Activating Mucosal-Associated Invariant T Cells Induces a Broad Antitumor Response

Mucosal-associated invariant T (MAIT) cells are a unique population of innate-like T cells (1) that share properties of both innate immune cells and conventional T cells. MAIT cells express a limited repertoire of T-cell receptors (TCR; 2) and are restricted by MR1 (3), a monomorphic class I–related MHC molecule that has a highly conserved ligand-binding groove. Unlike conventional T cells, which are reactive to peptide antigens, MAITs recognize microbial non-peptide antigens such as riboflavin derivates, suggesting they have a role in controlling microbial infection. To date, 5-(2-oxopropylideneamino)-6-D-ribitylaminouracil (5-OP-RU) is the most potent antigen known to activate MAITs (4). In combination with a Toll-like receptor 9 (TLR9) agonist CpG, 5-OP-RU causes a strong activation and expansion of MAIT cells in an MR1-dependent manner (5, 6). Activation of MAIT cells by various species of bacteria and yeast can cause the cells to secrete proinflammatory cytokines and lyse bacterial and virus infected cells (7). Homozygous point mutations in Mr1 resulting in a lack of MAITs can cause primary immunodeficiency in humans, which manifests with viral and bacterial infections that are refractory to therapy (8). This indicates that MAIT cells may play an early role in protection against infections (9). Consistent with a proposed antimicrobial function, MAIT cells are enriched in many organs with high microbial exposure, such as the intestine, lung, and liver. In humans, MAIT cells can represent the most common T-cell subpopulation (up to ∼50%) in the liver (10).

MAITs can be detected in both primary cancers and metastasis (11–15), but their exact function in cancer remains to be elucidated. Emerging data suggest that MAIT cells are critical for tumor progression, however, both pro- and antitumor functions have been reported (16–18). Decreased MAIT cell frequencies are seen in tumor lesions of patients with hepatocellular carcinoma (HCC) compared with normal liver tissue (14). Importantly, high numbers of MAIT cells in HCC lesions are associated with a poor prognosis, suggesting the cells have a tumor-promoting function (14). Consistent with the human data, tumor initiation, growth, and metastasis are reduced in MAIT-deficient Mr1^−/− mice (17). In that study, MAITs are shown to promote tumor growth by suppressing T cells and natural killer (NK) cells through an IL17Amediated mechanism of action (17). In contrast, MAITs exhibit cytolytic activity in vitro against multiple myeloma cell lines pulsed with 5-OP-RU (18), suggesting that MAITs also possess antitumor properties. We suggest that context-dependent effector functions of MAITs explain the dichotomous observations.

MAITs can be selectively activated by exposure to 5-OP-RU and CpG, which increases their levels of IFNγ and cytotoxic effector molecules (granzyme B and perforin). This suggests that MAITs are potential targets for antitumor immunotherapy. However, in vivo data on activating MAIT cells to boost anticancer immunity are lacking.

Here, we report in vivo evidence that boosting MAIT cells can be deployed to treat tumors of various origins. Administration of 5-OP-RU and CpG led to robust systemic MAIT-cell activation and expansion with high expression of CD69, a pronounced effector memory phenotype, and upregulated effector molecules including IFNγ, granzyme B, and perforin. Activated and expanded MAITs induced a potent and broad antitumor response in various tumor models in two different mouse strains. This effect was not seen in MAIT-deficient Mr1^−/− mice. Loss of MR1 on tumor cells did not result in decreased antitumor function, indicating an indirect mechanism of action. Thus, our findings demonstrate that activating MAIT cells are an attractive target for cancer immunotherapy.

Seven- to 12-week-old, female C57BL/6 mice (strain code #556) were purchased from Charles River Laboratories. Female, 7- to 10-week-old albino B6(Cg)-Tyr<c-2J>/J mice (IMSR, catalog no. JAX:000058, RRID:IMSR_JAX:000058) were obtained from the Jackson Laboratory. Mr1^−/− mice were originally generated and described (3) by Susan Gilfillan (Washington University, St. Louis School of Medicine, St. Louis, MO) and were kindly provided by Daniel Barber (National Institute of Allergy and Infectious Diseases/NIH, Bethesda, MD). Male and female Mr1^−/− mice between 7 and 16 weeks of age were used for the experiments. Loss of MR1 was confirmed by genotyping of Mr1^−/− mice using PCR with the following primers: MR1 5 8763-8783: AGC TGA AGT CTT TCC AGA TCG, MR1 9188–9168 rev: ACA GTC ACA CCT GAG TGG TTG, MR1 10451-1043: GAT TCT GTG AAC CCT TGC TTC. All animals were randomized between cages at least 1 week before start of the experiments. Mice were housed at Center for Cancer Research animal facility at NCI in accordance with the National Research Council Guide for the Care and Use of Laboratory Animals. All technical procedures and experimental endpoints were approved by the NCI Division of Intramural Research Animal Care and Use Committee and listed in the animal study proposal TGOB-013.

