VSIG4 suppresses antitumor T cell immunity via the SLC3A2-dependent metabolic–ionic checkpoint Authors: Guoling Huang, Ruirui He, Wei Wang, Heping Wang, Yangyang Li, Bo Zeng, Panyin Shu, Ting Pan, Ming Yi, Lingyun Feng, Zhihui Cui, Xi Li, Yanyun Du, Hong Jiang, Xue Xiao, Yuan Wang, Chenhui Wang Published: 2026 ABSTRACT AbstractBackgroundImmune checkpoint blockade therapy aims to restore T-cell function within the tumor microenvironment (TME), eliciting durable antitumor responses. However, clinical response rates to PD-1/PD-L1 inhibitors remain limited, particularly in immunologically “cold” tumors such as pancreatic cancer, underscoring the need for alternative immunotherapeutic strategies. V-set and immunoglobulin domain-containing 4 (VSIG4) has been implicated in tumor progression, but its functional role in T-cell regulation and mechanisms of tumor immune evasion remain unclear.MethodsSyngeneic tumor models and human VSIG4 knock-in mice were employed to investigate the therapeutic effect of VSIG4 blockade. Mouse-specific and human-specific neutralizing antibodies against VSIG4 were administered across multiple tumor types, including pancreatic cancer. Tumor-infiltrating immune cells were analyzed by flow cytometry and functional assays. Mechanistic studies examined the interaction between VSIG4 and solute carrier family 3 member 2 (SLC3A2) and its impact on amino acid transport, ion flux, and T-cell activation.ResultsTherapeutic blockade of VSIG4 significantly suppressed tumor growth and prolonged survival in several cancer models, with particular efficacy in pancreatic cancer. VSIG4 expression correlated with tumor aggressiveness, and its blockade reactivated CD8+ T cells within the TME, enhancing intratumoral infiltration and effector function. Mechanistically, VSIG4 directly bound to SLC3A2, a chaperone for amino acid transporters, thereby impairing glutamine uptake, disrupting sodium and calcium flux, and ultimately suppressing T-cell activation. VSIG4 blockade restored nutrient and ion availability to CD8+ T cells, thereby reinvigorating antitumor immunity.ConclusionsThese findings define a previously unrecognized metabolic–ionic checkpoint axis by which VSIG4 restricts T-cell activation through SLC3A2. Blockade of VSIG4 reprograms the TME and enhances antitumor immunity, highlighting VSIG4 as a promising therapeutic target, particularly in metabolically repressive tumors such as pancreatic cancer.Keywords: T cell, Immunotherapy, Antibody, Immune Checkpoint Inhibitor, Macrophage WHAT IS ALREADY KNOWN ON THIS TOPIC Immune checkpoint inhibitors such as PD-1/PD-L1 antibodies can restore T-cell function but are largely ineffective in immunologically “cold” tumors like pancreatic ductal adenocarcinoma. A major barrier to antitumor immunity in such settings is metabolic restriction, as tumor cells outcompete lymphocytes for nutrients such as glutamine. V-set and immunoglobulin domain-containing 4 (VSIG4), a macrophage-expressed B7 family protein, has been reported to inhibit T-cell activation and is associated with poor prognosis, but its precise functional mechanism and receptor on T cells remained undefined. Immune checkpoint inhibitors such as PD-1/PD-L1 antibodies can restore T-cell function but are largely ineffective in immunologically “cold” tumors like pancreatic ductal adenocarcinoma. A major barrier to antitumor immunity in such settings is metabolic restriction, as tumor cells outcompete lymphocytes for nutrients such as glutamine. V-set and immunoglobulin domain-containing 4 (VSIG4), a macrophage-expressed B7 family protein, has been reported to inhibit T-cell activation and is associated with poor prognosis, but its precise functional mechanism and receptor on T cells remained undefined. WHAT THIS STUDY ADDS This study identifies SLC3A2 as the functional receptor for VSIG4 on T cells, forming a novel metabolic–ionic checkpoint axis. VSIG4 binding to SLC3A2 restricts glutamine uptake and ion flux, thereby suppressing T-cell activation and cytotoxicity. Antibody-mediated blockade of VSIG4 restored nutrient and ion access, reactivated CD8+ T cells, and produced potent antitumor effects, particularly in pancreatic cancer models, with synergy observed when combined with PD-1 blockade or chemotherapy. This study identifies SLC3A2 as the functional receptor for VSIG4 on T cells, forming a novel metabolic–ionic checkpoint axis. VSIG4 binding to SLC3A2 restricts glutamine uptake and ion flux, thereby suppressing T-cell activation and cytotoxicity. Antibody-mediated blockade of VSIG4 restored nutrient and ion access, reactivated CD8+ T cells, and produced potent antitumor effects, particularly in pancreatic cancer models, with synergy observed when combined with PD-1 blockade or chemotherapy. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings establish VSIG4 as a previously unrecognized immune checkpoint that suppresses T-cell activity through metabolic and ionic restriction. Therapeutically, VSIG4 neutralization provides a promising strategy to overcome resistance in cold tumors and to enhance current immunotherapies. The VSIG4–SLC3A2 axis also broadens the conceptual framework of checkpoint biology, suggesting new opportunities for metabolic–ionic interventions in cancer and potentially other immune-related diseases. These findings establish VSIG4 as a previously unrecognized immune checkpoint that suppresses T-cell activity through metabolic and ionic restriction. Therapeutically, VSIG4 neutralization provides a promising strategy to overcome resistance in cold tumors and to enhance current immunotherapies. The VSIG4–SLC3A2 axis also broadens the conceptual framework of checkpoint biology, suggesting new opportunities for metabolic–ionic interventions in cancer and potentially other immune-related diseases. Background Immune checkpoint blockade (ICB) therapies, such as programmed cell death protein-1 or ligand-1 (PD-1/PD-L1) inhibitors, have revolutionized cancer treatment by reactivating exhausted T cells.1 2 However, their efficacy varies substantially among tumor types and is particularly limited in immunologically “cold” malignancies such as pancreatic ductal adenocarcinoma (PDAC). These tumors exhibit minimal T-cell infiltration and a profoundly immunosuppressive tumor microenvironment (TME), leading to poor responses to ICB.3,7 Thus, identifying alternative mechanisms of immune evasion in “cold” tumors remains a critical goal in cancer immunotherapy. A major barrier to effective T-cell responses in the TME is nutrient competition. Tumor cells avidly consume glucose and amino acids, depriving infiltrating lymphocytes of critical metabolic substrates.8,15 Among these, glutamine (Gln) is essential for T-cell survival, mitochondrial function, and cytokine secretion.16,20 Activated T cells dramatically upregulate Gln uptake to support their proliferative and effector functions.18,21 Gln deprivation perturbs sodium (Na+) balance, impairs calcium (Ca²+) mobilization from the endoplasmic reticulum, and attenuates mechanistic target of rapamycin kinase (mTOR) signaling, collectively dampening T-cell activity.1922,24 Notably, immune checkpoint molecules such as PD-1 and cytotoxic lymphocyte-associated protein (CTLA)-4 suppress T-cell metabolism in part by downregulating Gln transporter expression, suggesting that nutrient access is tightly linked to immune regulation.25 26 Solute carrier family 3 member 2 (SLC3A2) forms heterodimeric complexes with SLC7 family transporters to mediate the uptake of Gln, leucine, and other neutral amino acids.27 28 In addition to nutrient transport, SLC3A2 regulates Na+-coupled fluxes that are critical for T-cell receptor (TCR) signaling.914 29,33 Transient Na+ influx promotes Gln metabolism and facilitates Ca²+ entry, thereby enhancing T-cell cytotoxicity.34 35 Moreover, SLC3A2 localizes to immune synapses via CD147 and contributes to downstream lymphocyte cell-specific protein-tyrosine kinase (LCK) and extracellular signal-regulated kinase (ERK) activation.36,38 However, the upstream regulatory inputs of SLC3A2 in T cells, particularly within the TME context, remain poorly defined. V-set and immunoglobulin domain-containing 4 (VSIG4), also known as CRIg, is a B7-related transmembrane protein highly expressed by tissue-resident macrophages.39 Since its initial conceptualization as a potential immune checkpoint within the B7 family framework, VSIG4 has been increasingly recognized as an immunosuppressive regulator in cancer and inflammatory settings.40,42 Previous studies have shown that VSIG4 suppresses proinflammatory macrophage polarization by modulating pyruvate metabolism and enhances M2-like states via lactate-signal transducer and activator of transcription 3 signaling and fatty acid oxidation.43,45 