M2 Macrophage-Based Classification Identifies DOK3 as a Driver of Pro-tumoral Polarization and Migration in Glioblastoma

M2 Macrophage-Based Classification Identifies DOK3 as a Driver of Pro-tumoral Polarization and Migration in Glioblastoma | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article M2 Macrophage-Based Classification Identifies DOK3 as a Driver of Pro-tumoral Polarization and Migration in Glioblastoma Chang-Yuan Ren, Ji Shi, Chang-Lin Yang, Jin-Hao Zhang, Wen-Lu Tan, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7534936/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: IDH-wildtype glioblastoma (GBM) is the most aggressive primary brain tumor, characterized by limited therapeutic options and dismal prognosis. Among the components of the immunosuppressive tumor microenvironment (TME), M2-polarized macrophages are pivotal mediators of tumor progression, yet the molecular mechanisms underlying their polarization and pro-tumoral functions remain inadequately understood. Methods: We integrated bulk and single-cell RNA sequencing datasets from CGGA, TCGA, and GSE131928. M2 macrophage infiltration was quantified using the xCell algorithm, and unsupervised clustering of M2-associated genes defined immune subtypes, further refined by XGBoost and LASSO modeling. A macrophage-associated risk score based on CTSB , LITAF , and DOK3 was constructed and validated across independent cohorts for prognostic and immune relevance. Functional validation was performed by silencing DOK3 in THP-1–derived macrophages, followed by co-culture with glioma cells to assess macrophage polarization and tumor cell behavior. Results Elevated M2 macrophage infiltration correlated with reduced tumor purity, spatial heterogeneity, and worse survival. Three immune subtypes (C1-C3) were identified; notably, the C1 subtype exhibited the highest M2 infiltration, strongest immunosuppressive features, and poorest prognosis. The macrophage-based risk score robustly stratified patient survival and correlated with CD163 expression and immune checkpoint activation. Single-cell analysis revealed predominant DOK3 expression in macrophages and microglia. Functional assays demonstrated that DOK3 knockdown reduced CD163 expression and attenuated glioma cell invasiveness, supporting its role in promoting M2 polarization and tumor aggressiveness. Conclusion: This integrative analysis identifies DOK3 as a pivotal regulator of M2 macrophage polarization and a driver of glioblastoma progression. The macrophage-based risk score provides a practical tool for prognostic stratification, and targeting DOK3 offers a promising therapeutic strategy to reprogram the TME and improve clinical outcomes in patients with IDH-wildtype GBM. glioblastoma DOK3 M2 macrophage pro-tumoral polarization tumor microenvironment Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Introduction Glioblastoma (GBM) is the most common and highly aggressive primary malignant tumor of the central nervous system (CNS), accounting for approximately 80% of primary malignant CNS tumors [ 1 , 2 ]. Despite advances in multimodal treatment–including maximal surgical resection, radiotherapy, and temozolomide-based chemotherapy–the prognosis remains dismal, with a median overall survival (OS) of only 14–18 months and a 5-year survival rate below 10% [ 3 ]. The highly infiltrative nature of GBM hampers completed surgical resection, contributing to inevitable recurrence and resistance to conventional therapies [ 4 ]. Recent molecular and immunological profiling has revealed that GBM harbors a profoundly immunosuppressive tumor microenvironment (TME), enriched with regulatory immune cells and inhibitory soluble factors that dampen the activity of cytotoxic T cells and natural killer (NK) cells, thereby promoting immune evasion and disease progression [ 5 – 7 ]. This immunosuppressive milieu not only accelerates tumor growth but also limits the efficacy of emerging immunotherapeutic strategies, highlighting the urgent need for precise prognostic stratification systems and robust molecular target to guide personalized treatment strategies. Tumor-associated macrophages (TAMs) are the dominant immune cell component within the GBM microenvironment, with the M2-polarized