The B16-F10 GFP and luciferase-expressing melanoma cell line (B16F10-GFP+LUC) was a gift from Ugur Sahin (Mainz, Germany; ref. 19). Cells were cultured in RPMI1640 GlutaMax (Thermo Fisher Scientific, cataog no. 61870127) +10% FCS (Corning, catalog no. MT35010CV) + 1% So-Pyruvate (Thermo Fisher Scientific, catalog no. 11360070) +1% 1 mol/L HEPES (Thermo Fisher Scientific, catalog no. 15630080) +1% NEAA (Thermo Fisher Scientific, catalog no. 11140050), and 500 ng/mL Puromycin (InvivoGen, catalog no. ant-pr-1). The CT26 colon cancer cell line (ATCC, catalog no. CRL-2638, RRID:CVCL_7256) was obtained from ATCC and cultured in RPMI GlutaMax (+10% FCS). The luciferase-expressing HCC cell line RIL-175 has been described previously (20) and was cultured in RPMI1640 GlutaMax (+10% FCS). P19 cells were obtained from the ATCC (catalog no. CRL-1825) and were cultured in standard DMEM Glutamax (Thermo Fisher Scientific, catalog no. 10564029) with 10% FBS (Thermo Fisher Scientific, catalog no. 16140071) and 1% Penstrep (Thermo Fisher Scientific, catalog no. 15140122).

All cell lines used in experiments were cultured for 3 to 10 days and for 1 to 5 passages prior to implantation. Routine testing for murine pathogens including Mycoplasma was last performed in March 2019 and was negative for all cell lines. The cell lines used in this article were not authenticated in the past year.

Bioluminescence imaging allowed us to monitor intrahepatic tumor growth of luciferase-expressing tumors (B16F10-GFP+LUC and RIL-175 LUC). Noninvasive visualization and intrahepatic tracking of tumor cell activity was performed using an in vivo bioluminescence imaging system (IVIS; Xenogen IVIS 100 Imaging System, RRID:SCR_020901). Briefly, mice received 150 mg/kg luciferin (REGIS Technologies, catalog no. 2591-17-5) intraperitoneal and bioluminescence of intrahepatic tumors was measured 8 minutes after injection and analyzed using Xenogen's IVIS living image program.

Intrahepatic tumor cell injections were performed as described in detail elsewhere (21). In brief, cell suspensions of luciferase-expressing B16F10-GFP+Luc and RIL-175-Luc or CT26 cell lines were prepared and resuspended in a 1:1 mixture of PBS and Matrigel (Corning, catalog no.354230). Intrahepatic tumor establishment was achieved by injecting 20 μL containing 2–2.5 × 10⁵ tumor cells/animal into the left lateral liver lobe.

For the subcutaneous tumor model, 1 × 10⁶ B16F10-GFP+LUC cells suspended in 100 μL PBS were injected into the right flank of 8-week-old C57BL/6 females or 8- to 20-week-old Mr1^−/− males. Mice were randomized prior to treatment initiation. Tumor size was measured as greatest diameter biweekly by a blinded examiner using a caliper. At day 21, tumors were excised and weighed.

For the lung metastasis model, 2 × 10⁵ B16F10-GFP+LUC cells suspended in 200 μL of PBS were injected intravenously into the tail vein of 8-week-old C57BL/6 mice. At day 14, lungs were flushed with PBS to remove circulating blood cells and surface metastasis were counted by a blinded examiner.

The riboflavin pathway–derived antigen 5-OP-RU was synthesized as described previously (6). A total of 10 μg of class B CpG Oligonucleotide (ODN 1826, Invivogen, catalog no. tlrl-1826-1) was diluted in 100 μL of sterile PBS ± 1 μmol/L 5-OP-RU, which was diluted from a 10 mmol/L stock solution immediately prior to intraperitoneal administration.