Consistently, multiple independent studies have demonstrated that VSIG4 promotes M2-like macrophage polarization and tumor immune escape across diverse malignancies, including colorectal cancer and renal cell carcinoma.43 46 47 VSIG4 also inhibits T-cell activation and proliferation, making it a candidate immune checkpoint molecule.48 Recent mechanistic studies from both academic and translational groups have further established VSIG4 as a functional macrophage immune checkpoint that suppresses T-cell cytotoxicity and limits immunotherapy efficacy, with genetic ablation or antibody-mediated blockade of VSIG4 restoring antitumor T-cell responses.49,51 Recent work has linked VSIG4+ tumor-associated macrophages to T-cell suppression and therapeutic resistance in multiple cancers, although the identity of the T-cell receptor for VSIG4 remains unknown.49 This immunosuppressive role of VSIG4 has been observed in a broad spectrum of tumor types, including PDAC, colorectal cancer, diffuse large B-cell lymphoma, and other solid tumors, and is frequently associated with poor prognosis and resistance to ICB. However, the mechanisms by which VSIG4 mediates antitumor immune evasion, as well as the identity of its cognate receptor, remain largely unknown.52,54 In this study, we identify SLC3A2 as a functional receptor for VSIG4 on T cells. We demonstrate that VSIG4 binds SLC3A2 to suppress Gln uptake and disrupt ion homeostasis, thereby impairing T-cell activation. Antibody-mediated VSIG4 blockade restores T-cell function and induces robust antitumor responses, especially in pancreatic cancer models. These findings define a novel VSIG4–SLC3A2 axis that integrates metabolic and ionic control to suppress immunity, offering a promising strategy to overcome resistance in “cold” tumors. Materials and methods Cell lines and cell culture The complete culture medium was prepared by adding 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin/streptomycin to the base medium. B16-F10, RM-1, EL4, HEK293T, and LLC cells were obtained from American Type Culture Collection (ATCC) and maintained in complete Dulbecco’s Modifed Eagle Medium (DMEM). MC38 and Jurkat cells were obtained from ATCC and cultured in complete RPMI 1640 supplemented with 1% sodium pyruvate. PANC-1 cells were procured from Shanghai Quicell Biotechnology, and incubated in complete DMEM. KPC cells were obtained from Dr. Hong Jiang’s laboratory at Sichuan University cultured in complete DMEM supplemented with 1% sodium pyruvate, 1% MEM nonessential amino acids, and 1% L-Gln. KPC cells were harvested from KPC mice (KrasG12D/+; TP53R172H/+; Ptf1a-CreERT2), which were bred by crossing KrasG12D mice with B6-P53R17H mice and pancreas-specific Ptfla-CreERT2 mice.55 Bone marrow-derived macrophages (BMDMs) were isolated and induced as described previously.44 HEK293F cells were grown in Freestyle 293 expression medium in a CO2 shaking incubator at 37°C, 5% CO2, and 130 rpm agitation. Mouse models and treatment All the mice were housed under SPF conditions at the University of Electronic Science and Technology of China’s animal facility, with controlled temperature (18°C–23°C), humidity (40%–60%), and a 12-hour light/dark cycle. Vsig4+/− on a C57BL/6 background were obtained from Cyagen Biosciences, Inc. To minimize the potential impact of untracked mutations or genetic background drift, experimental mice were generated using a littermate-controlled design.56 Briefly, the founder Vsig4+/− were expanded to establish a stable colony consisting of 10 breeding trios of Vsig4+/−, with each trio comprising one male and two females. Experimental Vsig4+/+ and Vsig4−/− mice were obtained as littermates from subsequent generations derived from these heterozygous intercrosses. All animals were genotyped by PCR before inclusion in the study, and age- and sex-matched littermates were used for all experimental procedures. For the subcutaneous tumor models, C57BL/6J (GemPharmatech), Vsig4+/+, Vsig4−/−, and human VSIG4 knock-in (hVSIG4-KI; Cyagen Biosciences) mice (8–10 weeks old, male or female) were injected in the right flank with MC38 (1×10⁶), B16-F10 (0.5×10⁶), or RM-1 (0.15×10⁶) cells suspended in 100 µL of phosphate-buffered saline (PBS). The tumor volume was measured via a Vernier caliper and calculated as (length) × (width)²/2. Tumors (volume was no more than 2000 mm3) were excised and weighed on euthanasia. For the orthotopic KPC model, 0.15×106 or 0.3×106 KPC cells in 50 µL of PBS with 40% Matrigel (Cat# 354234, Corning) were inoculated into the tail of the pancreas under anesthesia and harvested 11–18 days post-inoculation. To visualize and monitor tumor growth, KPC orthotopic mice under light anesthesia were depilated and then imaged using the high-contrast Vevo 3100 ultrasound system (Fujifilm Visualsonics/VevoLAZR-X).57 The tumor region was subjected to 2D scanning to obtain its area, followed by 3D reconstruction for volume calculation. For the humanized PDAC model, NCG mice (GemPharmatech) were injected subcutaneously with PANC-1 cells and then injected intravenously with human PBMCs when tumors reached 100–150 mm3.58 59 Mice were randomly assigned into two or four groups for treatments when body weight or volume of tumors reached an appropriate value. For antibody treatments, mouse/rat IgG (Cat# A7050/A7052, Beyotime; as control groups), anti-mVSIG4, or anti-hVSIG4 antibodies were administered intraperitoneally at 200 µg/mouse. Combination therapies included rat IgG or anti-mVSIG4 (200 µg/mouse) with ⍺PD-1 (Cat# A2122, Selleck; 100 µg/mouse) and rat IgG, saline, or anti-mVSIG4 (200 µg/mouse) with gemcitabine (Cat# S17132, Yuanye; 50 mg/kg). For the depletion of CD8+ T cells, macrophages, myeloid cells, or C-C motif chemokine ligand (CCL) 5, mouse/rat IgG or anti-mVSIG4 (200 µg/mouse) with ⍺CD8 (Cat# A2102, Selleck; 100 µg/mouse), ⍺CSF1R (Cat# A2159, Selleck; 400 µg/mouse), ⍺Gr-1 (Cat# RB6-8C5, Selleck; 100 µg/mouse), or ⍺CCL5 (Cat# #BE0449, BioXcell; 100 µg/mouse) was intraperitoneally administered. Western blotting To evaluate VSIG4 protein levels, cell lines and tissues were homogenized in RIPA buffer (Cat# P0013B, Beyotime). For phosphoprotein analysis, the cells were lysed in buffer containing 20 mM Tris-HCl (pH 7.5), 0.5% Nonidet P-40, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 2 mM DTT, and 1% phosphorylated protease inhibitors (Cat# SW107-02, Seven). Proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes. After blocking, the membranes were incubated overnight at 4°C with the following primary antibodies: #40-anti-mVSIG4 (1 µg/mL), anti-VSIG4 polyclonal (Cat# abs103696, Absin; 1:1000), anti-ERK1 (Cat# sc-271269, Santa Cruz; 1:1000), anti-p-ERK1/2 (Cat# F0007, Selleck; 1:2500), anti-p-LCK (Cat# AF3101, Affinity; 1:2500), anti-HSP90 (Cat# sc-13119, Santa Cruz; 1:2000), anti-β-Actin (Cat# AC026, ABclonal; 1:5000), anti-C3/C3b/C3c (Cat# 21337–1 AP, Proteintech; 1:8000), or anti-GAPDH (Cat# D190090-0100, BBI; 1:2500). The membranes were then incubated for 1 hour with HRP-conjugated secondary antibodies (goat anti-rabbit IgG, Cat# L153, ABP; goat anti-mouse IgG, Cat# L152, ABP; or goat anti-rat IgG, Cat# A21040, Abbkine; 1:10,000). The membranes were visualized with a highly sensitive enhanced chemiluminescence (ECL) detection kit (Cat# ECL001, bltlux). Single-cell RNA sequencing and data analysis All six C57BL/6J mice with orthotopic KPC tumors were randomly divided into two groups and treated with either IgG or #44-anti-mVSIG4 antibody. At 11 days post-inoculation, the tumor tissues were collected and sent to Shanghai Majorbio for single-cell RNA sequencing (scRNA-seq). Library construction was performed via the DNBelab C TaiM4 V3.0 platform, and sequencing was conducted via the MGI DNBSEQ-T7 system. Bioinformatics analysis was carried out via the proprietary platform of Shanghai Majorbio. The raw data were uploaded to the Gene Expression Omnibus (GEO) database (accession number GSE307934). Lentivirus-mediated gene knockout in Jurkat cells The guide RNA targeting the human Slc3a2 gene (5′-CAGGATCTGCTGAAGGTCGG-3′, sgSLC3A2) was designed via the Synthego online tool (https://design.synthego.com/#/) and cloned and inserted into the clustered regularly interspaced short palindromic repeats (CRISPR) plasmid (pLentiCRISPR-GFP vector). Lentiviral particles were produced by cotransfecting HEK293T cells with the packaging plasmids psPAX2 and VSV-G and either lenti-sgSLC3A2-GFP or an empty vector control. Supernatants containing lentivirus were collected at 48 and 72 hours post-transfection and used to infect Jurkat cells. After 1 week, GFP-positive cells were isolated via FACS and seeded into 96-well plates. Human SLC3A2 knock-out (hSLC3A2-KO) was validated by quantitative real-time PCR (qPCR) and Western blotting. The qPCR primers