Abstract
Mucosal-associated invariant T (MAIT) cells have a semi-invariant TCR Vα-chain, and their optimal development is dependent upon commensal flora and expression of the nonpolymorphic MHC class I–like molecule MR1. MAIT cells are activated in an MR1-restricted manner by diverse strains of bacteria and yeast, suggesting a widely shared Ag. Recently, human and mouse MR1 were found to bind bacterial riboflavin metabolites (ribityllumazine [RL] Ags) capable of activating MAIT cells. In this study, we used MR1/RL tetramers to study MR1 dependency, subset heterogeneity, and protective effector functions important for tuberculosis immunity. Although tetramer+ cells were detected in both MR1+/+ and MR1−/− TCR Vα19i-transgenic (Tg) mice, MR1 expression resulted in significantly increased tetramer+ cells coexpressing TCR Vβ6/8, NK1.1, CD44, and CD69 that displayed more robust in vitro responses to IL-12 plus IL-18 and RL Ag, indicating that MR1 is necessary for the optimal development of the classic murine MAIT cell memory/effector subset. In addition, tetramer+ MAIT cells expressing CD4, CD8, or neither developing in MR1+/+ Vα19i-Tg mice had disparate cytokine profiles in response to RL Ag. Therefore, murine MAIT cells are considerably more heterogeneous than previously thought. Most notably, after mycobacterial pulmonary infection, heterogeneous subsets of tetramer+ Vα19i-Tg MAIT cells expressing CXCR3 and α4β1 were recruited into the lungs and afforded early protection. In addition, Vα19iCα−/−MR+/+ mice were significantly better protected than were Vα19iCα−/−MR1−/−, wild-type, and MR1−/− non-Tg mice. Overall, we demonstrate considerable functional diversity of MAIT cell responses, as well as that MR1-restricted MAIT cells are important for tuberculosis protective immunity.
Introduction
Mucosal-associated invariant T (MAIT) cells display a limited TCR repertoire and are restricted by the nonpolymorphic MHC class I–like molecule MR1 (1). Human and mouse MR1 were recently shown to bind riboflavin (vitamin B2) and folic acid (vitamin B9) metabolites, but only riboflavin metabolites activated MAIT cells in vitro (2–7). Accumulating evidence predicts that MAIT cells are relevant for the control of microbial infection. First, there is a striking evolutionary conservation in mammals of both limited MAIT TCR usage and MR1 sequence, suggesting pathogen-driven purifying selection. More specifically, MAIT cells express structurally homologous invariant TCRα chains consisting of the TRAV1-2 segment (Vα7.2i in humans) and TRAV1 (Vα19i in mice) joined mostly to a TRAJ33 (Jα33) segment, resulting in a CDR3α of constant length (8). The Jα33-encoded CDR3α loop has three critical residues (Ser93α, Asn94α, Tyr95α) that engage both the α1 and α2 helices of MR1 (9). Of these, Tyr95α residue is the principal player in the invariant Jα33 use of the MAIT TCR, and it is conserved in non-TRAJ33 junctional genes (i.e., TRAJ20 and TRAJ12) expressed by a minor subset of human MAIT cells (3, 4, 10–12). In addition, the invariant TCRα of MAIT cells uses a broad TCR-β repertoire, but it is preferentially paired with limited Vβ segments TRBV6 (Vβ13) or TRBV20 (Vβ2) in humans and TRBV19 (Vβ6) or TRBV13 (Vβ8.1 and Vβ8.2) in mice (4, 8, 13–15). Interestingly, most of the residues of the MAIT TCR α-chain that contact MR1 are germline encoded, and the canonical CDR3α of MAIT cells is formed at a high frequency (3, 16). In addition, MR1 shares 80–98% amino acid sequence identity among mammals in its α1/α2 domains that interact with the MAIT TCR and/or antigenic riboflavin metabolites (3, 17). Thus, the MR1/MAIT cell Ag-presentation pathway has been strikingly conserved throughout mammalian evolution (18).
Vitamin B2 metabolites presented by MR1 appear to be the predominant Ags by which MAIT cells can detect a variety of microbes (2, 6). More specifically, Kjer-Nielsen et al. (2) found that the vitamin B9 metabolite 6-formylpterin (6-FP) bound human and mouse MR1, but it did not stimulate MAIT cells. In contrast, riboflavin intermediates, including reduced 6-hydroxymethyl-8-d-ribityllumazine (rRL-6HM), 7-hydroxy-6-methyl-8-d-ribityllumazine, and its precursor, 6,7-dimethyl-8-d-ribityllumazine, stimulated MAIT cells in an MR1-dependent manner. Structural studies showed that the form of Ag trapped by MR1 includes the relatively unstable adducts 5-(2-oxoethylideneamino)-6-d-ribitylaminouracil and 5-(2-oxopropylideneamino)-6-d-ribitylaminouracil (5-OP-RU), which are formed by the reaction between 5-amino-6-d-ribitylaminouracil and glyoxal or methylglyoxal, respectively (6). MR1 tetramers formed between MR1 and the synthetic preparation of rRL-6HM or 5-OP-RU give identical results (6).
Evidence that vitamin B2 metabolites are predominant MAIT cell Ags includes the observation that the diverse bacterial and yeast strains previously shown to activate MAIT cells in vitro have a vitamin B2 synthesis pathway, whereas microbes previously shown to not activate MAIT cells lack this synthesis pathway (19, 20). Expansion in response to commensal flora Ags explains why MAIT cells are abundant in mucosal tissues. Furthermore, human liver is also constantly exposed to bacterial products absorbed from the gut, likely explaining why MAIT cells can constitute as much as 45% of the total lymphocytes in human liver (21–23). In addition, MAIT cells represent up to 10% of the mature CD8+ and/or double-negative (DN) T cells in the blood of healthy individuals (8, 22).
Further supporting their antimicrobial activity, following in vitro TCR ligation, MAIT cells rapidly secrete the inflammatory cytokines IFN-γ, TNF-α, and IL-17 (24, 25). In addition, MAIT cells express chemokine receptors, indicating their migratory potential, and MAIT cell distribution is altered in several diseases (22). For example, patients infected with mycobacteria had increased numbers of MAIT cells in their infected lung and fewer MAIT cells in the blood compared with uninfected controls (19, 20, 26). In addition, sharp and nonreversible decreases in MAIT cells were found in the blood and tissues of patients with HIV monoinfection and HIV/tuberculosis (TB) coinfection (27–30). It was speculated that this loss of MAIT cells was caused by HIV infection inducing MAIT cell exhaustion from exposure to bacterial products and that loss of MAIT cells rendered HIV patients susceptible to opportunistic infections. MAIT cells also were implicated in other diseases secondarily affected by microbiota imbalance, such as multiple sclerosis, type 2 diabetes, and inflammatory bowel disease (31–34).
As a foundation for future studies of the clinical relevance of MAIT cells, genetically defined mouse models continue to provide key insights into the development and antimicrobial mechanism of MAIT cells. For example, MR1-knockout mice were used in a seminal report to show that optimal MAIT cell development is dependent on MR1 expression (35). In addition, MR1−/− mice were more susceptible than were wild-type (WT) mice to bacterial infection, providing, to our knowledge, the first in vivo evidence that MAIT cells control bacterial infection (20, 24, 36, 37). Because of the paucity of MAIT cells in laboratory mouse strains compared with human, three groups reported the production of Vα19i-transgenic (Tg) mice (13, 25, 38). These Vα19i-Tg mice recapitulate several features of the phenotypic heterogeneity of human MAIT cells. For example, Vα19i-Tg MAIT cells, like human MAIT cells, include CD8+ (CD8αα), CD4+, and CD4−CD8− DN subsets, although in different proportions (13, 28, 39). In addition, subsets of Vα19i-Tg mouse MAIT cells and human MAIT cells express a similar NK receptor (NK1.1 in mouse and CD161 in humans). Furthermore, like human MAIT cells, mouse Vα19i-Tg MAIT cells use diverse Vβ-chains in addition to the ones that are preferred (4, 19). These similarities in the phenotypic heterogeneity of human and Vα19i-Tg MAIT cells raise the largely unanswered question of the functional importance of different MAIT cell subsets.
In our initial mouse study, we showed that MR1-restricted MAIT cells confer early protection against intranasal Mycobacterium bovis infection, providing in vivo evidence that MAIT cells are critical for the prompt control of mycobacterial infection (24). In the current study, we used Ag-loaded native mouse MR1 tetramers to investigate the subsets of MAIT cells important for antimycobacterial immunity and gain insight into the mechanisms by which they mediate early protection against mycobacterial infection. We demonstrate that MR1/ribityllumazine (RL) tetramer+ mouse MAIT cells contain heterogeneous populations, as well as that MR1 expression leads to the development of more “classic” MAIT cell populations (increased Vβ6/8+NK1.1+ MAIT cells that can be activated by either MR1-restricted TCR signaling or innate cytokine stimulation). We further show that these classical MR1-restricted MAIT cells are capable of providing important in vivo antimycobacterial effects.
Materials and Methods
Mice
The Vα19i-Tg TCRα-chain deficient (Cα−/−) mice used in this study (on a C57BL/6 [B6] genetic background) were described previously (25). The Vα19i-Tg B6 MR1-sufficient mice (MR1+/+) were crossed with Cα−/− B6 mice (B6.129S2-Tcratm1Mom/J; The Jackson Laboratory) to eliminate the expression of endogenous TCRα-chains and generate Vα19iCα−/−MR+/+ mice. These Vα19iCα−/−-Tg mice were further crossed with B6 MR1−/− mice to generate Vα19iCα−/−MR1−/− mice (25). Therefore, the Vα19i-Tg MR1-sufficient and -deficient mice that we used in all of our experiments exclusively express a Vα19i transgene that is the canonical TCR Vα of mouse MAIT cells. As previously described, MAIT cells isolated from these Tg mice use endogenous heterogeneous TCR Vβ-chains; thus, Vα19i-Tg T cells have characteristics of polyclonal MAIT cells (25). Inbred B6 MR1−/−, Vα19iCα−/−MR1+/+, and Vα19iCα−/−MR1−/− mice were a generous gift from Dr. Susan Gilfillan (Washington University) to T.H.H. B6.129P2-H2-Kbtm1 H2-Dbtm1 N12 (H2-Kb−/− H2-Db−/−, homozygous H2-Kb H2-Db double-knockout) mice were a generous gift from Dr. Wayne Yokoyama (Washington University). H2-Kb−/− H2-Db−/− mice are virtually devoid of class Ia cell surface molecules (40). WT B6 mice were purchased from The Jackson Laboratory. This animal study was performed in strict compliance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health, Office of Laboratory Animal Welfare. Mice were bred and maintained under specific pathogen–free conditions, and experimental procedures used as described in the study protocol were approved by the Washington University’s Institutional Animal Care and Use Committee, the Animal Studies Committee, and the Saint Louis University’s Animal Care and Use Committee.
