Tissue-resident memory CD4⁺ T cells play a dominant role in the initiation of antitumor immunity

Abstract

Tumor immunology has been studied extensively. Tumor immunology-based cancer immunotherapy has become one of the most promising approaches for cancer treatment. However, one of the fundamental aspects of tumor immunology – the initiation of antitumor immunity – is not fully understood. Compared to that of CD8⁺ T cells, the effect of CD4⁺ T cells on antitumor immunity has not been fully appreciated. Using a gene knockout (KO) mouse model – these mice are deficient in TCRa repertoire, specifically lacking iNKT and MAIT cells - we found that the deficiency in TCRa repertoire diversity did not affect the antitumor immunity, at least to B16BL6 melanoma and EO771 breast cancer. However, after acquiring thymocytes or splenocytes from wild-type mice, these KO mice exhibited greatly enhanced and long-lasting antitumor immunity. This enhanced antitumor immunity depended on CD4⁺ T cells, especially CD4⁺ tissue-resident memory T (T_RM) cells, but not iNKT or CD8⁺ T cells. We also present evidence that CD4⁺ T_RM cells initiate antitumor immunity through IFN-γ, and the process is dependent on NK cells. The CD4⁺ T_RM/NK axis appears to control tumor formation and development by eliminating tumor cells and modulating the tumor microenvironment. Taken together, our results demonstrated that CD4⁺ T_RM cells play a dominant role in the initiation of antitumor immunity.

Introduction

Tumor immunoediting theory has clearly outlined the interaction between the immune system and tumor cells during tumor formation and development (1). Most clinically visible tumors are in the third phase of tumor immunoediting – escape. Tumor cells evade the immune system by decreasing immune cell recognition, inducing immune cell dysfunction, or inhibiting immune cell infiltration into the tumor. Most preclinical and clinical tumor immunology studies are focused on immune response in the third phase of immunoediting, specifically on CD8⁺ T cells, because CD8⁺ T cells are the primary killer cells to eliminate tumor cells, and the accumulation of tumor-infiltrating CD8⁺ T cells is associated with a favorable prognosis (2). Indeed, CD8⁺ T cell-based cancer immunotherapies have achieved unprecedented clinical success in cancer treatment (3). However, the low number of tumor-infiltrating CD8⁺ T cells and CD8⁺ T cell dysfunction in the tumor microenvironment (TME) lead to a low response rate to cancer immunotherapy (4). Type 1 conventional dendritic cells (cDC1) play an essential role in CD8⁺ T cell infiltration and sustaining CD8⁺ T cell function in the TME (5, 6). NK cells are crucial for cDC1 migration and maturation in the TME (7, 8). The mechanism of regulating NK cell function in the TME is not fully understood.

Compared to the extensive study on CD8⁺ T cell antitumor immunity, the CD4⁺ T cell antitumor function has not been fully appreciated. Most research has focused on investigating CD4⁺ T cell helper function in antitumor immunity, although new studies have indicated that cytotoxic CD4 T cells exist in the immune system and play an important role in antitumor immunity (4, 9, 10). Some preclinical and clinical studies indicate that CD4⁺ T cells can effectively control tumor development and progression independent of CD8⁺ T cells (11–13). The CD4⁺ T cell antitumor function has shown to be even more effective than CD8⁺ T cells (14, 15). It is important to note that these studies were performed by transferring transgenic or in vitro expanded tumor antigen-specific CD4⁺ T cells into CD4⁺ T cell-depleted mice or patients. The underlying mechanism remains unclear.

While it is crucial to study the antitumor immune response in clinically visible tumors, which allows us to develop a therapeutic strategy for cancer treatment, it is equally important to study the mechanism of the initiation of antitumor immunity. Most cancer patients die from metastasis rather than the primary tumor (16). Understanding the initiation of antitumor immune response would allow for developing a therapeutic strategy to prevent tumor metastasis. It is generally accepted that immune response, such as antiviral immune response, is triggered by the activation of innate immune cells, and the activated innate immune cells activate the adaptive immune system (17). Recent research in antiviral immunity has indicated that tissue-resident memory T (T_RM) cells – which are bona fide adaptive immune cells – play potent roles in activating innate immune cells and initiating an antiviral immune response (18).

T_RM cells are a linage of memory T cells that reside in the tissues, especially in non-lymphoid tissues, such as epithelial mucosa and skin, without recirculating (19). One of the most critical functions of T_RM cells is their quick response to viral infection. CD8⁺ T_RM cells can get activated within 6 hours after viral infection (18). Activated CD8⁺ T_RM cells produce IFN-γ, IL-2, and TNFα cytokines further to activate NK cells and DC (18). Therefore, CD8⁺ T_RM cells act as sentinels to initiate an antiviral immune response (20). Although the studies on CD4⁺ T_RM cells are not as extensive as the studies on CD8⁺ T_RM cells due to some of the CD4⁺ T_RM cells lacking the typical T_RM cell markers such as CD103, many elegant studies indicate that CD4⁺ T_RM cells play the same important role as, or even more dominant role than CD8⁺ T_RM in antiviral immunity (21, 22).

Only recently, the antitumor function of T_RM cells has drawn attention in preclinical and clinical studies (23, 24). T_RM cells have been found in many types of tumors, and more T_RM cells in tumors are correlated with a more favorable prognostic outcome (25–27). Most of these studies focus on CD8⁺ T_RM cells and find that CD8⁺ T_RM cells can not only mediate durable anti-melanoma immunity but also amplify antitumor immunity by triggering antigen spreading (28, 29). How CD4⁺ T_RM cells regulate antitumor immunity has not been well defined.

Using Ja281 KO mice, we found that the deficiency in iNKT cells and TCRa repertoire diversity did not cause an apparent difference in antitumor responses compared to wild-type (WT) mice, at least for B16BL6 melanoma and E0771 breast cancer. However, transferring thymocytes or splenocytes to the Ja281 KO mice and their WT counterparts, only the cell-transferred Ja281 mice acquired potent antitumor immunity. We further found that this enhanced antitumor immunity is mediated by CD4⁺ T but not iNKT or CD8⁺ T cells. More specifically, the enhanced antitumor immunity is dominated by CD4⁺ T_RM and depends on NK cells. CD4⁺ T_RM cells can initiate an antitumor immune response. CD4⁺ T_RM/NK axis orchestrates the formation of a TME in favor of antitumor immunity.

