PD-1 H (VISTA) drives immunosuppressive reprogramming of glioma-associated myeloid cells to promote glioma progression Authors: Yuduo Guo, Chao Zhao, Huafeng Liu, Weihai Ning, Jitao Zhao, Yan Chen, Hang Li, Ying Li, Shengdian Wang, Hongwei Zhang Published: 2026 ABSTRACT AbstractBackgroundGlioma is characterized by a profoundly immunosuppressive tumor microenvironment (TME) that severely limits immune checkpoint blockade efficacy. Glioma-associated myeloid cells (GAMs), including resident microglia and infiltrating macrophages, constitute the dominant immune population within gliomas and shape antitumor immune responses. While immune checkpoint pathways have been extensively studied in T lymphocytes, the contribution of myeloid-restricted immune checkpoints to glioma progression remains poorly defined. Programmed death-1 homolog (PD-1 H/VISTA) is an inhibitory immune checkpoint predominantly expressed by myeloid cells, yet its functional role in GAMs has not been systematically investigated.MethodsPD-1 H expression in human glioma was analyzed using immunohistochemistry, bulk transcriptomic datasets, and single-cell RNA sequencing. Associations between PD-1 H expression, immune cell composition, and patient survival were evaluated in clinical cohorts. The functional role of PD-1 H in GAMs was examined using genetically manipulated myeloid cells in vitro and orthotopic glioma mouse models in vivo. Transcriptomic profiling was performed to identify PD-1 H-associated biological pathways. The therapeutic potential of PD-1 H blockade was assessed using monoclonal antibody treatment in glioma-bearing mice.ResultsPD-1 H was highly expressed in human glioma tissues and was predominantly enriched in GAMs. Elevated PD-1 H expression was associated with an immunosuppressive tumor microenvironment characterized by increased infiltration of immunosuppressive myeloid populations, reduced cytotoxic T-cell presence, elevated expression of multiple immune checkpoint molecules, and unfavorable clinical outcomes. Functionally, PD-1 H promoted immunosuppressive myeloid polarization, enhanced malignant behaviors of glioma cells, and suppressed CD8+ T-cell proliferation in vitro. In orthotopic glioma models, PD-1 H expression in GAMs accelerated tumor progression, promoted T-cell exhaustion, and shortened survival, whereas PD-1 H deficiency restrained tumor growth and prolonged survival. Mechanistic studies revealed that the AKT/NF-κB pathway mediates PD-1 H-dependent myeloid reprogramming. Importantly, therapeutic blockade of PD-1 H significantly suppressed glioma growth and improved survival in vivo.ConclusionsPD-1 H functions as a myeloid-restricted immune checkpoint that drives immunosuppressive reprogramming of glioma-associated myeloid cells, promotes glioma progression, and impairs antitumor T-cell immunity. Targeting PD-1 H represents a promising myeloid-focused immunotherapeutic strategy to overcome immune suppression in glioma.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-026-08497-0. Introduction Glioma is the most common and aggressive primary malignancy of the central nervous system, with glioblastoma representing the highest-grade and most lethal subtype. Despite maximal surgical resection followed by radiotherapy and temozolomide chemotherapy, patient prognosis remains poor, with median survival rarely exceeding 15 months [1–3]. In recent years, immune checkpoint blockade has transformed the treatment of multiple systemic cancers. However, standard approaches targeting the PD-1/PD-L1 axis or CTLA-4 have demonstrated limited clinical efficacy in large unselected glioma cohorts [4–6]. While emerging trials indicate that clinical responses can be context-dependent, with potential benefits observed primarily in the neoadjuvant window or within specific molecular subpopulations-the overall therapeutic benefit remains constrained by extensive resistance mechanisms [7–9]. These heterogeneous but generally limited outcomes highlight the presence of profound immunosuppressive mechanisms within the glioma tumor microenvironment (TME) that constrain effective antitumor immunity [10]. A hallmark of the glioma TME is the extensive infiltration of glioma-associated myeloid cells (GAMs), including resident microglia and bone-marrow-derived macrophages, which together can comprise up to half of the tumor mass [11, 12]. Rather than exerting antitumor functions, GAMs are frequently skewed toward an immunosuppressive phenotype that supports tumor growth, enhances invasion, and suppresses cytotoxic T-cell activity [13, 14]. Importantly, immune checkpoint pathways are now recognized to operate beyond T lymphocytes and to play critical roles in myeloid cell polarization, cytokine production, and antigen presentation [15–17]. Nevertheless, the identity and functional relevance of myeloid-restricted immune checkpoints in glioma remain incompletely defined. Programmed death-1 homolog (PD-1 H), also known as V-domain Ig suppressor of T-cell activation (VISTA), is an inhibitory immune checkpoint molecule predominantly expressed by myeloid cells. Initial studies identified PD-1 H as a negative regulator of T-cell responses, functioning independently of the PD-1/PD-L1 axis [18–20]. Subsequent work demonstrated that PD-1 H plays a critical role in maintaining immune quiescence and regulating innate immune responses [21, 