Purification of splenic Vα19i-Tg T cells, conventional CD4+, CD8+ T cells, and innate type I NKT and NK cells
Single-cell suspensions were prepared from spleens aseptically removed from naive mice. After RBCs were lysed with ammonium chloride buffer, Vα19i-Tg T cells, CD4+ and CD8+ T cells, and NK cells were purified by negative selection using immunomagnetic isolation kits (Miltenyi Biotec), following the manufacturer’s instructions. The isolation of invariant NKT (iNKT) cells was performed in a two-step procedure. First, the non-NK1.1+ iNKT cells were labeled with a mixture of biotin-conjugated Abs and subsequently depleted by separation over a MACS column. In the second step, the NK1.1+ iNKT cells were positively isolated with anti-NK1.1–allophycocyanin and anti-allophycocyanin MicroBeads. The purity was checked by flow cytometry and was consistently >98%.
Isolation of liver T lymphocytes by cation chelator-containing buffer
Liver T lymphocytes were isolated by perfusion with 10 mM EDTA/PBS, according to the protocol described by Fang et al. (41). Briefly, 100 μl 14.8 mg/ml sodium pentobarbital (75 mg/kg final dosage) and 3–5 μl Heparin Sodium Injection were injected via the i.p. route. Once the mouse was unconscious, the abdomen and thorax were opened rapidly to expose the superior vein cava. The superior vein cava was tied carefully using A184 surgical suture. Both right and left kidney renal veins were tied carefully. With the intestines reflected to the mouse’s left, a 23G scalp vein set (BD 367285) was carefully inserted into the portal vein, and the scalp vein set was held in place with an artery clamp (Bulldog Serrefines, F.S.T cat. no. 18050-28). Through the mouse hind fat tissue, the inferior vena cava was cannulated with an 18G needle. The liver was perfused with 10 ml PBS with CaCl2 and MgCl2 via the scalp vein set, and the flow out blood was collected with a 100-mm petri dish under the 18G needle. Intrahepatic T cells were isolated by perfusion with 20 ml 10 mM EDTA/PBS without CaCl2 and MgCl2. This second flow out was collected and transferred into a 50-ml centrifuge tube and centrifuged for 5 min at 400 × g. The cell pellet was resuspended with 1 ml ACK Lysis buffer to lyse RBCs. Cells were washed with DMEM supplemented with 10% FBS and counted.
Preparation of murine bone marrow–derived macrophages
Bone marrow cells isolated from femora of B6 WT, B6 MR1−/−, and B6 H2-Kb−/− H2-Db−/− mice were cultured (2.5 × 104 cells/well) in 10% FBS DMEM with 20 ng/ml murine rM-CSF (PeproTech) for 7 d. After 7 d of culturing, bone marrow–derived macrophages (BMDMΦs) formed a confluent monolayer with an estimated density of 2.5 × 105 cells/well.
IFN-γ response to IL-12/IL-18 cytokines
Purified naive T cells were cultured in 96-well plates (2 × 105
Vitamin B2 metabolites
Activating and nonactivating synthetic vitamin B metabolites were described previously (2, 6). In this study, we used synthetic rRL-6HM to make first-generation (1G) MR1/RL tetramers and to stimulate MAIT cells in vitro; 5-OP-RU was used as the stimulatory vitamin B2 metabolite to make second-generation (2G) MR1/RL tetramers.
Generation of MR1(K43A)–rRL-6HM (1G) and MR1(WT)–5-OP-RU (2G) tetramers to detect MAIT cells
The generation of MR1(K43A) tetramers loaded with synthetic rRL-6HM and WT MR1 tetramers loaded with 5-OP-RU was described previously (4, 6). Briefly, refolded and purified empty C-terminal cysteine-tagged-MR1(K43A) was loaded with a 136 M excess of synthetic rRL-6HM for 4 h at room temperature in the dark. Cysteine-tagged WT MR1–5-OP-RU was reduced with 5 mM DTT for 20 min before buffer exchange into PBS using a PD-10 column (GE Healthcare). Negative-control MR1 tetramers were MR1(K43A)-empty tetramers (for 1G) and MR1(WT) loaded with the nonstimulatory 6-FP (for 2G). MR1(K43A)-empty, MR1(K43A)–rRL-6HM, MR1(WT)–6-FP, or MR1(WT)–5-OP-RU were biotinylated with Maleimide-PEG2 biotin (Thermo Scientific) with a 30:1 M ratio of biotin/protein at 4°C for 16 h in the dark. Biotinylated MR1 was subjected to S200 10/300 GL (GE Healthcare) chromatography to remove excess biotin. Biotinylated monomers were tetramerized with streptavidin conjugated to PE (BD Pharmingen).
MAIT cell activation by the soluble rRL-6HM vitamin B2 metabolite
Prior to coculture, CH27-mMR1 cells, a mouse B cell lymphoma transduced to overexpress mouse MR1 (42–44), were incubated with either medium (no Ab control) or 10 μg/ml anti-MR1–blocking Abs (clones 26.5 and 8F2.F9) for 1 h at 37°C. These CH27-mMR1 cells were cocultured with sorted tetramer+ Vα19i-Tg T cells in 96-well plates (Corning) at an APC/T cell ratio of 1:5 along with the synthetic vitamin B2 metabolite rRL-6HM (76.2 μM final concentration) for 60 h at 37°C in the absence or presence of 10 μg/ml anti-MR1–blocking Abs. MAIT cell activation was determined using Milliplex MAP mouse cytokine/chemokine panel I and CD8+ T cell assay kits (EMD Millipore).
MR1/RL tetramer staining and flow cytometry
Approximately 1 × 106 T cells were washed with cold FACS staining buffer (1× PBS with 2% FBS) and stained with either PE-conjugated control unloaded (empty) mouse MR1(K43A) tetramer or PE-conjugated rRL-6HM–loaded mouse MR1(K43A) tetramer at 20 μg/ml [1G tetramers (4)] or with PE-conjugated nonstimulatory control mouse MR1(WT)–6-FP tetramer or PE-conjugated mouse MR1(WT)–5-OP-RU tetramer at 1.4 μg/ml [2G tetramers (6
Mycobacteria
Mycobacterium bovis TMC 1010 (ATCC 35733; bacillus Calmette–Guérin [BCG] Danish) and M. tuberculosis strain Erdman (BEI Resources NR-15404 or ATCC 35801) were grown in Middlebrook 7H9 broth (BD Difco) supplemented with 10% oleic acid–albumin, dextrose, and catalase enrichment and 0.05% Tween 80. When the bacterial density reached log-phase growth, mycobacteria were harvested and stored in PBS at −80°C. The concentration (CFU/ml) was quantified on Middlebrook 7H10 agar (BD Difco) plates. Before being used for infection, bacteria were thawed, sonicated to obtain a single-cell suspension, and diluted appropriately in antibiotic-free complete RPMI 1640 medium.
In vivo mycobacterial infection
Eight- to ten-week-old B6 MR1−/−, B6 WT, Vα19iCα−/−MR1−/−, and Vα19iCα−/−MR1+/+ mice were infected with BCG Danish via the intranasal route. Groups of five mice were inoculated with 107 CFU/mouse of the BCG Danish suspension in 50 μl (25 μl/nostril) under anesthesia. On day 10 postinfection, mice were euthanized, and peripheral blood, bronchoalveolar lavage fluid (BALF) cells, and lungs were collected. The collected total BALF cells were enumerated by a trypan blue staining technique. The frequency of tetramer-reactive T cells in BALF and peripheral blood was determined by FACS, as described above. In the aerosol infection, M. tuberculosis strain Erdman was diluted to 107 CFU/ml in PBS with 0.04% Tween 80. A nebulizer (CH Technologies, Westwood, NJ) was set to liquid feed at 1 ml/min and airflow at 1 l/min to expose B6 WT or Vα19iCα−/−MR1+/+ mice to M. tuberculosis Erdman. A 20-min exposure to the aerosol resulted in the delivery of 100–300 infectious bacilli/mouse lung. Mycobacterial load was determined in lung homogenates on Middlebrook 7H10 agar plates supplemented with 10% oleic acid–albumin, dextrose, and catalase enrichment, as previously described (24).
Activation of MR1−/− Vα19i-Tg T cells by MHC class Ia–restricted epitopes
B6 WT, B6 MR1−/−, and B6 H2-Kb−/− H2-Db−/− BMDMΦs, prepared as described above, were infected with BCG Danish at a multiplicity of infection of 3:1 (bacteria/macrophage) in antibiotic-free 10% FBS DMEM at 37°C in 5% CO2. After overnight culture, infected BMDMΦs were washed with 10% FBS DMEM medium to remove the extracellular BCG before coculture with purified Vα19iCα−/−MR1−/−CD8+ T cells or total Tg T cells at a ratio of 1:1. Activation of Vα19iCα−/−MR1−/−CD8+ T cells was assessed by intracellular IFN-γ and MIP-1α production after overnight (24 h) coculture at 37°C in 5% CO2. Inhibition of intracellular BCG growth in BMDMΦs, by purified total T cells as a measure of their function, was determined at the end of a 72-h coculture. Cocultured cells were lysed with 0.2% saponin to release intracellular BCG. Mycobacteria were radiolabeled with tritiated uridine, as previously described (24, 45), to determine the viability of bacteria in B6 WT, B6 MR1−/−, and B6- H2-Kb−/− H2-Db−/− macrophages. Briefly, [5,6-3H]uridine (Perkin Elmer, Waltham, MA), at 1 μCi/ml and prepared in Middlebrook 7H9 broth supplemented with 10% ADC enrichment medium (BD BBL 211887; Becton Dickinson, Franklin Lakes, NJ), was added to saponin lysates. After incubation at 37°C for 72 h, the tritiated uridine–pulsed mycobacteria were harvested (TomTec Harvester 96 MACH III M; TomTec, Hamden, CT) onto glass fiber filters (Perkin Elmer), and radioactivity was quantitated by liquid scintillation counting (Wallac 1450 MicroBeta TriLux; Perkin Elmer). The percentages of BCG growth inhibition were calculated using the following formula: percentage inhibition 100 − [100 × (dpm from T cell–infected macrophage-cocultured wells/dpm from infected macrophage-alone wells)].
Statistical analysis
Statistical analyses of experimental data were performed using GraphPad Prism (GraphPad Software, San Diego, CA). For comparisons, the unpaired two-tailed t test with the Welch correction, Mann–Whitney U test, and two-way ANOVA were used.