phenotype constituting the predominant and functionally pro-tumoral subset [ 8 ]. High densities of M2 macrophages have been strongly associated with higher WHO grades, enhanced invasiveness, and poorer clinical outcomes [ 9 ]. Mechanistically, M2 macrophages exert potent immunosuppressive effects by secreting IL10, TGF-β, Arg-1, and VEGF, upregulating PD-L1 expression [ 10 ], promoting T-cell exhaustion, and expanding regulatory T-cells (Treg) populations [ 11 ]. Enrichment of M2 macrophages in glioma tissues has been correlated with reduced IFN-γ expression, indicative of impaired T-cell activity in vivo [ 12 ]. Furthermore, M2 macrophages suppress NK cells cytotoxicity through TGF-β–mediated downregulation of the activating receptor NKG2D [ 13 , 14 ], a key mechanism of tumor immune escape [ 15 ]. Notably, blocking TGF-β signaling in preclinical glioma models restores NKG2D expression and enhances NK cell-mediated tumor killing [ 16 ]. These findings collectively highlight the central role of M2 macrophages as orchestrators of GBM immunosuppression and tumor progression, positioning them as compelling targets for therapeutic intervention. Preclinical studies using CSF1R inhibitors, for example, have demonstrated the potential to reprogram M2 macrophages toward a pro-inflammatory M1 phenotype, thereby enhancing antitumor immunity [ 16 ]. In this study, we combined integrative multi-omics analyses with functional experiments to investigate the molecular drivers of M2 macrophage polarization in GBM. We identified DOK3 as a key regulator of pro-tumoral polarization and glioma cell migration, providing mechanistic insights into macrophage-mediated tumor progression and highlighting DOK3 as a promising immunotherapeutic target for patients with IDH-wildtype GBM. Methods and materials Data Acquisition Multiple glioblastoma (GBM) datasets were obtained from publicly available, authoritative databases. RNA sequencing data and corresponding clinical information from 205 GBM patients were retrieved from The Cancer Genome Atlas (TCGA, http://cancergenome.nih.gov/ ) [ 17 ] and used as the training cohort for feature selection and model construction. An additional 74 GBM samples from the Chinese Glioma Genome Atlas (CGGA, http://cgga.org.cn/ ) [ 18 , 19 ] served as an independent external validation cohort to evaluate the robustness and generalizability of the model. Clinical variables included, but were not limited to, age, sex, IDH mutation status, MGMT promoter methylation status, chemoradiotherapy, and survival outcomes such as overall survival (OS) and progression-free survival (PFS). To investigate the spatial distribution of M2 macrophages in different histological regions, we analyzed data from the Ivy Glioblastoma Atlas Project ( http://glioblastoma.alleninstitute.org/ )[ 20 ]. This dataset comprises gene expression profiles from five anatomically defined regions: (1) cellular tumor (CT), (2) infiltrating tumor (IT), (3) leading edge (LE), (4) microvascular proliferation (MP), and (5) pseudopalisading cells (PC). To further exploring the immune microenvironment at single-cell resolution, we incorporated single-cell RNA sequencing data (scRNA-seq) data from both the CGGA database and the GSE131928 dataset in the Gene Expression Omnibus (GEO) ( https://www.ncbi.nlm.nih.gov/geo/ ) [ 21 ]. These datasets include multiple primary and recurrent GBM samples, enabling high resolution characterization of immune cell heterogeneity and dynamics within the tumor microenvironment. All datasets were accessed in accordance with their respective usage policies and ethical guidelines. The study was conducted in compliance with the principles of the Declaration of Helsinki. Estimation of Immune Infiltrates Immune cell infiltration was quantified using the TIMER2.0 platform ( http://timer.cistrome.org/ ) [ 22 ] and xCell algorithm [ 23 ] based on bulk RNA-seq data from both the TCGA, CGGA, and IVY cohorts. TIMER2.0 integrates multiple state-of-the-art deconvolution algorithms, including TIMER, CIBERSORT, quanTIseq, and MCP-counter, enabling robust and comprehensive estimates of immune cell abundance across tumor samples. These estimates serves as the basis for subsequent immunological analyses and correlation studies. To