22]. In several systemic malignancies, PD-1 H expression on tumor-infiltrating myeloid cells has been associated with immunosuppression and resistance to immune checkpoint blockade [23–25]. Despite these findings, the landscape of myeloid-restricted inhibitory receptors within the numerically dominant myeloid compartment of gliomas represents a distinct knowledge gap. Mechanistically, PD-1 H regulates both lymphoid and myeloid subsets within tumor microenvironments, potentially operating alongside conventional pathways like PD-1/PD-L1 and CTLA-4. Given that standard single immune checkpoint therapies primarily target T-cell immunity and fail to overcome myeloid cell-dominated suppression, targeting PD-1 H provides a strategy to simultaneously modulate adaptive and innate immune dysfunction [12, 26]. Crucially, preliminary bioinformatic screening of public transcriptomic and single-cell datasets reveals that PD-1 H expression is markedly elevated in human glioma compared with non-neoplastic brain tissues, and its expression profile is uniquely restricted to glioma-associated myeloid cells (GAMs) rather than malignant subclusters. Based on this glioma-specific preliminary evidence, we hypothesized that myeloid-restricted PD-1 H functions as an active regulator of immunosuppression and tumor progression in glioma. In this study, we systematically characterized PD-1 H expression in human glioma using bulk and single-cell transcriptomic analyses, immunohistochemistry, and flow cytometry. We further investigated the functional consequences of PD-1 H modulation in GAMs on myeloid polarization, glioma cell malignant behaviors, and T-cell responses using in vitro co-culture systems and orthotopic glioma models. Finally, we explored the molecular pathways associated with PD-1 H signaling and evaluated the therapeutic potential of PD-1 H blockade in vivo. Materials and methods Human specimens Human glioma specimens (WHO grade III–IV, n = 27) and histologically validated non-focal, non-neoplastic brain tissues resected from epilepsy patients (n = 5) were obtained from Sanbo Brain Hospital, Capital Medical University between 2021 and 2022. Written informed consent was obtained from all participants. The study was approved by the Ethics Committee of Sanbo Brain Hospital, Capital Medical University (YJ-2021–023-02) and conducted in accordance with the Declaration of Helsinki. Bioinformatic analyses Bulk RNA-seq data and clinical information were obtained from TCGA (https://portal.gdc.cancer.gov/), while normal brain data were retrieved from GTEx (https://gtexportal.org/). Single-cell RNA-seq data (SCP2389) were downloaded from the Single Cell Portal (https://singlecell.broadinstitute.org/). Differential expressions, immune infiltration, correlation, and survival analyses were performed using R (V4.4) software. Flow cytometry Single-cell suspensions from tumors or cultured cells were stained with fluorochrome-conjugated antibodies against CD45, CD11b, F4/80, Tmem119, CD3, CD4, CD8, PD-1 H, CD206, CD86, PD-1, and TIM-3. Dead cells were excluded using viability dye. Data was acquired on a CytoFLEX LX flow cytometer (Beckman coulter) and analyzed using FlowJo (V.10.9) software. Genetic manipulation of PD-1 H PD-1 H-deficient BV2 cells were generated using CRISPR-Cas9-mediated genome editing. Single-guide RNAs targeting the murine Pd-1 h (Vsir) gene were designed (https://www.benchling.com) and cloned into lentiviral CRISPR vectors. sgRNA-1: 5’-CACCCGTACCTCAGCTCACGAGGCG-3’, sgRNA-2: 5’-ACCCTGACTCCGAAGGTTGCCGC-3’. Lentiviral particles were produced in HEK293T cells and used to transduce BV2 cells, followed by antibiotic selection. For PD-1 H overexpression, full-length murine Pd-1 h (Vsir) cDNA was cloned into a lentiviral expression vector. BV2 cells were transduced with PD-1 H-expressing or control lentiviruses, and stable populations were established by fluorescence-based sorting. PD-1 H expression levels were confirmed by WB and flow cytometry. In vitro CD8+ T-cell activation, expansion, and CFSE proliferation assays Unfractionated total lymph node cell suspensions from OT-I mice were used. For baseline quality control (Fig. S4), total lymph node cells were stimulated with OVA257-268 peptide (0.1 ng/ml) for 48 h, and CD25, CD44, and CD62L expression on T cells was analyzed via flow cytometry. For the antigen-priming phase (Fig. 3K), manipulated BV2 cells and 0.1 ng/ml OVA peptide were co-cultured with the lymph node cell suspension at a T-cell to BV2 ratio of 2:1 for 48 h. For the differentiation phase (Fig. 3L), lymph node cells were pre-primed with OVA peptide for 48 h, washed, and then co-cultured with BV2 cells at a T-cell to BV2 ratio of 2:1 in complete medium supplemented with 10 ng/ml IL-2 for an additional 48 h. PD-1 H regulates GAM polarization to promote glioma malignancy and suppress T-cell immunity. (A) Flow cytometric analysis of CD206 expression in BV2 cells, used as an in vitro model of GAMs, with PD-1 H knockout (PD-1 H-KO and PD-1 H-KO#1) or negative control (PD-1 H-NC) under basal conditions (blank) or following IL-4 stimulation (M2) for 48 h. (B) Flow cytometric analysis of CD86 expression in PD-1 H-deficient or control BV2 cells under basal conditions (blank) or after IFN-γ stimulation (M1) for 48 h. (C) Phagocytic activity of PD-1 H-deficient and control BV2 cells, assessed by uptake of