Results
Phenotypic characterization of MR1/RL tetramer+ T cells in uninfected Vα19i-Tg mice on both MR1+/+ and MR1−/− genetic backgrounds
Because of the low frequency of MAIT cells in standard laboratory strains of mice and the lack of specific Abs to the mouse Vα19i chain, identification of mouse MAIT cells in WT mice has been extremely difficult. Vα19i-Tg mice were shown to have an increased frequency of MAIT cells, as determined by the biased Vβ6 or Vβ8.1/8.2 (Vβ6/8) chain usage in Vα19iCα−/−MR1+/+ mice (25). However, Vβ6/8 chains can pair with the Vα19i chain in Vα19iCα−/−MR1−/− mice, and the Vα19i chain can pair with other Vβ-chains in Tg mice, obscuring identification of the classically described MR1-dependent MAIT cells. Furthermore, NK cell activation markers (e.g., NK1.1 in the mouse and CD161 in humans) have been associated with classic MAIT cell populations, but whether the expression of these NK markers can identify the functionally important MAIT cell lineage or simply indicates the previous activation status of peripheral MAIT cells remains unknown. The availability of mouse MR1/RL tetramers offers, for the first time to our knowledge, an approach to detect MAIT cells based on their RL-specific TCR and determine how this Ag specificity relates to Vβ6/8 and NK1.1 expression and MAIT cell functional responses.
Using 1G MR1/RL tetramers in which MR1 K43A mutated molecules were loaded with or without the vitamin B2 metabolite rRL-6HM, 16.3 ± 1.6 and 4.5 ± 1.0% of splenic CD3+ T cells from Vα19iCα−/−MR1+/+ and Vα19iCα−/−MR1−/− mice, respectively, were stained by RL-loaded MR1 tetramers but not by RL-empty tetramers (Fig. 1A, 1C, upper panels). These results indicated that mouse MR1/RL tetramer can specifically detect MR1/RL-reactive T cells in the Vα19i TCR-Tg mouse. To further characterize phenotypically and functionally MR1/RL-reactive cells, as well as to understand the requirements for MR1 and CD4/CD8 coreceptors in the development of mouse MAIT cells, a more stable MR1/RL tetramer (2G) was used in the current study (6). We used the gating strategy shown in Supplemental Fig. 1A, 1B to analyze the lymphocyte population for the frequency and phenotype of tetramer+CD3+ T cells in the thymus, two secondary lymphoid organs (mesenteric lymph nodes [mLNs] and spleen), blood, and the liver peripheral organ. Similar to what was observed using the empty 1G tetramer, the negative-control 2G tetramers loaded with nonantigenic 6-FP stained <1% of CD3+ splenocytes from both Vα19iCα−/−MR1+/+ and Vα19iCα−/−MR1−/− mice. However, the 2G tetramer loaded with antigenic 5-OP-RU stained a markedly higher percentage of splenic CD3+ T cells (50.2 ± 0.5% in Vα19iCα−/−MR1+/+ mice and 32.0 ± 0.6% in Vα19iCα−/−MR1−/−) (Fig. 1B, 1C, lower panels). Thus, the 2G tetramers were more sensitive while retaining antigenic specificity. Of note, although a higher proportion of tetramer+ splenic T cells was detected in Vα19iCα−/−MR1−/− mice with the 2G tetramers than with the 1G tetramers, both the percentages and the numbers stained were significantly lower than those in Vα19iCα−/−MR1+/+ mice (Fig. 1C, lower panels). Tetramer+ MAIT cells also were detected in the thymus, mLN, peripheral blood, and the liver, and they ranged from the lowest percentage (45 ± 1.2%) among mature (CD3high) thymocytes to the highest frequency (60 ± 1.9%) in the liver of Vα19i-Tg MR1-sufficient mice (Supplemental Fig. 1B, left panels, Supplemental Fig. 1C, top panel). In contrast, in Vα19i-Tg MR1-deficient mice, we observed a gradual decrease in the proportion of tetramer+ cells when comparing thymocytes and liver peripheral tissue (Supplemental Fig. 1C, top panel). Furthermore, the proportions of tetramer+ cells were significantly lower in all tissues of MR1-deficient Tg mice compared with MR1-sufficient Tg mice. Most impressively, the diminution in the proportion of liver tetramer+CD3+ T cells was 56% less compared with that in Vα19iCα−/−MR1+/+ mice. Therefore, our results are consistent with the previous findings that optimal peripheral development/expansion of MAIT cells depend on MR1 expression. Further demonstrating the specificity of the 2G tetramers, neither purified NK nor type I NKT cells were stained (Fig. 1D). A <5% staining of type I NKT cells is an acceptable background, because the negative control MR1–6-FP tetramer similarly stained <5% of these cells, demonstrating nonspecific low-level staining. Also, consistent with the known low frequency of MAIT cells in specific pathogen–free laboratory mice, the proportion of splenic T cells reactive with 2G MR1/RL tetramers was not different from the control MR1–6-FP tetramer-stained cells compared with Vα19iCα−/−MR1+/+ T cells (Fig. 1D). These data demonstrate that the tetramer staining of Vα19i-Tg T cells is not attributable to nonspecific binding by the MR1/RL tetramers. Thus, the specificity and increased sensitivity of 2G tetramers provide the unique potential for defining the phenotypic, developmental, and functional heterogeneity of MAIT cell subsets in Vα19i-Tg mice, including RL-reactive T cells that develop in the absence of MR1.
Mouse MR1/RL tetramers stain phenotypically diverse subsets of CD3+ cells in Vα19i-Tg mice on both MR1+/+ and MR1−/− genetic backgrounds. Purified splenic T cells from Ag-naive Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice were stained with anti-mouse CD3ε, CD4, CD8α, NK1.1 (clone PK136), Vβ6/8.1-2 TCR, and mouse MR1 tetramers and analyzed by flow cytometry. (A) Representative FACS plots of CD3ε (x-axis) versus tetramer (y-axis) gated on splenic CD3+ lymphocyte population for CD3+ T cells stained with 1G mouse MR1 tetramers loaded with “empty” as negative control or loaded with antigenic rRL-6HM. (B) Representative FACS plots of splenic CD3+ T cells stained with 2G mouse MR1 tetramers, loaded with nonantigenic 6-FP as negative control or loaded with antigenic 5-OP-RU. The percentages of stained cells are indicated. (C) Percentages (upper left panel) and absolute numbers (upper right panel) of 1G tetramer+ CD3+ cells in Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. Percentages (lower left panel) and absolute numbers (lower right panel) of 2G tetramer+ CD3+ cells in Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. (D) Staining of purified splenic NK, type I NKT, and bulk T cells from non-Tg WT B6 mice with murine MR1–6-FP or MR1–5-OP-RU tetramers compared with Vα19iCα−/−MR1+/+ T cells. (E) Representative FACS plots of percentage of Vβ6/8+ (left panels) or NK1.1+ (right panels) of 2G tetramer+ CD3+ T cells in Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. (F) Absolute number of Vβ6/8+ (left panels) or NK1.1+ (right panels) of 2G tetramer+ CD3+ T cells in Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. Data shown are representative of three separate experiments. **p < 0.01, Mann–Whitney U test (n = 5/group).
Because Vβ6 and Vβ8 are the major Vβ-chains previously found to pair with Vα19i, we next analyzed the expression of Vβ6/8 in tetramer+ MAIT cells. In the thymus, mLN, spleen, and peripheral blood of Vα19iCα−/−MR1+/+ mice, ∼31–42% of tetramer+ T cells coexpressed Vβ6/8 compared with 25–28% in Vα19iCα−/−MR1−/− mice (Fig. 1E, 1F, Supplemental Fig. 1B, middle column, Supplemental Fig. 1C, middle panel). A relatively higher level of Vβ6/8 expression on tetramer+CD3+ T cells developing in the presence of MR1 was seen in the liver (53 ± 3.1%) (Supplemental Fig. 1B, middle column, Supplemental Fig. 1C, middle panel). Vβ6/8 expression on liver tetramer+ MAIT cells was 1.7-fold higher than on tetramer+ mature thymocytes, and there also was a 3-fold reduction in Vβ6/8 pairing (17 ± 0.9%) in the liver tetramer+ T cells in Vα19iCα−/−MR1−/− mice. We also compared Vβ6/8 pairing in immature (CD3−CD4+CD8+ and CD3low/intermediateCD4+CD8+) and mature (CD3high) thymocytes (46) in the two Vα19i-Tg mice strains. The level of Vβ6/8 expression was very low (<2%) on both subsets of immature thymocytes (data not shown). However, marked increases in Vβ6/8 expression were seen on mature thymocytes. In Vα19iCα−/−MR1+/+ mice, 31 ± 0.7% of tetramer+ mature thymocytes expressed Vβ6/8 compared with 25 ± 0.5% in Vα19iCα−/−MR1−/− mice (Supplemental Fig. 1C, middle panel). Similar differences in Vβ6/8 coexpression by tetramer+ T cells were seen when comparing CD4+, DN, and CD8+ subsets in Vα19i-Tg MR1-sufficient and -deficient mice (data not shown).
In addition, we determined NK1.1 expression on tetramer+CD3+ cells in the thymus, mLN, spleen, peripheral blood, and liver tissue. In MR1+/+ and MR1−/− Vα19i-Tg mice, the level of NK1.1 expression was similar (<3%) on tetramer+ mature thymocytes, but it progressively increased in secondary lymphoid organs, blood, and liver (Fig. 1E, 1F, Supplemental Fig. 1B, right column, Supplemental Fig. 1C, bottom panel). This steady increase in the proportions of NK1.1+tetramer+ T cells from the most naive to the most peripherally expanded/activated populations was most remarkable in MR1-sufficient Tg mice. In these mice, the liver contained 63 ± 4.9% NK1.1+tetramer+ cells, followed by blood (48 ± 1.4%), spleen (18 ± 1.9%), and mLN (11 ± 1.1%). It is important to point out that large fractions of CD3+tetramer+ T cells in all tissues were both Vβ6/8− and NK1.1−. These results indicate that, as we (4) and other investigators (47) reported recently for human MAIT cells, murine MAIT cells, as defined by MR1/RL tetramer staining, are considerably more heterogeneous than previously thought. Nonetheless, our data confirm that the classical murine Vβ6/8+NK1.1+ MAIT cells are dependent on MR1 for optimal development.