Isolation of liver and tumor-infiltrating mononuclear cells was performed using previously reported protocols (22). Briefly, after necropsy, tumor-bearing livers were removed and tumor nodules excised. Tumors and livers were weighed separately before livers were homogenized and filtered through 70 μm nylon mesh. Mononuclear cells (MNC) were separated from hepatocytes by density-gradient centrifugation using 90% Percoll (Cytiva, catalog no. 17089109). Intrahepatic tumors were excised from the surrounding liver tissue and isolation of tumor-infiltrating lymphocytes followed a separate protocol to distinguish between hepatic and tumor-infiltrating lymphocytes. Briefly, tumors were dissociated using a gentleMACS Octo Dissociator (Miltenyi Biotech gentleMACS Octo Dissociator Tissue Dissociator, RRID:SCR_020272). Samples were filtered through 100 and then 40 μmol/L mesh filters prior to density-gradient centrifugation with Lympholyte Cell Separation Media (Cedarlane Laboratories, catalog no. CL5035). Red blood lysis was achieved using ACK-lysis buffer (Quality Biologicals, catalog no. 118-156-101).

For the isolation of pulmonary MNCs, lungs were flushed with PBS prior to harvest. Lungs were then minced, followed by enzymatic digestion in RPMI (Thermo Fisher Scientific, catalog no. 21875034) containing 1 mg/mL Collagenase IV (Sigma-Aldrich, catalog no. C5138-1G), 1 mg/mL hyaluronidase (Sigma-Aldrich, catalog no. H3506), and 50 U/mL DNase I (Stemcell Technologies, catalog no. 100-0762). After filtering through nylon mesh, pulmonary MNCs were separated from other parenchymal cells by density-gradient centrifugation using 37% Percoll.

MR1-tetramers loaded with 5-OP-RU were kindly provided by the NIH tetramer core facility (Emory University, Atlanta, GA). MHC class-I AH1_423–431/L^d tetramers (AH-1 tetramer) allowed us to detect antigen-specific CD8⁺ T cells reactive to the immunodominant antigen of the CT26 colon cancer cell line, gp70_423–431 (AH1; ref. 23). AH1 tetramers were also provided by the NIH tetramer core facility.

Fixable Viability dye ZOMBIE-UV (BioLegend, catalog no. 423108) was applied to cell suspensions for 20 minutes at 4°C. After Fc-blocking (BioLegend, catalog no. 101302) for 15 minutes at 4°C, surface staining including tetramers was performed by incubating 1–2 × 10⁶ cells at 37°C for 30 minutes in staining buffer (BD Biosciences, catalog no. 554656). The following commercially available antibodies were used: anti-CD62L-PerCP/Cy5.5 (BioLegend, catalog no. 104432, RRID:AB_2285839), anti-CD44-PE/Cy7 (BioLegend, catalog no. 103030, RRID:AB_830787), anti-CD69-BV650 (BioLegend, catalog no. 104541, RRID:AB_2616934), anti-TCRβ-APC/Fire 750 (BioLegend, catalog no. 109246, RRID:AB_2629697), anti-NK1.1-PE (BioLegend, catalog no. 108708, RRID:AB_313395), anti-F4/80-Alexa Fluor 700 (BioLegend, catalog no. 123130, RRID:AB_2293450), anti-B220-Alexa Fluor 700 (BioLegend, catalog no. 103232, RRID:AB_493717), anti-CD11b-Alexa Fluor 700 (BioLegend, catalog no. 101222, RRID:AB_493705), anti-CD3-Alexa Fluor 594 (BioLegend, catalog no. 100240, RRID:AB_2563427), anti-CD4-BV605 (BioLegend, catalog no. 100451, RRID:AB_2564591), anti-CD4-AF700 (BD Biosciences, catalog no. 557956, RRID: AB_396956), anti-CD8-BV786 (BD Biosciences, catalog no. 563332, RRID:AB_2721167), anti-CD11b-Pacific Blue (BioLegend, catalog no. 101224, RRID:AB_755986), anti-Ly-6G-Alexa Fluor 700 (BioLegend, catalog no. 127622, RRID:AB_10643269), anti-Ly-6C-APC/Cy7 (BioLegend, catalog no. 128026, RRID:AB_10640120), and anti-FOXP3-APC (Thermo Fisher Scientific, catalog no. 17-5773-82, RRID:AB_469457).

