Mice genetically deficient in the Lsp1 gene (Lsp1 KO) on the C57BL/6 background were kindly provided by Dr Laurent Sabbagh (University of Montreal, Montreal, Quebec, Canada).23 For the generation of Lsp1 transgenic (Lsp1 Tg) mice in which the Lsp1 gene was specifically overexpressed in T cells, mouse Lsp1 cDNA was cloned into a lymphocyte-specific expression cassette, including the human CD2 promoter. The construct was injected directly into the pronucleus of fertilized eggs and the transgenic founder was isolated by PCR of genomic DNA. To detect the Lsp1 transgene in Lsp1 Tg mice, genomic DNA was extracted from tails of WT and Lsp1 Tg mice, and then PCR analysis of the Lsp1 transgene was performed using the following primer sequences: forward, 5’-GGACTCCACCAGTCTCACTTCAG-3’ and reverse, 5’-CAGTTCAGAGGACTTCAGGCTGAT-3’. G protein signaling 7 gene (Rgs7) was used as an internal control with the primers 5’-CAACCACTTACAAGAGACCCGTA-3’ and 5’-GAGCCCTTAGAAATAACGTTCACC-3’.

For the adoptive transfer experiments using T cells, Rag1 KO mice were obtained from Jackson Laboratory (Bar Harbor, Maine, USA). All strains were in the C57BL/6 background, and age-matched and sex-matched WT C57BL/6 mice were used as a control.

The B16BL6 melanoma cell line (hereafter termed ‘B16 melanoma’) was purchased from the Korean Cell Line Bank (Seoul, Korea). The Lewis lung carcinoma (LLC) cell line was obtained from the American Type Culture Collection (Manassas, Virginia, USA). The MC38 colon adenocarcinoma cell line was kindly provided by Dr Tai-Gyu Kim (The Catholic University of Korea, Seoul, Korea). All cell lines were cultured in Dulbecco’s Modified Eagle Medium (Welgene, Gyeongsan, Korea) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL streptomycin and 0.25 μg/mL Fungizone (Gibco; Thermo Fisher Scientific, Waltham, Massachusetts, USA). All cell lines used in this study were negative for Mycoplasma, when tested using an e-Myco Mycoplasma PCR Detection Kit (iNtRON Biotec, Seongnam, Korea). After being resuspended in phosphate buffered saline (PBS), 5×10⁵ B16 melanoma cells, 1×10⁵ MC38 cells or 2.5×10⁵ LLC cells were injected subcutaneously into the right flank of mice aged 8–12 weeks. Tumor volumes were measured every 2–3 days with a caliper and calculated according to the following formula: V(mm³)=D×d²×0.52, where D (mm) and d (mm) are the largest and smallest perpendicular tumor diameters, respectively.

For isolation of TILs, the mice were sacrificed when the average WT tumor volume reached 500 or 700 mm³. After euthanizing mice, primary tumors were excised and dissociated by mechanical force. Tumor cell suspensions were obtained after filtration with a 70 μm cell strainer. TILs were isolated using Ficoll-Hypaque (GE Healthcare, Chicago, Illinois, USA) density gradient centrifugation. The single-cell suspensions were washed in PBS and then subjected to flow cytometry analysis or in vitro re-stimulation for intracellular cytokine staining.