To identify CD4+, CD8+, and DN subsets of splenic MR1/RL-reactive MAIT cells in uninfected control mice, we used the gating strategy shown in Fig. 2A and 2B. Representative FACS plots in Fig. 2A show the gating of splenic CD3+ T subsets and indicate that the fraction of Va19i-Tg T cells that expresses neither CD4 nor CD8 coreceptors represents ∼50–55% of total CD3+ T cells. As previously shown by Gilfillan and colleagues (25), most mature CD3+ T cells in both MR1+/+ and MR1−/− Va19i-Tg mice are the DN subset. This was not unexpected because of the decreased thymic DN-to-CD4/CD8 double-positive (DP) developmental transition seen in most TCR-Tg mouse systems due to early expression of the TCRα-chain (48, 49). Despite this limitation, the expressed TCRα-chain can pair with a diverse repertoire of endogenous TCR β-chains for cell surface expression (49). The 2G MR1/RL tetramers stain ∼50% of these total T cells, indicating that about half of splenic DN Vα19i-Tg T cells are RL-reactive MAIT cells. These data may suggest that, although the vitamin B2 metabolites are the predominant Ags, other MAIT cell ligand(s) have yet to be discovered. In Vα19iCα−/−MR1+/+ mice, the 2G MR1/RL tetramers stained the following percentages of CD3+ T cell subsets: 43–50% in the thymus, 47–56% in secondary lymphoid tissues/blood, and 54–62% in the liver peripheral organ (Fig. 2C). We found that the proportions of tetramer+ cells were higher in Vα19iCα−/−MR1+/+ mice than in Vα19iCα−/−MR1−/− mice in all CD3+ T cell subsets, with the exception of thymic CD8+ T cells (Fig. 2C). Most remarkably, tetramer+ DN T cells were reduced in Vα19iCα−/−MR1−/− mice (Fig. 2C, 2D). These data indicate that MR1 is important for the optimal development of all MAIT cell coreceptor subsets, but it has the biggest impact on the development of the DN subset of tetramer+ MAIT cells. Tetramer+ T cells in the tested tissues in Vα19iCα−/−MR1+/+ mice were mainly DN, with fewer CD4+ and CD8+ cells (Fig. 2D, upper row). This relative distribution of MR1/RL-reactive MAIT cell coreceptor expression is similar to that reported by Gilfillan and colleagues (25), identifying MAIT cells based on biased Vβ6/8 expression in Vα19iCα−/−MR1+/+ mice. Interestingly, we found an increase in the relative frequency of the tetramer+CD8+ subset in Vα19i-Tg MAIT cells developing in the absence of MR1 (Fig. 2D, lower row), suggesting the possibility that, in the absence of MR1, MR1/RL-reactive cells can be selected/expanded by interacting with classical MHC class Ia molecules.
Frequency of CD3+ tetramer+ MAIT cell subsets in the thymus (CD3high thymocytes), mLN, spleen, blood, and liver in Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. (A) Representative FACS plots of CD4 versus CD8α gated on CD3+ lymphocyte population for CD4−CD8− (DN), CD4+CD8−, and CD4−CD8+ coreceptor CD3+ subsets in the spleen of Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. (B) Staining of tetramer+ cells among CD4+ (left panels), DN (middle panels), or CD8+ (right panels) CD3+ splenocytes from Vα19iTgMR1+/+ (upper panels) or Vα19iCα−/−MR1−/− (lower panels) mice. (C) Percentages (mean ± SEM) of tetramer+CD4+ T cells (left panel), tetramer+ DN T cells (middle panel), and tetramer+CD8+ T cells (right panel) in the indicated tissues. (D) Pie charts comparing the relative distribution of absolute numbers (AN) of tetramer+CD3+ T cell subsets in the indicated tissues of Vα19iCα−/−MR1+/+ (upper panels) and Vα19iCα−/−MR1−/− (lower panels) mice. Data are from more than three separate experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Mann–Whitney U test and two-way ANOVA (multiple comparison test), Vα19iCα−/−MR1+/+ versus Vα19iCα−/−MR1−/− mice. ns, not significant.
Innate functions of tetramer+ T cells in Vα19iCα−/−MR1+/+ and Vα19iCα−/−MR1−/− mice
A pivotal function of MAIT cells is their ability to rapidly release cytokines in response to innate stimuli. We reported that mouse and human MAIT cells can respond to IL-12 and IL-12 plus IL-18 in a TCR-independent manner (24, 50). In this study, we extend these findings to investigate the APC and MR1 dependence of these innate cytokine responses of MAIT cells. After IL-12 treatment, Vα19i-Tg T cells, cultured or not with uninfected BMDMΦs, secreted comparable amounts of IFN-γ. However, Vα19i-Tg T cells that developed in MR1-expressing mice secreted significantly higher levels of IFN-γ than did Vα19i-Tg T cells from MR1-deficient mice (representative experiment shown in Fig. 3Ai). We next extended these findings to test the response of Vα19i-Tg T cells to IL-12 plus IL-18, which are known to synergistically enhance IFN-γ responses of NK and type I NKT cells (51, 52). IFN-γ was not induced in Vα19i-Tg T cells from MR1-sufficient mice following stimulation with IL-18 alone (data not shown), whereas the combined stimulation with 500 pg/ml of IL-12 resulted in strikingly higher levels of IFN-γ production (Fig. 3Aii); as reported previously, anti-MR1–blocking mAb did not inhibit these innate responses (24, 50). Of note, the responses of Vα19iCα−/−MR1+/+ T cells to IL-12 plus IL-18 were comparable to the responses of splenic NK and type I NKT cells purified from WT B6 mice (Fig. 3Aii). In contrast, purified splenic CD4+ or CD8+ T cells from B6 mice secreted very low levels of IFN-γ after IL-12 plus IL-18 treatment. Interestingly, Vα19i-Tg T cells from MR1-knockout mice responded to IL-12 plus IL-18 significantly better than did conventional CD4+ and CD8+ T cells; however, the responses to IL-12 and IL-18 by Vα19i-Tg T cells developing in the absence of MR1 were significantly lower than from Vα19i-Tg T cells developing in the presence of MR1.
Responsiveness of Vα19i-Tg T cells to IL-12 plus IL-18. (Ai) Splenic T cells from naive Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice were stimulated with the indicated doses of IL-12 alone in the presence or absence of BMDMΦs (w/MΦ or w/o MΦ, respectively). Data shown are from one of two independent experiments with similar results. (Aii) Splenic T cells (Tc) from naive Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice, conventional CD4+ and CD8+ αβ T cells, type I NKT (iNKT) cells, and NK cells from naive B6 WT mice were stimulated with 500 pg/ml IL-12 plus IL-18 (5–1000 pg/ml) in the absence of BMDMΦs for 24 h. IFN-γ in triplicate culture supernatants was determined by ELISA. (B) FACS plots showing the percentage of intracellular IFN-γ produced by activated (CD69+) sort-purified tetramer+ T cells after 24 h of stimulation with 500 pg/ml IL-12 plus 1000 pg/ml IL-18. Data are representative of three experiments with similar results. (C) Data from two separate experiments showing mean fluorescent intensity (MFI) for CD69 staining on Vβ6/8+ and Vβ6/8− bulk Vα19i T cells (top panel), FACS-sorted tetramer+ cells (middle panel), or FACS-sorted tetramer− cells (bottom panel) from naive Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. (D) Percentage of intracellular IFN-γ produced by tetramer+Vβ6/8+NK1.1+ cells (upper panel) or tetramer+Vβ6/8−NK1.1− cells (lower panel) after 24 h of stimulation with 500 pg/ml IL-12 plus 1000 pg/ml IL-18. Tetramer+Vβ6/8+NK1.1+ cells are the best responder subset to IL-12 plus IL-18, and the response depends on the presence of MR1 during development. *p < 0.05, **p < 0.01, ***p < 0.001, unpaired two-tailed t test.
We observed that tetramer+CD3+ T cells in Vα19iCα−/−MR1+/+ mice expressed more CD44 than did tetramer+CD3+ T cells in Vα19iCα−/−MR1−/− mice in secondary lymphoid organs and liver peripheral tissue (Supplemental Fig. 2A). CD44 expression indicates a memory-like cell phenotype, which is more likely to readily produce IFN-γ in response to IL-12 and IL-18 and/or Ag TCR signals. Perhaps surprisingly, Vα19iCα−/−MR1−/− T cells produced more IFN-γ relative to CD4+ and CD8+ αβ T cells from naive B6 WT mice. However, it was shown that naive conventional αβ T cells, unlike previously activated cells, are less responsive to either IL-12 or IL-18 because of low levels of cell surface cognate cytokine receptors (53). Activation of CD4/8 T cells with anti-CD3 and anti-CD28 induces IL-12R expression and responsiveness to IL-12 (53). Tomura and colleagues (53) showed that TCR triggering and CD28 costimulation followed by IL-12 stimulation induces IL-18R expression and responsiveness to IL-18. Although to a significantly lesser extent than Vα19iCα−/−MR1+/+ T cells, MR1−/−Vα19i-Tg T cells exhibited a more memory-like phenotype (CD44+) than did naive CD4/8 αβ T cells. Because of this phenotypic difference, a higher basal level of IL-12 and/or IL-18 receptors on Vα19iCα−/−MR1−/− T cells may explain their more rapid and greater responsiveness to IL-12 plus IL-18. The relevance of these results is that, like other innate T cells, MAIT cells can display immediate function upon infection and participate in the shaping of the adaptive immune response. The detection of lower-level responses in Vα19iCα−/−MR1−/− T cells compared with Vα19iCα−/−MR1+/+ T cells confirms that the presence of MR1 is important for optimal development of MAIT cell innate responsiveness to these proinflammatory cytokines.
Indeed, tetramer+ MAIT cells in MR1-sufficient mice were found to have a preactivated phenotype (i.e., >75% expressed CD69 even before in vitro stimulation). Moreover, CD44 and CD69 were coexpressed at 2-fold greater levels on Vα19iCα−/−MR1+/+ tetramer+CD3+ T cells than on Vα19iCα−/−MR1−/− T cells (data not shown). To demonstrate that previously activated MAIT cells developing with MR1 responded optimally to innate cytokines, we performed intracellular IFN-γ staining on sort-purified tetramer+ T cells. CD69+tetramer+ cells likely represent a more recently preactivated effector memory–like population capable of producing IFN-γ after the addition of IL-12 plus IL-18 (Fig. 3B). We also tested for a panel of cytokines (including IFN-γ) and chemokines in culture supernatant using a Multiplex Beads Array kit (EMD Millipore). Tetramer+ T cell subsets from Vα19iCα−/−MR1+/+ mice produced higher amounts of IFN-γ than did MR1−/− cells after stimulation with IL-12 plus IL-18 (Supplemental Fig. 2B). In Vα19iCα−/−MR1+/+ and Vα19iCα−/−MR1−/− mice, tetramer+ DN and CD8+ T cell subsets were the predominant producers of IFN-γ. In addition, MR1+/+tetramer+ DN subsets produced small amounts of GM-CSF, IL-2, IL-4, IL-13, and IL-17, whereas CD8+ subsets produced some GM-CSF, IL-2, and IL-17 (Supplemental Fig. 2B, upper panel). Furthermore, MR1-sufficient mice had significantly higher levels of CD69 and higher proportions of CD69+ preactivated MAIT cells compared with MR1-deficient mice (Fig. 3C, Supplemental Fig. 2C). As illustrated in Fig. 3C (middle and bottom panels), the levels of CD69 were much higher in MR1+/+tetramer+ MAIT cells than in MR1−/− tetramer+ cells or MR1+/+ tetramer− (non–RL-reactive) cells, and the levels of CD69 in the tetramer+ cells from MR1−/− mice are similar to the levels observed in tetramer− cells from either Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. Therefore, acquisition of this semiactivation status in vivo and optimal responsiveness to innate cytokines in vitro are likely dependent upon the interaction of TCR with MR1/Ag. In addition, the development of innate cytokine responsiveness in MR1-sufficient mice was greatest in the classic tetramer+ MAIT cell populations that coexpressed Vβ6/8 and NK1.1 (Fig. 3D). Taken together, these findings demonstrated that the presence of MR1 is important for the development of MAIT cell optimal responses to innate cytokines, presumably because of a previous in vivo MR1/RL activation. However, after in vivo activation, the responsiveness to innate cytokines is independent of TCR–MR1 engagement.