For intracellular cytokine staining, cells were stimulated for 4 hour at 37°C with a commercially available leukocyte activation cocktail (BD Biosciences, catalog no. 550583) containing phorbol 12-myristate 13-acetate (PMA) and ionomycin as well as brefeldin A at a concentration of 2 μL/mL. After surface staining, cells were fixed and permeabilized using a Fixation/Permeabilization Solution Kit (eBioscience, catalog no. 88-8824-00). Cells were incubated in permeabilization buffer (part of the abovementioned kit) containing the following fluorochrome-labeled antibodies for half an hour at 4°C: anti-IFNγ-APC (BioLegend, catalog no. 505810, RRID:AB_315404), anti-Perforin-APC (BioLegend, catalog no. 154304, RRID:AB_2721463), anti-Granzyme B-FITC (BioLegend, catalog no. 515403, RRID:AB_2114575), anti-IL17A-FITC (BioLegend, catalog no. 506908, RRID:AB_536010).

Data for all samples were collected on a CytoFLEX LX flow cytometer (Beckman Coulter CytoFLEX Flow Cytometer, RRID:SCR_019627) and analyzed using FlowJo software (FlowJo, RRID:SCR_008520).

Three weeks after implantation of orthotopic RIL-175, tumor-bearing mice were injected intravenously with 1 μmol/L 5-OP-RU, 10 μg class B CpG ODN 1826 and brefeldin A (500 μg/mouse) diluted in 200 μL PBS or DMSO control + brefeldin A (also 500 μg/mouse) diluted in 200 μL PBS. Three hours after injection, mice were sacrificed and liver-infiltrating MNCs were isolated and stained for intracellular cytokines as outlined above.

Tumor cell lines CT26, RIL-175, and B16F10-GFP+LUC were stimulated with 5-OP-RU at indicated molarities (see figure legends) dissolved in RPMI containing 10% FCS or left unstimulated in cell culture media for 4 hours prior to harvest as described previously (17). MR1 surface expression was determined by flow cytometry using anti-MR1-APC (BioLegend, catalog no. 361108, RRID:AB_2563204) or isotype control (mouse IgG2a, κ, BioLegend, catalog no. 400222, RRID:AB_2891178) for surface staining.

The B16F10-GFP+LUC melanoma cell line was selected for targeted knockout of MR1 due to increased surface expression upon 5-OP-RU stimulation. Single-guide RNAs (sgRNA) targeting protein coding exons of Mr1 (ENSMUST00000027744.9) were designed using sgRNA Scorer 2.0 (24). Five candidates were selected for in vitro testing (Supplementary Table S1). Briefly, in vitro transcribed sgRNAs were generated using the NEB HiScribe T7 Quick High Yield Synthesis Kit (NEB, catalog no. E2050S), complexed with recombinant Cas9 protein (gift from Protein Expression Laboratory, Frederick National Lab), and subsequently transfected into P19 cells using Lipofectamine 2000 (Thermo Fisher Scientific, catalog no. 11668019). Seventy-two hours later, genomic DNA was harvested using Quick Extract (Lucigen, catalog no. QE09050) and assessed for editing efficiency using Illumina sequencing, as described previously (24, 25). The analysis pipeline used is available here: https://github.com/rajchari2/ngs_amplicon_analysis. Oligonucleotides encoding the top candidate, Mr1-1913, as well as a nontargeting sgRNA (GTGTCGTGATGCGTAGACGG) were annealed and then ligated into the LentiCRISPRv2-mCherry vector using the T4 rapid ligase according to the manufacturer's instructions (Enzymatics, catalog no. L6030-HC-L). LentiCRISPRv2-mCherry was a gift from Agata Smogorzewska (Addgene plasmid #99154; http://n2t.net/addgene:99154; RRID:Addgene_99154). The cloned guide RNA plasmid targeting Mr1 as well as the nontargeting plasmid have been deposited into Addgene (Addgene #172525 and #172526, respectively).

Lentiviral vectors encoding the CRISPR/Cas9 construct, the sgRNA (either targeting the Mr1 gene locus or control sgRNA) and mCherry were transfected into B16F10-GFP+LUC cells using Lipofectamine3000 (Invitrogen, catalog no. L3000015) according to the manufacturer's instructions. A total of 72 hours after transfection, cells were harvested and single, live, mCherry⁺ cells were sorted into single wells of a 96-well plate on a BD Influx (BD Biosciences). Monoclonal B16F10-GFP+LUC-Mr1^−/− or B16F10-GFP+LUC-Mr1^WT cell lines were generated and loss of MR1 protein surface expression was confirmed by flow cytometry after stimulation with 5-OP-RU (see “MR1 expression on tumor cell lines”).