Materials and Methods

Experimental animals:

C57BL/6N mice at the age of 6–7 weeks were purchased from NIH/Charles River Laboratories (Wilmington, MA) or Envigo (Indianapolis, IN). Ja281^−/− iNKT KO (Ja281 KO) mice with C57BL/6 background were obtained from the National Institutes of Health and bred at Washington State University Wegner Hall Vivarium in Pullman and PBS Vivarium in Spokane. Traj18 KO mice (B6(Cg)-Traj18^tm1.1Kro/J), CD45.1 mice (B6.SJL-Ptprc^a Pepc^b/BoyJ), IFN-γ KO mice (B6. 129S7-Ifng^tm1Ts/J), IL-4 KO mice (C57BL/6-Il4^tm1Nnt/J), CXCR6 KO mice (B6.129P2-Cxcr6^tm1Litt/J), and IL-17A KO mice (Il17a^{tm1.1(icre)Stck}/J) were purchased from Jackson Laboratories (Bar Harbor, ME). IFN-γ/iNKT double KO (dKO) mice were created at Washington State University Wegner Hall Vivarium through crossing Ja281 KO mice with IFN-γ KO mice. Mice were housed in a specific pathogen-free room, in plastic cages with micro-filter tops and CareFresh beddings, and were allowed free access to Purina 5001 rodent laboratory chow and sterilized Milli-Q water. The animal protocols used in this study were approved by the Institutional Animal Care and Use Committee at Washington State University.

Antibodies and reagents:

The following FITC-, PE-, APC-, PE-Cy5.5-, PE-Cy7, Biotin-, PE-eFluor 610-labeled anti-mouse antibodies were purchased from eBiosciences (San Diego, CA) or BioLegend (San Diego, CA): anti-CD3 (145-2C11), anti-CD4 (GK1.5), anti-CD8 (53–6.7), anti-CD11c (N418), anti-CD11b (M1/7), anti-Gr-1 (RB6-8C5), anti-NK1.1 (PK136), anti-CXCR6 (SA051D1), anti-PD-1 (29F.1A12), anti-Tim-3 (B8.2C12), anti-CD19 (6D5), anti-CD16/32 (Clone 93), anti-IFN-γ (XMG1.2), anti-TNFα (MP6-XT22), anti-IL-13 (eBio13A), anti-IL-10 (JES5-16E3), anti-TGF-β1 (TW7-16B4). The following antibodies were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX): anti-arginase (sc-271430), anti-NOS2 (sc-7271), anti-actin (sc-47778). MojoSort™ Streptavidin Nanobeads, MojoSort™ Mouse CD4 T Cell Isolation Kit (Cat#480033), and MojoSort™ Mouse CD8 T Cell Isolation Kit (Cat#480035) were purchased from BioLegend (San Diego, CA). FTY720 was purchased from Sigma-Aldrich (St. Louis, MO).

Cell transfer:

Thymocytes or splenocytes were isolated from the thymus or spleen of the indicated donor mice following our previously published methods (30, 31). Cells were suspended in sterilized PBS at 1.6 × 10⁸ cells/ml for thymocytes and 1 × 10⁸ cells/ml for splenocytes. Cells were transferred into recipient mice via tail vein injection at 8×10⁷ cells/mouse for thymocytes or 5×10⁷ cells/mouse for splenocytes. CD4⁺ T and CD8⁺ T cells were isolated from splenocytes using MojoSort™ Mouse CD4 T Cell Isolation Kit and MojoSort™ Mouse CD8 T Cell Isolation Kit. CD4⁺CXCR6⁺ and CD4⁺CXCR6⁻ cells were purified from splenic CD4⁺ T cells by incubating with Biotin-conjugated anti-mouse CXCR6 antibody and isolated with MojoSort™ Streptavidin Nanobeads. Purified CD4⁺, CD8⁺, CD4⁺CXCR6⁺, and CD4⁺CXCR6⁻ cells were transferred into recipient mice via tail vein injection at 2–5 ×10⁶ cells/mouse.

FTY720 administration:

FTY720 was dissolved in sterilized Milli-Q water at 5 mg/ml to make a stock solution. The mouse was injected i.p. with 25 ug of FTY720 in 100 ul of sterilized Milli-Q water daily for five consecutive days, started two days before tumor inoculation.

NK, CD4, and CD8 T cell depletion with respective antibodies:

Two weeks after cell transfer and one day before tumor inoculation, Ultra-LEAF™ Purified anti-Asialo-GM1 Antibody (BioLegend, San Diego, CA) was injected i.p. into mice at 250 ug/mouse in 200 ul of PBS to deplete NK cells. For CD4⁺, and CD8⁺ T cell depletion, 250 ug of anti-mouse CD4 (GK1.5, BioLegend, San Diego, CA) or anti-mouse CD8 (53–6.7, BioLegend, San Diego, CA) antibody was injected i.p. into mice two weeks after cell transfer five days before tumor inoculation.

Depletion of circulating CD45.1 cells:

Two weeks after CD45.1⁺ cell transfer, anti-CD45.1 antibody (Clone: A20, Biolegend, San Diego, CA) was injected i.p. into mice at 10 ug/mouse. Tumor cells were inoculated two days after anti-CD45.1 antibody injection.

Tumor cell culture, tumor inoculation, measurement of subcutaneous tumor growth, and lung melanoma growth:

B16BL6 and EO771 cells were cultured in Dulbecco′s Modified - Eagle′s medium‎ supplemented with 10% FBS, 1% ampicillin, and streptomycin, in an incubator at 37°C and 5% CO₂. Cells were harvested when they reached 70% confluence and suspended in sterilized PBS. For B16BL6 melanoma subcutaneous (sc) injection, tumor cells were inoculated in the right hip area at 2 × 10⁵ cells in 200 ul of PBS. For EO771 breast cancer cell inoculation, 1×10⁶ tumor cells in 100 ul of PBS were injected into the 4^th mammary gland fat pad. Tumor size was measured by a caliper every other day, one week after tumor inoculation when the tumor was palpable and calculated by the formula: $v=\frac{a*b^2}{2}$, where v is the volume of the tumor, a is the length of the tumor, and b is the width of the tumor. For lung melanoma growth, B16BL6 tumor cells were injected via tail vein at 5 × 10⁴ cells in 200 ul of PBS. Mice were euthanized three weeks after tumor inoculation. The lungs were fixed with Fekete’s solution. The tumor colonies on five lobes of the lung were counted using a stereomicroscope. Without specific indication, all the tumor cells were inoculated two weeks after cell transfer.