Vα19i-Tg cells that develop in MR1+/+ mice functionally respond to MR1-restricted RL Ag more robustly than those developing in MR1−/− mice
To address the question of whether tetramer+ cells from MR1-sufficient and -deficient mice are functionally responsive to vitamin B2 metabolites, FACS-sorted tetramer+ T cell subsets from the blood of Vα19iCα−/−MR1+/+ and Vα19iCα−/−MR1−/− mice were tested for their response to the potent rRL-6HM Ag (2). The sorted tetramer+ cells were rested in medium overnight before the addition of CH27-mMR1 cells as APCs. This CH27 cell line overexpressing murine MR1 was shown in previous work (2, 18, 20) to optimize the stimulation of MAIT cells in vitro. The vitamin B2 metabolite rRL-6HM, with or without anti-MR1 Abs, was added for 60 h. A panel of cytokines and other effector molecules secreted into the supernatants of these cultures were studied using a multiplex cytokine assay.
Tetramer+ T cells from Vα19iCα−/−MR1+/+ mice responded to rRL-6HM Ag much more robustly than did tetramer+ T cells from Vα19iCα−/−MR1−/− mice by secreting much higher levels of cytokines, chemokines, and cytolytic effector molecules (Fig. 4A). These rRL-6HM–specific responses were completely blocked with mAb to MR1, confirming MR1 presentation of the active vitamin B2 metabolite to MAIT cells. As shown in Fig. 4B, stimulation of purified splenic MR1+/+ Vα19i-Tg T cells by rRL-6HM Ag presented by CH27-mMR1 was blocked with anti-MR1 Ab, but not mouse IgG isotype control, confirming the requirement for MR1 presentation. Interestingly, the soluble effectors produced by different coreceptor subsets of tetramer+ T cells that developed in the presence of MR1 were markedly diverse, further indicating that tetramer+ MAIT cells include functionally heterogeneous subsets. The CD4+ subset predominantly produced IL-4, followed by IL-2, whereas the DN and CD8+ subsets produced IFN-γ, MIP-1α/CCL3, and MIP-1β/CCL4, as well as increased levels of granzyme B (GZMB) but not IL-2 (Fig. 4A, upper panels). The MR1+/+ DN subset also produced increased amounts of IL-4 and a small amount of IL-13. Thus, tetramer+CD4+ MAIT cells produced cytokines characteristic of Th1/Th2 responses, the DN subset produced a pattern of Th1/Th2/Tc1 T cell responses, and the tetramer+CD8+ subset produced a Tc1-like cytokine profile. Mature thymocytes from both MR1+/+ and MR1−/− Vα19i-Tg mice did not produce appreciable amounts of IFN-γ in response to M. bovis BCG–infected APCs (data not shown). These latter results confirm that thymic Vα19i+ MAIT cells have an unbiased phenotype that is characteristic of naive unactivated T cells. These experimental findings demonstrate that peripheral tetramer+ Vα19i-Tg cells developing in MR1-sufficient mice are functionally capable MAIT cells that, like human MAIT cells, respond to RL compounds in an MR1-restricted manner (2). On the contrary, tetramer+ Vα19i-Tg cells from MR1-deficient mice displayed markedly reduced effector functions after MR1/RL in vitro activation (Fig. 4A, lower panels). These combined results suggest that MR1 is necessary for the development of MAIT cell effector functions. However, another possibility suggested by the fact that MR1/RL tetramer+ MAIT cells develop in the absence of MR1, albeit in significantly lower frequencies, is that peripheral activation by MR1/RL is necessary in vivo for the development of MAIT cell effector functions and peripheral expansion. Furthermore, to our knowledge, this is the first report that MIP-1α/CCL3, MIP-1β/CCL4, and GZMB are produced by mouse MAIT cells. Perhaps more importantly, the differences in effector functions observed in different MAIT cell coreceptor subsets predict that these subsets also may provide functional differences in controlling microbial infections.
Vα19iCα−/−MR1+/+ T cells respond to RL Ag better than Vα19iTgMR1−/− T cells. (A) PBMCs from 22 Vα19iCα−/−MR1+/+ (upper panels) or Vα19iCα−/−MR1−/− (lower panels) mice were stained with Abs specific for CD3ε, CD4, CD8α, and MR1/RL tetramer, and tetramer+ cells were sorted by FACS. Sorted tetramer+ cells were rested overnight in medium and then cocultured in triplicates (2.5 × 103 cells/well) with CH27-mMR1 APCs (500 cells/well) along with medium alone, rRL-6HM (rRL) Ag, or rRL Ag plus 10 μg/ml anti-MR1 for 60 h at 37°C. The indicated cytokines, chemokines, and GZMB were measured in supernatants using Multiplex bead array assays, according to the manufacturer’s instructions (Milliplex MAP Assays; EMD Millipore). Tetramer+Vα19iCα−/−MR1+/+ T cells are functionally capable MAIT cells. (B) Anti-MR1 Ab, but not mouse IgG isotype control, specifically blocks Ag presentation by APCs. Activation of Vα19iCα−/−MR1+/+ T cells by rRL-6HM Ag. Purified MR1+/+ splenic Vα19i-Tg T cells (2 × 105/well) were cocultured with CH27-mMR1 APCs (4 × 104 cells/well) alone (Nil) or with 76.2 μM (final concentration) rRL-6HM in triplicate wells. APCs and Tg T cells were cocultured overnight (24 h) at 37°C in the absence or presence of anti-MR1–blocking Abs or mouse IgG isotype control. MAIT cell activation was determined by IL-2 secretion using ELISA.
The Vβ6/8+NK1.1+ subpopulation of Vα19i-Tg MAIT cells is highly enriched in peripheral sites
Among tetramer+CD3+ mature thymocytes and mLN, splenic, and blood cells from uninfected Vα19iCα−/−MR1+/+ mice, we found that less than half expressed Vβ6/8 (Fig. 1E, Supplemental Fig. 1B, middle column, Supplemental Fig. 1C, middle panel). Also, <25% expressed NK1.1 in the thymus, mLN, and spleen (Fig. 1E, Supplemental Fig. 1B, right column, Supplemental Fig. 1C, bottom panel). Furthermore, as shown in Fig. 5A and Supplemental Fig. 3A–C, <20% of all three coreceptor subsets of tetramer+ MAIT cells in thymus and secondary lymphoid organs harvested from Vα19i-Tg MR1-sufficient mice expressed both Vβ6/8 and NK1.1. In contrast, in the blood from these same mice, approximately half of all tetramer+ MAIT cells expressed both Vβ6/8 and NK1.1, and >80% of tetramer+Vβ6/8+ MAIT cells coexpressed NK1.1 (Fig. 5A, 5B). An even higher percentage of tetramer+CD3+ MAIT cell subsets coexpressing Vβ6/8 and NK1.1 was observed in liver peripheral tissue (Supplemental Fig. 3A–C, 3E). Additionally, >84% of tetramer+ Vβ6/8, NK1.1 DP cells coexpressed CD44 and CD69 in peripheral blood and the liver (data not shown). In contrast, <11% of tetramer+ Vβ6/8, NK1.1 DN cells expressed both CD44 and CD69. These results suggest that the tetramer+Vβ6/8+NK1.1+ cells are the subset of MAIT cells most highly reactive with MR1; as a consequence, they optimally expand and express a more activated phenotype in the peripheral sites. In support of this notion, despite significantly lower levels of CD3 expression, the Vβ6/8, NK1.1 DP subset of tetramer+ cells that developed in the presence of MR1 was shown to have significantly higher levels of MR1/RL tetramer staining than did the Vβ6/8, NK1.1 DN subset in the spleen and liver (Fig. 5C, Supplemental Fig. 3D). The combination of lower CD3 and higher tetramer staining indicates that the Vβ6/8, NK1.1 DP subsets of tetramer+ cells express TCRs with higher affinity. These combined results further indicate that the Vβ6/8 and NK1.1 DP subset of tetramer+ cells constitute the most highly functional MAIT cell subset in vivo. We speculate that this circulating Vβ6/8, NK1.1 DP MAIT cell population may be a further differentiated effector subset because of previous MR1/RL activation, consistent with better responsiveness to IL-12 plus IL-18 cytokines (Fig. 3D). As shown in Supplemental Fig. 3E, NK1.1 was minimally expressed on mature Vβ6/8+tetramer+ MAIT cells in the thymus, but it progressively increased with peripheral activation in both MR1-sufficient and -deficient Tg mice, most strikingly in MR1-sufficient mice. In the absence of MR1, there was a 76 and a 84% relative reduction in the proportion of peripherally expanded/activated Vβ6/8 NK1.1 DP tetramer+CD3+ T cells in blood and the liver, respectively. Taken together, our data indicate that peripheral MR1 is necessary for optimal expansion/activation of MAIT cells, and NK1.1 is likely to represent an activation, rather than a developmental, marker of MAIT cells.
Peripheral expansion of Vβ6/8+NK1.1+ population of tetramer+CD3+ T cells in Vα19i-Tg mice. (A) Surface Vβ6/8.1-2 (x-axis) and NK1.1 (y-axis) expression on tetramer+ CD4+ (upper panels), DN (middle panels), or CD8+ (bottom panels) T cells in spleen and blood from uninfected Vα19iCα−/−MR1−/− (left panels) and Vα19iCα−/−MR1+/+ (right panels) mice. Numbers in the upper right quadrants are NK1.1+ and Vβ6/8.1-2+ cells. (B) Increasing enrichment of tetramer+ NK1.1 lineage in Vβ6/8+ cells more than in Vβ6/8− subset from spleen to blood. (C) Vα19i-Tg cells developing with MR1 have higher affinity for MR1/RL complexes. Vβ6/8 (x-axis) versus tetramer (y-axis) FACS plots of splenic CD3+ MAIT cells from Vα19iCα−/−MR1−/− or Vα19iCα−/−MR1+/+ mice (left panels). Mean fluorescence intensity (MFI) of CD3ε (upper right panel) and tetramer (lower right panel) staining of splenic tetramer+ Vβ6/8+NK1.1+ or Vβ6/8−NK1.1− CD3+ T cells from Vα19iCα−/−MR1+/+ or Vα19iCα−/−MR1−/− mice. Despite lower CD3 expression, Vα19i-Tg T cells from MR1+/+ mice display higher affinity for MR1/RL tetramers than do Vα19i-Tg cells from MR1−/− mice. Furthermore, tetramer+CD3+ MAIT cells expressing both Vβ6/8 and NK1.1 in Vα19i-Tg MR1-sufficient mice display the highest affinity for tetramer binding. Data shown are from two separate experiments. **p < 0.01, ***p < 0.001, Mann–Whitney U test (n = 5/group).