In addition, indel occurrence was confirmed by targeted next-generation sequencing (NGS) of genomic DNA (gDNA). Isolation of gDNA from tumor cells was performed using a AllPrep DNA/RNA Mini Kit (Qiagen, catalog no. 80204) according to the manufacturer's instruction. A total of 20 ng of gDNA from knockout clones and control clones were amplified using the primers listed in Supplementary Table S2 and analyzed using Illumina sequencing to confirm the presence of insertions/deletions using the analysis pipeline described earlier (24, 25). Illumina sequencing was performed using the Illumina MiSeq machine using the MiSeq Nano 300 cycle kit and 2 × 150 bp read length configuration. The summary of edited alleles is listed in Supplementary Table S3.

Sample sizes for animal studies were guided by previous studies with similar or identical tumor models. For all readouts, examiners were blinded. Statistical analysis was performed using GraphPad Prism 9 (GraphPad Prism, RRID:SCR_002798). The differences between groups were tested either by an unpaired Student t test or one-way ANOVA (with a Tukey multiple comparison posttest). P < 0.05 was considered statistically significant.

To stimulate and expand MAIT cells in vivo, we used a combination of 5-OP-RU and the TLR9 agonist CpG (ODN 1826; ref. 6). Naïve mice received intraperitoneal 5-OP-RU + CpG or each reagent alone twice a week for 3 weeks. MAIT cells were detected using mouse MR1 tetramers loaded with 5-OP-RU and defined as B220⁻F4/80⁻CD11b⁻CD3⁺TCRβ^intMR1-5-OP-RU tetramer⁺ by flow cytometry analysis (Fig. 1A, gating strategy displayed in Supplementary Fig. S1A). We focused our study primarily on hepatic MAIT cells given the relative abundancy of MAIT cells in human livers. Consistent with previous reports (5, 6), neither 5-OP-RU nor CpG alone expanded MAIT cells, but the 5-OP-RU + CpG combination resulted in up to 70-fold expansion of hepatic MAIT cells (Fig. 1B). The hepatic MAIT cells significantly upregulated expression of the surface activation marker CD69 (Fig. 1C) and showed a pronounced effector memory (CD44⁺CD62L⁻) phenotype (Fig. 1D). These findings indicated a change of functional status upon MAIT-cell activation. MAIT cells produce both antitumor molecules such as granzymes, perforin, and IFNγ (26), and potentially tumor-promoting factors such as IL17A, which has been reported to lead to NK-cell dysfunction and mediate MAIT cell–driven tumor progression (17). Therefore, we studied the expression of these molecules in MAIT cells. After treatment with 5-OP-RU + CpG, hepatic MAIT cells significantly increased production of IFNγ, granzyme B, and perforin, whereas IL17A expression decreased (Fig. 1E,–H). These findings point toward MAIT cells switching from a Th₁₇-like phenotype to a Th₁/NK-like phenotype as a result of costimulation using 5-OP-RU + CpG.

MAIT-cell expansion and activation were also assessed in other organs. The lung is another site with an enrichment of MAIT cells, and 5-OP-RU + CpG treatment significantly increased MAIT-cell frequency in the lung (Fig. 1I). Like expanded hepatic MAIT cells, expanded pulmonary MAIT cells had higher CD69 expression (Fig. 1J). Taken together, our results demonstrate that 5-OP-RU + CpG intraperitoneal administration results in systemic expansion and activation of MAIT cells in vivo.

Next, we tested whether stimulating MAIT cells using 5-OP-RU + CpG can be used to treat malignant tumors. An established liver metastasis model was used in which B16F10-GFP+LUC melanoma cells were introduced by intrahepatic implantation into C57BL/6 wild-type (WT) mice (22). The B16F10-GFP+LUC tumor cells express luciferase, which allowed us to monitor tumor burden using a noninvasive imaging assay (19). The treatment schedule of tumor-bearing mice is illustrated in Fig. 2A. Endpoint measurement of tumor weight showed a significant reduction of liver tumor burden upon combination therapy but not with CpG or 5-OP-RU monotherapy (Fig. 2B). We saw that growth of intrahepatic B16F10-GFP+LUC tumors was significantly delayed in 5-OP-RU + CpG treated mice when tumor growth was measured by in vivo bioluminescence imaging over time (Fig. 2C). 5-OP-RU + CpG treatment also significantly prolonged survival of tumor-bearing mice (Fig. 2D). To investigate whether the antitumor function of 5-OP-RU + CpG treatment was dependent on MAIT cells, we repeated the study including MAIT cell–deficient Mr1^−/− mice. As hypothesized, in the absence of MAIT cells, 5-OP-RU + CpG treatment failed to inhibit hepatic B16F10-GFP+LUC tumor growth (Fig. 2E). In accordance with previous reports (17), tumor burden in untreated Mr1^−/− mice was significantly lower compared with tumor burden in untreated WT mice, indicating that without activation, MAIT cells may have tumor-promoting functions.