Single-cell suspensions were prepared from the tumor tissues and spleens of WT, Lsp1 KO and Lsp1 Tg mice after tumor inoculation. Surface staining was performed for 30 min at 4°C with the following fluorochrome-labeled antimouse Abs: CD45 (30-F11, BD Pharmingen, Franklin Lakes, New Jersey, USA), CD3 (145-2 C11, Invitrogen, Carlsbad, California, USA), CD4 (GK1.5, Biolegend, San Diego, California, USA), CD8 (53–6.7, Biolegend), CD19 (eBio 1D3, Invitrogen), NK1.1 (PK136, Biolegend), CD11b (M1/70, Biolegend), Ly-6C (HK1.4, Invitrogen), Ly-6G (1A8, Biolegend), F4/80 (BM8, Invitrogen), CD206 (C068C2, Biolegend), major histocompatibility complex (MHC) class II (M5/114.15.2, Biolegend) and CXCR3 (S18001A, Biolegend). To detect cytokine production, splenocytes and TILs were re-stimulated in vitro with phorbol-12-myristate-13-acetate (50 ng/mL, Sigma, St. Louis, MO) plus ionomycin (500 ng/mL, Sigma) for 4 hours in the presence of GolgiPlug (BD Bioscience). After surface staining, cells were fixed, permeabilized and stained with the following fluorochrome-labeled Abs: interleukin (IL)-2 (JES6-5H4, Biolegend), TNF-α (MP6-XT22, Biolegend), IFN-γ (XMG1.2, BD Biosciences) and Foxp3 (NRRF-30, Invitrogen) for 1 hour at 4°C. Intracellular expression of LSP1 in T cells also was detected by flow cytometry. In brief, cells were stained with surface markers for 30 min at 4°C. After surface staining, cells were fixed, permeabilized and incubated with rabbit anti-LSP1 Ab (Cell Signaling Technology, Danvers, Massachusetts, USA) or recombinant rabbit IgG (Abcam, Cambridge, UK) for 1 hour and then stained with a fluorescein isothiocyanate (FITC)-conjugated secondary Ab (Invitrogen) for 30 min to detect LSP1 expression in B16-challenged TILs, or they were stained with FITC-conjugated mouse anti-LSP1 Ab (mouse IgG₁; BD Biosciences) for 1 hour to detect LSP1 expression in in vitro stimulated T cells. FITC-conjugated mouse IgG₁ (Santa Cruz Biotechnology, Dallas, Texas, USA) was used as an isotype control. Cells were resuspended in fluorescence-activated cell sorting (FACS) buffer and acquired through FACS Canto II (BD Biosciences) or LSR Fortessa (BD Biosciences) with DIVA software. All data were analyzed using FlowJo software (FlowJo, Franklin Lakes).

For immunohistochemical staining, 7 μm sections of Optimal Cutting Temperature (OCT)-embedded tumor tissues were fixed with cold acetone for 10 min at −20°C, endogenous peroxidase was quenched by incubating the sections in 0.3% H₂O₂ for 30 min at room temperature and then tissues were blocked with 10% normal donkey serum for 1 hour at room temperature. Tissue sections were incubated with rat antimouse CD4 (1:1000, Biolegend) or rat antimouse CD8 (1:1000, Biolegend) Ab overnight at 4°C. Each slide was washed three times with PBS and detected using an antirat secondary Ab (1:100, Vector Laboratories, Burlingame, California, USA) with the VECTASTAIN Elite ABC HRP kit (Vector Laboratories). Positive cells were detected using 3,3′-diaminobenzidine tetrahydrochloride (Vector Laboratories) and counterstained with hematoxylin. Images were obtained using a Pannoramic MIDI slide scanner (3DHISTECH). Positive cells were counted manually in six fields per slide, with two different slides per mouse.

Total RNA was isolated from T cells of mouse spleen using the RNeasy Mini kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. For real-time quantitative PCR (qPCR), RNA was reverse transcribed to cDNA using RevertAid Reverse Transcriptase (Thermo Fisher Scientific), and qPCR was performed on the CFX96 real-time PCR system using SYBR Green PCR premix (Bio-Rad, Hercules, California, USA). The primer sequences used for detection of Lsp1 mRNA expression were 5’-CCAGCCCTTTGGCCTTAGAA-3’ and 5’-TGGAAATGGGCAAGGTTGGT-3’. All samples were normalized to Gapdh expression detected using the primers 5’-AGGTCGGTGTGAACGGATTTG-3’ and 5’-TGTAGACCATGTAGTTGAGGTCA-3’ and relative fold-change was calculated using the 2^−ΔΔCt method.

Expression of LSP1 and phosphorylated Akt (p-Akt) in T cells was detected by western blot analysis. Briefly, stimulated T cells were lysed in a lysis buffer, and final protein concentrations were determined using a Bradford protein assay (Bio-Rad). Total protein was separated on 12% SDS-PAGE gels and transferred to a polyvinylidene fluoride membrane by electroblotting. The membranes were incubated with Abs against LSP1 (1:1000), p-Akt (1:500), Akt (1:1000, all from Cell Signaling Technology) or β-tubulin (1:1000, Abcam), followed by horseradish peroxidase-conjugated antirabbit IgG (Thermo Fisher Scientific). The membranes were visualized using an enhanced chemiluminescent detection system (Thermo Fisher Scientific).

