Targeting Immunosuppressive Tumor-Associated Macrophages Using Innate T Cells for Enhanced Antitumor Reactivity

Simple Summary This study seeks to evaluate innate T cells for antitumor therapies that can target both tumor cells and immunosuppressive tumor-associated macrophages (TAMs). With their innate immune-like nature, mucosal-associated invariant T (MAIT) cells, invariant natural killer T (iNKT) cells, and gamma delta T (γδT) cells do not demonstrate the harmful graft-versus-host effects associated with many conventional αβ T cell trials, and could be engineered to target both tumor cells and anti-inflammatory TAMs. The success of these trials could suggest potent new avenues for the field of cell-based cancer immunotherapy. Abstract The field of T cell-based and chimeric antigen receptor (CAR)-engineered T (CAR-T) cell-based antitumor immunotherapy has seen substantial developments in the past decade; however, considerable issues, such as graft-versus-host disease (GvHD) and tumor-associated immunosuppression, have proven to be substantial roadblocks to widespread adoption and implementation. Recent developments in innate immune cell-based CAR therapy have opened several doors for the expansion of this therapy, especially as it relates to allogeneic cell sources and solid tumor infiltration. This study establishes in vitro killing assays to examine the TAM-targeting efficacy of MAIT, iNKT, and γδT cells. This study also assesses the antitumor ability of CAR-engineered innate T cells, evaluating their potential adoption for clinical therapies. The in vitro trials presented in this study demonstrate the considerable TAM-killing abilities of all three innate T cell types, and confirm the enhanced antitumor abilities of CAR-engineered innate T cells. The tumor- and TAM-targeting capacity of these innate T cells suggest their potential for antitumor therapy that supplements cytotoxicity with remediation of tumor microenvironment (TME)-immunosuppression.

To generate iNKT cells, PBMCs were MACS-sorted via anti-iNKT MicroBead (Miltenyi Biotec) labeling to enrich iNKT cells, which were then stimulated with donor-matched irradiated αGC-PBMCs at the ratio of 1:1, and cultured in C10 medium supplemented with human IL-7 (10 ng/mL) and IL-15 (10 ng/mL) for 2-3 weeks. If needed, the resulting iNKT cells could be further purified using FACS via human iNKT TCR antibody (Clone 6B11; BD Biosciences) staining.

In Vitro Tumor Cell Killing Assay
Tumor cells (1 × 10 4 cells per well) were co-cultured with effector cells (at ratios indicated in figure legends) in C10 medium in Corning 96-well clear-bottom black plates. After 24 h, live tumor cells were quantified by adding D-luciferin (150 mg/mL; Caliper Life Science, Hopkinton, MA, USA) to cell cultures and reading out luciferase activities using an Infinite M1000 microplate reader (Tecan, Mannedorf, Switzerland).

In Vitro Mixed Lymphocyte Reaction (MLR) Assay
Healthy donor PBMCs were irradiated at 2500 rads and used as stimulators to study the graft-versus-host (GvH) response of conventional αβT and innate T cells as responders. Stimulators (5 × 10 5 cells/well) and responders (2 × 10 4 cells/well) were co-cultured in 96-well round-bottom plates in C10 medium for 4 days; the cell culture supernatants were then collected to measure IFN-γ production using ELISA.