To test whether the observed antitumor effect was limited to the liver, we extended our study to other tumor models. Lung metastases were induced by intravenous delivery of B16F10-GFP+LUC tumor cells. Consistent with our findings in the liver, 5-OP-RU + CpG treatment significantly reduced the number of lung surface metastases (Fig. 3A). Beyond this, tumor growth of subcutaneous B16F10-GFP+LUC tumors was significantly suppressed by 5-OP-RU + CpG combination treatment, but not by CpG monotherapy (Fig. 3B, left). Such reduction of tumor growth was absent in Mr1^−/− mice (Fig. 3B, right).

HCCs comprise the majority (75%–85%) of primary liver malignancies. As such, we used a murine orthotopic HCC model to test the efficacy of MAIT cell–activating immunotherapy to treat liver cancer. Murine HCC tumor cells, RIL-175, were injected into livers of C57BL/6 mice, and tumor-bearing mice were given either control treatment, CpG monotherapy, or combination 5-OP-RU + CpG. Consistently, 5-OP-RU + CpG combination treatment caused a robust reduction of RIL-175 tumor weight, whereas CpG monotherapy had very little effect on intrahepatic RIL-175 tumor growth (Fig. 3C, left). In accordance with the findings in our other tumor models, addition of 5-OP-RU to CpG treatment did not significantly reduce RIL-175 tumor burden in mice that lack MAIT cells (Fig. 3C, right).

To rule out strain-specific effects, a similar study was performed using BALB/c mice. These mice received intrahepatic injection of syngeneic colon cancer CT26 cells. Again, activating MAIT cells using 5-OP-RU + CpG decreased CT26 intrahepatic tumor weights significantly (Fig. 3D).

Taken together, these findings suggest that activating MAIT cells using a combination of 5-OP-RU + CpG but not either agent alone results in broad antitumor function in several tumor models and across different mouse strains.

Immune-cell monitoring using flow cytometric analysis of tumor-infiltrating cells was performed to monitor numeric and functional changes of potential effector cells. As expected, an increased infiltration of MAIT cells into hepatic B16F10-GFP+LUC tumors was found after 5-OP-RU + CpG treatment in C57BL/6 mice but not in Mr1^−/− mice (Fig. 4A; Supplementary Fig. S1B). In addition, the number of tumor-infiltrating MAIT cells per gram of tumor was negatively correlated with tumor weight (Fig. 4B) supporting their antitumor role. Next, MAIT cells were stimulated in vivo with 5-OP-RU + CpG in the presence of brefeldin A to study cytokine production in vivo in tumor-bearing mice. We found that IFNγ production by MAIT cells derived from mice treated with 5-OP-RU + CpG was increased compared with IFNγ production by MAIT cells derived from mice not treated with 5-OP-RU + CpG (Fig. 4C). No change in IFNγ production was observed in CD8⁺ T cells (Supplementary Fig. S1C).