Vα19i-Tg MAIT cells developing in MR1-sufficient mice are optimally recruited to mycobacteria-infected lungs and provide maximal early protection
We showed previously that purified Vα19i-Tg MAIT cells developing with MR1 inhibit intracellular growth of mycobacteria in vitro, whereas control Vα19i-Tg T cells from MR1−/− mice showed no inhibition of bacterial growth in M. bovis BCG–infected macrophages (24). This observation indicated that the maximal antibacterial function of Vα19-Jα33 TCR-bearing T cells requires MR1 selection during development. T cells from naive (unvaccinated and uninfected) B6 MR1−/− and B6 WT mice also failed to inhibit the intracellular growth of M. bovis BCG in macrophages. We also demonstrated in vivo that, on day 10 following aerosol infection with M. bovis BCG, B6 mice deficient in MR1 had a greater mycobacterial burden in the lungs than did WT mice, indicating that MR1-restricted MAIT cells are important for optimal protection against mycobacterial infection (24).
Of note, previous studies of mouse and human MAIT cells showed that MAIT cells increase at the site of infections (19, 20, 35, 37). In the current study, we used MR1/RL tetramers to identify Ag-specific MAIT cell subsets at the site of mycobacterial infection in the lung. Groups of B6 MR1−/−, B6 WT, Vα19iCα−/−MR1−/−, and Vα19iCα−/−MR1+/+ mice were infected intranasally with M. bovis BCG Danish. MR1/RL-specific MAIT cells were increased in the infected lung airways at 10 d postinfection, a time point when MAIT cells were shown previously to be important for control of the infection (24). Moreover, the early accumulation in mouse lungs of a variety of innate cells, including MAIT cells, after mycobacterial infection appears to occur between days 10 and 15 (54). Mycobacterial burden in the lung at day 10 postinfection was significantly lower in Vα19i-Tg MR1-sufficient mice than in Vα19i-Tg mice lacking MR1, suggesting an MR1-dependent in vivo function for Vα19i-Tg T cells (Fig. 6Ai). Confirming our previous results, mycobacterial burden in the lungs of B6 WT mice was significantly lower than in B6 MR1−/− mice. Interestingly, Vα19iCα−/−MR1−/− and B6 WT mice had similar mycobacterial loads, and they were lower than those found in B6-MR1−/− mice. We were surprised that Vα19i-Tg T cells developing in the absence of MR1 conferred protection against mycobacterial infection that was comparable to that of non-Tg MAIT cells in B6 WT mice. As shown in Fig. 6D, it is likely that the slightly increased number of tetramer+ Vα19i-Tg T cells developing in the absence of MR1 provides partial protection in vivo. In addition, we found that tetramer+ MAIT cells developing in the absence of MR1 can respond to RL Ags in an MR1-dependent manner (Fig. 4A, lower right panel) and respond in an MHC class Ia–dependent manner (Supplemental Fig. 4Ai). Indeed, the inhibitory effects of purified total T cells from Vα19iCα−/−MR1−/− mice were completely abrogated in macrophages deficient in MR1 or classical MHC class Ia molecules (Supplemental Fig. 4Aii). These combined results demonstrate that the tetramer+ Vα19i T cells generated in MR1-knockout mice can develop effector functions in vivo in response to infection, despite the absence of MR1. Most importantly, after aerosol lung infection with virulent M. tuberculosis Erdman strain, lung mycobacterial load was significantly lower in Vα19iCα−/−MR1+/+ mice than in B6 WT control mice (Fig. 6Aii). These differences strongly corroborate a role for MAIT cells in controlling pulmonary mycobacterial infections.
MR1/RL tetramer+ T cells accumulate in the lung during mycobacteria primary infection. Groups of 5–11 B6-MR1−/−, B6 WT, Vα19iCα−/−MR1−/−, or Vα19iCα−/−MR1+/+ mice were infected with M. bovis BCG Danish or M. tuberculosis Erdman, as described in Materials and Methods. At day 10 postinfection, BALF collected from uninfected and infected groups was pooled for staining with tetramer and other phenotypic markers. The bacterial burden was determined by CFU plating of lung tissue homogenates. Bar graphs show M. bovis BCG Danish (Ai) and M. tuberculosis Erdman CFU (Aii) in lungs of infected mice. Results are mean ± SEM of mycobacteria CFU and are from three independent experiments. (B) Absolute numbers of BALF CD3+ lymphocytes in uninfected or infected mice. (C) Data from three independent experiments showing total leukocyte counts in BALF from uninfected and infected mice. (D) Bar graphs showing absolute number of BALF tetramer+CD3+ T cells in the indicated mice. (E) Pie charts showing relative distribution of absolute numbers of tetramer+CD3+ MAIT cell subsets in infected Vα19iCα−/−MR1+/+ (upper panel) or Vα19iCα−/−MR1−/− (lower panel) mice. *p < 0.05, **p < 0.01, ****p < 0.0001, Mann–Whitney U test.
MR1/RL tetramers were used to enumerate MAIT cells in the BALF of uninfected and infected mice. Representative plots of CD3+ lymphocytes and tetramer+CD3+ T cells in BALF from uninfected or BCG-infected mice are shown in Supplemental Fig. 4B. BALF from all uninfected mouse groups had very low numbers of detectable T cells (Fig. 6B). Although the total number of BALF mononuclear cells increased in the lungs of all four infected groups of mice compared with uninfected controls (Fig. 6C), the highest numbers of all CD3+ T cell subsets (DN > CD4+/CD8+) were detected in infected Vα19iCα−/−MR1+/+ mice (Fig. 6B). Consistent with their expected low frequency of MAIT cells, very few tetramer+ cells were detected in the BALF of uninfected or infected B6 WT or B6 MR1−/− mice (Fig. 6D). However, remarkably higher numbers of tetramer+ T cells were detected in BALF from infected Vα19iCα−/−MR1+/+ mice than from all other groups of mice (Fig. 6D). DN tetramer+ cells were the predominant subset (43%), followed by the CD8+ (33%) and the CD4+ (24%) subsets (Fig. 6E, upper panel). This hierarchy of tetramer+ coreceptor subset distribution was similar to that found in the blood of infected mice (data not shown). Importantly, more MAIT cells were identified in Vα19i-Tg mice lacking MR1 than in B6 WT and B6 MR1−/− mice, and these increased numbers of MAIT cells were associated with better mycobacterial control (Fig. 6Ai). Tetramer+ BALF T cells detected in Vα19iCα−/−MR1−/− mice were CD8+ (38%), with comparable fractions of CD4+ (31%) and DN (31%) subsets (Fig. 6E, lower panel). Furthermore, B6 WT mice had more tetramer+ MAIT cells (∼0.25 × 104 tetramer+CD3+ cells) than did B6 MR1−/− mice, which explains the better protection in WT B6 mice than in non-Tg MR1-knockout mice and confirming our previous report (24). We conclude that MAIT cells are recruited to the lung early after mycobacterial challenge and provide important protective effects. In addition, MAIT cells developing in the presence of MR1 are optimally protective.
Phenotype of tetramer+ MAIT cell subsets in lung airways associated with optimal protection against primary mycobacterial challenge
Vβ6/8+NK1.1+ and Vβ6/8−NK1.1− subpopulations of T cells each represented ∼50% of the tetramer+ MAIT cells accumulating in the airways of Vα19i-Tg MR1-sufficient mice (Fig. 7A). Only minor populations (<6%) of tetramer+ cells were NK1.1+Vβ6/8− or NK1.1−Vβ6/8+ in the infected lung. This was true for all CD4, CD8, and DN coreceptor subsets of tetramer+ T cells. Further investigation is required to determine which of these NK1.1+Vβ6/8+ and Vβ6/8−NK1.1− subsets is most important for protection against mycobacterial infection. Gating on NK1.1+Vβ6/8+ and Vβ6/8−NK1.1− subsets, we found that the majority of tetramer+ Vα19i-Tg cells in BALF of infected MR1-sufficient mice also expressed α4β1 integrin and the chemokine receptor CXCR3 (Fig. 7B), molecules that were associated previously with trafficking activated Ag-specific T cells to infected lungs. Integrin α4β1 is required for cells to cross from blood to lung airway, where its ligand VCAM-1 is upregulated upon infection (55); CXCR3 mediates the recruitment of effector T cells into inflammatory tissues in response to its ligands induced by infection (56) and is known to be coexpressed with α4β1 in blood and BALF samples after mycobacterial infection. Thus, these combined data demonstrate that DN > CD8+ > CD4+, Vβ6/8, NK1.1 DP and DN tetramer+ cells were associated with lung protection in Vα19iCα−/−MR1+/+ mice. Moreover, most MAIT cells recruited and/or expanded in the infected lung express both α4β1 integrin and chemokine receptor CXCR3. Interestingly, in the BALF from infected WT B6 mice, almost all MAIT cell subsets (DN > CD4+ > CD8+) were Vβ6/8+NK1.1+ DP and expressed both α4β1 integrin and CXCR3 (Fig. 7D–F). These latter results are consistent with a previous study reporting that, in response to Francisella tularensis lung infection, CD4/8 DN Vβ6/8+ MAIT cells were associated with optimal early control of bacterial infection in unvaccinated mice (37). Because of the limited cell numbers available in BALF, and the fact that MR1/RL tetramer staining does not survive fixation/permeabilization treatment, we have not been able to identify cytokine or in situ proliferation profiles after mycobacterial infection.
Vβ6/8 and NK1.1 expression by BALF tetramer+CD3+ MAIT cells in Vα19iCα−/−MR1+/+ and B6 WT mice after mycobacterial infection. (A–C) Phenotype of tetramer+CD3+ MAIT cell subsets in BALF from infected Vα19iCα−/−MR1+/+ mice. (A) Representative FACS plots from three separate experiments showing percentages of tetramer+CD3+ Vα19i-Tg T cell subsets (left panels); the numbers in the upper right quadrants indicate tetramer+CD4+ (top panel), tetramer+ DN (middle panel), and tetramer+CD8+ (bottom panel) MAIT cells. Representative plots showing Vβ6/8.1-2 (x-axis) and NK1.1 (y-axis) expression on tetramer+ CD4+ (top right panel), DN (middle right panel), or CD8+ (bottom right panel) MAIT cells; the numbers in the quadrants indicate NK1.1−Vβ6/8.1-2−tetramer+ cells (lower left quadrants), NK1.1+Vβ6/8.1-2−tetramer+ cells (upper left quadrants), NK1.1+Vβ6/8.1-2+tetramer+ cells (upper right quadrants), and NK1.1−Vβ6/8.1-2+tetramer+ cells (lower right quadrants). The NK1.1−Vβ6/8−tetramer+ and NK1.1+Vβ6/8+ cells were the most predominant subpopulations. (B) Representative data from three separate experiments showing percentage of CXCR3 (x-axis) and α4β1 (y-axis) expression on NK1.1+Vβ6/8+ or NK1.1−Vβ6/8− tetramer+ CD4+ (top panels), DN (middle panels), or CD8+ (bottom panels) MAIT cells in BALF from infected Vα19iCα−/−MR1+/+ mice. Numbers in the upper right quadrants are proportions of cells that expressed both CXCR3 and α4β1 integrin. (C) Representative data from three separate experiments showing absolute numbers of CXCR3- and α4β1-expressing NK1.1+Vβ6/8+ or NK1.1−Vβ6/8− tetramer+ CD4+ (top panel), DN (middle panel), or CD8+ (bottom panel) MAIT cells. (D–F) Frequency and phenotype of tetramer+CD3+ non-Tg MAIT cell subsets in BALF from infected B6 WT mice. (D) The pie chart shows the relative distribution of absolute numbers of tetramer+CD3+ MAIT cell subsets in the BALF from infected B6 WT mice. (E) Representative FACS plots showing percentages of tetramer+CD3+ MAIT cell subsets (left panels). Numbers in the upper right quadrants indicate tetramer+CD4+ (top panel), tetramer+ DN (middle panel), and tetramer+CD8+ (bottom panel) MAIT cells. Representative FACS plots showing Vβ6/8 (x-axis) and NK1.1 (y-axis) expression on tetramer+CD3+ MAIT cell subsets (right panels). Numbers in the upper right quadrants are proportions of tetramer+ cells that expressed both Vβ6/8 and NK1.1. (F) CXCR3 and α4β1 integrin expression on Vβ6/8+NK1.1+tetramer+ MAIT cell populations in the airways of infected B6 WT mice.