Chemotaxis of WT, Lsp1 KO and Lsp1 Tg CD8⁺ T cells was performed in 24-well plates with 5 μm pore size Transwell inserts (Corning, Corning, New York, USA). Murine CXCL9 and CXCL10 (R&D Systems, Minneapolis, Minnesota, USA) were diluted to the indicated concentrations in migration medium (0.1% FBS in RPMI1640) and placed in the lower chamber. A half million CD8⁺ T cells were loaded into the upper chamber in migration medium. After 4 hours of incubation at 37°C, the cells that had migrated to the lower chamber were counted using a hemocytometer. In some experiments, WT and Lsp1 KO CD4⁺ T cells were co-cultured with carboxyfluorescein succinimidyl ester (CFSE, 1 μM, Thermo Fisher Scientific)-labeled WT CD8⁺ T cells in the absence or presence of anti-CD3 Ab. After 3 days, cells were harvested and resuspended in migration medium and loaded in the upper chamber after CXCL9 or CXCL10 were added in the bottom chamber. After 4 hours of incubation, CFSE-labeled CD8⁺ T cells that had migrated to the lower chamber were calculated by flow cytometry.

Splenic T cells of WT mice were cultured to examine the major stimuli and their signaling pathways to induce LSP1 expression. Briefly, T cells were isolated from the spleens and prepared as single-cell suspensions. CD4⁺ T cells or CD8⁺ T cells were purified by magnetic separation using anti-CD4 beads or anti-CD8 beads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. Purified CD4⁺ T cells or CD8⁺ T cells were stimulated with recombinant IFN-γ (10 ng/mL, R&D Systems), transforming growth factor-β (TGF-β, 2 ng/mL, R&D Systems), IL-10 (10 ng/mL, R&D Systems) or antimouse CD3ε Ab (1 µg/mL, 145-2 C11, Invitrogen) plus antimouse CD28 Ab (1 µg/mL, 37.51, Invitrogen) in complete media for 72 hours. In some experiments, ciclosporin A (Sigma), tacrolimus (FK506, Sigma) and rapamycin (Sigma) were treated to the T cells stimulated with anti-CD3/anti-CD28 Abs for 72 hours to determine whether the calcineurin pathway is involved in LSP1 expression. The cultured cells were harvested and stained to detect intracellular LSP1 expression by flow cytometry and/or western blot analysis.

T cell proliferation was assessed by flow cytometry analysis of CFSE-stained cells according to the manufacturer’s instructions. Briefly, isolated CD4⁺ or CD8⁺ T cells were resuspended in PBS at a density of 1×10⁷ cells/mL and incubated with 1 μM of CFSE (Invitrogen) for 10 min at room temperature in the dark. Stained cells were quenched using FBS for 10 min on ice. The cells were washed twice and resuspended in complete RPMI 1640 medium and then analyzed by flow cytometry. Apoptosis was measured using the FITC-Annexin V Apoptosis Detection Kit (BD Bioscience) according to the manufacturer’s protocol. In brief, cultured T cells were harvested, washed with PBS and resuspended in annexin V-binding buffer. The cells were gently mixed with FITC-annexin V and propidium iodide (PI) and then incubated for 15 min at room temperature in the dark. Subsequently, annexin V⁺ and/or PI⁺ cells were analyzed by flow cytometry.