fluorescent microspheres (FluoSpheres) after 4 h incubation. (D) Cell viability of GL261 glioma cells cultured with conditioned media derived from PD-1 H-NC, PD-1 H-KO, PD-1 H-KO#1, lenti-NC, or PD-1 H-overexpressing BV2 cells, measured by CCK-8 assay at 24, 48, and 72 h (***p < 0.001, ****p < 0.0001). (E, F) Transwell invasion assays of GL261 cells co-cultured with BV2 cells expressing different levels of PD-1 H. Representative images (E) and quantification of invading cells per field (F) Are shown (*p < 0.05). (G, H) Colony formation assay of GL261 cells cultured with conditioned medium from BV2 cells as indicated. Representative images (G) and quantification of colony numbers (H) are shown (*p < 0.05, **p < 0.01). (I, J) Wound-healing assays of GL261 cells at 0 h and 48 h after scratch following treatment with BV2 cell–derived conditioned media. Representative images (I) and quantification of wound closure (J) Are shown (*p < 0.05, **p < 0.01). (K) Antigen-specific proliferation of OT-I CD8+ T cells in the priming stage. Total lymph node cells were stimulated with OVA257-268 peptide and co-cultured with BV2 cells for 48 h. Representative CFSE histograms are shown. (L) Proliferation analysis of CD8+ T cells in the differentiation stage. Pre-primed T cells were co-cultured with BV2 cells under IL-2 stimulation for another 48 h. Relative proliferative capacity was quantified via derivative analysis of normalized CFSE signals. Data are presented as mean ± SD (**p < 0.01) For proliferation tracking, cells (1×107) were labeled with 5 μM CFSE (37 °C, 10 min), washed, and analyzed by flow cytometry. Multi-generation division histories in Fig. 3K were captured via dilution histograms. For Fig. 3L, raw CFSE fluorescence profiles were normalized against baseline controls, and a mathematical derivative analysis (first derivative of the normalized signal) was performed to calculate the relative proliferation index of CD8+ T cells. Animal experiments C57BL/6 mice (6–8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. OT-I transgenic mice were obtained from The Jackson Laboratory. PD-1 H knockout mice (C57BL/6PD-1H-/-) were provided by the laboratory of Prof. Lieping Chen (Yale University). All mice were on a C57BL/6 genetic background and were maintained under specific pathogen-free conditions at the animal facility of the Institute of Biophysics, Chinese Academy of Sciences. Orthotopic glioma models were established by stereotactic intracranial implantation of GL261-luc cells, either alone or together with BV2 cells expressing PD-1 H knockout, overexpression, or corresponding controls. Tumor growth was monitored longitudinally by bioluminescence imaging. Survival was assessed using Kaplan-Meier analysis. For in vivo immune profiling, tumor-bearing mice were euthanized at indicated time points, and brain tumors were harvested for preparation of single-cell suspensions. Tumor-infiltrating immune cells were analyzed by flow cytometry. For therapeutic studies, PD-1 H-deficient mice bearing orthotopic gliomas were randomized on day 7 post-implantation into two groups based on baseline tumor bioluminescence radiance values to ensure equivalent initial tumor burden between groups. Mice were subsequently treated with PBS or anti-PD-1 H monoclonal antibody (13F3, 300 μg/mouse, HY-P990145, MCE) every two days, a dosing regimen based on previous studies [18, 27]. Tumor signal acquisition (bioluminescence imaging), quantification, and endpoint evaluation were performed by an investigator blinded to the allocation of treatment groups. Statistical analysis All data are presented as mean ± SD unless otherwise indicated. Statistical analyses were performed using GraphPad Prism (V10.1) and R (V4.4) software. For comparisons between two groups, unpaired two-tailed Student’s t tests were used. For comparisons among multiple groups, one-way ANOVA followed by appropriate post hoc tests was applied. Correlations between variables were assessed using Spearman correlation analysis. Survival analyses were performed using the Kaplan-Meier method and compared using the log-rank test. Results PD-1 H is highly expressed in glioma and predominantly enriched in GAMs We first examined PD-1 H expression in human glioma tissues and normal brain. Immunohistochemical (IHC) staining revealed markedly higher PD-1 H expression in glioma compared with normal brain tissue, which was confirmed by quantitative IHC scoring (**p < 0.01; Fig. 1A–B). Consistently, bulk RNA-seq analysis of TCGA and GTEx datasets showed significantly elevated PD-1 H mRNA levels in glioma samples (n = 702) relative to normal brain tissues (n = 423; ****p < 0.0001; Fig. 1D). Comparison of immune checkpoint molecules in the TCGA glioma cohort further demonstrated that PD-1 H was highly expressed relative to other inhibitory receptors, including TIM-3, LAG-3, TIGIT, CTLA-4, PD-L1, and PD-1 (Fig. 1C). PD-1 H expression and distribution in glioma. (A) Representative IHC images of PD-1 H in human glioma vs. normal brain tissue (scale bars, 200 μm). (B) Quantification of IHC scores for PD-1 H in glioma (n = 27) and normal brain (n = 5) samples (**p < 0.01). (C) Violin plots showing expression distribution of PD-1 H and other immune checkpoint molecules (TIM-3, LAG-3, TIGIT, CTLA-4, PD-L1, and PD-1) in glioma samples from the TCGA dataset. (D) Bulk RNA-seq of PD-1 H