identify genes associated with M2 macrophage activity, Pearson correlation analysis was performed between gene expression levels and the M2 macrophages score in each dataset, with genes showing strong positive correlation (r > 0.5) in both cohorts retained for downstream analyses. Unsupervised Clustering and Subtype Identification Genes significantly correlated with M2 macrophage infiltration ( R > 0.5 and p < 0.05) in both the TCGA and CGGA datasets were subjected to unsupervised clustering using the ConsensusClusterPlus R package. To ensure the stability and predictive utility of the identified subtypes, a random forest classifier was trained using the TCGA cohorts and subsequently applied to the CGGA cohort for external validation. The robustness and separability of the resulting clusters were further assessed using principal component analysis (PCA). Risk Model Construction To identify key genes predictive of the immunosuppressive subtype, we applied Extreme Gradient Boosting (XGBoost) and least absolute shrinkage and selection operator (LASSO) regression to the TCGA cohort. Three candidate genes–CTSB, LITAF, and DOK3–were selected to construct a macrophage-associated risk score model. The prognostic value of this model was evaluated using Kaplan-Meier survival analysis and univariate and multivariate Cox regression. In addition, correlations between the risk score and CD163 expression, immune checkpoint markers, and immune infiltration patterns were systematically analyzed to elucidate the immunological and clinical relevance of the model. Single-cell Transcriptomic Analysis Single-cell RNA sequencing (scRNA-seq) data were processed using the Seurat R package (version 5.0.0) [ 24 ]. Low-quality cells and lowly expressed genes were filtered out based on standard quality control metrics, followed by normalization and batch-effect correlation. Dimensionality reduction and clustering were performed using the Uniform Manifold Approximation and Projection (UMAP) algorithm. The expression patterns of key marker genes (CD68, TMEM119, CD163, and DOK3) was analyzed across different immune and non-immune cell clusters to characterize their cellular distribution and functional relevance. THP-1 Cell Culture and Polarization Induction THP-1 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FPS), 1% penicillin-streptomycin, and 0.1 mM β-mercaptoethanol at 37°C in a humidified atmosphere containing 5% CO₂. To induce differentiation into M0 macrophages, cells were treated with 185 ng/mL phorbol 12-myristate 13-acetate (PMA, dissolved in DMSO) for 12 hours, resulting in an adherent phenotype [ 25 ]. For M2 polarization, adherent cells were co-incubated with 20 ng/mL IL-4 and 20 ng/mL IL-13 in the continued presence of PMA for 48 hours [ 26 ]. For gene silencing experiments, siRNA negative control (siRNA-NC) or siRNA targeting DOK3 (siRNA-DOK3) was transfected into the respective groups, followed by 48 hours of incubation. Western-blot Assay After polarization to the M2 phenotype, THP-1 cells were transfected with siRNA targeting DOK3. Total protein was extracted using RIPA buffer, and protein concentrations were measured by BCA assay. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% skim milk for 1 hour at room temperature, followed by overnight incubation at 4°C with primary antibodies: DOK3 (Immunoway, YT1397, 1:1000), CD163 (Abcam, ab156769, 1:1000), and GAPDH (Immunoway, YM3029, 1:20000). After washing, membranes were incubated with HRP-conjugated secondary antibodies for 1 hour at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) and imaged with a Bio-Rad imaging system. Immunofluorescence Cells were fixed with 4% paraformaldehyde, permeabilized with Triton X-100, and blocked with goat serum. Cells were incubated overnight with primary antibodies against DOK3 and CD163, followed by incubation with Alexa Fluor-conjugated secondary antibodies. Nuclei were counterstained with DAPI, and fluorescence imaging was performed using a confocal microscope. Transwell Migration Assay Transwell assays were performed using 8-µm pore size polycarbonate membrane inserts. LN229 glioma cells (4 × 10 4 cells in 200 µL serum-free medium) were seeded into the upper chamber, while M2-polarized macrophages (1 × 10 5 cells in 