Leukocyte isolation and flow cytometry analysis:

Leukocytes from the thymus, spleen, blood and lymph nodes (LN) were isolated following our previously published methods (31). The protocol for tumor-infiltrating leukocytes isolation was modified from a method for immune cell isolation developed by Blom et al. (32). Briefly, tumor tissue was rinsed with ice-cold PBS and crushed with a plastic syringe plunger to pass through a stainless wire mesh strainer. Cells were suspended in 50 ml of PBS. The cell suspension was centrifuged at 60 g for 1 min. The supernatant was transferred into a fresh tube and centrifuged at 480 g for 10 min. The cell pellet was re-suspended in 10 ml of 37.5% Percoll in PBS and centrifuged at 850 g for 30 min. The cell pellet was re-suspended in 5 ml of RBC lysis buffer for 5 min. Cells were washed with PBS+0.1% BSA and collected by centrifugation. Cell phenotype was analyzed by flow cytometry following our previously published methods (33). Fluorescence Minus One (FMO) controls, including isotype controls, were used to set the gate on the interested cells. For tumor samples, anti-mouse CD45 was used to gate on leukocytes for further analysis. Beckman Coulter Gallios Flow Cytometer and Kaluza acquisition and analysis software (Beckman Coulter, Inc. Miami, FL.) were used to collect and analyze the cell phenotype data.

Lymphocyte activation and intracellular cytokine staining:

Lymphocytes were activated by 50 ng/ml PMA and 500 ng/ml ionomycin in RPMI 1640 medium supplemented with 10% FBS and 5 ug/ml Brefeldin A at 37°C for 4 h. Cytokine-producing cells were determined by flow cytometry-based intracellular staining following our previously published method (30). FMO gate strategy, Beckman Coulter Gallios Flow Cytometer, and Kaluza acquisition and analysis software (Beckman Coulter, Inc. Miami, FL.) were used to collect and analyze the cell phenotype data.

Western blotting:

Western blotting was employed to study protein expression in the tumor tissue. The method was the same as we reported previously (34).

RT-PCR:

Trizol reagent was used to extract total RNA from tumor tissue following the manufacturer’s instructions. The RNA concentration was determined by NanoDrop. cDNA was synthesized by using M-MLV reverse transcriptase (Promega, Madison, WI). SsoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad, Hercules, CA) was used to run real-time PCR with an Analytik Jena qTOWER 3 Real-Time PCR Thermal Cycler. Actin and GAPDH were used as internal controls.

Statistical analysis:

Data were analyzed by Microsoft Excel and GraphPad Prism 7. A normality test was performed. A power analysis on the number of animals to be used in each experiment was performed. The confidence level was set to 95% and the type I error was 0.05. Unpaired, two-tailed Student’s t-test or one-way ANOVA with Tucky’s test was used to determine the significance of the difference between or among groups where appropriate. Kaplan-Meier survival analysis and Log-rank (Mantel-Cox) test (conservative) were used to analyze the survival. The difference was considered significant when p<0.05.

Results

Ja281 KO mice have a normal response to B16BL6 melanoma and E0771 breast cancer; transferring WT thymocytes or splenocytes into these mice enhances their antitumor immunity significantly.

Ja281 KO mice were originally developed to target Va14Ja18 invariant NKT (iNKT) cells and have been used to study iNKT cell function, including antitumor immunity (35–37). A later study found that the PGK-neo cassette that was used to replace the Ja18 gene fragment in the Ja281 KO mice severely affected the transcription of the Ja genes upstream of Ja18, which led to about a 60% loss of the diversity of TCRa repertoire, including deficient in iNKT and mucosal-associated invariant T (MAIT) cells (38–40). Since both iNKT and MAIT cells play critical roles in antitumor immunity (35, 41), we wanted to know how the deficiency in iNKT, MAIT, and TCRa diversity in Ja281 mice and reconstitution of these T cells by transferring WT thymocytes or splenocytes into Ja281 KO mice affect antitumor immunity. We found that neither subcutaneous B16BL6 melanoma development and growth nor survival of the tumor-bearing Ja281 KO mice were affected compared to their wild-type counterparts (Fig. 1A, 1B). Surprisingly, the cell transfer significantly inhibited tumor development and growth in Ja281 KO mice (Fig. 1A). The cell transfer also greatly extended the survival of the tumor-bearing Ja281 KO mice (Fig. 1B). We also transferred the same amount of WT thymocytes or splenocytes into WT mice. The tumor development, growth, and survival of the tumor-bearing WT mice were not significantly affected by the cell transfer (Fig. 1A, 1B). These results suggest that the transfer of WT thymocytes or splenocytes greatly enhanced antitumor immunity in the Ja281 KO, but not the WT mice.

Figure 1. Transferring thymocytes or splenocytes into Ja281 KO mice significantly increases the antitumor immune response to B16BL6 melanoma and EO771 breast cancer.

A. Tumor growth of B16BL6 melanoma in female wildtype C57BL/6 or Ja281 KO mice transferred with or without thymocytes or splenocytes. a is different from b (p<0.01). B. Survival of B16BL6 melanoma-bearing female wildtype C57BL/6 or Ja281 KO mice transferred with or without thymocytes or splenocytes. Numbers in the parentheses are the days of median survival time. a is different from b (p<0.001). C. Lung colonies of B16BL6 melanoma in female Ja281 KO mice transferred with or without thymocytes or splenocytes. *p<0.05, ***p<0.001. D. Survival of EO771 tumor-bearing female wildtype C57BL/6 and Ja281 KO mice transferred with or without thymocyte. Numbers in the parentheses are the days of median survival time. In each independent experiment, each group contained 8–10 mice. Experiments are repeated at least once with similar results. C57BL/6 (WT)=wildtype C57BL/6 mice; C57BL/6_Thy=C57BL/6 mice transferred with wildtype thymocytes; C57BL/6_Spl=C57BL/6 mice transferred with wildtype splenocytes; J281 KO=Ja281 KO mice; Ja281 KO_Thy=Ja281 KO mice transferred with wildtype thymocytes; Ja281 KO_Spl=Ja281 KO mice transferred with wildtype splenocytes;

We used the intravenous injection of B16BL6 melanoma to determine whether the cell transfer-enhanced antitumor immunity in the Ja281 KO mice is only limited to a subcutaneous tumor or may also affect tumor growth in the lung. Results indicated that both thymocyte and splenocyte transfer significantly inhibited the formation of melanoma colonies in the lung in the Ja281 KO mice compared to the WT mice (Fig. 1C). However, the effect of inhibition on tumor formation induced by thymocyte transfer was more significant than splenocyte transfer did (Fig. 1C). To further determine if the cell-transfer-induced antitumor immunity in Ja281 KO mice was melanoma-specific, we tested it in the EO771 breast cancer model. Results indicated that there was no difference in survival of non-cell-transferred, tumor-bearing mice between the Ja281 KO mice and the WT mice (Fig. 1D). However, the thymocyte-transfer dramatically suppressed EO771 breast cancer development in Ja281 KO mice. In the group of 10 thymocyte-transferred Ja281 KO mice, only one mouse developed a tumor within the period of the experiment; the survival time of this mouse was significantly longer than any of the WT or Ja281 KO counterparts without cell transfer (Fig.1D). Taken together, these results indicated that the deficiency of iNKT, MAIT, and TCR repertoire diversity in Ja281 KO mice did not significantly affect antitumor immunity to melanoma and breast cancer. However, transferring thymocyte or splenocytes into Ja281 KO mice dramatically enhanced antitumor immunity compared to the WT mice that possessed an intact T cell immune system, which implies the loss of TCR repertoires in Ja281 KO mice facilitates the formation of acquired antitumor immunity by reconstitution of T cells.