expression levels in glioma (n = 702) and normal brain (n = 423) samples based on RNA-seq data from TCGA and GTEx datasets (****p < 0.0001). (E) UMAP plot showing single-cell RNA-seq data of glioma, colored by cell type (5.27 × 105 cells from 85 glioma samples, single-cell portal, SCP2389). (F) UMAP plot with PD-1 H expression intensity (color scale). (G) Bar chart showing the expression of PD-1 H in different cell clusters Single-cell RNA-seq analysis revealed that PD-1 H expression was not uniformly distributed across cell types within the glioma microenvironment. UMAP visualization showed enrichment of PD-1 H expression in specific cellular populations, with the highest expression observed in myeloid cells compared with malignant cells, T cells, and other stromal or immune populations (Fig. 1E–G, p < 0.0001). Given that these myeloid populations primarily consisted of microglia and infiltrating macrophages, we hereafter refer to glioma-associated myeloid cells collectively as GAMs. Consistent with these findings, PD-1 H expression positively correlated with macrophage enrichment in glioma samples from the TCGA dataset (Spearman’s R = 0.328, p < 0.001; Fig. S1A). To further validate PD-1 H expression at the cellular level, flow cytometric analysis was performed in a mouse glioma model. PD-1 H expression was highest on Tmem119+ microglia and CD11b+ macrophages, while CD3+ T cells exhibited relatively low expression. Quantification of mean fluorescence intensity confirmed significantly higher PD-1 H expression in GAMs compared with T cells and isotype controls (p < 0.0001; Fig. S1B–C). Collectively, these results demonstrate that PD-1 H is upregulated in glioma and is predominantly expressed by GAMs within the tumor microenvironment. High PD-1 H expression in GAMs associates with an immunosuppressive microenvironment and poor prognosis As PD-1 H was predominantly expressed by GAMs (Fig. 1), we next evaluated its relationship with immune cell phenotypes in a cohort of glioma specimens. Two representative cases are shown: a glioma with low PD-1 H expression (Fig. 2A–E) exhibited reduced infiltration of macrophages (CD68+) and M2-like macrophages (CD163+), but higher densities of antigen-presenting cells (APCs), M1-like macrophages (CD80+) and cytotoxic T lymphocytes (CTLs, CD8+). In contrast, a glioma with high PD-1 H expression (Fig. 2F–J) displayed abundant CD68+ and CD163+ cells but diminished CD80+ cells and CD8+ CTLs. Consistent with these observations, correlation analysis in our cohort (n = 27) demonstrated that PD-1 H expression positively correlated with CD68 and CD163 expression, while showing negative correlations with CD80 and CD8 expression (Fig. 2K–N). Furthermore, analysis of TCGA glioma samples revealed that PD-1 H expression was positively associated with multiple immune checkpoint molecules, including TIM-3, PD-1, PD-L1, and CTLA-4 (Fig. 2O–R). Clinically, Kaplan–Meier survival analysis showed that glioma patients stratified by PD-1 H expression levels in GAMs exhibited significantly different outcomes, with high PD-1 H expression associated with shorter progression-free survival and overall survival compared with low PD-1 H expression (Fig. 2S–T, log-rank p < 0.0001). Together, these data indicate that elevated PD-1 H expression in GAMs is associated with distinct immunological features of the glioma microenvironment and unfavorable clinical prognosis. PD-1 H expression in GAMs is associated with immune cell composition and clinical outcome in glioma. (A–E) Representative IHC staining of PD-1 H, CD68, CD163, CD80, and CD8 in a glioma specimen with low PD-1 H expression (patient-1). (F–J) Corresponding IHC staining of PD-1 H, CD68, CD163, CD80, and CD8 in a glioma specimen with high PD-1 H expression (patient-2). Scale bars, 200 μm. (K–N) Correlation analyses between PD-1 H and immune cell markers, including CD68, CD163, CD80 and CD8 in glioma samples (n = 27). (O–R) Correlation analyses of expression levels between PD-1 H and immune checkpoint molecules, including TIM-3, PD-L1, PD-1, and CTLA-4 in glioma samples from the TCGA dataset (n = 702, Spearman correlation, p < 0.001). (S, T) Kaplan–Meier survival curves of glioma patients stratified by PD-1 H expression levels in GAMs, including progression-free survival (PFS, S) and overall survival (OS, T) (Log-rank test) PD-1 H in GAMs regulates myeloid polarization, promotes glioma cell malignant behaviors, and modulates T-cell proliferation To investigate the functional role of PD-1 H in GAMs, BV2 cells were used as an in vitro model with genetic manipulation of PD-1 H expression. Efficient PD-1 H knockout (PD-1 H-KO, PD-1 H-KO#1) and overexpression (PD-1 H-OE) in BV2 cells were confirmed by flow cytometry and immunoblotting (Fig. S2A, B, F, G). Manipulation of PD-1 H expression did not significantly affect BV2 cell viability or apoptosis, as assessed by CCK-8 assays and Annexin V/7-AAD staining (Fig. S2C–E, H–J), indicating that subsequent functional changes were not attributable to altered cell survival. We next examined whether PD-1 H influences myeloid polarization. Under IL-4 stimulation, PD-1 H-deficient BV2 cells exhibited reduced CD206 expression compared with control cells, whereas PD-1 H overexpression enhanced CD206 expression (Fig. 3A; Fig. S3B). Conversely, IFN-γ–induced CD86 expression was increased in PD-1 H-deficient