HIV than other T cells. Finding that E. coli exposure induced caspase‐3‐mediated apoptosis in MAIT cells, they proposed that translocation of microbial products due to injury to the intestinal mucosa during acute HIV infection could result in uncontrolled activation and activation‐induced MAIT cell death.49 In support of this hypothesis, which was also put forth by Leeansyah et al.53, activation markers (CD38 and HLA‐DR) are elevated on MAIT cells from HIV‐infected individuals compared with healthy controls and levels of MAIT cell activation inversely correlate with MAIT cell frequency.50, 51, 56 Despite the fact that multiple groups have put forward the hypothesis that MAIT cell depletion is due to microbial translocation, no study to date has directly linked plasma markers of microbial translocation to MAIT cell loss or dysfunction. An alternative, and perhaps related, hypothesis to explain the absence of peripheral MAIT cells in HIV is that they redistribute to mucosal sites and become sequestered there. The human data on non‐peripheral MAIT cells in HIV are very limited. Eberhard et al.50 found that MAIT cells were reduced in the lymph nodes of HIV patients compared with uninfected controls. MAIT cell levels in the colon have been assessed by three groups and were also found to be lower in HIV‐infected compared with uninfected people during chronic infection.49, 52, 53 Recently, human MR1‐5‐OP‐RU tetramers were used to evaluate MAIT cell distribution in simian immunodeficiency virus‐infected rhesus macaques. Although limited by lack of species‐specific reagents, this study found low frequencies of MAIT cells in all examined tissue sites (peripheral blood, liver, mesenteric lymph nodes, jejunum and bronchoalveolar lavage) of simian immunodeficiency virus‐infected animals.57 There are therefore currently no data to support the redistribution hypothesis; however, with only two non‐blood sites having been investigated to date human data are limited.

Studies suggest that otherwise effective antiretroviral therapy that suppresses HIV viral load and restores peripheral CD4⁺ T cells does not appear to restore circulating MAIT cell frequencies.49, 50, 53, 54, 58, 59 Greathead et al.52 have reported that colonic MAIT cells were restored to levels comparable to those of healthy controls after long term ART. Although there is no consensus about the effect of HIV on residual MAIT cell function, the group that reported that cytokine and cytolytic function decreased during chronic untreated HIV infection has found that antiretroviral therapy partially restores functional capacity.53 Importantly there has been a single report of a patient who received antiretroviral therapy very early after HIV infection (Fiebig Stage III) who had relatively preserved MAIT frequencies.50

In summary, HIV infection is associated with early and largely irreversible depletion of MAIT cells from the peripheral circulation. The mechanism for this depletion is currently unknown and does not appear to be direct viral infection. Whether or not this peripheral depletion is reflective of universal MAIT cell depletion or whether it may be due to redistribution to mucosal sites has not been definitively addressed, although there is some intriguing data indicating that mucosal MAIT cells may be more responsive to antiretroviral therapy than circulating MAIT cells. The impact that HIV‐induced MAIT cell alterations have on overall immune function is unknown. Gaardbo et al.60 have found that low MAIT cells are associated with a constellation of immunological and metabolic derangements that are linked to immune activation and poor outcomes, but the specific role of MAIT cells in this is unclear. Based on the previous discussion of the role of MAIT cells as pre‐armed mucosal effectors that respond to bacterial and fungal infections early and contribute to the eventual quality of the adaptive immune response, it is hypothesized that derangements of normal MAIT cell frequencies, function and distribution in HIV may play a role in the increased vulnerability to mucosal bacterial and fungal infections that is seen both early in and throughout the course of HIV disease. Assessing this hypothesis and determining mechanisms and potential strategies for intervention are critical next steps for the field.

MAIT cells and viral hepatitis

MAIT cells are significantly enriched in the liver, where they have been found to comprise up to 50% of liver‐resident lymphocytes. They are found primarily in the biliary tract and in the context of liver infection can be activated by MR1‐presented bacterial ligand or indirectly via IL‐12 and IL‐18 produced by antigen‐presenting cells in response to Toll‐like receptor 8 signalling triggered by viral RNA.18, 61 Jo et al.18found that MAIT cells were enriched in normal and chronically hepatitis B virus‐infected livers but were relatively depleted in end‐stage liver disease. Billerbeck et al.17 found that CD161⁺⁺ CD8⁺ cells (probably MAIT cells) had higher capacity for IL‐17 and IL‐22 production in the liver compared with the peripheral blood and that dual functionality (ability to co‐express IL‐17 and IFN‐γ upon stimulation) of these cells was inversely correlated with clinical disease score in chronic untreated hepatitis C. In the peripheral compartment, Barathan et al.62 found that patients with chronic HCV had lower MAIT cells frequencies compared with healthy controls. Likely reflecting a state of chronic inflammation and infection, MAIT cells in these patients had higher levels of markers of activation and exhaustion (HLA‐DR, CD38, PD‐1, TIM‐3 and CTLA‐4) compared with healthy controls. van Wilgenburg et al.48 confirmed the finding that people with chronic HCV infection have depleted and activated peripheral MAIT cell populations; and interestingly demonstrated that MAIT cells have the ability to inhibit HCV replication in an IFN‐γ‐dependent manner. As discussed above, in patients with HIV infection, peripheral MAIT cell depletions are not improved by otherwise effective antiretroviral therapy. Two recent studies in HCV‐infected patients have similarly found that the MAIT cell depletion and dysfunction observed in chronic HCV is not reversible even after successful interferon therapy or interferon‐free therapy.63, 64 The role of MAIT cells in liver inflammation and the impact of MAIT cell perturbations on the general state of mucosal immunity in patients with these infections require further investigation.