Total RNA was isolated from splenic T cells of Lsp1 KO and WT mice, which were stimulated with anti-CD3/anti-CD28 Abs for 6 hours. The RNA was reverse-transcribed, amplified according to standard Agilent protocols, and then hybridized to an array chip (SurePrint G3 Mouse GE 8×60K Microarray, Agilent) containing 62 976 probes for 24 241 annotated genes (GSE75123). Briefly, after normalization, the log2 fold-change values and p values of each gene were calculated as previously described.13 The cut-off values of differentially expressed genes (DEGs) in Lsp1 KO T cells were as follows: |fold-change values| > |the fold-change values of 2.5th and 97.5th percentile of the empirical null distribution| and p values <0.05. Functional enrichment analysis of DEGs was performed using DAVID Bioinformatics Resources 6.8 (https://david.ncifcrf.gov/). The GOBP terms of leukocyte related were first selected based on their titles and definitions. Gene set enrichment analysis (GSEA) was performed by clusterProfiler (R package, V.3.4.6)24 and GSEA plots were generated with enrichrplot (R package).

Rag1 KO mice were inoculated subcutaneously with 1×10⁵ B16 melanoma cells in the right flank. Next day, T cells were isolated from spleens of non-tumor-bearing Lsp1 KO and Lsp1 Tg mice by magnetic separation using a Pan T cell isolation kit (Miltenyi Biotec) according to the manufacturer’s instructions. PBS as a vehicle or 1×10⁷ T cells from Lsp1 KO or Lsp1 Tg mice were injected intravenously into B16 melanoma-bearing Rag1 KO mice. Tumor growth was recorded every other day.

For anti-PD-1 blockade therapy, 5×10⁵ B16 melanoma cells were inoculated subcutaneously into the right flank of WT and Lsp1 KO mice. On days 3, 6, 9 and 12 after tumor inoculation, the tumor-bearing mice were treated intraperitoneally with 10 mg/kg anti-PD-1 Ab (RMP1-14, Bio X cell, Lebanon, New Hampshire, USA) or matched rat IgG2a isotype control (2A3, Bio X cell). Tumor growth was monitored every other day.

Statistical analysis was conducted using GraphPad Prism software. The unpaired t-test was used for statistical evaluations as indicated in each experiment, while two-way analysis of variance analysis was performed to determine the significance of tumor growth. Data are shown as the mean±SD. P values <0.05 were considered to be statistically significant.

The role of LSP1 for tumor immunity remains to be defined. To address this issue, we subcutaneously inoculated syngeneic B16 melanoma cells into WT and Lsp1 KO mice and then observed tumor growth over 3 weeks. We found that Lsp1 KO mice had significant reductions in tumor growth as compared with WT mice (figure 1A). Moreover, tumor volume and weight in Lsp1 KO mice were lower than in WT mice as determined on day 14 after tumor inoculation (figure 1B,C). To explore whether the inhibitory effect of Lsp1 deficiency on tumor growth is reproduced in other types of solid tumor, we assessed the growth of MC38 colon cancer in WT and Lsp1 KO mice. Similar to the results obtained in the B16 melanoma model, the growth of MC38 colon cancer cells was significantly diminished in Lsp1 KO mice (online supplemental figure 1A). However, there was no difference in the growth of LLC between WT and Lsp1 KO mice (online supplemental figure 1B).

To characterize the effects of Lsp1 deficiency on immune contexture in the TME, TILs in tumor-bearing WT and Lsp1 KO mice were first analyzed using flow cytometry when the average tumor volume in WT mice reached approximately 700 mm³. The results showed that tumors derived from Lsp1 KO mice had a greater number of infiltrated CD45⁺, CD3⁺, CD4⁺ and CD8⁺ T cells than those from WT mice (figure 1D) and that the frequencies of those cells were not different between WT and Lsp1 KO mice (online supplemental figure 2A,B). These results imply that reduced melanoma growth in Lsp1 KO mice may be related to the increased numbers of T cells rather than composition. A growing body of evidence suggests that the spatial distribution of TILs, specifically whether the cells are located at the center or invasive margin of a tumor, as well as immune heterogeneity of TILs, critically determines the responsiveness to antitumor immunotherapy.1 25 26 In this melanoma model, infiltration of CD8⁺ T cells, but not that of CD4⁺ T cells, was significantly higher in both the center and edge regions of the tumors in Lsp1-deficient mice than in WT mice, as assessed by immunohistochemical staining (figure 1E,F), indicating that Lsp1 deficiency promotes CD8⁺ T cell infiltration into the tumor center.