used included human β-actin (forward, 5′-TCACCATGGATGATGATATCGC-3′; reverse, 5′-GGAATCCTTCTGACCCATGC-3′) and human Slc3a2 (forward, 5′-GGGACTAACTCCTCCGACCT-3′; reverse, 5′-GGCCAATCTCATCCCCGTAG-3′). ELISA of cytokines To measure IFN-γ levels in the TME and serum, tumor tissues were minced and cultured in complete DMEM at 37°C for 24 hours. The IFN-γ concentrations in the supernatants and mouse sera were determined via a mIFN-γ ELISA kit (Cat# 430804, BioLegend) and normalized to the tumor weight. To evaluate T-cell function under specific conditions, Jurkat cells were pretreated as follows: for Gln studies, the cells were cultured in Gln-free RPMI 1640 (AlfaMedX) supplemented with 10% dialyzed FBS and 1% sodium pyruvate for 2–3 days; for Na+ studies, the cells were suspended in 70 mM N-methyl-D-glucamine (NMDG; Cat# ST1551, Beyotime), pH 7.4; for Ca²+ studies, the cells were treated with RPMI 1640 containing BAPTA-AM (Cat# HY-100545, MCE) for 30 min at 37°C. Pretreated Jurkat cells (5×10⁴ cells/well) were seeded into anti-hCD3/CD28 antibody-coated 96-well plates with Fc or hVSIG4-Fc. hIL-2 levels in supernatants were quantified via a hIL-2 ELISA kit (Cat# 431804, BioLegend) at designated time points. Coimmunoprecipitation Fc, hVSIG4-WT-Fc, or mutation of hVSIG4-Fc (hVSIG4-Mut-Fc) and hSLC3A2-His (Cat# HY-P75368, MCE) were mixed with protein G in IP buffer and incubated overnight at 4°C. After being washed with IP buffer, the complexes were boiled in 1×SDS–PAGE loading buffer at 95°C for 5 min, followed by Western blot analysis. Gln detection Jurkat cells were seeded into high-affinity 6-well plates precoated with anti-hCD3/CD28 antibodies (5 µg/mL), Fc, and hVSIG4-Fc and cultured for 16 hours at 37°C. The intracellular Gln concentration was determined via a Gln microplate assay kit (Cat# CAK1206, Cohesion).60 Intracellular Ca²+ flux detection The cells were stained with 5 µM Rhod-2 AM (Cat# 40 776ES50, Yeasen) and 7.5% Pluronic F-127 (Cat# ST501, Beyotime) at 37°C for 30 min, washed twice with PBS or NMDG solution, and incubated for 30 min. Then, the cells were seeded into black 96-well plates precoated sequentially with poly-D-lysine (Cat# ST508, Beyotime; 0.1 mg/mL), anti-hCD3/CD28 antibodies (5 µg/mL), and Fc/hVSIG4-Fc protein. After supplementation with 2 mM CaCl₂, fluorescence was measured via a Varioskan LUX microplate reader (Thermo Fisher) at an excitation wavelength (Ex)/emission wavelength (Em) of 549/578 nm. Intracellular Na+ flux detection The cells were stained with 5 µM sodium-binding benzofuran isophthalate acetoxymethyl ester (SBFI-AM; Cat# abs42027097, Absin) and 7.5% Pluronic F-127 in NMDG-supplemented medium at 37°C for 60 min. After probe deesterification, the cells were seeded into black 96-well plates precoated sequentially with poly-D-lysine and Fc/hVSIG4-Fc. After each well was supplemented with 70 mM NaCl, the fluorescence intensity was measured at Ex/Em 389/580 nm. Bioinformatics and data mining VSIG4 mRNA expression data for PDAC patients were obtained from GEO datasets (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi) and analyzed via R software. The correlation between overall survival and VSIG4 expression was assessed via the Kaplan-Meier survival plot tool (https://kmplot.com/analysis/). Immunohistochemistry and multiplex immunohistochemistry The microarray and sections of human PDAC were heated, deparaffinized in xylene, and rehydrated through a graded ethanol series, as described previously.61 For Immunohistochemistry (IHC), antigen retrieval was performed in 1×sodium citrate solution (Cat# C1032, Solarbio) at 96°C for 30 min. Then, the slide was blocked with 5% goat serum (Cat# SL038, Solarbio) and incubated with an anti-hVSIG4 primary antibody (Cat# ME012547S, Abmart; 1:200) overnight at 4°C. After washing with PBS, the slides were incubated with an HRP-conjugated goat anti-mouse secondary antibody (Cat# RGAM011, Proteintech) for 1 hour at room temperature. Peroxidase reactivity was visualized via a DAB substrate kit (Cat# 550880, BD Pharmingen), with positive staining appearing brown. Nuclei were counterstained with hematoxylin (Cat# G1004, Servicebio). Positive areas were quantified via Fiji software (ImageJ). In multiplex IHC (mIHC), antigen retrieval was performed on tissue sections using heated EDTA buffer (pH 9.0). After blocking (3% H₂O₂ and 10% goat serum), slides were stained for CD8 (Cat# ZA-0508, ZSGB-BIO) using an HRP-secondary antibody and TSA. Between subsequent rounds, slides underwent microwave-assisted stripping/retrieval in citrate buffer (pH 6.0), followed by staining for VSIG4 (Cat# m044899s, Abmart) and SLC3A2 (Cat# 13 180S, CST). Nuclei were visualized with 4’, 6-diamidino-2-phenylindole (DAPI; Cat# C0060, Solarbio). Statistical analysis All the graphs and statistical analyses were generated via Prism V.10 software. Data are presented as the mean±SEM. Statistical comparisons were performed via unpaired two-tailed Student’s t tests for two groups, one-way analysis of variance (ANOVA) with Šídák’s test, or two-way ANOVA with Bonferroni correction. The number of independent experiments (n) is detailed in the figure legends. Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. Results VSIG4 facilitates tumor progression by impairing T-cell cytotoxicity To investigate the role of VSIG4 in tumor progression, MC38 colorectal cancer cells were subcutaneously implanted into Vsig4+/+ or Vsig4−/− mice. Genetic deletion of Vsig4 significantly suppressed MC38 tumor growth, reducing tumor weight by nearly 80% on day 25 postinjection (online supplemental figure S1A–C). Similar growth inhibition was observed in Vsig4−/− mice inoculated with B16-F10 melanoma (online supplemental figure S1D–F), RM-1 prostate cancer (online supplemental figure S1G–I), or KPC pancreatic cancer cells (online supplemental figure S1J–L). We next used flow cytometry to evaluate the immune cell composition in both B16-F10 subcutaneous and KPC orthotopic tumors derived from Vsig4+/+ or Vsig4−/− mice (online supplemental figure S1M). Notably, a significant increase in the percentage of IFN-γ+CD8+ T cells (online supplemental figure S1N) was observed in the tumors of Vsig4−/− mice. Consistently, the levels of IFN-γ were elevated in the supernatants of cultured MC38 and B16-F10 tumor tissues (online supplemental figure S1O), and depletion of CD8+ T cells abrogated the difference in tumor growth between Vsig4+/+ and Vsig4−/− mice (online supplemental figure S1P), suggesting that VSIG4 promotes tumor progression by suppressing CD8+ T-cell-mediated immunity. Together, these data establish a strong association between VSIG4 expression and immune exclusion across multiple cancer types, including PDAC. Screening and functional validation of anti-mouse VSIG4 blocking monoclonal antibodies As VSIG4 promotes tumor progression through CD8+ T cells (online supplemental figure S1P), we examined whether VSIG4 could directly interfere with the activation of T cells. Notably, we found that the recombinant mouse VSIG4 protein significantly inhibited CD4+ or CD8+ T-cell activation induced by anti-CD3/CD28 stimulation (online supplemental figure S2A). These data suggest that VSIG4 could directly impair T-cell activation. To explore the possible binding of VSIG4, we generated a panel of mouse VSIG4-neutralizing monoclonal antibodies (mAbs), which were screened on the basis of their ability to bind mVSIG4 (online supplemental figure S2B). To further characterize the screening and functional validation of mVSIG4-neutralizing antibodies, additional analyses were performed. Binding of individual hybridoma clones to mVSIG4 was confirmed by flow cytometry using mVSIG4-Flag-transfected HEK293T cells, demonstrating clone-specific recognition of surface-expressed mVSIG4 (online supplemental figure S2C). Moreover, in activated primary CD4+ and CD8+ T cells, selected anti-mVSIG4 antibody supernatants effectively blocked the binding of mVSIG4-Fc, as assessed by flow cytometry (online supplemental figure S2D). Consistent with this observation, these antibodies restored IL-2 production in both CD4+ and CD8+ T cells suppressed by mVSIG4-Fc treatment (online supplemental figure S2E), further supporting their neutralizing capacity. The #44 antibody exhibited high production as assessed by Coomassie blue staining (online supplemental figure S2F). Among several candidates with comparable production yields, clone #44 was selected for further study because of its superior binding and functional performance (online supplemental figure S2B–F). To further assess the specificity of the selected anti-mVSIG4 mAb (#44 antibody), cross-reactivity analyses were performed using recombinant proteins. The #44 antibody selectively blocked the mVSIG4-Fc binding to activated CD8+ T cells, with