BV2 cells but suppressed upon PD-1 H overexpression (Fig. 3B; Fig. S3A). In addition, PD-1 H-deficient BV2 cells displayed enhanced phagocytic activity, while PD-1 H overexpression reduced uptake of fluorescent microspheres (Fig. 3C; Fig. S3C). Consistent with these phenotypic changes, quantitative PCR analysis demonstrated that PD-1 H knockout attenuated the induction of M2-associated genes (Cd206, Cd163, and Arg1) under IL-4 stimulation, whereas PD-1 H overexpression markedly upregulated their expression (Fig. S3D–F). In contrast, expression of M1-associated genes (Cd86, Tnf-α, and Nos2) induced by IFN-γ was enhanced by PD-1 H deficiency and suppressed by PD-1 H overexpression (Fig. S3G–I). We then assessed the impact of PD-1 H–manipulated GAMs on glioma cell behaviors. Conditioned media derived from PD-1 H-deficient BV2 cells significantly reduced GL261 glioma cell viability, whereas conditioned media from PD-1 H-overexpressing BV2 cells promoted GL261 cell growth (Fig. 3D). In Transwell co-culture assays, PD-1 H-deficient BV2 cells markedly decreased the invasive capacity of GL261 cells, while PD-1 H overexpression exerted the opposite effect (Fig. 3E, F). Similarly, conditioned media from PD-1 H-deficient BV2 cells suppressed colony formation and delayed wound closure in GL261 cells, whereas PD-1 H-overexpressing BV2 cells significantly enhanced clonogenic growth (Fig. 2G, H) and migratory capacity (Fig. 3I, J). These phenotypic shifts observed in the BV2 cell line were further evaluated and replicated in primary bone marrow-derived macrophages (BMDMs) isolated from WT and PD-1 H-/- mice (Fig. S5A-H). Finally, to determine whether PD-1 H expression in GAMs dynamically regulates CD8+ T-cell responses, unfractionated total lymph node cell suspensions from OT-I mice were utilized across two distinct functional phases. In the initial antigen-priming phase, baseline quality control verified that the unfractionated lymph node system successfully primed CD8+ T cells independently (Fig. S4). Upon this validation, CFSE-based proliferation assays revealed that the introduction of PD-1 H-OE BV2 cells significantly restricted the initial proliferative responses of CD8+ T cells, whereas PD-1 H-KO enhanced the proliferation (Fig. 3K). In the subsequent differentiation phase, where pre-primed T cells were expanded via IL-2 stimulation, T cells were co-cultured with manipulated BV2 cells to evaluate the specific checkpoint effects of myeloid PD-1 H. Consistent with the initial priming stage, proliferation analysis of the CFSE signals confirmed that while PD-1 H-OE BV2 cells significantly slowed down CD8+ T-cell expansion velocity, PD-1 H-KO accelerated it (Fig. 3L). Together, these results demonstrate that PD-1 H expression in GAMs skews myeloid polarization toward an immunosuppressive phenotype, enhances malignant behaviors of glioma cells, and systematically attenuates CD8+ T-cell proliferative responses in vitro. PD-1 H expression in GAMs promotes glioma progression and is associated with impaired antitumor T-cell responses in vivo To determine the in vivo relevance of PD-1 H expression in GAMs, we established an orthotopic glioma model in which GL261-luc cells were intracranially implanted together with GAMs expressing different levels of PD-1 H (Fig. 4A). Longitudinal bioluminescence imaging revealed that mice receiving PD-1 H-overexpressing GAMs exhibited markedly accelerated tumor growth compared with control mice, whereas implantation with PD-1 H-deficient GAMs significantly restrained tumor progression (Fig. 4B, C). Consistent with these findings, Kaplan–Meier survival analysis demonstrated that PD-1 H overexpression in GAMs was associated with significantly reduced survival, while PD-1 H deficiency prolonged survival in glioma-bearing mice (Fig. 4D, log-rank p < 0.001). PD-1 H expression in GAMs promotes glioma progression and impairs antitumor T-cell responses in vivo. (A) Schematic illustration of the in vivo experimental design. C57BL/6 mice were intracranially implanted with GL261-luc glioma cells together with GAMs expressing PD-1 H overexpression (PD-1 H-OE), negative control (PD-1 H-NC), or PD-1 H knockout (PD-1 H-KO). Tumor growth was monitored by bioluminescence imaging at the indicated time points. (B) Representative bioluminescence images of tumor-bearing mice from each group at days 7, 14, and 21 after implantation. (C) Quantification of tumor bioluminescence radiance over time in mice receiving PD-1 H-OE, PD-1 H-NC, or PD-1 H-KO GAMs (*p < 0.05, ***p < 0.001). (D) Kaplan–Meier survival analysis of glioma-bearing mice in the indicated groups. Statistical significance was determined by log-rank test (p < 0.001). (E) Flow cytometric analysis of GAM phenotypes isolated from glioma tissues. Representative histograms show expression of CD206 and MHC-II on GAMs from PD-1 H-OE, PD-1 H-NC, and PD-1 H-KO groups. (F, G) Representative flow cytometry histograms showing expression of PD-1 and TIM-3 on tumor-infiltrating CD8+ T (F) and CD4+ T (G) cells from each group We next analyzed the phenotype of tumor-infiltrating GAMs. Flow cytometric analysis showed that PD-1 H-overexpressing GAMs displayed increased expression of CD206 and MHC-II, whereas PD-1 H-deficient GAMs exhibited reduced expression of both markers (Fig. 4E), indicating that PD-1 H modulates the immunophenotype of GAMs in vivo. To