The cell transfer-induced antitumor immunity in Ja281 KO mice is mediated by CD4⁺ T cells but not iNKT or CD8⁺ T cells.

The Ja281 KO mice are deficient not only in iNKT and MAIT cells but also in 60% of TCRα repertoire diversity (38). We next determined if the cell transfer-induced antitumor tumor immunity is mediated by iNKT cells. Traj18 KO mice were developed in Dr. Kronenberg’s lab by depleting only the Ja18 gene (39). The usage of all Ja genes, except for Ja18, in Traj18 KO mice is comparable to that of WT mice. Therefore, the Traj18 KO mice are deficient only in iNKT cells but have a normal repertoire of other TCRαs (39). We transferred WT thymocytes into respective Ja281 and Traj18 KO mice and transferred Traj18 KO thymocytes into Ja281 KO mice. Two weeks after the cell transfer, the mice were inoculated s.c. with B16BL6 cells. Results indicated that transferring WT thymocytes into Ja281 KO mice significantly inhibited tumor progression and increased the survival of the tumor-bearing Ja281 KO mice compared to the WT control (Fig. 2A, 2B). Instead of enhancing antitumor immunity, transferring WT thymocytes into Traj18 KO mice enhanced the tumor growth and decreased the survival of the tumor-bearing mice (Fig. 2A,2B). Transferring Traj18 KO thymocytes, which lack iNKT cells, into Ja281 KO mice generated the same effective antitumor immunity as transferring WT thymocytes in Ja281 KO mice (Fig. 2A, 2B). These results indicated that cell transfer-induced antitumor immunity in Ja281 KO mice was not mediated by iNKT cells.

Figure 2. The cell transfer-induced antitumor immunity in Ja281 KO mice is independent of iNKT or CD8⁺ T cells.

A. Tumor growth of B16BL6 melanoma in female wildtype C57BL/6 and Ja281 KO mice transferred with wildtype or iNKT KO thymocytes, or in female Traj18 KO mice transferred wildtype thymocytes. a is different from b (p<0.05) and c (p<0.001). b is different from c (p<0.01). B. Survival of B16BL6 melanoma-bearing female wildtype C57BL/6 and Ja281 KO mice transferred with wildtype or iNKT KO thymocytes, and Traj18 KO mice transferred with wildtype thymocytes. Numbers in the parentheses are the days of median survival time. ns: difference not significant; *p<0.05, **p<0.01, ***p<0.001. C: Survival of B16BL/6 melanoma-bearing female wildtype C57BL/6 and Ja281 KO mice transferred with CD4⁺ T or CD8⁺ T cells. Numbers in the parentheses are the days of median survival time. *p<0.05, ***p<0.001. D. Survival of B16BL/6 melanoma-bearing male wildtype C57BL/6 and Ja281 KO mice transferred with CD4⁺ T or CD8⁺ T cells. Numbers in the parentheses are the days of median survival time. *p<0.05, **p<0.01. In each independent experiment, each group contained 8–10 mice. Experiments are repeated once with similar results. Ja281 KO Traj18_Thy= Ja281 KO mice transferred with iNKT KO thymocytes from Traj18 KO mice; Traj18 KO WT_Thy= Traj18 KO mice transferred with wildtype thymocytes; Ja281 KO WT_Thy=Ja281 KO mice transferred with wildtype thymocytes; C57BL/6=Wildtype C57BL/6 mice; F_Ja281 KO_CD8=Female Ja281 KO mice transferred with wildtype CD8⁺ T cells from female C57BL/6 mice; F_J281 KO_CD4=Female Ja281 KO mice transferred with wildtype CD4⁺ T cells from female C57BL/6 mice; F_C57BL/6=Female wildtype C57BL/6 mice. M_Ja281 KO_CD8=Male Ja281 KO mice transferred with wildtype CD8⁺ T cells from male C57BL/6 mice; M_J281 KO_CD4=Male Ja281 KO mice transferred with wildtype CD4⁺ T cells from male C57BL/6 mice; M_C57BL/6=Male wildtype C57BL/6 mice.

Thymocytes primarily include CD4⁺ T cells and CD8⁺ T cells. NK cells in thymocytes are negligible. To further determine if CD4⁺ or CD8⁺ T cells mediate the cell transfer-induced antitumor immunity in Ja281 KO mice, we isolated CD4⁺ and CD8⁺ T cells from the spleen of WT mice and separately transferred them into Ja281 KO mice. Two weeks after the cell transfer, B16B16 melanoma cells were inoculated s.c. The survival of tumor-bearing mice was observed. This experiment was conducted in both male and female mice. Results indicated that only CD4⁺ T cell-transfer significantly increased the survival of tumor-bearing mice in both female and male mouse experiments (Fig. 2C, 2D). CD8⁺ T cell-transfer did not affect tumor-bearing mice survival compared to their WT counterparts (Fig. 2C, 2D).

To further verify that CD4⁺ T cells but not CD8⁺ T cells mediated cell-transfer-induced antitumor immunity, we used anti-CD4 and anti-CD8 antibodies to deplete CD4⁺ and CD8⁺ T cells, respectively, in the Ja281 KO mice transferred with thymocytes and determined the tumor growth and survival of the B16BL6 tumor-bearing mice. Results indicated that depletion of CD8⁺ T cells did not affect cell transfer-induced antitumor immunity in Ja281 mice. Depletion of CD4⁺ cell partially compromised cell transfer-induced antitumor immunity in Ja281 KO mice (Fig. S1).

These results indicated that cell-transfer-induced antitumor immunity in Ja281 KO mice is mediated by CD4⁺ T, but not iNKT or CD8⁺ T cells.

The antitumor immunity acquired by cell transfer in Ja281 KO mice takes time to develop but can be long-lasting once developed.