200 µL medium containing 5% FBS) were placed in the lower chamber. After 12 hours of incubation, non-invading cells on the upper surface of the membrane were carefully removed and invading cells on the lower surface were fixed and stained with crystal violet. Statistical Analysis All statistical analyses were conducted using R software (version 4.2.3) and IBM SPSS Statistics (version 27.0.1). R packages included ConsensusClusterPlus, simplifyEnrichment [ 27 ], ggplot2, pheatmap, GSVA, ggpubr, survival, survminer, dplyr, Seurat, patchwork, and pROC. Pearson correlation, Student’s t-test, and one-way ANOVA were applied to assess group differences. Cox regression analysis was performed for survival analysis. P < 0.05 were considered statistically significant. Results M2 Macrophages Are Spatially Enriched and Associated with Poor Prognosis in GBM To clarify the clinical and biological significance of M2 macrophage in glioblastoma (GBM), we quantified their infiltration using xCell-based deconvolution in two independent transcriptomic cohorts: TCGA ( n = 205) and CGGA ( n = 74). In both cohorts, higher M2 macrophage scores were strongly correlated with reduced tumor purity ( R = -0.72, p < 0.0001; CGGA: R = -0.73, p < 0.0001; Fig. 1 A, Fig. S1A ), indicating preferential accumulation of these cells within the non-neoplastic tumor microenvironment (TME) at the expense of tumor cell proportion. To explore their spatial context, we analyzed anatomically annotated transcriptional profiles from the Ivy Glioblastoma Atlas Project (Ivy GAP, n = 270). M2 macrophage scores were markedly elevated in microvascular proliferation (MVP) and pseudopalisading cells around necrosis (PAN), two histological niches strongly linked to angiogenesis, hypoxia and aggressive tumor behavior (Fig. 1 B). Consistent with these findings, these regions were also exhibited in anti-inflammatory and pro-angiogenic pathways (Fig. 1 C, Fig. S1B ), suggesting that M2 macrophages may be recruited to these specialized niches to promote vascular remodeling and immune evasion. Functionally, gene set enrichment analysis (GSEA) further revealed that M2 macrophage infiltration positively correlated with pathways involved in complement activation (e.g., C1QA , C1QB , C3AR1 ) and immune modulation, including inflammatory response , interferon gamma response , and TNFα signaling via NF-κB (Fig. 1 D, Fig. S1C-E ). These results indicate that beyond supporting angiogenesis, M2 macrophages actively shape an immunosuppressive microenvironment that suppresses anti-tumor immunity and facilitating tumor immune evasion. Survival analyses underscored the clinical importance of these observations: across both cohorts, patients with high M2 macrophage infiltration exhibited significantly shorter overall survival (OS) and progression-free survival (PFS) compared with those with lower infiltration (Fig. 1 E, Fig. S1F ). Collectively, these findings demonstrate that M2 macrophages are spatially enriched in aggressive tumor niches, where they drive an immunosuppressive and pro-tumoral microenvironment, ultimately contributing to poor clinical outcomes in patients with GBM. M2 Macrophage–Related Gene Signatures Define Three Prognostically Relevant Immune Subtypes in GBM To characterize the immune heterogeneity of GBM, we first performed Pearson correlation analysis to identify genes strongly correlated with the M2 macrophages infiltration ( R > 0.5 and p < 0.05) in both TCGA and CGGA cohorts. This analysis yielded 299 overlapping M2 macrophage-related genes (Fig. 2 A). Using these genes, unsupervised consensus clustering of the TCGA cohort stratified patients into three robust immune subtypes (C1, C2, and C3) (Fig. 2 B). A random forest classifier trained on the TCGA cohort was subsequently applied to the CGGA cohort, confirming the reproducibility of the three-subtype classification (TCGA: C1, n = 90; C2, n = 55; C3, n = 60; CGGA: C1, n = 32; C2, n = 29; C3, n = 13) (Fig. 2 C). Principal component analysis (PCA) demonstrated clear separation among the three immune subtypes in both datasets (Fig. 2 D-E), supporting the robustness of the clustering. Survival analysis revealed that patients in the C1 subtype exhibited significantly short overall survival compared with