assess the impact of PD-1 H–manipulated GAMs on antitumor immunity, we examined tumor-infiltrating T cells. Both CD8+ and CD4+ T cells from tumors harboring PD-1 H-overexpressing GAMs showed elevated expression of exhaustion-associated markers PD-1 and TIM-3 compared with controls, whereas PD-1 H deficiency in GAMs was associated with reduced expression of these inhibitors (Fig. 4F, G). Collectively, these results demonstrate that PD-1 H expression in GAMs promotes glioma progression in vivo and is closely associated with an immunosuppressive tumor microenvironment characterized by enhanced T-cell exhaustion. PD-1 H reshapes transcriptional programs and identifies AKT/ NF-κB signaling in GAMs To further elucidate the molecular mechanisms underlying PD-1 H-mediated functional changes in GAMs, transcriptomic profiling was performed in BV2 cells with PD-1 H overexpression (OE), PD-1 H knockout (KO), or negative control (NC). Principal component analysis (PCA) revealed clear separation among the three groups, indicating that modulation of PD-1 H expression induces distinct global transcriptional states in GAMs (Fig. 5A). Differential expression analysis identified substantial numbers of genes that were differentially expressed in both PD-1 H-OE and PD-1 H-KO cells compared with controls (Fig. 5B–E), suggesting that PD-1 H profoundly influences the transcriptional landscape of GAMs. PD-1 H regulates transcriptional programs and signaling pathways in GAMs. (A) Principal component analysis (PCA) of transcriptomic profiles from BV2 cells with PD-1 H-KO (KO), PD-1 H-NC (NC), or PD-1 H-OE (OE), showing distinct clustering among groups. (B, C) Volcano plot (B) and heatmap (C) showing differentially expressed genes (DEGs) in PD-1 H-OE vs. PD-1 H-NC. Upregulated and downregulated genes are highlighted. (D, E) Volcano plot (D) and heatmap (E) displaying DEGs in PD-1 H-KO vs. PD-1 H-NC. (F, G) Gene ontology (GO) enrichment analysis of DEGs from PD-1 H-OE vs. PD-1 H-NC (F) and PD-1 H-KO vs. PD-1 H-NC (G), with dot size indicating gene count and color representing statistical significance. (H) Western blot analysis of NF-κB, phosphorylated AKT (p-AKT), total AKT, in BV2 cells with different expression levels of PD-1 H. The PD-1H-blocking antibody 13F3 was applied as indicated (−/+) To identify biological processes associated with PD-1 H-dependent transcriptional changes, Gene Ontology (GO) enrichment analysis was performed. In PD-1 H-overexpressing cells, differentially expressed genes were significantly enriched in immune- and inflammation-related processes, including myeloid leukocyte–mediated immunity, cytokine-mediated signaling, regulation of phagocytosis, and notably, positive regulation of NF-κB signaling (Fig. 5F). Conversely, GO analysis of PD-1 H-deficient cells revealed enrichment of pathways associated with immune activation, such as regulation of cytokine production, T-cell activation, and Toll-like receptor signaling, and prominently highlighted negative regulation of NF-κB transcription factor activity (Fig. 5G). Importantly, NF-κB-related regulatory processes were consistently enriched in both gain- and loss-of-function analyses, albeit in opposite directions, indicating that NF-κB signaling represents a central pathway associated with PD-1 H-mediated transcriptional reprogramming in GAMs. Guided by these GO analysis results, we next examined intracellular signaling pathways associated with PD-1 H. Immunoblot analysis revealed that PD-1 H overexpression was associated with reduced NF-κB expression but increased AKT phosphorylation, whereas PD-1 H deficiency resulted in elevated NF-κB expression accompanied by decreased AKT phosphorylation (Fig. 5H and Fig. S5I). Notably, the total protein level of AKT remained largely unchanged across different conditions, indicating that PD-1 H primarily modulates AKT activation rather than AKT expression. Moreover, treatment with 13F3, a PD-1 H–blocking monoclonal antibody, resulted in similar changes in NF-κB expression and AKT phosphorylation. To evaluate whether these downstream phenotypes are directly driven by the identified axis, rescue experiments were performed using the AKT activator SC79. Reactivation of AKT successfully rescued its suppressed phosphorylation in PD-1 H-KO cells and concurrently counteracted the upregulation of NF-κB (Fig. S6E). Consistently, this molecular rescue restored glioma proliferation, colony formation, migration, and invasion (Fig. S6A–D), confirming that the PD-1 H-dependent effects are mediated via the AKT/NF-κB pathway. Together, these findings demonstrate that PD-1 H orchestrates the immunomodulatory and tumor-promoting phenotypes of GAMs through the coordinated regulation of the AKT/NF-κB signaling axis. Therapeutic blockade of PD-1 H suppresses glioma growth in PD-1 H-deficient host mice To further evaluate the therapeutic potential of PD-1H blockade while excluding the contribution of host-derived PD-1H, an orthotopic glioma model was established in C57BL/6PD-1H-/- mice. In this compartment-defined setup, GL261-luc glioma cells were intracranially implanted together with PD-1 H-expressing GAMs (PD-1 H-NC BV2 cells), which represented the exclusive source of PD-1 H within the host organism. Consequently, systemic administration of the anti-PD-1 H monoclonal antibody (13F3) was restricted to targeting the local