In Vitro Mixed Mϕ/T Reaction Assay
Healthy donor PBMC-derived M2-polarized macrophages were cocultured with donor-mismatched αβT, MAIT, iNKT, or γδT cells at a ratio of 1:1 in 96-well round-bottom plates in C10 medium for 24 h. At the end of cell culture, these cells were collected to study surface marker expression using flow cytometry, and the cell culture supernatants were collected to measure cytokine production using ELISA. In the mixed Mϕ/MAIT reaction assay, 50 nM 5-OP-RU was added to co-cultures to activate MAIT cells; 10 µg/mL LEAF TM purified anti-human MR1 (Clone 26.5, BioLegend, San Diego, CA, USA) or LEAF TM purified mouse IgG2aκ isotype control antibody (Clone MOPC-173, BioLegend, San Diego, CA, USA) was added to co-cultures to study MAIT TCR-mediated target cell killing mechanism. In the mixed Mϕ/iNKT reaction assay, 100 ng/mL αGC was added to co-cultures to activate iNKT cells; 10 µg/mL LEAF TM purified anti-human CD1d (Clone 51.1, BioLegend, San Diego, CA, USA) or LEAF TM purified mouse IgG2bκ isotype control antibody (Clone MG2b-57, BioLegend, San Diego, CA, USA) was added to co-cultures to study iNKT TCR-mediated target cell killing mechanism. In the mixed Mϕ/γδT reaction assay, 5 µM zoledronate was added to co-cultures to activate γδT cells.

Statistical Analysis
GraphPad Prism 6 (GraphPad software) was used for statistical data analysis. Student's two-tailed t test was used for pairwise comparisons. Ordinary one-way ANOVA followed by Tukey's or Dunnett's multiple comparisons test was used for multiple comparisons. Data are presented as the mean ± SEM for more than three independent experiments, unless otherwise indicated. In all figures and figure legends, "n" represents the number of samples utilized in the indicated experiments. A p value of less than 0.05 was considered significant. ns, not significant; * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.
To evaluate the antitumor capacity and TCR function of these innate T cells, we set up an in vitro tumor cell killing assay (Figure 2A,D,G). Two human tumor cell lines were used as targets, including a melanoma cell line A375 and a multiple myeloma cell line MM.1S. The two parental tumor cell lines were engineered to overexpress the firefly luciferase and enhanced GFP dual reporters (FG), and/or human CD1d, resulting in A375-FG, A375-CD1d-FG, MM.1S-FG, and MM.1S-CD1d-FG cell lines. Both A375 and MM.1S cells express MR1, which presents antigenic bacterial metabolites (e.g., 5-OP-RU) to MAIT TCR ( Figure 2B) [28], and CD277, which plays a key role in phosphorylated metabolite (e.g., zoledronate)-induced activation of γδT cells ( Figure 2H) [31]. The engineered A375-CD1d-FG and MM.1S-CD1d-FG cells express CD1d, which presents glycolipids (e.g., αGC) to iNKT TCR and activates iNKT cells ( Figure 2E and Supplementary Figure S1A) [32]. In the presence of 5-OP-RU, αGC, or zoledronate, MAIT, iNKT, and γδT cells killed tumor cells with an improved efficacy ( Figure 2C,F,I), confirming the potent antitumor capacity of these innate T cells mediated by their TCR recognition. Certain levels of tumor cell killing were observed in the absence of TCR agonists ( Figure 2F,I), indicating the intrinsic NK activating receptor (e.g., NKG2D, NKp30, and NKp44)-mediated tumor cell killing of innate T cells [4,8,28,33,34].  Graft-versus-host (GvH) response is the foremost safety concern of off-the-shelf allogeneic cellular products [1,4]. MAIT, iNKT, and γδT cells do not recognize mismatched human leukocyte antigen (HLA) molecules and protein alloantigens; as a result, these innate T cells are considered not to induce GvH responses [4,7,8,[35][36][37][38]. To verify the GvH-free feature of the innate T cells, we performed an in vitro mixed lymphocyte reaction (MLR) assay ( Figure 2G). Different from conventional αβ T cells, innate T cells, including MAIT, iNKT, and γδT cells, did not react to the mismatched healthy donor PBMCs, evidenced by the lack of IFN-γ secretion ( Figure 2K). Therefore, the GvH-free feature of innate T cells grant them a high safety profile and make them suitable for off-the-shelf allogeneic cell therapy.