In addition to T cells, other immune cell populations contribute to the TME. We found that there were no differences in the frequencies of intratumoral NK1.1⁺ natural killer (NK) cells and CD19⁺ B cells (online supplemental figure 2C), even though the absolute numbers of TILs were higher in the tumors derived from Lsp1 KO mice. The frequency of regulatory T cells (T_reg cells: Foxp3⁺ CD4⁺ T cells), a representative subset of immunosuppressors in the tumor milieu, also showed no difference (online supplemental figure 2C). Among TILs, the myeloid cell population is another substantial component of the TME that regulates tumor growth.1 8 Interestingly, the frequency of CD11b⁺Ly6C^lowF4/80^high tumor-associated macrophages (TAMs) was also significantly decreased in the tumors of Lsp1 KO mice compared with those of WT mice, although the frequency of CD11b⁺ myeloid cells was similar between Lsp1 KO and WT mice (online supplemental figure 2D and figure 1G). Of note, among TAMs, there was a significant increase of pro-inflammatory M1-like (CD206^low MHCII^high) TAMs and a substantial decrease of anti-inflammatory M2-like (CD206^high MHCII^low) TAMs in the tumors of Lsp1 KO mice (online supplemental figure 2D and figure 1G).

Collectively, these results demonstrate that Lsp1-deficient mice establish a more favorable antitumor immune milieu by enhancing infiltration of pro-inflammatory M1-like rather than M2-like TAMs as well as that of CD8⁺ T cells.

To extend our understanding of the T cell-specific effects of LSP1 during tumor development, we generated transgenic (Tg) mice that specifically overexpress Lsp1 in T cells using CD2 promoter, as described in ‘Materials and methods’ section (figure 2A). We first confirmed that CD4⁺ and CD8⁺ T cells in Lsp1 Tg mice expressed much higher levels of Lsp1 mRNA than those in WT mice (figure 2B,C). WT and Lsp1 Tg mice were subcutaneously challenged with B16 melanoma cells, as shown in figure 1. In sharp contrast with Lsp1 KO mice, Lsp1 Tg mice showed substantial acceleration of tumor growth over a period of 3 weeks in comparison to WT mice (figure 2D). At day 14 after tumor inoculation, the volume and weight of the tumors derived from Lsp1 Tg mice were also significantly higher than those of WT mice (figure 2E,F).

Since CD2 promoter drives early expression of the transgene in the double negative stage of thymocytes, it is conceivable that high LSP1 expression in CD2⁺ T cells would affect T cell development in the thymus. Here, we found no differences in the number of total thymocytes and the frequencies of cells in thymus developmental stages from double negative cells to single positive cells between WT and Lsp1 Tg mice, as determined by flow cytometry (online supplemental figure 3A). Additionally, no differences were observed in the peripheral T cell numbers and the ratio of CD4⁺ or CD8⁺ T cell population in blood and spleens of WT versus Lsp1 Tg mice (online supplemental figure 3B,C). Together, these data suggest that CD2-driven Lsp1 overexpression does not affect T cell education and homeostasis. In support of this notion, it has been demonstrated that Lsp1 deficiency does not affect T cell development and homeostasis in the central and peripheral tissues.11

To evaluate the immune cell population in the TME of Lsp1 Tg mice, tumors were collected when tumor volume derived from WT mice reached approximately 500 mm³ and then the number and frequency of TILs were analyzed using flow cytometry (figure 2G). The results showed that the frequencies of tumor-infiltrating CD3⁺ and CD4⁺ T cells were similar between WT and Lsp1 Tg mice. However, in contrast to the results in Lsp1 KO mice, the frequency of CD8⁺ T cells was reduced in Lsp1 Tg as compared with WT mice (figure 2G). The absolute number of intratumoral CD45⁺, CD3⁺ and, especially, CD8⁺ T cells was also significantly lower in Lsp1 Tg mice than in WT mice (figure 2H). As expected, since Lsp1 overexpression is specific to T cells, no difference was found in the frequency of intratumoral NK1.1⁺ NK cells, CD19⁺ B cells, CD11b⁺ myeloid cells and even TAMs between the two types of mice (online supplemental figure 4). Collectively, these results show that specific overexpression of Lsp1 in T cells enhances B16 melanoma growth, which is associated with a decrease in the number and frequency of TILs, particularly CD8⁺ T cells.