GAM compartment without confounding engagement of host immune cells. Crucially, with both groups sharing an identical genetic host background, the single variable evaluated was the administration of either PBS or the anti-PD-1 H antibody according to the indicated schedule (Fig. 6A) Tumor progression was monitored longitudinally by bioluminescence imaging. Therapeutic blockade of PD-1 H suppresses glioma progression in PD-1 H-/- host mice. (A) Schematic illustration of the in vivo experimental design. C57BL/6PD-1 H-/- mice were intracranially implanted with GL261-luc glioma cells together with PD-1 H-NC GAMs. Mice were treated with PBS or anti-PD-1 H monoclonal antibody at the indicated time points, and tumor growth was monitored by serial bioluminescence imaging. (B) Representative bioluminescence images of mice treated with PBS or anti-PD-1 H antibody at days 7, 14, 21, and 28 after implantation. (C) Quantification of tumor bioluminescence radiance over time in PBS- and anti-PD-1 H-treated mice, *p < 0.05. (D) Kaplan–Meier survival curves of glioma-bearing mice receiving PBS or anti-PD-1 H treatment, log-rank test, *p < 0.05 Serial in vivo imaging revealed that glioma-bearing PD-1 H− /− mice treated with anti-PD-1 H antibody exhibited markedly reduced tumor growth compared with PBS-treated controls (Fig. 6B). Quantitative analysis of bioluminescence radiance demonstrated a significant suppression of tumor burden in the anti-PD-1 H–treated group over time (Fig. 6C). Consistent with these findings, Kaplan-Meier survival analysis showed that anti-PD-1 H treatment significantly prolonged survival of glioma-bearing mice compared with PBS-treated mice (Fig. 6D, log-rank p = 0.0492). Collectively, these results demonstrate that therapeutic blockade of PD-1 H effectively restrains glioma progression and improves survival even in the absence of host PD-1 H, supporting the notion that PD-1 H within the tumor microenvironment, particularly on GAMs, plays a critical role in driving glioma growth and represents a viable therapeutic target. Discussion This study defines the expression pattern, functional role, and therapeutic relevance of the immune checkpoint molecule PD-1 H in the glioma tumor microenvironment, with a focus on glioma-associated myeloid cells (GAMs). PD-1 H is predominantly enriched within the myeloid compartment, where it functions as a myeloid-restricted immune checkpoint that promotes immunosuppressive myeloid polarization, supports glioma progression, and impairs antitumor T-cell immunity [18, 20, 21]. PD-1 H–expressing GAMs enhance tumor-supportive activities while suppressing CD8+ T-cell proliferation and promoting T-cell exhaustion in vivo, providing a mechanistic basis for its association with unfavorable clinical outcomes and aligning with emerging models of myeloid-driven immune suppression in cancer [10, 15, 16, 28]. Transcriptomic and signaling analyses further link PD-1 H–dependent myeloid reprogramming to coordinated activation of NF-κB- and AKT-associated pathways, which are central regulators of myeloid activation, inflammatory signaling, and metabolic adaptation [29–31]. PD-1 H (also known as VISTA) has been previously characterized as an inhibitory immune checkpoint preferentially expressed by myeloid cells, with established roles in maintaining immune quiescence and regulating immune responses in peripheral tissues and systemic malignancies [25, 32, 33]. In multiple solid tumors, including melanoma, lung cancer, and colorectal cancer, elevated PD-1 H expression on tumor-infiltrating myeloid cells has been associated with immunosuppressive tumor microenvironments, impaired T-cell effector function, and resistance to T-cell-directed immune checkpoint blockade [23, 34, 35]. These observations support a broader role for PD-1H in shaping innate immune regulation across diverse tumor contexts. In glioma, however, immune checkpoint research has historically focused on PD-1/PD-L1 and CTLA-4 signaling in T lymphocytes, despite the numerical and functional dominance of myeloid cells within the tumor microenvironment [6, 12, 36]. Extensive studies have demonstrated that GAMs actively promote glioma progression through cytokine secretion, angiogenic support, metabolic reprogramming, and suppression of cytotoxic T-cell activity [11, 14, 17]. More recent single-cell transcriptomic analyses have revealed pronounced heterogeneity and functional plasticity within the GAM compartment, indicating that distinct myeloid states contribute differentially to glioma growth and immune regulation [16, 37, 38]. Despite these advances, the molecular determinants that actively drive immunosuppressive reprogramming of GAMs have remained incompletely defined, with most prior studies relying on descriptive or correlative analyses [37, 39]. Although PD-1H expression has been reported in myeloid populations across tumor types, direct functional evidence linking PD-1H to GAM-mediated immunosuppression and glioma progression has been lacking [19, 24, 25, 32]. By integrating expression profiling, functional manipulation, mechanistic interrogation, and in vivo therapeutic validation, the present study demonstrates that PD-1H is not merely a marker of immunosuppressive myeloid states but an active regulator that orchestrates GAM polarization, tumor-supportive functions, and suppression of antitumor T-cell