Comparisons of blood and BALF tetramer+ T cell subsets in uninfected and infected Vα19iCα−/−MR1+/+ mice
We further characterized the phenotypes and numbers of CD3+tetramer+ T cell subsets in the blood and BALF of uninfected and infected Vα19iCα−/−MR1+/+ mice. The relative distribution of CD4/8 coreceptor subsets of Vβ6/8-positive and -negative, tetramer+ cells in the blood of uninfected and infected mice was DN > CD8+ > CD4+ (data not shown), similar to infected BALF populations. In addition, comparable percentages of tetramer+, Vβ6/8 and NK1.1 DP and DN subsets were present in uninfected and infected blood (Fig. 8A). However, the frequencies of most of the tetramer+ subsets in blood were reduced by ≥50% in infected mice compared with uninfected mice (Fig. 8B). Thus, our data indicate that comparable numbers of tetramer+, Vβ6/8/NK1.1 DP and DN populations appear to exit the blood postinfection, and both subsets accumulate in the infected lungs of Vα19iCα−/−MR1+/+ mice (Fig. 8C, 8D). In WT B6 mice, only the Vβ6/8+NK1.1+ population of tetramer+ MAIT cells was detected at the site of infection (Fig. 7E). In summary, our tetramer studies of WT B6 mice further confirm the biological significance of our Tg mouse studies. Our combined findings allow us to propose a model that MAIT cells developing in the presence of MR1 are optimally protective, the population relevant for protection is heterogeneous, and that MAIT cell subsets controlling murine mycobacterial pulmonary infection traffic from blood to the lungs.
Vβ6/8 and NK1.1 phenotype of Vα19i-Tg T cells in blood and BALF from uninfected or mycobacterial-infected Vα19iCα−/−MR1+/+ mice. Vα19iCα−/−MR1+/+ mice were infected with 107 CFU/mouse of BCG Danish, as described in Fig. 6. At day 10 postinfection, peripheral blood T cells were stained with tetramer and a panel of other phenotypic markers and analyzed by FACS. (A) FACS plots from one of the two separate experiments showing percentages of Vβ6/8.1-2 (x-axis) and NK1.1 (y-axis) expression on tetramer+ CD4+ (top panels), DN (middle panels), or CD8+ (bottom panels) MAIT cells in blood from uninfected or infected Vα19iCα−/−MR1+/+ mice. (B) Data from two separate experiments showing absolute number (AN) of tetramer+NK1.1+ or tetramer+NK1.1− cells that were either Vβ6/8+ (upper panels) or Vβ6/8− (lower panels) cells in blood from uninfected or infected Vα19iCα−/−MR1+/+ mice. (C) Representative data from one of three separate experiments showing percentages of Vβ6/8.1-2 and NK1.1 expression on tetramer+ CD4+ (top panels), DN (middle panels), or CD8+ (bottom panels) MAIT cells in BALF from uninfected or infected Vα19iCα−/−MR1+/+ mice. (D) Data from two separate experiments showing AN of tetramer+NK1.1+ or tetramer+NK1.1− cells that were either Vβ6/8+ (upper panels) or Vβ6/8− (lower panels) cells in BALF from uninfected or infected Vα19iCα−/−MR1+/+ mice.
Discussion
Our novel MR1/RL Ag tetramer characterization of the Vα19i Tg mouse model offers new insights into the development and function of MAIT cells, as well as their role in TB protective immunity. We demonstrate that mouse MAIT cells contain heterogeneous populations distinguished by the expression or nonexpression of diverse Vβ TCR, NK1.1 and CD4/8 coreceptors. In addition, we show that there are more MR1/RL tetramer+ MAIT cells in Vα19i-Tg mice expressing MR1 than in those lacking MR1. The relative distributions of MAIT cell subsets are similar among thymus, mLN, spleen, blood, liver, and mycobacterial-infected lung airways, although the relative numbers in blood versus lung airways decrease postinfection. CD44 and CD69 expression on MAIT cells was higher in liver peripheral tissue than on MAIT cells in the thymus and naive secondary lymphoid organs in MR1-expressing mice. This likely represents a progressive activation of peripheral MAIT cells. The patterns of soluble effector molecules released by peripheral blood tetramer+ MAIT cell subsets after stimulation with rRL-6HM Ag were characteristic of Th1 (IL-2, IFN-γ) Th2 (IL-4), and Tc1 (IFN-γ, GZMB) immune responses. In contrast to the total Vα19i-Tg T cells used by Gilfillan and colleagues (25) to determine cytokine-production profiles, our sort-purified blood tetramer+ Tg T cell subsets did not produce IL-5 and IL-10 in response to rRL-6HM Ag. Also, contrary to our earlier work in which we stimulated purified total Vα19iCα−/−MR1+/+ T cells with M. bovis BCG–infected BMDMΦs (24), these circulating tetramer+ MAIT cells did not produce IL-17A after rRL-6HM Ag stimulation in vitro. A similar observation was reported recently for human MAIT cell clones stimulated with riboflavin-producing Escherichia coli–infected APCs (47). Tang and colleagues (57) speculated that the failure to produce IL-17A by MAIT cells following Ag/TCR stimulation may reflect the requirement for additional environmental signals, such as IL-7 signaling. The results reported in this article impact on several outstanding issues that are discussed below.
Is MR1 required for the thymic selection of MAIT cells or only for their peripheral expansion?
Similar to type I NKT cells, MAIT cells were demonstrated recently to be positively selected by CD4+CD8+ thymocytes (58). However, distinct from CD1d, MR1 is very poorly expressed at the cell surface, even on DP thymocytes. Furthermore, the thymic Ag presented by MR1, which was speculated to be of endogenous origin, remains unknown. In addition, the number of MAIT cells in the thymus of both humans and mice is very low (8, 13, 59). Thus, conditions for thymic selection of MR1-restricted MAIT cells may not be optimal. In contrast, Ag presentation in the periphery is clearly important for MAIT cell expansion and function. MAIT cells in the periphery remain naive until exposed to the commensal flora, resulting in their dramatic expansion and conversion to an effector/memory phenotype (13, 60). Thus, the role for MR1 in thymic selection may be of lesser importance than its role in peripheral expansion. Relevant to this hypothesis, we found that tetramers stained a significant proportion of mature T cells in the thymus from Vα19iCα−/−MR1−/− mice, indicating that MR1 is not absolutely required for MAIT cell thymic selection (Fig. 2C, 2D, Supplemental. Fig. 1B, 1C). This finding is similar to a previous report that type I NKT cells develop in Vα14i-Tg mice in the absence of CD1d (61). Because the TCR Vα-chain of type I NKT cells is predominant in CD1d/lipid recognition (62), the overexpression of Vα14i transgene was considered sufficient to bypass the requirement for type I NKT selection on CD1d in Tg mice. Because TCR Vα19i is also predominant in MAIT cell detection of MR1/RL (6), a similar mechanism might explain our findings of tetramer+ cells in Vα19iCα−/−MR1−/− mice. Alternatively, given that the α1 and α2 domains of MR1 share high similarity in both sequence and structure with classical MHC class I molecules and that the footprint of MAIT TCR on MR1 closely resembles the typical TCRs on classical MHC class I (3, 42, 63, 64), we speculate that, in the absence of MR1, tetramer+ T cells can be selected by a weak cross-reaction on MHC class Ia. Whether this type of selection occurs in nontransgenic mice or is physiologically relevant remains to be demonstrated.
Is Vβ6/8 pairing required for MAIT cell generation?
Previous investigations suggested that most MAIT cells expressed Vα19i paired with Vβ6/8. However, we found that large proportions of tetramer+ MAIT cells among both mature thymocytes and peripheral T cells did not express Vβ6/8. Furthermore, only half of the tetramer+ MAIT cells recruited to the lungs after mycobacterial infection in Vα19iCα−/−MR1+/+ mice coexpressed Vβ6/8. MR1/RL tetramers stained Vβ6/8+ T cells, as well as a significant number of other Vβ-chains in Vα19i-Tg mice. Furthermore, we demonstrated previously that tetramer+Vβ6/8− MAIT cells in Vα19i-Tg MR1-sufficient mice produced IFN-γ and TNF-α after rRL-6HM Ag stimulation in an MR1-restricted fashion (figure 8 in Ref. 4). These results suggest that functionally important MAIT cells are more heterogeneous in terms of Vβ expression than previously thought. In contrast, we demonstrate in this study that Vβ6/8+ MAIT cells that developed in the presence of MR1 bound MR1/RL tetramers with higher affinity (Fig. 5C, Supplemental Fig. 3D) and exhibited higher levels of CD69 expression, indicative of a previously activated state (Fig. 3B, 3C). In addition, in our work cited above, Vβ6/8+tetramer+ MAIT cells secreted higher levels of inflammatory cytokines in response to rRL-6HM restricted by MR1 (4). Moreover, we also show in this study that Vβ6/8+NK1.1+tetramer+ MAIT cells responded to IL-12 plus IL-18 better than did the Vβ6/8−NK1.1− population (Fig. 3D). Furthermore, in WT B6 mice that were significantly better protected than non-Tg MR1-knockout mice, only Vβ6/8+ MAIT cells were recruited into the lungs after mycobacteria challenge. Therefore, further studies are needed to determine whether tetramer+Vβ6/8− MAIT cells can provide functionally important, MR1-restricted protective responses against mycobacterial and other infectious pathogens.