The decreased infiltration of Lsp1 Tg CD8⁺ T cells could be due to the reduction of T cell survival and proliferation. To test this possibility, we checked the proliferation and survival of B16-challenged Lsp1 Tg versus WT T cells in the presence or absence of anti-CD3/anti-CD28 Abs using flow cytometry. As shown in online supplemental figure 5, no differences were found in the frequencies of CFSE-diluted cells or annexin V⁺ and/or PI⁺ cells in CD4⁺ and CD8⁺ T cells between the two groups of mice, suggesting that other mechanism(s) than cell proliferation and survival are responsible for the decrease of TILs in the melanoma of Lsp1 Tg mice.

Previously, we demonstrated that Lsp1 deficiency directly increases CD4⁺ T cell migration in response to stromal cell-derived factor-1, a major chemokine that is known to be involved in CD4⁺ T cell migration under arthritis-associated conditions.13 In this study, we tested whether LSP1 affects CD8⁺ T cell migration under tumor-associated conditions. It is widely accepted that the CXCR3-CXCL9/CXCL10 axis has a crucial role in driving the trafficking of CD8⁺ T cells to tumor sites.27 Moreover, activation of that axis promotes the interaction between tumor-specific T cells and dendritic cells in the TME during anti-PD-1 therapy.28 We, therefore, investigated whether LSP1 controls CD8⁺ T cell migration in response to CXCL9 and CXCL10. We found that Lsp1-deficient CD8⁺ T cells showed a greater chemotactic response to CXCL9 and CXCL10 than WT CD8⁺ T cells (figure 3A). By contrast, Lsp1-overexpressing CD8⁺ T cells displayed diminished chemotactic migration compared with WT and Lsp1-deficient CD8⁺ T cells (figure 3A), demonstrating that LSP1 negatively regulates the migration of CD8⁺ T cells. However, migration of CD8⁺ T cells stimulated with 10% FBS did not differ among the three types of CD8⁺ T cells (figure 3A), suggesting that LSP1 regulation of CD8⁺ T cell migration is specific to CXCL9 and CXCL10.

To mechanistically understand how T cell migration is modulated by LSP1 expression, we first examined the expression level of CXCR3, a specific receptor of CXCL9 and CXCL10, on T cells of WT, Lsp1 KO and Lsp1 Tg mice. As shown in online supplemental figure 6A and B, no differences in CXCR3 expression levels were observed among the three genotypes. Additionally, our microarray data showed that there was a paucity of chemokine receptor-related genes in the DEGs (online supplemental figure 6C), implying that expression of chemokine receptors, particularly with CXCR3, is not relevant to LSP1 control of T cell migration. Earlier studies demonstrated that CXCL9/10-CXCR3 axis transmits its signals through the Akt.29 Therefore, we next investigated to determine if Akt is a downstream target of LSP1 for T cell migration. As shown in figure 3B, the expression level of p-Akt was reduced in Lsp1 Tg CD8⁺ T cells on CXCL10 stimulation, as determined by western blot analysis. Conversely, CXCL10-triggered-p-Akt expression was higher in CD8⁺ T cells of Lsp1 KO mice than in those of WT mice (figure 3B), indicating that LSP1 is a negative regulator of Akt activation. CXCL9-stimulated Lsp1 Tg CD8⁺ T cells showed similar results (data not shown). Moreover, the CXCL9/10-induced increase in CD8⁺ T cell migration was almost completely abrogated by the Akt inhibitor LY294002, but not by the Erk inhibitor PD98059 or p38 inhibitor SB203580 (figure 3C). Overall, these results suggest that LSP1 inhibits CXCL9/10-induced T cell migration by regulating the extent of Akt phosphorylation.