no effect on hCD209-Fc binding, as determined by flow cytometry (online supplemental figure S3A,B). Structural modeling using AlphaFold 3 further predicted a defined interaction interface between the extracellular domain (ECD) of mVSIG4 and the #44 antibody (online supplemental figure S3C). In addition, mutation of key residues to alanine within the mVSIG4 ECD almost abolished the binding of the #44 antibody in transfected HEK293T cells, confirming epitope specificity (online supplemental figure S3D,E). Importantly, #44 antibody binding was markedly reduced to background levels in BMDMs and splenic macrophages from Vsig4−/− mice (online supplemental figure S3F–H), and failed to further suppress tumor growth in Vsig4−/− mice bearing orthotopic KPC tumors (online supplemental figure S3I–L), indicating that its antitumor activity is dependent on VSIG4 expression. VSIG4 blockade elicits potent antitumor responses across multiple tumor types Next, we examined whether blocking VSIG4 had any therapeutic effect on mouse tumors and found that treatment with anti-mVSIG4 clone #44 mAb significantly inhibited tumor growth in subcutaneous MC38 (figure 1A–D), RM-1 (figure 1E–H), and B16-F10 models (figure 1I–L), as well as in orthotopic KPC tumors (figure 1M–P). Consistent with these observations, longitudinal imaging and tumor burden analyses further confirmed the sustained tumor growth inhibition following anti-mVSIG4 treatment in the orthotopic KPC model (online supplemental figure S4A–J). The most robust therapeutic effect was observed in the KPC model, likely due to its relatively high VSIG4 expression (online supplemental figure S4K). Flow cytometry further confirmed increased frequencies of IFN-γ+CD4+ and IFN-γ+CD8+ T cells within the TME of B16-F10 and KPC tumors (figure 1Q,R). Moreover, IFN-γ levels were markedly elevated in tumor supernatants and mouse serum (online supplemental figure S4L,M). Collectively, these results indicate that the antibody-mediated blockade of VSIG4 enhances T-cell cytotoxicity and exerts significantly delayed tumor progression in vivo. Blockade of VSIG4 has significant therapeutic effects on multiple solid tumors. (A–D) MC38 colorectal cancer cells (1×106) were suspended in 100 µL of PBS and subcutaneously injected into wild-type mice, followed by treatment with either the anti-control (IgG) or anti-mVSIG4 (#44) mAb as indicated in (A). A representative tumor image (B), growth curve (C), and tumor weight (D) are shown (n=10/11). (E–H) RM-1 prostate cancer cells (0.15×106) were suspended in 100 µL of PBS and subcutaneously injected into wild-type mice, followed by treatment with either the IgG or #44 mAb as indicated in (E). A representative tumor image (F), growth curve (G), and tumor weight (H) are shown (n=18/17). (I–L) B16-F10 melanoma cells (0.5×106) were suspended in 100 µL of PBS and subcutaneously injected into wild-type mice, followed by treatment with either the IgG or #44 mAb as indicated in (I). A representative tumor image (J), growth curve (K), and tumor weight (L) are shown (n=13/13). (M–P) KPC pancreatic cancer cells (0.15×106) were suspended in 50 µL of PBS with 40% Matrigel and orthotopically injected into the pancreas of wild-type mice, followed by treatment with either the IgG or #44 mAb as indicated in (M). Representative tumor image (N), tumor volume (O), and tumor weight (P) are shown (n=11/12). (Q, R) The percentages of IFN-γ+CD4+ T cells (Q) and IFN-γ+CD8+ T cells (R) in B16-F10 (left) or KPC (right) tumors were analyzed via flow cytometry following the gating strategy shown in (online supplemental figure S1M). The data are quantitated and presented sequentially. All data are represented as the mean±SEM. The data are representative of two or three independent experiments and were analyzed via two-way ANOVA with Bonferroni’s multiple comparisons test (C, G, K) or two-tailed unpaired t-test (D, H, L, O–R). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. ANOVA, analysis of variance; IFN, interferon; mAb, monoclonal antibodies; PBS, phosphate-buffered saline; s.c., subcutaneously; VSIG4, v-set and immunoglobulin domain-containing 4. VSIG4 blockade reverses the immunosuppressive TME To elucidate the cellular mechanisms underlying the therapeutic efficacy of the anti-mVSIG4 clone #44 antibody, scRNA-seq was performed on KPC tumors. After stringent quality control, 12 major cell clusters were identified via dimensionality reduction, manual annotation, and SingleR analysis (online supplemental figure S5A). T cells, monocyte-macrophages, and neutrophils accounted for approximately 70% of the total cells (online supplemental figure S5B). While the overall proportion of T cells slightly decreased on VSIG4 blockade, a distinct expansion of effector CD8+ T cells was observed (figure 2A,B), which was consistent with the flow cytometry results (figure 1R). The proportion of monocyte-macrophages remained stable, whereas that of neutrophils slightly increased (online supplemental figure S5B). Gene set variation analysis and gene set enrichment analysis revealed that the anti-VSIG4 antibody group was enriched for pathways associated with IFN-γ and IFN-α responses, oxidative phosphorylation, DNA repair, p53 signaling, and Myc targets in T cells (figure 2C). Differential gene expression (DGE) analysis revealed the upregulation of cytotoxic effector molecules, including Gzmb, Gzmc, Gzmd, Gzme, Gzmf, and Gzmk, as well as Ifng in T cells (figure 2D,E), indicating enhanced T-cell effector function following VSIG4 blockade. Blockade of VSIG4 enhances the antitumor immune response in the TME. (A–E) KPC tumor cells (0.15×106) were orthotopically implanted into wild-type mice, followed by treatment with either the IgG or #44 mAb, as described in figure 1M. scRNA-seq was subsequently performed on three pairs of tumors. (A) Dimension reduction analysis identified eight distinct subgroups of T cells, which are visualized in t-SNE plots. (B) The proportions of these T-cell subgroups are shown (n=3/3). (C) Pathway enrichment analysis of significantly altered genes in T cells was conducted via a one-sided Fisher’s exact test. (D) DGEs between the IgG and #44 mAb treatment groups in T cells were analyzed via the DGE method and are displayed. (E) A bubble heatmap displays the expression of cytotoxicity-associated genes across the identified T cell cluster. (F–I) Wild-type mice were intraperitoneally injected with isotype control antibody, anti-CD8 antibody, or anti-CSF1R antibody prior to the in situ injection of KPC cells (0.15×106). Subsequently, either the IgG or #44 mAb was intraperitoneally injected daily as indicated (F). A representative tumor image (G, n=10/9/11/11/11/10), tumor volume (H), and tumor weight (I) are shown. (J–M) Wild-type mice were intraperitoneally injected with either an isotype control antibody or an anti-Gr-1 antibody once every 2 days for a total of four doses, beginning 1 day prior to the in situ injection of KPC cells (0.15×106). Thereafter, mice received daily intraperitoneal injections of either the IgG or #44 as specified (J). A representative tumor image (K, n=7/8/8/8), tumor volume (L), and tumor weight (M) are shown. (N–Q) Wild-type mice received intraperitoneal injection of either an isotype control antibody or an anti-CCL5 antibody five times every other day on the day after in situ injection of KPC cells (0.15×106). Then, either the IgG or #44 mAb was intraperitoneally injected every day as indicated (N). A representative tumor image (O, n=11/11/9/10), tumor volume (P), and tumor weight (Q) are shown. All data are represented as the mean±SEM. The data are representative of two independent experiments and were analyzed by one-way ANOVA with Šídák’s multiple comparisons test (H, I, L, M, P, Q). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. ANOVA, analysis of variance; CCL5, C-C motif chemokine ligand 5; CSF1R, colony stimulating factor 1 receptor; DGE, differential gene expression; Gzm, granzyme; mAb, monoclonal antibodies; NES, normalized enrichment score; Pro T, proliferating T; scRNA-seq, single-cell RNA sequencing; Teff, effector T; Tem, effector memory T; Tex, exhausted T; Th1, T helper 1; Th17, T helper 17; TME, tumor microenvironment; Tnf, tumor necrosis factor; Treg, regulatory CD4+ T; t-SNE, t-distributed stochastic neighbor embedding; VSIG4, v-set and immunoglobulin domain-containing 4. Single-cell transcriptomic analysis further revealed that Vsig4 expression was predominantly enriched within myeloid populations in the TME, with minimal expression in lymphoid lineages (online supplemental figure S5C). Notably, subclustering of monocytes and macrophages demonstrated that Vsig4 expression was selectively concentrated in monocytes and macrophage subsets (online supplemental figure S5D). We hypothesized that macrophages expressing VSIG4 impaired the activation of CD8+ T cells. To test this hypothesis, we employed CD8+ T-cell and macrophage depletion experiments and found that the administration of anti-CD8 or anti-CSF1R antibodies abolished the therapeutic benefit of VSIG4 blockade (figure 2F–I and online supplemental figure S5E,F), indicating a dependency of the antitumor effect on both cell types. In addition, depletion of Gr-1+ myeloid cells significantly diminished the antitumor efficacy of VSIG4 blockade in the KPC model (figure 2J–M and online supplemental figure S5G,H). Given that Vsig4 is not expressed in neutrophils (online supplemental figure S5C), these findings suggest that besides macrophages, VSIG4-expressing monocytes also play an essential role in mediating tumor immune evasion. To further identify mediators of enhanced CD8+ T-cell recruitment, transcription factor and chemokine profiles were analyzed across cell subsets. CCL5, a chemokine known to promote CD8+ T-cell migration and activation,62 was upregulated in T-cell subclusters (online supplemental figure S5I,J). Neutralizing CCL5 almost completely abolished the antitumor efficacy of VSIG4 blockade in a pancreatic cancer model (figure 2N–Q), highlighting its essential role in mediating the therapeutic effect of VSIG4 blockade. VSIG4 blockade enhances antitumor efficacy in pancreatic cancer PDAC is a lethal malignancy with a limited response to ICB therapies such as anti-PD-1 antibodies.63 Currently, gemcitabine (GEM) remains the standard first-line chemotherapy for treating PDAC. Coadministration of GEM with the #44 anti-mVSIG4 antibody significantly reduced the tumor burden in the KPC model (figure 3A–C), demonstrating a synergistic therapeutic effect. To evaluate whether VSIG4 blockade could augment ICB efficacy, combination therapy with anti-PD-1 and anti-VSIG4 antibodies was tested. Anti-VSIG4 treatment resulted in significant tumor control in KPC-bearing mice, and the combination with anti-PD-1 did not further improve efficacy compared with anti-VSIG4 monotherapy, but showed more pronounced tumor suppression than anti-PD-1 alone (figure 3D–F). Consistently, the combination was associated with increased infiltration of IFN-γ+CD8+ T cells (figure 3G) and elevated IFN-γ levels in tumor supernatants (figure 3H). These results suggest that targeting VSIG4 may sensitize pancreatic tumors to PD-1 blockade. VSIG4 blockade synergistically enhances the antitumor efficacy of clinical therapeutic agents. (A–C) KPC tumor cells (0.3×106) were orthotopically implanted into wild-type mice, followed by treatment with saline or GEM in combination with either the IgG or #44 mAb, as outlined (A, left panel). Representative tumor image (A, right panel; n=12/12/13/13), tumor volume (B), and tumor weight (C) are presented. (D–H) KPC tumor cells (0.15×106) were orthotopically implanted into wild-type mice, followed by treatment with or without PD-1 antibody in combination with either the IgG or #44 mAb, as described previously (D, left panel). A representative tumor image (D, right panel; n=10/11/11/11), tumor volume (E), and tumor weight (F) are shown. (G) The percentages of IFN-γ+CD8+ T cells were quantified via flow cytometry. (H) KPC tumors were individually cultured in DMEM for 24 hours, and the IFN-γ concentrations in the tumor culture supernatants were measured via ELISA. All data are represented as the mean±SEM. The data are representative of two independent experiments and were analyzed via one-way ANOVA with Šídák’s multiple comparisons test (B, C, and E–H). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. ANOVA, analysis of variance; GEM, gemcitabine; IFN, interferon; mAb, monoclonal antibodies; VSIG4, v-set and immunoglobulin domain-containing 4. SLC3A2 is a functional receptor for VSIG4-mediated T-cell suppression Although VSIG4 is known to inhibit T-cell activation,42 its functional receptor on T cells has remained elusive. To investigate the molecular mechanism underlying this T-cell suppression, we next explored the potential binding partners of VSIG4 on T cells. To identify potential receptors, immunoprecipitation followed by mass spectrometry was performed. Among the candidates, SLC3A2 had the highest binding affinity for VSIG4 (figure 4A). Notably, SLC3A2 expression was elevated in T cells after anti-VSIG4 antibody treatment (figure 4B), suggesting that SLC3A2 plays a role in mediating the T-cell immune response. To test the function of SLC3A2, we generated hSLC3A2-KO Jurkat cells via CRISPR-Cas9, and the knockout efficiency was validated via qPCR and Western blotting (figure 4C). Loss of SLC3A2 largely abrogated the inhibitory effects of VSIG4 on TCR signaling and cytokine secretion (figure 4D,E), which indicates that SLC3A2 is the functional receptor of VSIG4 on T cells. SLC3A2 functions as the receptor mediating the inhibitory effect of VSIG4 on T cells. (A) Immunoprecipitation followed by mass spectrometry was performed using the cell membrane lysate of Jurkat cells, as described in the Materials and Methods. The proteins that interact with VSIG4 are shown. (B) A bubble heatmap illustrates the expression of SLC superfamily members in T-cell cluster identified by scRNA-seq, as indicated in figure 2A. (C) An sgRNA targeting exon 9 of human SLC3A2 was constructed (upper panel). The knockout efficiency was validated by qPCR (lower left) and Western blotting (lower right). (D) Vector or hSLC3A2-KO Jurkat cells were preincubated with Fc or hVSIG4-Fc recombinant proteins and then stimulated with plate-coated anti-hCD3/CD28 mAbs for the indicated times. The expression of the indicated proteins was analyzed by Western blotting. (E) Vector control or three distinct hSLC3A2-knockout Jurkat clones (KO1, KO2, and KO3) were preincubated with Fc or hVSIG4-Fc recombinant protein, respectively, and then seeded into 96-well plates precoated with anti-hCD3/CD28 mAbs. IL-2 production was quantified by ELISA. (F) The interaction mode between the ECDs of hVSIG4 and hSLC3A2 was predicted via AlphaFold 3, and the interacting amino acids are highlighted. (G) A schematic representation of the wild-type and mutant constructs of the ECD of hVSIG4 is shown. (H) Immunoprecipitation was performed using protein G resin in the presence of recombinant hSLC3A2-His, hVSIG4-WT-Fc, or hVSIG4-Mut-Fc, as described in the Materials and Methods. hSLC3A2-His was immunoprecipitated with agarose beads. (I) Jurkat cells were left untreated or stimulated with anti-hCD3/CD28 mAbs in the presence of Fc, hVSIG4-WT-Fc, or hVSIG4-Mut-Fc for 48 hours, followed by measurement of IL-2 secretion via ELISA. All data are represented as the mean±SEM. The data are representative of two or three independent experiments and were analyzed via one-way ANOVA with Šídák’s multiple comparisons test (E, I) or two-tailed unpaired t test (C). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.01, highly significant; p<0.001, extremely significant. ANOVA, analysis of variance; Atp6ap2, ATPase H+ transporting accessory protein 2; CRISPR, clustered regularly interspaced short palindromic repeats; ECD, extracellular domain; ERK, extracellular signal-regulated kinase; Fc, crystallizable fragment; HRP, horseradish peroxidase; IL-2, interleukin 2; KO, knock-out; LCK, lymphocyte cell-specific protein-tyrosine kinase; mAb, monoclonal antibodies; Mut, mutation; p-, phosphor-; qPCR, quantitative real-time PCR; scRNA-seq, single-cell RNA sequencing; sgRNAs, single-guide RNAs; SLC, solute carrier family; VSIG4, v-set and immunoglobulin domain-containing 4; WT, wild-type. AlphaFold 3 modeling predicted a direct interaction between the ECDs of VSIG4 and SLC3A2, likely resulting in the formation of a heterodimer or multimer complex to stabilize binding.64 Nine key amino acid residues (D52, S54, I58, E85, D87, T173, S175, K180, and Q197) in VSIG4’s IgV and IgC domains were predicted as essential contact points with SLC3A2 (figure 4F). Mutation of these residues to alanine in human VSIG4 completely disrupted the interaction with SLC3A2 and eliminated VSIG4-mediated inhibition on IL-2 production (figure 4G–I), further supporting SLC3A2 as a functional receptor for VSIG4-mediated suppression. VSIG4 inactivates T-cell activation via SLC3A2-mediated Gln and ion transport SLC3A2 is a molecular chaperone essential for the assembly of amino acid transporters, including those that mediate Gln uptake.14 30 Reanalysis of the scRNA-seq dataset revealed that treatment with the anti-VSIG4 antibody restored metabolic pathways in T cells, including the citric acid cycle and glucose, pyruvate, and Gln metabolism (figure 5A). To examine whether VSIG4 affects Gln uptake in T cells, Gln uptake was quantified in activated T cells via a Gln detection kit. Notably, the