We next determined if the cell transfer-induced antitumor immunity acquired by Ja281 KO mice was transient or long-lasting. We inoculated three groups of Ja281 KO mice with B16BL6 melanoma at one week, three weeks, or three months, respectively, after they had received thymocyte transfer. Ja281 KO mice without thymocyte transfer were used as control. Results indicated no difference in tumor growth and survival between Ja281 KO mice that had received no thymocyte transfer and those that had received the transfer one week before (Fig. 3A, 3B). However, the tumor growth was significantly slower in the mice that had received thymocyte transfer three-week or three-month before than those that had received no thymocyte transfer (Fig. 3A). Consistent with tumor growth, the survival of the tumor-bearing mice that had received the thymocyte transfer three-week or three-month before was significantly longer than those that received no cell transfer (Fig. 3B). Since more than 85% of the transferred thymocytes were immature T cells, they may need time for further maturation after transferring into the recipient mice. We used the fully matured splenic CD4⁺ T cells to test this possibility. B16BL6 cells were inoculated s.c. into male Ja281 KO mice after one day, one week, or two weeks of splenic CD4⁺ T cell transfer. WT C57BL/6 mice were used as control. Results indicated no difference in tumor growth among the control, one-day, and one-week cell transferred groups. However, the tumor progression was significantly inhibited in the two-week cell-transferred group compared to the other three groups (Fig. 3C). These results indicated that cell transfer-induced antitumor immunity in Ja281 KO mice needs time to develop.

Figure 3. Cell transfer-induced antitumor immunity in Ja281 KO mice is time-dependent and long-lasting.

A. Tumor growth of B16L6 melanoma in female Ja281 KO mice without cell-transfer (Ja281 KO) and in those with the tumor cells inoculated one week (Ja281 KO Thy-1w), three weeks (Ja281 KO Thy-3w) or three months (Ja281 KO Thy-3m) after WT thymocyte transfer. B. Survival of B16BL6 tumor-bearing mice shown in A. The numbers in the parentheses are the days of median survival time.* p<0.05; ** p<0.01. C. Tumor growth of B16BL6 melanoma in male WT C57BL/6 mice without cell-transfer and in male Ja281 KO mice with tumor cells inoculated at one day (Ja281 KO CD4_1d), one week (Ja281 KO CD4_1w), or two weeks (Ja281 KO CD4_2w) after WT splenic CD4⁺ T cell transfer. a is different from b (p<0.01). D. Survival of female Ja281 KO mice inoculated with B16BL6 melanoma without (Ja281 KO) or with prior WT thymocyte transfer (two weeks before the melanoma inoculation). One year after the first tumor inoculation, the mice without tumor development were inoculated with B16BL6 cells for the second time (Ja281 KO Thy 2°tumor). The control mice were two months of age at the time of tumor inoculation. The numbers in the parentheses are the days of median survival. Each group contained 6–10 mice. ***p<0.001.

The Ja281 KO mice are deficient in not only iNKT cells, but also TCRα repertoires. This might decrease T cell cellularity to induce the transferred T cell homeostatic proliferation. T cell homeostatic proliferation could enhance antitumor immunity (42). To test this possibility, we determined the T cell numbers in the different organs of Ja281 KO mice and WT mice. Results indicated that the numbers of CD4⁺ T and CD8⁺ T cells in the spleen, lymph nodes (LN), and liver were comparable between Ja281 KO mice and their WT control (Fig. S2).

It should be noted that about 20% of mice with a two-week or longer time of cell transfer did not develop a tumor after B16BL6 melanoma inoculation. We used these mice a second time and conducted a tumor inoculation one year after the first tumor inoculation. The tumor growth was significantly slower, and the survival time was significantly longer than the control mice, although these mice with the second time of tumor injection were much older than the control group (Fig. 3D).

Collectively, these results indicated that the transferred thymocytes or splenocytes need at least two weeks to develop and acquire antitumor immunity in Ja281 KO mice. The acquired antitumor immunity is long-lasting.

Cell transfer-induced antitumor immunity in Ja281 KO mice is mediated by tissue-resident CD4⁺ T cells.

The CD4⁺ cell transfer-induced antitumor immunity in Ja281 KO mice has shown these features: 1) partial resistance to antibody depletion (Fig. S1), 2) needing at least two weeks to develop (Fig. 3A, B, C), 3) long-lasting effect (Fig. 3D). These features are the hallmarks of tissue-resident memory T cells (43). Therefore, we used FTY720-mediated lymphocyte egress blockade and low-dose antibody depletion of circulating lymphocytes to verify that the cell transfer-induced antitumor immunity in Ja281 KO mice is mediated by tissue-resident CD4⁺ T cells.

FTY720 is an antagonist of S1PR1 (44). FTY720 treatment blocks lymphocyte egress from lymph nodes to circulation (44). We treated the cell-transferred mice with FTY720 (Fig. 4A) in the EO771 tumor model. Results indicated that all WT C57BL/6 mice developed a tumor; cell-transferred Ja281 KO mice and inoculated with 100 ul of water as control did not develop a tumor; two-thirds of cell-transferred Ja281 KO mice injected with FTY720 developed a tumor. However, the tumor growth was significantly slower than the tumor growth in C57BL/6 mice (Fig. 4B). FTY720 treatment inhibits T cell IFN-γ and Granzyme B production (45). Compared to the untreated cell-transferred mice, the development of tumors in FTY720 treated mice could be associated with impaired CD4⁺ T cell function induced by FTY720 treatment. These results suggest that blocking circulating lymphocyte egress from lymphoid organs compromises the cell transfer-induced antitumor immunity.

Figure 4. Effects of FTY720 treatment or antibody depletion of circulating transferred CD4⁺ T cells on cell transfer-induced antitumor immunity in Ja281 KO mice.

A. Scheme shows the timeline of cell transfer, FTY720 injection, and tumor inoculation. B. Growth of EO771 breast tumor in WT C57BL/6 mice (C57BL/6), Ja281 KO mice transferred with WT thymocytes (Ja281 KO_Thy), or Ja281 KO mice transferred with WT thymocytes and treated with FTY720 (Ja281 KO _Thy_FTY720). C. Scheme shows the timeline of CD45.1⁺thymocyte transfer, anti-CD45.1 mAb injection, and tumor inoculation. D. Growth of EO771 breast tumor in female C57BL/6 WT mice and Ja281 KO mice transferred with CD45.1⁺ thymocytes and treated with anti-CD45.1 mAb (10 ug/mouse) (Ja281 KO_CD45.1⁺Thy_aCD45.1). Each line stands for one mouse. Four out of five cell-transferred mice treated with anti-CD45.1 mAb did not develop a tumor (the Green squires on the x-axis are the mice without tumor).