Targeting Immunosuppressive M2-Polarized Macrophages by MAIT Cells
To induce M2 polarization, human PBMCs were first cultured with M-CSF for 6 days and subsequently differentiated into M2 polarized macrophages by adding the antiinflammatory stimuli IL-4 and IL-13 for another 2 days ( Figure 3A,B). Indeed, after polarization, we observed upregulation of M2 macrophage markers CD163 and CD206 ( Figure 3A-C) [39][40][41]. Graft-versus-host (GvH) response is the foremost safety concern of off-the-shelf logeneic cellular products [1,4]. MAIT, iNKT, and γδT cells do not recognize mismatch human leukocyte antigen (HLA) molecules and protein alloantigens; as a result, these nate T cells are considered not to induce GvH responses [4,7,8,[35][36][37][38]. To verify the Gv free feature of the innate T cells, we performed an in vitro mixed lymphocyte react (MLR) assay ( Figure 2G). Different from conventional αβ T cells, innate T cells, includ MAIT, iNKT, and γδT cells, did not react to the mismatched healthy donor PBMCs, e denced by the lack of IFN-γ secretion ( Figure 2K). Therefore, the GvH-free feature of nate T cells grant them a high safety profile and make them suitable for off-the-shelf logeneic cell therapy.

Targeting Immunosuppressive M2-Polarized Macrophages by MAIT Cells
To induce M2 polarization, human PBMCs were first cultured with M-CSF for 6 d and subsequently differentiated into M2 polarized macrophages by adding the anti flammatory stimuli IL-4 and IL-13 for another 2 days ( Figure 3A,B). Indeed, after pol zation, we observed upregulation of M2 macrophage markers CD163 and CD206 ( Fig  3A-C) [39][40][41].  To study the macrophage-targeting capacity of innate T cells, we set up an in vitro mixed Mϕ/T reaction assay ( Figure 3D,F). M2 macrophages that upregulated MR1 could be recognized by MAIT TCR ( Figure 3G). Unlike conventional αβ T cells, which did not target M2 macrophages ( Figure 3E), MAIT cells killed M2 macrophages effectively in the presence of 5-OP-RU, and the killing capacity was dampened by blocking MR1, attesting to the macrophage-targeting potency of MAIT cells in an MR1-antigen-MAIT TCR recognition ( Figure 2H) [42]. Interestingly, even without the addition of 5-OP-RU, MAIT cells could still effectively kill M2 macrophages, which may account for their intrinsic NK activating receptor-mediated function ( Figure 2H) [43]. The macrophage-targeting and killing by MAIT cells were correlated with the upregulation of activation markers (i.e., CD25; Figure 3I,J) and generation of pro-inflammatory cytokines (i.e., IFN-γ; Figure 3K). In the mixed Mϕ/MAIT assay, the residual M2 macrophages, after co-culturing with MAIT cells, displayed a significant reduction in MR1 expression level (Supplementary Figure S1B). One possible reason was that anti-inflammatory stimuli (i.e., IL-4 and IL-13) induced macrophage polarization and upregulation of MR1, and the MR1 high macrophages were more easily recognized and killed by MAIT cells, resulting in a residual MR1 low macrophage population ( Figure S1). Collectively, MAIT cells can potentially limit the macrophage-modulated immunosuppression by targeting and killing macrophages.

Targeting Immunosuppressive M2-Polarized Macrophages by iNKT Cells
One attractive feature of iNKT cells is that iNKT cells can alter the solid tumor immunosuppressive TME via inhibition of immunosuppressive TAMs and myeloid-derived suppressive cells (MDSCs); these cells express CD1d, and thereby can be recognized by iNKT cells (Figure 4B) [24,[44][45][46]. In an in vitro mixed Mϕ/iNKT reaction assay, iNKT cells effectively killed M2 macrophages in the absence of αGC by virtue of their NK activating receptor-mediated function ( Figure 4A,C). The addition of αGC significantly enhanced iNKT cell-induced killing of macrophages, which could be blocked using the anti-CD1d antibody ( Figure 4C), validating CD1d-antigen-iNKT TCR-mediated macrophage recognition and killing by iNKT cells. In addition, macrophage killing by iNKT cells was correlated with their enhanced cytokine secretion and activation marker expression ( Figure 4D-F). Similar to the residual M2 macrophages in the mixed Mϕ/MAIT assay ( Figure S1), these cells, after co-culturing with iNKT cells, also displayed a significant reduction in CD1d expression level (Figure 4G), indicating a prior targeting of CD1d high macrophage population by iNKT cells.