recombinant hVSIG4-Fc protein inhibited Gln uptake in T cells, whereas SLC3A2 knockout abolished this effect (figure 5B). IL-2 production was also assessed under Gln-free conditions. Under these conditions, VSIG4 failed to suppress IL-2, confirming that Gln is a key mediator of VSIG4-induced T-cell inactivation (figure 5C). Together, these data indicate that the ligation of VSIG4 and SLC3A2 impairs Gln uptake in T cells. VSIG4 inhibits T-cell activation through the SLC3A2-mediated metabolic signaling pathway. (A) Metabolic pathway enrichment analysis of significantly altered genes identified by scRNA-seq, as shown in figure 2A, was conducted via a one-sided Fisher’s exact test. (B) Jurkat cells were either left untreated or stimulated with anti-hCD3/CD28 mAbs, followed by pretreatment with Fc or hVSIG4-Fc recombinant protein for 48 hours, after which the intracellular Gln levels were measured. (C) Jurkat cells were pretreated with Fc or hVSIG4-Fc recombinant protein and subsequently either left untreated or stimulated with anti-hCD3/CD28 mAbs in Gln-deficient medium or Gln-replenished medium for 48 hours. IL-2 secretion was then quantified via ELISA. (D, E) Vector or hSLC3A2-KO Jurkat cells were stained with SBFI-AM dye, and the intracellular Na+ was immediately measured after treatment with Fc or hVSIG4-Fc recombinant protein. Representative images of Na+ influx (D) and quantification (E) are shown. (F, G) Vector or hSLC3A2-KO Jurkat cells were stained with Rhod-2 AM dye, and the intracellular Ca²+ was immediately measured after treatment with Fc or hVSIG4-Fc recombinant protein. Representative images of Ca²⁺+ influx (F) and quantification (G) are shown. (H, I) Jurkat cells were stained with SBFI-AM dye and then cultured in Gln-deficient medium or Gln-replenished medium, as indicated. The intracellular Na+ was immediately measured after treatment with Fc or hVSIG4-Fc recombinant protein. Representative images of Na+ influx (H) and quantification (I) are shown. (J,K) Jurkat cells were stained with Rhod-2 AM dyes and then cultured in Gln-deficient medium or Gln-replenished medium, as indicated. The intracellular Ca²⁺+ was immediately measured on treatment with Fc or hVSIG4-Fc recombinant protein. Representative images of Ca²+ influx (J) and quantification (K) are shown. (L, M) Jurkat cells were stained with Rhod-2 AM dye and then cultured in control medium or NMDG-supplemented medium, as indicated. The intracellular Ca²+ was immediately measured after treatment with Fc or hVSIG4-Fc recombinant protein. Representative images of Ca²+ influx (L) and quantification (M) are shown. (N) Jurkat cells were left untreated or stimulated with anti-hCD3/CD28 mAbs following pretreatment with Fc or hVSIG4-Fc recombinant protein in control medium or NMDG-supplemented medium for 48 hours. IL-2 secretion was quantified via ELISA. (O) Jurkat cells were treated in a specified concentration gradient of BAPTA-AM for 30 min, and then were left untreated or stimulated with anti-hCD3/CD28 mAbs following pretreatment with Fc or hVSIG4-Fc recombinant protein for 48 hours. IL-2 secretion was quantified via ELISA. All data are represented as the mean±SEM. The data are representative of two or three independent experiments and were analyzed via one-way ANOVA with Šídák’s multiple comparisons test (B, C, E, G, I, K, M, N, O). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. ANOVA, analysis of variance; AUC, area under curve; Ca²⁺, calcium; F, fluorescence; Fc, crystallizable fragment; Gln, glutamine; IL-2, interleukin 2; KO, knock-out; mAb, monoclonal antibodies; Na+, sodium; NMDG, N-methyl-D-glucamine; SBFI-AM, sodium-binding benzofuran isophthalate acetoxymethyl ester; scRNA-seq, single-cell RNA sequencing; SLC3A2, solute carrier family 3 membrane 2; VSIG4, v-set and immunoglobulin domain-containing 4. Given that Gln transport via SLC3A2 is partially coupled with Na+ and Ca²+ exchange,29 we examined the intracellular Na+ and Ca²+ levels in Jurkat cells transfected with Vector or SLC3A2-targeting single-guide RNAs (sgRNAs) after VSIG4 treatment. hVSIG4-Fc treatment suppressed both Na+ (figure 5D,E) and Ca²+ (figure 5F,G) influx in Vector cells but had no effect on SLC3A2-deficient cells, indicating that VSIG4 suppresses ion flux in an SLC3A2-dependent manner. To explore the interdependency of Gln and ion flux, T cells were cultured in Gln-deprived conditions. The absence of Gln abolished VSIG4-induced inhibition of Na+ (figure 5H,I) and Ca²+ influx (figure 5J,K), suggesting that Gln uptake is upstream of these ionic changes. Furthermore, replacing extracellular Na+ with NMDG eliminated VSIG4-induced Ca²+ suppression (figure 5L,M) while impairing Gln uptake and IL-2 suppression by VSIG4 (figure 5N). Chelation of intracellular Ca²+ via BAPTA-AM also blocked VSIG4-mediated IL-2 inhibition (figure 5O). Together, these results indicate that VSIG4 suppresses T-cell function by simultaneously impairing Gln and Na+ influx through SLC3A2, subsequently limiting Ca²+-dependent T-cell activation. Development and validation of a human VSIG4-blocking monoclonal antibody Next, we explored whether blocking human VSIG4 had any therapeutic effect on tumors. We developed a blocking antibody targeting human VSIG4, and the mice were immunized with recombinant human VSIG4-Fc protein. To identify candidate anti-hVSIG4 antibodies, hybridoma supernatants were first screened for their ability to recognize hVSIG4 in HEK293T cells transfected with hVSIG4-Flag by flow cytometry (online supplemental figure S6A). We next evaluated the functional blocking capacity of these candidate clones in activated primary mouse and human T cells and found that supernatants from clone #88 effectively reversed VSIG4-mediated T-cell suppression in both species, as evidenced by restored IL-2 production (online supplemental figure S6B,C). The clone #88 exhibited high production as assessed by Coomassie blue staining (online supplemental figure S6D). To further validate the specificity and functional selectivity of the anti-hVSIG4 clone #88 antibody, we performed a series of comparative binding and functional assays. Flow cytometric analysis showed that clone #88 selectively blocked the binding of hVSIG4-Fc, but not hCD209-Fc, to activated primary mouse CD8+ T cells and Jurkat T cells, indicating high target specificity without cross-reactivity to the closely related lectin CD209 (online supplemental figure S7A–D). Structural modeling further revealed a defined interaction interface between hVSIG4 and the heavy and light chains of clone #88, providing a structural basis for its high-affinity and specificity (online supplemental figure S7E). In addition, mutational analysis confirmed that clone #88 specifically recognized the ECD of wild-type hVSIG4, without binding to VSIG4 mutants lacking the predicted epitope (online supplemental figure S7F,G). To further characterize the molecular specificity of the anti-hVSIG4 clone #88 antibody, we examined whether it interferes with the interaction between VSIG4 and its receptor SLC3A2. Flow cytometric analysis revealed that clone #88 effectively blocked the binding of both mouse and human VSIG4 to SLC3A2-expressing cells, indicating conserved receptor-blocking activity across species (online supplemental figure S7H,I). These findings indicate that the antitumor activity of clone #88 is strictly dependent on the presence of human VSIG4, thereby confirming the target specificity of this antibody in vivo. Together, these data support clone #88 as a specific and functionally effective blocking antibody targeting human VSIG4. Anti-human VSIG4 neutralizing antibody suppresses tumor growth in hVSIG4-KI mice Clone #88 was selected for further study. To assess its efficacy in vivo, we generated hVSIG4-KI mice, in which human VSIG4 expression is driven by the endogenous mouse Vsig4 promoter (figure 6A,B). Subcutaneous MC38 and orthotopic KPC tumors established in hVSIG4-KI mice were significantly suppressed by #88 anti-hVSIG4 antibody treatment (figure 6C–H). Flow cytometry of the TME revealed increased frequencies of IFN-γ+CD8+ T cells following antibody administration (figure 6I,J). To assess the in vivo specificity of clone #88, we next evaluated its therapeutic efficacy in wild-type mice bearing orthotopic KPC tumors. In contrast to the robust antitumor effect observed in hVSIG4-KI mice (figure 6A–J), treatment with clone #88 failed to suppress KPC tumor growth in wild-type mice (figure 6K–M). Together, these data indicate that blockade of human VSIG4 exerts significant therapeutic effects in tumors such as KPC. Anti-hVSIG4 neutralizing antibodies have shown therapeutic efficacy in multiple tumor models. (A) Schematic representation of the strategy for generating hVSIG4-knock-in (KI) mice. KI mice were