To further demonstrate the pathological relevance of LSP1 in T cells to tumor conditions, we investigated LSP1 expression levels in TILs of B16 melanoma by flow cytometry. Specificity of the anti-LSP1 Ab was validated, as shown in figure 3D. As shown in figure 3E, CD4⁺ T cells infiltrated into B16 melanoma tissue, but not splenic CD4⁺ T cells in the same mice, exhibited substantially higher LSP1 expression than splenic T cells of non-tumor-bearing mice. CD8⁺ T cells infiltrated into B16 melanoma tissue showed similar results, demonstrating that high levels of LSP1 in T cells are possibly induced by B16 melanoma. To better understand how the upregulation of LSP1 expression occurs in tumor-infiltrating T cells in vivo, we investigated which kinds of tumor-associated stimuli can induce LSP1 expression. As shown in figure 3F, stimulation of T cells with anti-CD3/anti-CD28 Abs or IFN-γ, which is known as a pro-inflammatory cytokine abundantly produced in the TME,30 strongly increased LSP1 expression in both CD4⁺ and CD8⁺ T cells. The increase in LSP1 expression by T cell receptor (TCR) activation and IFN-γ was confirmed by western blot analysis (figure 3G). By contrast, TGF-β and IL-10, anti-inflammatory cytokines derived from the TME,8 31 failed to upregulate LSP1 expression (figure 3F,G). Ciclosporin A and tacrolimus (FK506), specific calcineurin inhibitors, markedly suppressed anti-CD3/anti-CD28 Abs-induced increase in LSP1 expression, while rapamycin failed to do so, indicating that LSP1 induction by TCR activation is calcineurin-dependent (figure 3H). Given that LSP1 in T cells negatively regulates T cell migration, these results suggest that B16 melanoma can evade the antitumor activity of host T cells by upregulating LSP1 expression in T cells within the TME.

We next questioned whether LSP1 regulation of tumor growth originates entirely from its effect on T cell migration. To answer this question, we unbiasedly analyzed the global transcriptome profile of Lsp1 KO T cells (GSE75123), which was generated by our group in a previous study.13 As compared with WT T cells, 1256 DEGs (721 upregulated and 535 downregulated DEGs) were identified in Lsp1 KO T cells under media or anti-CD3/anti-CD28 Abs stimulated conditions (figure 4A). We next tried to define the major cellular processes represented by the DEG (online supplemental table 1; all supplemental tables are available at https://www.cirad-catholic.com/supplementary-figures-data). Functional enrichment analysis demonstrated that the immune system process, response to stimulus, cell adhesion, localization, developmental process, cell signaling and cell killing were significantly enriched (p<0.01) by the DEGs in Lsp1 KO T cells (figure 4B and online supplemental table 2). Among the 17 parent GOBP terms in figure 4B, we further analyzed the child GOBP terms related to leukocyte biology and then calculated their enrichment scores. As a result, ‘leukocyte-mediated migration’ and ‘leukocyte-mediated cytotoxicity’ were significantly enriched by the DEGs upregulated in Lsp1 KO T cells, but leukocyte proliferation and survival failed to show such significance (figure 4C and online supplemental table 3). GSEA also revealed that the biological processes of ‘cell killing’ and ‘leukocyte-mediated cytotoxicity’ were increased in Lsp1 KO T cells (figure 4D), whereas ‘negative regulation of cell killing’ was decreased (data not shown), which confirms that Lsp1 governs the genes associated with T cell-mediated cytotoxicity. The volcano plots in online supplemental figure 7 show the 18 upregulated DEGs (78.2%) of the 23 involved in ‘leukocyte-mediated migration’ and 27 upregulated DEGs (81.8%) of the 33 involved in ‘cell killing’.

Cytotoxicity is one of the essential steps by which tumor-infiltrating T cells suppress tumor growth.32 Based on our microarray data, we sought to assess if loss of Lsp1 promotes the cytotoxic effector function of T cells. To this end, we measured the expression levels of IFN-γ and TNF-α, representative antitumor effector cytokines,28 33 in splenic and tumor-infiltrating T cells in WT and Lsp1 KO mice by flow cytometry. Spleen size, the number of splenocytes and the proportion of splenic CD4⁺ and CD8⁺ T cells did not differ between tumor-bearing WT and Lsp1 KO mice (online supplemental figure 8A–C). In the spleen, the frequencies of TNF-α⁺ and IFN-γ⁺ cells in CD4⁺ and CD8⁺ T cells were also similar between the two groups (online supplemental figure 8D,E). In the tumor, however, they were significantly higher in the infiltrated CD4⁺ and CD8⁺ T cells of Lsp1 KO mice than in those of WT mice (figure 5A,B), suggesting that Lsp1 deficiency increases antitumor immunity by inducing TNF-α⁺ and IFN-γ⁺ expression in tumor-infiltrating T cells.