Targeting Immunosuppressive M2-Polarized Macrophages by γδT Cells
The functions of γδT cells in TME modulation are controversial. In recent years, there have been a number of ongoing reports claiming that γδT cells can raise the population of myeloid-derived suppressor cells (MDSCs) and facilitate cancer progression [47][48][49], while other studies demonstrate that zoledronic acid can induce powerful γδT cell-mediated antitumor responses, and trigger γδT cells to target monocytes and downregulate inflammatory homing [50]. Here, we utilized an in vitro Mφ/γδT cell assay to study the killing of M2 macrophages, wherein γδT cells were co-cultured with M2 macrophages with or without the addition of zoledronate ( Figure 5A). Similar to MAIT and iNKT cells, γδT cells by themselves demonstrated a strong killing capacity towards macrophages, which may have been a result of their NK function [51,52]. This macrophage killing was further enhanced by the addition of zoledronate ( Figure 5B) [31,53]. The enhanced killing capacity was correlated with the increased CD25 expression and IFN-γ secretion of γδT cells ( Figure 5C-E). Previous studies showed the generation of activating or inhibitory anti-CD277 monoclonal antibodies, which can induce or inhibit γδT cell activation and proliferation, respectively [54,55]. These antibodies can also have similar activating or suppressive effects on the γδT cells' macrophage-targeting capacity.

Targeting Immunosuppressive M2-Polarized Macrophages by γδT Cells
The functions of γδT cells in TME modulation are controversial. In recent years, there have been a number of ongoing reports claiming that γδT cells can raise the population of myeloid-derived suppressor cells (MDSCs) and facilitate cancer progression [47][48][49], while other studies demonstrate that zoledronic acid can induce powerful γδT cell-mediated antitumor responses, and trigger γδT cells to target monocytes and downregulate inflammatory homing [50]. Here, we utilized an in vitro Mϕ/γδT cell assay to study the killing of M2 macrophages, wherein γδT cells were co-cultured with M2 macrophages with or without the addition of zoledronate ( Figure 5A). Similar to MAIT and iNKT cells, γδT cells by themselves demonstrated a strong killing capacity towards macrophages, which may have been a result of their NK function [51,52]. This macrophage killing was further enhanced by the addition of zoledronate ( Figure 5B) [31,53]. The enhanced killing capacity was correlated with the increased CD25 expression and IFN-γ secretion of γδT cells ( Figure 5C-E). Previous studies showed the generation of activating or inhibitory anti-CD277 monoclonal antibodies, which can induce or inhibit γδT cell activation and proliferation, respectively [54,55]. These antibodies can also have similar activating or suppressive effects on the γδT cells' macrophage-targeting capacity.