bred to homozygosity for subsequent tumor implantation experiments. (B) Spleens and BMDMs were isolated, and hVSIG4 expression was assessed by Western blotting using #88 anti-hVSIG4 mAb supernatant. HSP90 was used as the loading control. (C–E) MC38 colorectal cancer cells (1×106) were suspended in 100 µL of PBS and subcutaneously injected into hVSIG4-KI mice. The mice were treated with either the IgG or #88 mAb as indicated in (C, upper panel). A representative tumor image (C, lower panel), growth volume (D) and tumor weight (E) are shown (n=6/6). (F–H) KPC pancreatic cancer cells (0.15×106) were suspended in 50 µL of PBS with 40% Matrigel and orthotopically injected into the pancreas of hVSIG4-KI mice, followed by treatment with either the IgG or #88 mAb as indicated in (F, upper panel). Representative tumor image (F, lower panel), tumor volume (G) and tumor weight (H) are shown (n=16/16). (I) Gating strategies for flow cytometry analysis of IFN-γ+CD8+ T cells. (J) The percentages of IFN-γ+CD8+ T cells were quantified via flow cytometry. (K–M) KPC pancreatic cancer cells (0.15×106) were suspended in 50 µL of PBS with 40% Matrigel and orthotopically injected into the pancreas of wild-type mice, followed by treatment with either the IgG or #88 mAb as indicated in (K, upper panel). Representative tumor image (K, lower panel), tumor volume (L) and tumor weight (M) are shown (n=12/13). (N–Q) Pancreatic cancer cells PANC-1 cells (3×10⁶) were suspended in 100 µL PBS and subcutaneously injected into NCG mice. When tumors reached an average volume of 150 mm³, mice received intravenous injection of human PBMCs (3×10⁶) and were then treated with either IgG or the #88 mAb as indicated in (N). A representative tumor image (O), growth volume (P), and tumor weight (Q) are shown (n=10/9). All data are represented as the mean±SEM. The data are representative of two independent experiments and were analyzed via two-way ANOVA with Bonferroni’s multiple comparisons test (D, P) or two-tailed unpaired t test (E, G, H, J, L, M, Q). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. ANOVA, analysis of variance; APC, allophycocyanin; BMDMs, bone marrow-derived macrophages; CDS, coding sequences; BGH pA, bovine growth hormone polyA; IFN, interferon; mAb, monoclonal antibodies; PBMCs, peripheral blood mononuclear cells; PBS, phosphate-buffered saline; PE, phycoerythrin; s.c., subcutaneously; UTR, untranslated regions; VSIG4, v-set and immunoglobulin domain-containing 4; WPRE, woodchuck hepatitis virus posttranscriptional regulatory element. To further evaluate the translational relevance of VSIG4 blockade in a human immune context, we employed a human PBMC-reconstituted xenograft model. NCG mice bearing established PANC-1 tumors were intravenously infused with human PBMCs and subsequently treated with either control IgG or the anti-hVSIG4 neutralizing antibody #88. Treatment with the anti-hVSIG4 antibody significantly suppressed tumor growth, as evidenced by reduced tumor volumes and tumor weights compared with the control group (figure 6N–Q), demonstrating that blockade of human VSIG4 elicits antitumor activity in a humanized immune microenvironment. VSIG4 has been reported to function as a receptor for activated complement C3 fragments, such as C3b and iC3b.39 We, therefore, investigated the impact of C3b on VSIG4 blockade. Recombinant human C3 and C3b were generated and validated (online supplemental figure S7J,K). Binding assays showed that hC3b did not interfere with the binding of the anti-hVSIG4 clone #88 antibody to hVSIG4 (online supplemental figure S7L). Moreover, the presence of hC3b did not impair the ability of clone #88 antibody to reverse hVSIG4-Fc-mediated inhibition of T-cell activation, as evidenced by comparable levels of IL-2 production (online supplemental figure S7M). Together, these results indicate that the immunomodulatory effects of VSIG4 blockade are largely independent of C3 engagement. VSIG4 expression is correlated with cancer aggressiveness in patients Previous studies have reported that elevated VSIG4 expression correlates with poor prognosis and immunosuppressive TMEs in several cancer types, including pancreatic, colorectal, renal, and hematologic malignancies.41 43 52 53 To explore the clinical relevance of VSIG4, we analyzed publicly available datasets. In the GEO cohort, VSIG4 mRNA expression was significantly elevated in pancreatic tumors compared with adjacent normal tissues (figure 7A). Kaplan-Meier analysis revealed that high VSIG4 expression was associated with poor overall survival across multiple tumor types, including PDAC, colon adenocarcinoma, ovarian serous cystadenocarcinoma (OV), LUAD, acute myeloid leukemia, breast invasive carcinoma, liver hepatocellular carcinoma, myeloma (MM) and stomach adenocarcinoma (figure 7B). We further confirmed these findings in clinical samples. Western blot analysis of fresh LUAD tissues revealed significantly higher VSIG4 expression in tumors than in adjacent tissues in 9 out of 16 paired samples (figure 7C,D). Additionally, analysis of a pancreatic cancer tissue microarray revealed increased VSIG4 expression in more than 50% of tumors relative to adjacent tissues (figure 7E,F). To investigate the spatial relationship between VSIG4-expressing cells and CD8+ T cells in human pancreatic cancer, mIHC staining was performed on patient tumor samples. VSIG4+ cells were frequently found in close physical proximity to CD8+ T cells expressing SLC3A2, with direct cell-cell contacts observed within the TME (figure 7G). These spatial associations provide in situ evidence supporting a potential interaction between VSIG4-expressing cells and SLC3A2+CD8+ T cells in human pancreatic cancer. Collectively, these data support a strong correlation between elevated VSIG4 expression and tumor malignancy in human cancers, especially in human pancreatic cancer. VSIG4 expression is positively correlated with malignancy in human cancer samples. (A) VSIG4 mRNA expression in PDAC and adjacent normal tissues was analyzed across multiple GEO datasets, including GSE15471, GSE183795, GSE49641, GSE62165, and GSE71989. (B) The survival rates of patients with various cancers, including PDAC, COAD, OV, LUAD, AML, BRCA, LIHC, MM and STAD, in the VSIG4-high and VSIG4-low groups were analyzed via the Kaplan-Meier survival plot tool. (C, D) VSIG4 expression levels in fresh cancer tissues and corresponding paracancerous tissues from LUAD patients were assessed by Western blot (C), with actin serving as the loading control. VSIG4 protein levels were quantified (D). (E, F) Immunohistochemical staining for VSIG4 was performed on PDAC patient samples (E), and VSIG4 expression levels were quantified (F) (n=50 in each group; scale bar, 100 or 200 µm). (G) Multiplex immunohistochemistry staining for SLC3A2 (green), CD8 (red) and VSIG4 (purple) was performed on PDAC patient-derived samples (n=3; scale bar, 5 or 50 µm). All data are represented as the mean±SEM. The data were analyzed via a two-tailed unpaired t test (A, D, F). Statistical significance was defined as follows: p>0.05, not significant (ns); p<0.05, significant; p<0.01, highly significant; and p<0.001, extremely significant. AML, acute myeloid leukemia; BRCA, breast invasive carcinoma; COAD, colon adenocarcinoma; DAPI, 4’, 6-diamidino-2-phenylindole; GEO, Gene Expression Omnibus; LIHC, liver hepatocellular carcinoma; LUAD, lung adenocarcinoma; MM, myeloma; N, normal tissues; OV, ovarian serous cystadenocarcinoma; PDAC, pancreatic ductal adenocarcinoma; SLC3A2, solute carrier family 3 membrane 2; STAD, stomach adenocarcinoma; T, tumor tissues; VSIG4, v-set and immunoglobulin domain-containing 4; VSIG4, v-set and immunoglobulin domain-containing 4. Discussion This study reveals a previously unrecognized immunosuppressive mechanism in the TME, wherein monocyte- and macrophage-expressed VSIG4 directly binds SLC3A2 on T cells, impairing their activation through disrupted Gln uptake and ionic homeostasis (online supplemental figure S8). While SLC3A2 has been classically defined as a chaperone for amino acid transporters that facilitate Gln and leucine uptake to support mTOR-driven T-cell proliferation,28 recent evidence highlights its additional role in mediating the Na+ and Ca²+ fluxes essential for TCR signaling.32 Our results show that VSIG4 engagement restricts Gln-dependent Na+ influx, subsequently suppressing the Ca²+ signaling required for full T-cell activation. In contrast to canonical checkpoints such as PD-1 and CTLA-4, which primarily modulate intracellular signaling pathways,25 VSIG4 functions through extracellular nutrient and ion deprivation. Thus, the VSIG4–SLC3A2 interaction defines a novel metabolic–ionic checkpoint, highlighting a novel checkpoint logic wherein metabolic and