those in the C2 and C3 subtypes (TCGA: p = 0.042; CGGA: p = 0.021) (Fig. 2 F-G), indicating that the C1 immune subtype is associated with the poorest prognosis. Further molecular profiling revealed that mesenchymal transcriptional subtype was enriched within the C1 immune subtype (Fig. 2 H-I), suggesting that the high M2 macrophage activity is linked to mesenchymal transition and aggressive tumor biology. Collectively, these findings demonstrate that M2 macrophage-related gene signature define three biologically and clinically distinct immune subtypes of GBM, providing a framework for understanding immune heterogeneity and informing the development of subtype-specific therapeutic strategies. C1 Subtype Exhibits an Immune-Infiltrated but Immunosuppressive Microenvironment To better characterize the tumor microenvironment (TME) across M2 macrophage-base subtype, we applied the ESTIMATE algorithm to calculate immune score, stromal score, and tumor purity in both the TCGA and CGGA cohorts. In the TCGA cohort, the C1 subtype–which showed the poorest prognosis–exhibited the highest immune and stromal scores and markedly reduced tumor purity. Conversely, the C2 subtype, linked to the most favorable outcomes, displayed the lowest immune and stromal scores and the highest tumor purity. These patterns were consistently reproduced in the CGGA cohort (Fig. 3 A-C), confirming the robustness of subtype-specific microenvironmental features. Importantly, the elevated immune and stromal scores in C1 did not indicate effective antitumor immunity. Instead, this subtype reflected a profoundly immunosuppressive TME dominated by M2-like tumor-associated macrophages (TAMs). Quantitative analysis revealed that the C1 subtype harbored the highest levels of M2 macrophage infiltration and elevated expression of canonical M2 markers, including CD163 and IL10 (Fig. 3 D-F). Immune cell composition profiling further highlighted the tumor-supportive and dysfunctional nature of the C1 immune landscape. This subtype was enriched with monocytes, neutrophiles, resting NK cells, and a population of CD8 + T cells with an exhausted phenotype (Fig. 3 G-H). In contrast, effector cell populations with recognized antitumor activity–such as Th1-polarized CD4 + T cells and activated NK cells–were depleted in C1 and relatively enriched in the C2 subtype. Collectively, these findings indicate that the C1 subtype represents an immune-infiltrated but immunosuppressive TME, where high immune cell content is driven by dysfunctional or tumor-supportive subsets rather than effective antitumor effectors. This immunological architecture, likely orchestrated by M2-polarized macrophages, may underlie the aggressive clinical behavior of C1 tumors and highlight the need for therapeutic strategies targeting immunosuppressive myeloid populations. Enrichment Analysis Reveals Immune Response-Associated but Potentially Immunosuppressive Signatures in the C1 Subtype To explore the molecular underpinnings of the distinct tumor microenvironment observed in the C1 subtype, we performed functional enrichment analysis on genes significantly upregulated in C1 compared to C2 and C3 (logFC > 1, p < 0.05). This analysis revealed significant enrichment of immune-related biological processes, such as inflammatory response and immune response (Fig. 4 A-B), indicating that the C1 subtype is characterized by a strong immune response–associated signature. Consistently, Gene Set Enrichment Analysis (GSEA) in both the TCGA and CCGA cohorts demonstrated that C1-upregulated genes were significantly enriched in hallmark pathways, including allograft rejection , IL6-JAK-STAT3 signaling , inflammatory response , and TNFα signaling via NF-κB (Fig. 4 C-D). These pathways are well known to drive chronic inflammation, cytokine signaling, and immune cell recruitment–processes that often foster immune dysregulation rather that effective antitumor immunity. Taken together, these findings suggest that although the C1 subtype exhibits a transcriptomic signature heightened immune responses, this activity primarily reflects an inflammation-driven, immunosuppressive microenvironment. This is consistent with the high abundance of M2 macrophages and dysfunctional effector immune subsets identified in