Targeting Tumor-Associated Macrophages (TAMs) by Mesothelin-Targeting CAR (MCAR)-Engineered MAIT (MCAR-MAIT) Cells
To mimic TME and study the immunosuppressive function of TAMs, we set up a 3D tumor/TAM/T-cell organoid culture [19]. Mesothelin (MSLN) was used as the model tumor antigen. MSLN-targeting CAR-αβT (MCAR-αβT), and MSLN-targeting CAR-MAIT (MCAR-MAIT) cells were generated by transducing healthy donor PBMCs with a Lenti/MCAR lentivector and sorting for CAR + populations ( Figure 6A-C and Supplementary Figure S2A). A human ovarian cancer cell line OVCAR3-FG overexpressing MSLN and Fluc-GFP was used as a target ( Figure 6E). As a proof-of-principle, here we chose MCAR-MAIT cells as an innate T cell type to study the interactions between tumor cells TAMs, and immune cells. MACR-engineered iNKT and γδT cells would also be very interesting to study in the future. Both MCAR-αβT and MCAR-MAIT cells effectively killed OVCAR3-FG tumor cells, and 5-OP-RU could further enhance the tumor cell killing capacity of MCAR-MAIT cells, indicating their CAR/TCR dual mechanisms for targeting OVCAR3 (Figure 6D,F). In the 3D tumor organoid culture ( Figure 6G), M2-polarized macrophages suppressed MCAR-αβT-mediated killing of tumor cells ( Figure 6H). Accordingly, MCAR-αβT co-cultured with M2-polarized macrophages, compared to those that were not co-cultured with macrophages, showed decreased expression of the T cell activation marker (i.e., CD25; Figure 6I). However, MCAR-MAIT cells sustained their potent antitumor capacity in the presence of macrophages, evidenced by the comparable or even increased OVCAR3 tumor cell killing ( Figure 6K). In addition, unlike MCAR-αβT cells MCAR-MAIT cells displayed a sustained expression of T cell activation marker CD25 (Figure 6L). The targeting and killing of TAMs by MCAR-MAIT cells may account for their sustained tumor killing capacity and activation ( Figure 6J,M, and Supplementary Figure  S2B). Collectively, these results support a cancer therapy potential of human CAR-MAIT cells, which can target TAMs in TME and maintain their antitumor activity. as the mean ± SEM. ns, not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, by one-way ANOVA.

Targeting Tumor-Associated Macrophages (TAMs) by Mesothelin-Targeting CAR (MCAR)-Engineered MAIT (MCAR-MAIT) Cells
To mimic TME and study the immunosuppressive function of TAMs, we set up a 3D tumor/TAM/T-cell organoid culture [19]. Mesothelin (MSLN) was used as the model tumor antigen. MSLN-targeting CAR-αβT (MCAR-αβT), and MSLN-targeting CAR-MAIT (MCAR-MAIT) cells were generated by transducing healthy donor PBMCs with a Lenti/MCAR lentivector and sorting for CAR + populations ( Figure 6A-C and Supplementary Figure S2A). A human ovarian cancer cell line OVCAR3-FG overexpressing MSLN and Fluc-GFP was used as a target ( Figure 6E). As a proof-of-principle, here we chose MCAR-MAIT cells as an innate T cell type to study the interactions between tumor cells, TAMs, and immune cells. MACR-engineered iNKT and γδT cells would also be very interesting to study in the future. Both MCAR-αβT and MCAR-MAIT cells effectively killed OVCAR3-FG tumor cells, and 5-OP-RU could further enhance the tumor cell killing capacity of MCAR-MAIT cells, indicating their CAR/TCR dual mechanisms for targeting OVCAR3 ( Figure 6D,F). In the 3D tumor organoid culture ( Figure 6G), M2-polarized macrophages suppressed MCAR-αβT-mediated killing of tumor cells ( Figure 6H). Accordingly, MCAR-αβT co-cultured with M2-polarized macrophages, compared to those that were not cocultured with macrophages, showed decreased expression of the T cell activation marker (i.e., CD25; Figure 6I). However, MCAR-MAIT cells sustained their potent antitumor capacity in the presence of macrophages, evidenced by the comparable or even increased OVCAR3 tumor cell killing ( Figure 6K). In addition, unlike MCAR-αβT cells, MCAR-MAIT cells displayed a sustained expression of T cell activation marker CD25 ( Figure 6L). The targeting and killing of TAMs by MCAR-MAIT cells may account for their sustained tumor killing capacity and activation ( Figure 6J,M, and Supplementary Figure S2B). Collectively, these results support a cancer therapy potential of human CAR-MAIT cells, which can target TAMs in TME and maintain their antitumor activity.