ionic restrictions converge to suppress T cell immunity. Although SLC3A2 deficiency reduced TCR-induced LCK phosphorylation, it did not abolish LCK or ERK activation on anti-CD3/CD28 stimulation (figure 4D), indicating that SLC3A2 is not an obligate component of canonical TCR/CD28 signaling initiation. Instead, SLC3A2 appears to support optimal proximal signaling through metabolic and ionic regulation, while acting as the functional receptor required for VSIG4-mediated suppression. Accordingly, residual LCK and ERK phosphorylation in SLC3A2-deficient cells likely reflects intact SLC3A2-independent TCR/CD28 signaling pathways, whereas the loss of VSIG4-mediated inhibition in these cells supports a specific requirement for SLC3A2 in the VSIG4 suppressive axis. Despite these findings, several limitations of the present study should be acknowledged. First, although we identify SLC3A2 as a functional receptor mediating VSIG4-induced T-cell suppression, our data do not exclude the possibility that additional receptors or signaling pathways may also contribute to VSIG4-mediated immunosuppression. Moreover, SLC3A2 is broadly expressed across multiple cell types, and thus our conclusions primarily reflect its role in intratumoral CD8+ T cells rather than a universal mechanism across all contexts. Second, while our in vitro and in vivo data support a requirement for SLC3A2 in VSIG4-mediated T-cell inhibition, the precise structural basis of the VSIG4–SLC3A2 interaction remains to be elucidated. Third, although we employ multiple tumor models and humanized systems, the long-term safety and tissue-specific consequences of sustained VSIG4 blockade warrant further investigation, particularly given the physiological roles of VSIG4 and SLC3A2 in immune homeostasis. From a biological perspective, the VSIG4–SLC3A2 axis may serve as a contact-dependent myeloid–T cell brake that tunes T-cell activation in metabolically constrained tissues. Despite the broad expression of SLC3A2, multiplex imaging reveals frequent physical contacts between VSIG4+ myeloid cells and SLC3A2+CD8+ T cells in human PDAC, providing an anatomical basis for functional engagement in vivo. Conceptually analogous to PD-1–PD-L1, which restrains TCR signaling, VSIG4–SLC3A2 may impose a local metabolic–ionic checkpoint by limiting Gln-linked Na+ influx and downstream Ca²+ signaling, thereby preventing excessive or sustained activation. In tumors, this homeostatic module appears to be co-opted and amplified to dampen cytotoxic T-cell function and promote immune evasion. Importantly, we find that the chemokine CCL5 is indispensable for the antitumor efficacy of VSIG4 blockade. Single-cell transcriptomic analysis revealed marked upregulation of CCL5 in both effector and exhausted CD8+ T-cell subsets following anti-VSIG4 therapy, which is consistent with previous findings that activated CD8+ T cells secrete CCL5 (also known as RANTES) to promote further immune cell recruitment and inflammatory amplification.65 66 Notably, exhausted CD8+ T cells also retain the ability to produce CCL5.67 Functionally, neutralization of CCL5 completely abolished the antitumor effects of VSIG4 blockade, indicating that CCL5 acts not only as a downstream effector but also as an upstream amplifier of immune reactivation. This finding aligns with prior studies showing that CCL5 is a key initiator of CD8+ T-cell infiltration into tumors and that it acts synergistically with IFN-γ and CXCL9/CXCL10 to sustain a cytotoxic immune milieu.68 69 Together, these findings suggest that VSIG4 blockade initiates a CCL5-mediated feedforward loop in which reactivated CD8+ T cells recruit additional cytotoxic cells to sustain antitumor immunity. Given the availability of CCR5 (the receptor of CCL5) inhibitors in clinical trials, this axis may be therapeutically exploitable. Furthermore, our results have significant implications for treating immunologically “cold” tumors, particularly PDAC. PDAC is characterized by low T-cell infiltration, a dense stromal architecture, and metabolic suppression, all of which contribute to immune evasion and resistance to conventional ICB.3,6 We show that VSIG4 is highly expressed in aggressive PDACs and contributes to T-cell exclusion by impairing Gln and ion access. Blockade of VSIG4 not only reactivated CD8+ T cells but also sensitized tumors to anti-PD-1 therapy and chemotherapy. These data support the potential of VSIG4 inhibition as a strategy to remodel the suppressive TME and enhance the efficacy of existing therapies in immune-refractory cancers. Although the blocking activity of the VSIG4 antibody clone #44 against the VSIG4–SLC3A2 interaction is partial rather than complete (online supplemental figure S2D and S3F), it nevertheless provides functional evidence that interfering with this pathway can relieve VSIG4-mediated immunosuppression and enhance antitumor immunity. In this study, clone #44 therefore serves primarily as a proof-of-concept reagent to demonstrate the therapeutic potential of targeting VSIG4. Importantly, the observed immunological and antitumor effects despite its moderate blocking activity further support the functional relevance of the VSIG4–SLC3A2 axis. These findings suggest that next-generation antibodies with improved epitope selection and stronger blocking potency may achieve even greater therapeutic efficacy, highlighting the potential of VSIG4 as a promising target for cancer immunotherapy. In summary, our study reveals a previously uncharacterized VSIG4–SLC3A2 metabolic–ionic checkpoint axis that links extracellular checkpoint signaling to intracellular nutrient and ion restriction. This axis operates independently of PD-1/CTLA-4 and provides a complementary mechanism of immune suppression. Future studies should explore the structural basis of the VSIG4–SLC3A2 interaction, evaluate the long-term safety and therapeutic synergy of anti-VSIG4 antibodies, and investigate whether similar metabolic–ionic strategies are exploited by other myeloid checkpoints in settings such as infection, autoimmunity, and transplantation. Ultimately, targeting VSIG4 and its SLC3A2-associated suppressive pathway may provide a powerful and tractable approach to reinvigorate antitumor immunity in metabolically constrained and therapy-resistant tumors. These insights not only establish a mechanistic basis for VSIG4 targeting but also suggest a broader paradigm of metabolic–ionic immunoregulation exploitable across disease contexts. Supplementary material Footnotes Funding: This investigation was supported by grants from the Science and Technology Innovation Key R&D Program of Chongqing (CSTB2024TIAD-STX0005, to CW); the Department of Science and Technology of Sichuan Province (2026NSFSC1558, to ZC); the National Science Fund for Distinguished Young Scholars (82225029, to CW); the Special Project of Natural Science Foundation of China (82441031, to CW); the Key Project of the National Natural Science Foundation of China (82430076, to CW); the General Program of the National Natural Science Foundation of China (82572023, to RH); the Youth Fund of the National Natural Science Foundation of China (82301987, to BZ; 82302628, to YD; 82301989, to RH; 82402704, to YL; and 82502806, to MY); the Postdoctoral Foundation of China (2024M750363, to BZ); the Postdoctoral Foundation of Sichuan Provincial People’s Hospital (2022BH018, to LF), and the Postdoctoral Foundation of Sichuan Province (TB2023092, to LF). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: For pancreatic tissue analysis, a human tissue microarray (CHPanA120CS01) was purchased from the National Engineering Center for Biochip (Shanghai Outdo Biotech), with approval from the Clinical Research Ethics Committee (Approval no. SHYJS-CP-240105). Paired tumor and adjacent normal tissues from 16 lung adenocarcinoma (LUAD) patients and human pancreatic cancer tissues (n=3) were obtained from Sichuan Provincial People’s Hospital. This study followed the guidelines set forth by the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China (Approval no. 2023-663). Human peripheral blood mononuclear cells (PBMCs) were provided by Milecell Biological Science & Technology, with approval from the Clinical Research Ethics Committee (Approval no. Z-ZJMS-21-08-001). All the animal experiments were performed in strict accordance with the relevant ethical guidelines and were approved by the Department of the Ethics Committee of Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China (Approval no. 2022-756). Data availability free text: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Data availability statement Data are available on reasonable request. References Associated Data Supplementary Materials Data Availability Statement Data are available on reasonable request.