Conversely, the frequencies of TNF-α⁺ and/or IFN-γ⁺ cells in splenic CD4⁺ and CD8⁺ cells were significantly lower in Lsp1 Tg mice than in WT mice after tumor inoculation (online supplemental figure 9D,E). In tumor tissue, Lsp1-overexpressing CD4⁺ and CD8⁺ T cells also showed markedly reduced frequencies of TNF-α⁺ and/or IFN-γ⁺ cells compared with WT CD4⁺ and CD8⁺ T cells, respectively (figure 5C,D). As a control, the frequency of IL-2⁺ cells was not different in Lsp1 Tg and WT mice. Spleen size and splenocyte numbers were also similar between the two groups of mice (online supplemental figure 9A,B). Interestingly, splenic CD8⁺ T cells of Lsp1 Tg mice were less expanded, while splenic CD4⁺ T cells of Lsp1 Tg mice were more expanded than those of WT mice (online supplemental figure 9C). Overall, these results suggest that Lsp1 overexpression in T cells promotes melanoma growth through downregulation of TNF-α⁺ and IFN-γ⁺ production by CD8⁺ T cells, in addition to a marked decrease in infiltrated CD8⁺ T cells.

On the basis of the aforementioned data, we postulated that Lsp1-deficient T cells more effectively suppress tumor growth due to their increased capacity for cytotoxicity as well as T cell trafficking. To address whether Lsp1-manipulated (eg, gene-edited) T cells have therapeutic potential and to confirm that the Lsp1 gene in T cells is required for melanoma growth, we performed adoptive transfer experiments using Lsp1 KO and Lsp1 Tg T cells in Rag1 KO mice, which are deficient in mature T and B cells.34 Prior to the adoptive transfer, we confirmed that Lsp1 KO T cells and Lsp1 Tg T cells had a similar ratio of CD4/CD8 in CD3⁺ T cells (figure 6A). As shown in figure 6B, Lsp1-deficient T cells more strongly repressed tumor progression in Rag1 KO mice challenged with B16 melanoma as compared with Lsp1-overexpressing T cells and vehicle alone (without mature T cells), which confirms that loss of Lsp1 in T cells specifically mediates the antitumor effect.

Anti-PD-1 blockade has been successfully used as an immunotherapy for a variety of advanced cancers, including melanoma.35 Finally, we investigated whether the antitumor effect of Lsp1 deficiency can be further improved by the administration of anti-PD-1 Ab since the two approaches have different antitumor mechanisms: improvement of T cell trafficking versus blockade of inhibitory immune checkpoints, respectively. To this end, WT and Lsp1 KO mice were subcutaneously challenged with B16 melanoma cells on day 0 followed by anti-PD-1 Ab or matched isotype control Ab on days 3, 6, 9 and 12 (figure 6C). As expected, treatment with anti-PD-1 Ab substantially reduced melanoma growth in WT mice (figure 6D). Of note, Lsp1 KO mice treated with anti-PD-1 blockade showed a greater antitumor effect than WT mice without or with anti-PD-1 Ab (figure 6D), indicating that the extent of the antitumor effect of Lsp1 deficiency is maintained irrespective of anti-PD-1 Ab treatment. In Lsp1 KO mice treated with anti-PD-1 Ab, the contribution of anti-PD-1 Ab treatment and Lsp1 deficiency to tumor suppression was 59.6% and 40.4%, respectively, as estimated by the degree of decrease in mean tumor volume 14 days after the melanoma inoculation (figure 6D).

Taken together, these observations suggest that genetic ablation of Lsp1 in T cells is a promising strategy to boost the therapeutic efficacy of immune checkpoint inhibitors for melanoma, including anti-PD-1 Ab.