To further verify that cell transfer-induced antitumor immunity in Ja281 KO mice is mediated by CD4⁺ T_RM, we used a low dose of antibody to deplete circulating lymphocytes. T_RM cells are resistant to low-dose antibody depletion (46). We transferred CD45.1⁺ thymocytes to the Ja281 KO mice. Two weeks after cell transfer, 10 ug of anti-CD45.1 antibody was injected into the cell-transferred Ja281 KO mice (Fig. 4C). EO771 tumor cells were inoculated two days after antibody administration. Results indicated that circulating CD45.1 cells were successfully depleted by the antibody (Fig. S3), and only one-fifth of the antibody-treated mice grew EO771 breast tumor, whereas all the control mice grew a tumor (Fig. 4D). These results further support that the cell transfer-induced antitumor immunity was not from the circulating portion of the transferred cells.

Taken together, these results indicated that cell transfer-induced antitumor immunity in Ja281 KO mice was mediated by CD4⁺ T_RM cells.

The expression of IFN-γ and CXCR6 in donor cells is essential for cell transfer-induced antitumor immunity in Ja281 KO mice.

T_RM cells initiate immune response through producing cytokines (18). CD4 T cells are a complicated T cell family. Based on cytokine production, CD4⁺ cells can at least be divided into Th1, Th2, and Th17 cells that produce their representative cytokine: IFN-γ, IL-4, and IL-17A, respectively. We next determined if these cytokines were critical for cell transfer-induced antitumor immunity in Ja281 KO mice. We transferred thymocytes from these specific cytokine KO mice into Ja281 KO mice to study B16BL6 tumor growth and the survival of tumor-bearing mice. Results indicated that transferring IFN-γ KO thymocytes to Ja281 KO mice abrogated cell-transfer-induced antitumor immunity in terms of tumor growth and host survival (Fig. 5A, 5B). Transferring IL-4 KO thymocytes into Ja281 KO mice did not significantly affect tumor growth compared to the WT thymocyte-transferred group but impaired the cell transfer-induced host survival (Fig. 5A, 5B). Transferring IL-17A KO thymocytes did not affect cell-transfer-induced inhibition on tumor growth but even enhanced the cell-transfer-induced host survival (Fig. 5C, 5D). These results indicated that IFN-γ produced by the transferred cells was essential for cell-transfer-induced antitumor immunity.

Figure 5. IFN-γ and CXCR6 in donor cells are essential for cell transfer-induced antitumor immunity in Ja281 KO mice.

A. Tumor growth of B16BL6 melanoma in female Ja281 KO mice without cell transfer (Ja281 KO), transferred with WT thymocytes (Ja281 KO_Thy), IFN-γ KO thymocytes (Ja281 KO_IFN-γ KO-Thy), or IL-4 KO thymocytes (Ja281 KO_IL-4 KO-Thy). a is different from b (p<0.05). B. Survival of female Ja281 KO mice with or without cell transfer shown in A. The numbers in the parentheses are the days of median survival time. ** p<0.01. C. Growth of B16BL6 melanoma in female Ja281 KO mice without cell transfer (Ja281 KO), transferred with WT thymocytes (Ja281 KO_Thy) or IL-17A KO thymocytes (Ja281 KO_Thy-IL-17A). a is different from b (p<0.001). D. Survival of B16BL6 melanoma-bearing mice shown in C. * p<0.05, ** p<0.01. E. Growth of B16BL6 melanoma in male WT C57BL/6 control (C57BL/6), and Ja281 KO mice transferred with Male WT thymocytes (Ja281 KO_Thy), WT splenic CD4⁺CXCR6⁺ (Ja281 KO_ CD4⁺CXCR6⁺), or WT splenic CD4⁺CXCR6⁻ cells (Ja281 KO_CD4⁺CXCR6⁻). a is different from b (P<0.05).

CXCR6 is an important cell surface marker governing T_RM cell retention and localization (47, 48). We sought to determine if CXCR6 on donor cells was required for cell transfer-induced antitumor immunity in Ja281 mice. We isolated CD4⁺CXCR6⁺ and CD4⁺CXCR6⁻ splenic cells from male WT C57BL/6 mice, transferred them into male Ja281 KO mice and studied the antitumor immunity. Results indicated that transferring CD4⁺CXCR6⁺, but not CD4⁺CXCR6⁻, cells into Ja281 KO mice significantly inhibited tumor growth (Fig. 5E).

Taken together, these results indicated that the expression of IFN-γ and CXCR6 in donor cells are essential for cell transferring-induced antitumor immunity in Ja281 KO mice.

Host endogenous IFN-γ and NK cells are required for cell transfer-induced antitumor immunity in Ja281 mice.

IFN-γ plays an essential role in antitumor immunity. We next determined if the endogenous IFN-γ produced by recipient Ja281 KO mice play a role in cell transfer-induced antitumor immunity. To this end, we created a strain of Ja281/IFN-γ double KO mice by crossing Ja281 KO mice and IFN-γ KO mice. We transferred WT thymocytes into Ja281/IFN-γ double KO mice and Ja281 KO mice, respectively. Non-transferred Ja281/IFN-γ KO and Ja281 KO mice were used as controls. The tumor growth and survival of B16BL6 tumor-bearing mice were determined. Results indicated that the tumor growth in Ja281/IFN-γ KO mice was significantly faster than in Ja281 KO mice, and the survival of tumor-bearing Ja281/IFN-γ KO mice was significantly decreased as opposed to those of Ja281 KO mice (Fig. 6A, 6B). Transferring thymocytes into Ja281/IFN-γ KO mice did not affect antitumor immunity compared to the non-transferred Ja281/IFN-γ KO mice and Ja281 KO mice (Fig. 6A, 6B). In line with the results shown above, transferring thymocytes into Ja281 KO mice significantly inhibited tumor growth and increased the survival of the tumor-bearing mice (Fig. 6A, 6B). These results indicated that IFN-γ in the recipient Ja281 KO mice was required for the induction of cell transfer-induced antitumor immunity.

Figure 6. Cell transfer-induced antitumor immunity is host IFN-γ- and NK cell-dependent.

A. Tumor growth of B16BL6 melanoma in female Ja281 KO mice without cell transfer (Ja281 KO), or with WT thymocyte transfer (Ja281 KO_Thy), thymocyte transfer and NK cell depletion by anti-asialo GM1 antibody (Ja281 KO_Thy-Anti-G), or female Ja281 and IFN-γ double KO mice without cell transfer (Ja281/IFN-γ KO) or with WT thymocyte transfer (Ja281/IFN-γ KO_Thy). a is different from b and c (p<0.001). b is different from c (p<0.05). B. Survival of B16BL6 tumor-bearing mice shown in A. The numbers in the parentheses are the days of median survival. a is different from b (p<0.001) and c (p<0.001). b is different c (p<0.01).