immunity. Therapeutic strategies aimed at targeting the myeloid compartment in glioma have largely relied on indirect approaches, such as inhibition of CSF1R signaling or modulation of macrophage differentiation and education [40–42]. While these strategies have provided important proof-of-concept evidence, their efficacy has been limited by adaptive resistance mechanisms and incomplete or transient myeloid reprogramming [40, 42]. In contrast, our findings suggest that direct targeting of PD-1H, a myeloid-restricted immune checkpoint, may provide a more specific and potentially durable strategy to disrupt myeloid-mediated immune suppression in glioma. Notably, therapeutic blockade of PD-1H effectively restrained tumor growth and prolonged survival within the PD-1H-deficient host environment. This compartment-specific efficacy is informative because the PD-1 H-/- host background isolates the tumor-infiltrating myeloid checkpoint from systemic host immunity, thereby restricting the pharmacological target of the 13F3 antibody to the GAM compartment. This experimental framework demonstrates that blocking PD-1 H specifically within the local myeloid tumor microenvironment is sufficient to mitigate immune evasion, bypassing the requirement for systemic host checkpoint manipulation and suggesting its role as a targetable immunotherapeutic axis. Collectively, these results support a conceptual shift from a predominantly T-cell-centric view of immune checkpoint regulation toward a framework that recognizes myeloid cells as dominant and targetable regulators of immune suppression in glioma [15, 16]. From a therapeutic perspective, targeting PD-1 H may complement existing immune checkpoint therapies by modulating the innate immune compartment rather than directly targeting T cells, consistent with emerging interest in innate immune checkpoints as anticancer targets [43–45]. Several limitations should be acknowledged. First, although our co-implantation model using PD-1 H-/- host mice controlled for exclusive antibody target engagement, global knockout strains possess baseline immunological dysregulation. Therefore, we cannot entirely exclude the systemic influence of a pre-altered immune background on the local therapeutic effects. Definitive validation of cell-type-specific PD-1 H functions under physiological conditions will require conditional knockout models, such as Cx3cr1CreERT2×Pd1hfl/fl strains on a wild-type background [46, 47]. Furthermore, our exploratory clinical analysis is inherently limited by a relatively constrained local cohort size and the reliance on epilepsy-derived non-neoplastic tissue controls, though cross-validated by large-scale public cohorts, broader multi-center verification remains necessary. Finally, the clinical analyses presented here are correlative in nature and do not establish PD-1 H as a predictive biomarker for therapeutic response, which will require prospective clinical validation. In summary, this study identifies PD-1H as a myeloid-restricted immune checkpoint that orchestrates immunosuppressive reprogramming of glioma-associated myeloid cells, promotes glioma progression, and limits antitumor T-cell immunity. These findings expand current understanding of immune checkpoint regulation in glioma and provide a strong rationale for targeting PD-1H as part of myeloid-focused immunotherapeutic strategies. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary material 1 Supplementary material 2 Supplementary material 3 Supplementary material 4 Supplementary material 5 Supplementary material 6 Supplementary material 7 Supplementary material 8 Acknowledgements We thank Prof. Lieping Chen (Yale University) for kindly providing the PD-1 H knockout mice. Author contributions YG performed the data analysis and wrote the manuscript, YG and CZ performed most of the experiments; HZ and SW designed and supervised the study. All authors contributed to the article and approved the submitted version. Funding This work was supported by CFH-brain (No. 2024-1-6013), National Natural Science Foundation of China (No. 82101932), Xuanwu Hospital Elite Cultivation Program (No. YC20250105), and Tianjin Key Medical Discipline Construction Project (No. TJYXZDXK-3-002A). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data Availability The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors. Declarations Ethics approval and consent to participate The study involving human specimens was approved by the Ethics Committee of Sanbo Brain Hospital, Capital Medical University (No. YJ-2021-023-02). Written informed consent was obtained from all participants. All animal experiments were performed in accordance with guidelines of the Institutional Laboratory Animal Care and Use Committee, Institute of Biophysics, Chinese Academy of Sciences. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yuduo Guo and Chao Zhao contributed equally to this work. Contributor Information Shengdian Wang, Email: sdwang@ibp.ac.cn. Hongwei Zhang, Email: zhanghongwei@ccmu.edu.cn. References Associated Data Supplementary Materials Supplementary material 1 Supplementary material 2 Supplementary material 3 Supplementary material 4 Supplementary material 5 Supplementary material 6 Supplementary material 7 Supplementary material 8 Data Availability Statement The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.