Increased numbers of CD8⁺ T cells but not CD4⁺ cells were detected in hepatic B16F10-GFP+LUC tumors after 5-OP-RU + CpG treatment in C57BL/6 mice (Fig. 4D and E). CD8⁺ T cells play a critical role in antitumor immunity. Using the CT26 tumor model, tumor antigen gp70_423–431 (AH1)–specific CD8⁺ T cells were detected via MHC class-I AH1_423–431/L^d tetramers (AH-1 tetramer; ref. 23). The frequency of tumor-specific, AH1-tetramer⁺ CD8⁺ T cells in the liver increased after 5-OP-RU + CpG treatment (Fig. 4F). Moreover, the presence of these hepatic AH1-tetramer⁺ CD8⁺ T cells strongly correlated with reduced tumor burden in this model (Fig. 4G), suggesting that CD8⁺ T cells may contribute to the antitumor effects of 5-OP-RU + CpG combination treatment at least in the CT26 tumor model. Another important and potential cytotoxic effector cell that is abundant in murine livers, invariant NK T (iNKT) cells, did not accumulate in hepatic B16F10-GFP+LUC tumors upon combination 5-OP-RU + CpG treatment in C57BL/6 mice (Fig. 4H). We also investigated the role of hepatic NK cells in our model, because the cross-talk between CD8⁺ T cells, NK cells, and MAIT cells has been reported to modulate tumor progression (17). In our study, tumor-infiltrating NK cells accumulated in hepatic B16F10-GFP+LUC tumors after combination 5-OP-RU + CpG treatment in C57BL/6 mice (Fig. 4I) and hepatic NK cells produced more perforin in tumor-bearing mice treated with 5-OP-RU + CpG, indicating an enhanced antitumor cytotoxicity of these cells (Fig. 4J). Immune-cell monitoring of potentially immunosuppressive cells revealed that regulatory T cells were increased in the livers of mice treated with 5-OP-RU + CpG, whereas monocytic myeloid-derived suppressor cells (M-MDSC) and polymorphonuclear MDSCs (PMN-MDSC) showed no changes (Supplementary Fig. S2A). We also investigated the immune-cell tumor microenvironment of Mr1^−/− tumor-bearing mice upon 5-OP-RU + CpG treatment. Here, we detected no changes in CD4⁺ T cell, CD8⁺ T-cell, iNKT- or NK-cell infiltrates compared with mock treatment, pointing toward a MAIT cell–mediated effect that was observed in the WT mice (Supplementary Fig. S2B).

Taken together, our results indicate that MAIT cell–mediated protection against tumors is accompanied by complex changes in the immune tumor environment with an accumulation of a variety of potential cytotoxic cells, including CD8 T cells and NK cells.

To investigate whether the observed inhibition of tumor growth by combination 5-OP-RU + CPG treatment was mediated by a direct interaction of the MAIT-cell TCR with MR1 on tumor cells, we first measured baseline expression of MR1 on the cancer cell lines used in this article. We found all three cell lines (B16F10-GFP+LUC, RIL-175, and CT26) expressed low but detectable levels of MR1 on the cell surface as determined by flow cytometry (Fig. 5A). As described previously (17), stimulation with increasing concentrations of 5-OP-RU led to upregulation of MR1 on the surface of the B16F10-GFP+LUC melanoma cell line in a dose dependent manner. Such effects were not be seen for the RIL-175 and CT26 cell lines (Fig. 5A). As reported previously (17, 27, 28), we found baseline surface expression of MR1 on other potential antigen-presenting cells, including macrophages, B cells, MDSCs and dendritic cells, hard to detect by flow cytometry (Supplementary Fig. S2C).

Thus, we decided to knockout Mr1 in B16F10-GFP+LUC cells using CRISPR/Cas9-mediated gene editing. We generated a clonal MR1-deficient B16F10-GFP+LUC-Mr1^−/− [sgRNA(MR1^−/−)] tumor cell line as well as a B16F10-GFP+LUC-Mr1^WT [sgRNA(MR1^WT)] clonal cell line using an empty-vector sgRNA. Successful knockout was confirmed by targeted NGS of the Mr1 locus (Supplementary Tables S1–S3). Loss of MR1 surface expression on clonal cells was confirmed by flow cytometry after 5-OP-RU stimulation and compared with the parenteral B16F10 cell line (Fig. 5B). Next, we tested the effect of 5-OP-RU + CpG treatment on Mr1-knockout tumor cells. 5-OP-RU + CpG effectively decreased disease burden in both B16F10-GFP+LUC-Mr1^−/− and B16F10-GFP+LUC-Mr1^WT tumor models (Fig. 5C), indicating that the MAIT cell–mediated antitumor effect was independent of direct engagement of the MAIT-cell TCR and MR1 on tumor cells.

Tumor-infiltrating MAIT cells are found in various cancers (11, 12, 14). Although the function of MAIT cells in cancer remains unclear, in vitro stimulated MAIT cells produce increased levels of antitumor effector molecules and possess a tumor-lytic function suggesting that activating MAIT cells could be potentially used to treat cancer. The MAIT cell–activating approach is particularly interesting because MAITs can be selectively stimulated with a combination of 5-OP-RU and a TLR9 agonist. In this study, we have provided in vivo evidence that activated and expanded MAIT cells promote broad antitumor immunity.