In antiviral immunity, T_RM cells usually act as a sentinel to initiate an antiviral immune response and need to activate other effector cells to complete the antiviral immune response. Since iNKT and CD8⁺ T cells were dispensable (Fig, 2) in cell transfer-induced antitumor immunity in Ja281 KO mice, we sought to determine if NK cells were required in this cell transfer-induced antitumor immunity. We first used the anti-asialo GM1 antibody to deplete NK cells from the cell-transferred Ja281 KO mice and then studied the B16BL6 melanoma growth. Results indicated that the depletion of NK cells abrogated cell transfer-induced antitumor immunity in Ja281 KO mice (Fig. 6A, 6B).

These results indicated that the endogenous IFN-γ and NK cells are required for cell transfer-induced antitumor immunity in the Ja281 KO mice.

Cell transfer alters TME in Ja281 KO mice.

We next investigated how cell transfer affected the TME in Ja281 KO mice. Two weeks after cell transfer, mice were inoculated with B16BL6 melanoma. Tumor samples were collected three weeks after tumor inoculation. Tumor-infiltrating lymphocytes (TIL) and myeloid cells were characterized by flow cytometry. Results indicated that the number of TIL per gram of tumor tissue in the cell-transferred Ja281 KO mice was 5-fold more than that in the non-transferred Ja281 KO mice (Fig. 7A). The increased TIL in cell-transferred Ja281 KO mice was mainly composed of CD4⁺ T, CD8⁺ T, and NK cells. (Fig. 7B). The number of total CD19⁺ B cells per gram of tumor did not alter (Fig.7B). However, the frequency of B cells in the CD45⁺ leukocyte population in the tumor was significantly lower in the cell-transferred mice than in the non-transferred mice (Fig. S4A, 4B). The percentage of IFN-γ-producing T cells and NK cells was significantly higher in the tumor of cell-transferred mice than that of the non-transferred mice (Fig. 7C). Cell-transferred mice had fewer TNFα-producing cells than the non-cell-transferred control (Fig. S4C). Therefore, the ratio of IFN-γ producing cells to TNFα-producing cells was significantly higher in the tumor of cell-transferred mice than that in the tumor of non-cell-transferred mice (Fig. S4D). TNFR1 expression was significantly higher in the tumor of non-cell-transferred Ja281 KO mice compared to the cell-transferred Ja281 KO mice (Fig. S4E). The frequencies of Foxp3⁺ CD4⁺ T (Treg) and T-bet⁺CD4⁺ T cells (Th1 cells) in the cell-transferred mice were significantly higher than those in the non-cell-transferred control mice (Fig. 7D and E). The frequencies of CD69⁺CD103⁺ (T_RM) cells in both CD8⁺ and CD4⁺ T cells were much higher in the cell-transferred mice than in the non-transferred control mice (Fig. 7F). Cell transfer significantly enhanced the expression of inhibitory receptors PD-1 and Tim3 on CD8⁺ T cells (Fig. 7G). Cell-transfer significantly decreased CD11b⁺Gr1⁺ myeloid-derived suppressor cells (MDSC) and increased tumor-associated macrophages (TAM) (Fig. 7J), specifically CD11c⁺ type I TAM in the tumor (Fig. 7H, 7I). The frequency of DC in the tumor of cell-transferred mice was also significantly increased compared to their non-cell-transferred counterparts (Fig. S4F).

Figure 7. Cell transfer significantly alters the contents, phenotype and function of the lymphocytes and myeloid cells in the tumor.

Female Ja281 KO mice were transferred with thymocytes and inoculated with B16BL6 two weeks after the transfer. Leukocytes were isolated from the tumor and analyzed by flow cytometry three weeks after tumor inoculation. A. The number of TIL per gram of tumor tissue. B. The number of the indicated lymphocytes per gram of tumor tissue. C. Percentage of IFN-γ-producing cells in CD8⁺, CD4⁺ or NK cells. Cell were stimulated with PMA+ ionomycin for 4 h, the cytokine producing cells were determined by intracellular staining and flow cytometry. D. Percentage of Foxp3⁺ Treg in CD4⁺ T cells. E. Percentage of T-bet⁺ cells in CD4⁺ T cells. F. Percentage of CD69⁺CD103⁺ cells in CD8⁺ or CD4⁺ T cells. G. PD1⁺ or Tim3⁺ cells in CD8⁺ T cells. H. Dot plots show the gate strategy of CD45⁺ leukocytes (Gate M, Q), Gr1⁺CD11b⁺ MDSC (Gate N), F4/80⁺Gr1⁻ tumor associated-macrophages (TAM: Gate P), and CD11c⁺ TAM. I. Percentage of MDSC in myeloid cells (Gate M). J. Percentage of TAM in myeloid cells. K. Percentage of CD11c⁺ TAM in TAM. Each group contained 9–13 mice. Data=Mean ± SD. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Effector molecules, such as enzymes, cytokines, and chemokines, in the TME greatly affect TIL function (49). Arginase and inducible nitric oxide synthase 2 (iNOS2) are produced by MDSC and other cell types and inhibit T cell and NK cell function (50). We found that the expression of Arginase and iNOS2 in the tumor of cell-transferred Ja281 KO mice was significantly lower than that in the tumor of non-transferred Ja281 KO mice (Fig. S4G). TGFβ, IL-10, and IL-13 are the inhibitory cytokines that inhibit TIL antitumor immunity (51). The expression of these cytokines was significantly downregulated in the tumor of the cell-transferred Ja281 KO mice (Fig. S4H). XCL1 and FLT3L are important growth factors for cDC1 development and maturation (7, 8). CCL4 and CCL5 play a key role in attracting cDC1 migration to the tumor (7). CXCL9 and CXCL10 are critical chemoattractants for T cell migration into the tumor (6). The expression of CXL1, FLT3L, CCL4, CCL5, CXCL9, and CXCL10 was significantly upregulated in the tumor of the cell transferred Ja281 KO mice compared to their non-cell transferred counterparts (Fig. S4I).

These results indicated that cell transfer in Ja281 KO mice had altered the TME by inducing a TME favoring antitumor immunity.

Discussion

Using a TCRα repertoire deficient mouse model, we found that CD4⁺ T_RM cells could initiate a potent antitumor immune response – significantly inhibit B16BL6 melanoma progression in the subcutaneous tissue and in the lung. The antitumor immune response was even more pronounced in the triple-negative EO771 breast tumor model. This CD4⁺ T_RM-initiated antitumor immunity was dependent on NK cells and IFN-γ but not iNKT and CD8⁺ T cells. In addition, the CD4⁺ T_RM/NK cell axis could orchestrate the formation of TME that favors antitumor immunity.