Mouse NK1.1 may be an important phenotypic marker associated with peripheral expansion of Vβ6/8-expressing functional MAIT cells
Defining MAIT cells by their preferential Vβ6/8 expression (classical MAIT cells) in Vα19i-Tg mice, a good proportion of MAIT cells from the spleen and mLNs were shown previously to express NK1.1 (25, 65). However, NK1.1 was not considered a reliable phenotypic marker for MAIT cells because only a subset of Vβ6/8+ cells was found to express NK1.1 in Vα19i-Tg MR1-sufficient mice (13, 25). Indeed, in the current study, only ∼3 and <20% of the CD3+tetramer+ cells from the thymus and secondary lymphoid tissues, respectively, of Vα19iCα−/−MR1+/+ mice were NK1.1+ (Supplemental Fig. 1C, bottom panel). Nevertheless, NK1.1 expression was substantially increased on the tetramer+Vβ6/8+ fraction compared with the Vβ6/8− population, especially in the most peripheral sites (Fig. 5A, 5B, Supplemental Fig. 3B, 3C, 3E). These peripheral Vβ6/8+NK1.1+tetramer+ cells represent effector memory cells, based on their higher coexpression of CD44 and CD69 and better functional response to in vitro rRL-6HM or cytokine stimulation. We propose that this segregation of MAIT cell phenotypes is driven by MR1-restricted presentation of bacterial flora RL Ags. In support of this model, we show that bulk Vβ6/8+NK1.1+ Tg T cells bind MR1/RL tetramers with higher affinity than Vβ6/8−, NK1.1− Tg T cells (Fig. 5C, Supplemental Fig. 3D). These findings support the model that presumed exposure to commensal flora–derived RL Ags promotes the peripheral expansion of Vβ6/8+NK1.1+ cells driven by increased affinity for MR1/RL complexes. However, in light of recent findings that human MAIT cells develop and mature in the fetus, prior to exposure to the commensal flora (59), the presence of commensal flora may not be absolutely necessary for acquisition of innate-like microbial reactivity or expansion/maturation of MAIT cells. In contrast, it is worth noting that Sandberg and colleagues (59) did not rule out the possibility of fetal mucosal exposure to components of the maternal microbial vitamin B2 metabolites delivered via the amniotic fluid. They further concluded that the association between maturation and gain of effector functions argues for the involvement of an MR1-presented ligand. They also found that there was a progressive increase in the proportion of MAIT cells detected in the fetal thymus, spleen, and peripheral sites including small intestine, liver, and lung.
Is NK1.1 expression on mouse MAIT cells a developmental lineage and/or activation marker?
Mouse NK1.1 (also known as KLRB1C and NKR-P1C) is expressed on NK and NKT cells in B6 mice. Although cross-linking of the receptor using an NK1.1-specific mAb induces NK cell–mediated cytotoxicity and effector cytokine secretion, the in vivo function of NK1.1 and the nature of its ligand(s) remain unknown (66–70). Thus, whether NK1.1 is a developmental lineage marker and/or an activation receptor is an important question. Of note, in the current study, we found that NK1.1 was marginally expressed on tetramer+ thymocytes, but it was gradually enriched with progressive peripheral activation in Vα19iCα−/−MR1+/+ mice. An increase in the proportional expression of NK1.1 from the most naive (thymus) to the most peripherally expanded/activated MAIT cell populations (Figs. 5A, 5B, 7A, Supplemental Fig. 3B, 3C, 3E) suggests an upregulation of NK1.1 expression upon progressive exposure to Ags. Thus, NK1.1 likely represents activation rather than developmental MAIT cell markers. Our findings further suggest that NK1.1 is a reliable marker for MR1-dependent classical MAIT cells with optimal innate-like cell function. However, both Vβ6/8, NK1.1 DP and DN tetramer+ MAIT cells were recruited into the lungs of Vα19i-Tg MR1-sufficient mice that were found to be optimally protected against mycobacterial challenges (see below). We conclude that mouse NK1.1 may be an important phenotypic marker associated with peripheral expansion/activation of classical Vβ6/8-expressing functional MAIT cells, rather than a MAIT cell lineage marker.
MR1-restricted MAIT cells have important antimycobacterial effects
MAIT cells are known to accumulate at the site of infection (19, 20). Because in some of these same studies an accompanying decrease in MAIT cell number was seen in the blood of infected individuals, MAIT cell migration was implicated. Consistent with these data, we show in this study that all tetramer+ subsets (DN/CD8+/CD4+ and both Vβ6/8+ and Vβ6/8−) decreased appreciably in blood after BCG infection (Fig. 8B). We also noted that the majority of all MAIT cell blood subsets expressed both α4β1 integrin and the chemokine receptor CXCR3 (data not shown). The expression of α4β1 integrin is known to localize Ag-specific T cells to the airway where its ligand VCAM-1 is upregulated upon infection (55, 71, 72). CXCR3 is a dominant chemokine receptor strongly expressed on activated Th1 cells, thus directing them toward the lung in mycobacterial and other respiratory infections (54, 73–75). CXCR3-expressing lymphocytes migrate toward three chemokines: CXCL9/monokine induced by IFN-γ, CXCL10/IFN-γ–inducible protein 10, and CXCL11/IFN-γ–inducible T cell α-chemoattractant (76–78). In the lung, CXCR3 ligands are known to be produced by bronchial epithelial cells following infection or vaccination in response to IFN-γ (54, 79, 80). Thus, the expression of α4β1 and CXCR3 on all MAIT subsets could explain why all MAIT subsets are diminished in the blood after pulmonary mycobacterial infection.
In Vα19iCα−/−MR1+/+ mice after pulmonary mycobacterial challenge, the disappearance of all subsets of MAIT cells from blood was associated with their accumulation among infected BALF cells and early protection against mycobacterial infection (Figs. 6, 7A, 8C, 8D). Interestingly, MAIT cell subsets expressing only Vβ6/8 TCR and NK1.1 accumulated in the airways of infected B6 WT mice (Fig. 7E). This is consistent with a recent report by Cowley and colleagues (37) that Vβ6/8+, but not Vβ6/8−, MAIT cells were detected in lung homogenates of WT B6 mice after F. tularensis infection. However, only 12.6% of these MAIT cells recruited in response to F. tularensis infection expressed NK1.1. Nevertheless, the concordance of their findings with F. tularensis infection of WT mice is strikingly similar to our findings with BCG infection of WT mice, strongly supporting the unique ability of Vβ6/8+, but not Vβ6/8−, MAIT cells to accumulate at the site of infection. Our studies further demonstrate that all MAIT cell subsets leave the blood after mycobacteria infection of the lung, presumably based on their expression of α4β1 and CXCR3. Although there are differences between Tg and WT mice, our findings demonstrate that the tetramer+ DN MAIT cell subset was the predominant population in the lungs of both MR1-expressing Vα19i-Tg and non-Tg B6 mice after mycobacterial pulmonary infection. These combined results suggest that the tetramer+ DN MAIT cell subset is the most important for TB immunity. This possibility will be confirmed in future studies comparing the protective effects of adoptively transferred, sort-purified MAIT cell subsets. However, both Vβ6/8+NK1.1+ and Vβ6/8−NK1.1− subsets were detected in the lung airways of Tg mice, whereas only the Vβ6/8+NK1.1+ subpopulation accumulated at the site of infection in WT B6 mice after mycobacterial infection and is associated with early protection. In the future, we will use adoptive-transfer studies to characterize which of these tetramer+CD3+ Vβ6/8+NK1.1+ and Vβ6/8−NK1.1− MAIT cell subpopulations is more important in protective antimycobacterial immunity. Moreover, we will perform TCRβ analysis to determine the CDR3β repertoire after vaccination/challenge to gain a better understanding of the biology of mycobacteria-specific MAIT cells.
In conclusion, our study provides insights into how MAIT cells activated by innate cytokines and/or MR1/RL-TCR signals function to control mycobacterial infection in vivo. MAIT cells recruited to the lung airways have the potential to produce IL-2, IFN-γ, MIP-1α/CCL3, and MIP-1β/CCL4 chemokines (signal through CCR1 and CCR5 chemokine receptors) and GZMB. These MAIT cell–derived soluble mediators could facilitate the early production of NO and the timely recruitment of CCR1+ and CCR5+ macrophages, immature dendritic cells, granulocytes, and Ag-specific CD8+ and Th1 CD4+ T cells to the lungs during mycobacteria infection. This study also revealed a previously unrecognized potential role for MAIT cells in the killing of mycobacteria-infected cells in the lung via the cytotoxic granule exocytosis effector pathway. The observation by Lantz and colleagues (60) that MAIT cells are cytotoxic in vitro is fitting with our novel findings that these cells produce large amounts of the granule serine protease GZMB. Thus, we infer that MAIT cell–mediated control of mycobacterial infection in vivo mechanistically involves the NOS2/NO and granule serine protease exocytosis effector pathways. Furthermore, MAIT cells can facilitate the recruitment of CCR1+ and CCR5+ immune cells to the lung during mycobacterial infection. Our study provides evidence that future TB therapeutic/prophylactic vaccine efforts could target the induction of MAIT cell responses. Such vaccines may require the incorporation of vitamin B2 metabolites as MAIT cell immunogens and/or IL-12– plus IL-18–expressing plasmid combinations to augment the immune response. The use of cytokine-expressing plasmids as a strategy to augment T cell responses was demonstrated in HIV vaccine research (81–83).
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Drs. Susan Gilfillan and Wayne Yokoyama for the generous gift of Vα19i-Tg and H2-Kb−/− H2-Db−/− mouse strains, respectively. We thank Dr. Liping Yang (Yokoyama Laboratory, Washington University School of Medicine) for assistance with isolation of lymphocytes by perfusion and Dr. Dale I. Godfrey (University of Melbourne) for critical advice.
Footnotes
This work was supported by National Institutes of Health Grants R01-AI046553 (to T.H.H.) and R01-A148391 (to D.F.H), a National Health and Medical Research Council of Australia Senior Principal Research Fellowship (to D.P.F.), a National Health and Medical Research Council of Australia Fellowship (to J.R.), and National Health and Medical Research Council of Australia Program Grant 1016629 (to J.R. and J.M.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- B6
- C57BL/6
- BALF
- bronchoalveolar lavage fluid
- BCG
- bacillus Calmette–Guérin
- BMDMΦ
- bone marrow–derived macrophage
- DN
- double negative
- DP
- double positive
- 6-FP
- 6-formylpterin
- 1G
- first generation
- 2G
- second generation
- GZMB
- granzyme B
- iNKT
- invariant NKT
- MAIT
- mucosal-associated invariant T
- mLN
- mesenteric lymph node
- 5-OP-RU
- 5-(2-oxopropylideneamino)-6-d-ribitylaminouracil
- RL
- ribityllumazine
- rRL-6HM
- reduced 6-hydroxymethyl-8-d-ribityllumazine
- TB
- tuberculosis
- Tg
- transgenic
- WT
- wild-type.
- Received October 17, 2014.
- Accepted May 13, 2015.
- Copyright © 2015 by The American Association of Immunologists, Inc.
References
In this issue