Discussion
TAMs have been demonstrated to promote tumor angiogenesis and tumor cell survival [16]. A variety of studies showed that depleting TAMs or inhibiting their functions could limit tumor progression, making tumor cells targets for cancer immunotherapy. Methods include (1) reducing the TAM population using bisphosphonates and their derivatives [56], or targeting CSF-1/CSF-1R signaling [57]; (2) reverting M2 TAMs to an M1-
This study successfully established CAR-engineered innate T cells as both antitumor effector cells and remediators of TME immunosuppression through the targeting of antiinflammatory TAMs. Compared to the conventional CAR-T cell therapy, which should be coupled with TAM-targeting drugs, the CAR-engineered innate T cell therapy itself is able to effectively target both tumor cells and TAMs, improving antitumor efficacy and reducing the potential adverse side effects induced by drugs. All three innate T cell types, including MAIT, iNKT, and γδT cells, exhibited potent Mϕ-killing capacity (Figures 3-5), and these innate T cells could be engineered to express tumor-targeting CARs that maintain this function in a 3D organoid culture mimicking the TME ( Figure 6). This development could be incredibly beneficial for the advancement of the field of CAR therapy for several key reasons. As innate immune cells, the therapies examined in this paper do not have the same proclivity for GvH responses as conventional αβ T cell-based CAR therapy, making them much more promising as candidates for universal "off-the-shelf" allogeneic therapy [3,5,65]. This would greatly expand the scalability of antitumor CAR therapy, making it much more accessible to patients with limited resources and to countries with poor or unevenly distributed medical infrastructure, as well as patients with truncated time scales that cannot accommodate the time limitations of autologous CAR-T generation [66]. Additionally, the Mϕ-killing abilities established by this study provide a multifaceted approach to tumor killing, expanding the scope of CAR therapy from solely antitumor cytotoxicity to the re-engineering of the TME towards a less anti-inflammatory state [67].
These results are incredibly promising for the future of innate immune cell-based CAR therapy; however, there is much work that needs to be done before they can be applied at a large scale. The low numbers and high variabilities of these innate T cells in cancer patients, as well as the limited in vitro expansion of these cells, are considered the major factors limiting innate T cell-based translational and clinical applications. These innate T cells were tested exclusively in vitro, and require in vivo replication before their TAM-targeting nature can be fully verified. While organoid culture can mimic the TME, animal models and clinical replication are necessary to characterize the full, on-site efficacy of CAR-innate immune cells as both antitumor effector cells and TAM-targeting cells. Of note, in the TME, M1 macrophages have been shown to suppress tumor growth by engulfing the target tumor cells and through antibody-dependent cell-mediated cytotoxicity (ADCC). The proposed CAR-engineered innate T cell therapy could also potentially target M1 macrophages in the TME and limit their antitumor reactivity [68]. Further explorations on the M1 macrophagetargeting capacity of innate T cells and the corresponding changes in the antitumor function are necessary. Additionally, this study only demonstrated Mϕ-and tumor-killing abilities of MCAR-MAIT cells in an organoid culture. While there have been many studies on CAR-iNKT and γδT cells, more representative studies are needed to verify their improved antitumor efficacy as a result of Mϕ depletion [4,69].

Conclusions
Overall, these results suggest an impressive new pathway via which CAR therapy can target tumors. The efficacy of MAIT, iNKT, and γδT cells in targeting tumor-associated macrophages in vitro lays the groundwork for future studies to expand the scope of antitumor CAR therapy and target one of the main defenses of tumors against immune infiltration. This is a promising development for the expansion of CAR therapy into solid tumors, and could improve both the efficacy and accessibility of one of the most exciting fields in cancer immunotherapy.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers14112749/s1. Figure S1: CD1d expression on A375 and MM.1S parental cell lines, and phenotype changes of macrophages after co-culturing with MAIT cells; Figure   Institutional Review Board Statement: Healthy donor human peripheral blood mononuclear cells were obtained from the UCLA/CFAR Virology Core Laboratory, without identification information, under federal and state regulations. Protocols using these human cells were exempted by the UCLA Institutional Review Board (IRB).

Informed Consent Statement: Not applicable.
Data Availability Statement: Data supporting reported results are available on request from the corresponding author.