The Ja281 KO mice were created to study iNKT cell function, including iNKT cell antitumor immunity (35). Using this strain of mice combined with iNKT cell transfer, several groups showed that iNKT cells play important roles in antitumor immunity by regulating DC and NK cell functions (52, 53). Because Ja281 KO mice are deficient in iNKT, MAIT, and other T cells, we used thymocytes – which contain all types and developmental stages of T cells, and splenocytes – which contain well-differentiated T cells, to reconstitute T cells in Ja281 KO mice to study antitumor immunity. Indeed, both splenocyte and thymocyte transfer could induce potent antitumor immunity. However, the results from Traj18 KO mice – which are only deficient in iNKT cells, as donor or recipient, unequivocally indicated that cell transfer-induced antitumor immunity in Ja281 KO mice was not mediated by iNKT cells. Antibody depletion and purified CD4⁺ and CD8⁺ T cell transfer experiments further confirmed that cell transfer-induced antitumor immunity in Ja281 KO mice was mediated by CD4⁺ T but not CD8⁺ T cells.

Several elegant studies have demonstrated that CD4⁺ T cells can orchestrate potent antitumor immunity, which may be even more effective than CD8⁺ T cell-mediated antitumor immunity (14, 15, 54). Because these CD4⁺ T cell-based antitumor studies were performed in CD4⁺ T cell-deficient mice transferred with tumor-specific CD4⁺ transgenic cells or TIL, lymphopenia and depletion of Treg might play a key role in the observed antitumor immunities. The Ja281 KO mice we used here are deficient in TCRa repertoire (38) but have no CD4⁺ T cell lymphopenia as their CD4⁺ T cell number is comparable to that of WT mice (Fig. S2). Ja281 KO mice with or without cell transfer had higher level Treg cells (Fig. 7). Thus, this CD4⁺ cell-mediated antitumor immunity in the cell-transferred Ja281 KO mice unlikely resulted from lymphopenia-induced CD4⁺ T cell activation or the lack of Treg inhibition. FTY720 treatment and low-dose antibody depletion experiments showed that the cell transfer-induced antitumor immunity in Ja281 KO mice was mediated by CD4⁺ T_RM cells.

One crucial question is why transferring the same type of donor cells could only generate antitumor immunity in Ja281 KO but not in WT mice. We speculate that there is a population of T cells that can regulate T_RM cell function, and this population of T cells is absent from the immune system of Ja281 KO mice because of the deficiency in TCR repertoire. Besides iNKT cells, MAIT cells are also missing in Ja281 KO mice (40). How the lack of MAIT cells in the Ja281 KO mice affects the outcome of antitumor immunity is unknown. It is reported that MAIT cells facilitate tumor initiation, growth, and metastasis through MR1 on tumor cells (41). The deficiency of MAIT cells in the Ja281 KO mice should enhance the antitumor immunity of these mice. However, our results indicated that Ja281 KO mice exhibited comparable antitumor immunity to WT mice. This implicates that Ja281 KO mice may lack not only protumoral but also antitumoral cells. As we mentioned above, these antitumoral cells are likely the population of cells regulating T_RM function. Cell transfer might only recover the antitumoral partitions.

CXCR6 is an important chemokine receptor governing T_RM cell retention and localization (47, 55). Our results indicated that the expression of CXCR6 on splenic CD4⁺ T cells is essential for cell transfer-induced antitumor immunity in Ja281 KO mice.

It should be noted that both male and female Ja281 KO mice exhibited cell-transfer induced antitumor immunity. However, the magnitude of the response was different. Female mice showed a stronger response compared to male mice. This phenomenon is consistent with the anti-melanoma response in WT mice (56).

In both antitumor and antiviral immune responses, usually, CD8⁺ cells play the dominant role in eliminating tumor or virally infected cells. CD4⁺ T cells play a regulatory role to help CD8⁺ T cell function. CD8⁺ T_RM cells can effectively control viral infection or tumor progression (57, 58). In this Ja281 KO mouse model, CD4⁺ T_RM cells play a decisive role in controlling antitumor immunity, and this CD4⁺ T_RM-mediated antitumor immunity is independent of CD8⁺ T cells. This provides us with a new strategy to develop tumor immunotherapy. CD4⁺ T cells regulate immune response mainly through producing cytokines to modulate other effector cells. Indeed, IFN-γ and NK cells are essential for CD4⁺ T_RM-mediated antitumor immunity. Interestingly, transferring IFN-γ KO thymocytes significantly, but not completely, abrogated cell transfer-induced antitumor immunity in Ja281 KO mice. This suggested that other cytokines produced by the transferred CD4⁺ T cells may also play a role in this CD4⁺ T_RM-mediated antitumor immunity. In antiviral immunity, besides IFN-γ, T_RM-produced IL-2 and TNFa also play a key role in activating NK cells and DC (18). The transferred CD4⁺ T cells initiate the immune response by activating NK cells. Depletion of host NK cells and IFN-γ completely abrogated CD4⁺ T cell transfer-induced antitumor immunity, suggesting that the CD4⁺ T_RM/NK axis can effectively fulfill antitumor immunity.

Cell transfer dramatically altered the TME of B16BL6 tumor in Ja281 KO mice: significantly increased TIL, specifically IFN-γ producing CD8⁺, CD4⁺, and NK cells; decreased MDSC; downregulated immunoinhibitory cytokine and enzyme expression. It was reported that the NK/DC axis defines TME in melanoma (8). Our results support that CD4⁺ T_RM/NK axis orchestrates the formation of TME. Indeed, the expression of XCL1 and Flt3L – cytokines governing DC1 proliferation and maturation – was significantly upregulated in the tumor of cell-transferred Ja281 KO mice. Interestingly, cell transfer in Ja281 KO mice increased TIL, Treg, and the expression of inhibitory receptors PD1 and Tim3 on CD8⁺ T cells. Therefore, combining cell transfer with immune checkpoint blockade might significantly enhance the efficacy of cancer immunotherapy.

In summary, using the Ja281 KO mouse model, we demonstrate that CD4⁺ T_RM cells can initiate antitumor immunity. The CD4⁺ T_RM/NK axis can effectively control tumor progression and orchestrate antitumor immunity by modulating TME formation. These results also imply that a population of T cells in the immune system controls CD4⁺ T_RM cell formation and function. These findings shed light on how CD4⁺ T_RM cells initiate and dominate antitumor immunity.

Key points:

1. CD4⁺ T_RM cells can initiate an antitumor immune response.

2. CD4⁺ T_RM/NK cell axis shapes TME and orchestrates antitumor immunity.

3. CD4⁺ T cells dominate antitumor immunity.

Abbreviations:

T_RM cells

Tissue-resident memory T cells

iNKT cells

invariant natural killer T cells

MAIT cells

Mucosal-associated invariant T cells

KO

Knockout

WT

Wild type

TME

Tumor microenvironment

TIL

Tumor-infiltrating lymphocytes