First, we used a combination of the riboflavin synthesis pathway–derived antigen 5-OP-RU and the TLR9 agonist CpG to activate these cells. Our findings confirmed that monotherapy 5-OP-RU or CpG was not sufficient to expand the MAIT-cell fraction in mice (5, 6). This may reflect a requirement for costimulation or the fact that 5-OP-RU is unstable in nature and might fail to produce sufficient MAIT-cell stimulation without adjuvant treatment. We showed that TCR-mediated and costimulation-dependent activation of MAIT cells licensed them to increase Th₁-cytokine production and enhanced expression of cytolytic perforin and granzyme B, a mechanism known from infectious disease models (29, 30). IL17A produced by MAIT cells has previously been attributed to mediating NK-cell dysfunction in murine tumor models (17). We found IL17A expression in activated MAIT cells to be downregulated upon combination treatment with 5-OP-RU and CpG, indicating that fine-tuned, context-dependent mechanisms determine MAIT-cell function in vivo.

Next, we found that in three different murine tumor models (liver tumors, lung metastases, and subcutaneous tumors) as well as two different mouse strains, MAIT cell–directed 5-OP-RU + CpG combination treatment, but not treatment with CpG or 5-OP-RU monotherapy, showed pronounced and consistent antitumor activity and prolonged survival. Antitumor effects of MAIT cells have thus far only been observed in in vitro experiments by coculturing activated MAIT cells with tumor cells (14, 18). The antitumor effects seen here are surprising given the fact that mice have extremely low overall MAIT-cell counts at baseline compared with humans. In mice, increased numbers of activated MAIT cells correlated with favorable outcome in our studies. Our findings have revealed the vast potential of licensed and educated MAIT cells for cancer immunotherapy. This is of interest, since unlike conventional T cells that are restricted by the HLA barrier, MAIT cell–directed, MR1-captured ligands could work beyond HLA restriction in genetically diverse populations (16).

In our study, we demonstrated that the efficacy of MAIT cell–directed immunotherapy using 5-OP-RU + CpG did not depend on MR1 expression on tumor cells, indicating an indirect mechanism of action, rather than MAIT–TCR engagement with MR1 on tumor cells. In a previous study, loss of MR1 expression on B16F10 cells decreased their potential to form lung metastasis (17), indicating that MAIT-cell inhibition might be mediated by tumor cells via MR1, whereas MAIT-cell activation might be mediated by another antigen-presenting cell that has not yet been determined.

Our results show that 5-OP-RU + CpG combination treatment of murine tumors increased immune-cell infiltration and improved cytotoxic function of various potential immune effector cells. NK cells and CD8⁺ T cells might, thus, be additional intermediaries of this MAIT cell–mediated therapeutic effect. It has previously been proposed that MAIT cell–NK cell interactions inhibit NK-cell function in the absence of MAIT-cell stimulation (17). Here, we provide further insight into this interaction and show that this interplay is context dependent and, importantly, could potentially be reversed by MAIT-cell activation.

To sum up, we provide a framework for how a TCR-dependent pathogenic role of MAIT cells in malignancies can be overcome using stimulatory agents. These findings pave the way for further development of MAIT cell–directed immunotherapeutic approaches for cancer.

No disclosures were reported.

B. Ruf: Conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. V.V. Catania: Investigation, writing–review and editing. S. Wabitsch: Investigation, writing–review and editing. C. Ma: Conceptualization, data curation, formal analysis, investigation, visualization, methodology, writing–original draft, writing–review and editing. L.P. Diggs: Investigation, writing–review and editing. Q. Zhang: Investigation, writing–review and editing. B. Heinrich: Investigation, writing–review and editing. V. Subramanyam: Investigation, writing–review and editing. L.L. Cui: Investigation, writing–review and editing. M. Pouzolles: Data curation, investigation, writing–review and editing. C.N. Evans: Data curation, investigation, writing–review and editing. R. Chari: Resources, formal analysis, investigation. S. Sakai: Investigation, writing–original draft. S. Oh: Resources, investigation, writing–review and editing. C.E. Barry III: Resources, funding acquisition, writing–review and editing. D.L. Barber: Resources, funding acquisition, writing–review and editing. T.F. Greten: Conceptualization, resources, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing.

Figure 2A presented in this article was created with http://biorender.com/.

The MR1 tetramer technology was developed jointly by James McCluskey, Jamie Rossjohn, and David Fairlie. and the material was produced by the NIH Tetramer Core Facility as permitted to be distributed by the University of Melbourne.

T.F. Greten was supported by the Intramural Research Program of the NIH, NCI (ZIA BC 011345). S. Wabitsch was funded by the Deutsche Forschungsgemeinschaft (WA-4610/1-1). D.L. Barber, S. Sakai, S. Oh, and C.E. Barry III are supported by the Intramural Research Program of National Institute of Allergy and Infectious Diseases.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.