our cellular infiltration analyses, underscoring the immune-evasive nature of the C1 subtype. XGBoost-LASSO–Derived Macrophage Signature Identified a High-Risk, Immunosuppressive C1 Subtype in Glioma To pinpoint genes most closely associated with the C1 subtype and macrophage-driven immunosuppression, we leveraged the 299 macrophage-related genes previously identified as strongly correlated with M2 macrophage infiltration. Using these candidate features, we constructed a XGBoost-based classifier in both the TCGA and CGGA cohorts. The models identified 200 genes with non-zero gain values in TCGA and 143 genes in CGGA (Fig. 5 A). In the XGBoost framework, gain value represents the average performance improvement contributed by a feature when splitting decision nodes; thus, genes with non-zero gain values were considered important contributors to identifying the C1 subtype classification. Notably, 102 overlapping genes were consistently identified across both datasets (Fig. 5 B), underscoring their stable predictive utility in independent cohorts. To further refine these candidates, we performed LASSO regression analysis to the 102 overlapping genes (Fig. 5 C), which yielded a concise three-gene macrophage signature composed of CTSB , LITAF , and DOK3 , with regression coefficients of 0.18895, 0.04699, and 0.00799, respectively (Fig. 5 D). Based on these coefficients, we calculated a composite macrophage-related risk model for each patient and stratified individuals within the C1 subtype in both cohorts into high- and low-risk groups according to the median score. Patients in the high-risk group exhibited significantly higher mortality than those in the low-risk group ( Fig. 5 E ). Kaplan-Meier survival analyses confirmed that higher risk scores were significantly shorter overall survival (OS) and progression-free survival (PFS) (Fig. 5 F-G ) . The risk score was markedly elevated in the C1 subtype compared with C2 and C3 ( Fig. 5 H ) and was also increased in the mesenchymal transcriptional subtype ( Fig. 5 I ) , indicating a strong association with more aggressive tumor phenotypes. Correlation analyses further revealed that the macrophage-based risk score was positively associated with M2 macrophage infiltration ( R = 0.72, p < 0.0001) ( Fig. 5 J ) and strongly correlated with the canonical M2 marker CD163 ( R = 0.75, p < 0.0001;) ( Fig. 5 K ) . These findings indicate that the risk score captures not only macrophage abundance but also the immunosuppressive activity of the tumor microenvironment. Importantly, these results were consistently validated in the CGGA cohort ( Fig. S2A-G ), demonstrating the robustness and reproducibility of this macrophage-associated risk model across independent datasets. High Risk Score Reflects an Immune-Infiltrated but Immunosuppressive Tumor Microenvironment To elucidate the biological significance of the macrophage-based risk score within the TME, we analyzed immune cell composition using the xCell algorithm in both the TCGA and CGGA cohorts. The high-risk group exhibited significantly elevated infiltration of B cells, cancer-associated fibroblasts, M2 macrophages, monocytes, and plasmacytoid dendritic cells (Fig. 6 A, Fig. S3A )–cell populations widely implicated in immune suppression and tumor progression [ 28 , 29 ]. In contrast, the low-risk group exhibited higher levels of CD4 + Th1 T cells, a subset known for potent antitumor activity [ 30 ], suggesting that these cells may contribute to more active immune response in this subgroup. We next assessed the overall immune and stromal contexture using the ESTIMATE algorithm. Both stromal and immune score were positively correlated with the risk score, whereas tumor purity showed a negative correlation (Fig. 6 B, Fig. S3B ). These findings indicate that tumors with higher risk scores harbor greater infiltration of non-tumor components, consistent with a highly infiltrated yet immunologically suppressive TME. To further explore potential mechanisms of immune suppression, we examined the relationship between the risk score and the expression of classical immune checkpoint molecules. Expression levels of TIMD3 , CSF1R , and PD-1 were significantly elevated in the high-risk group (Fig. 6 C, Fig. S3C ), suggesting enhanced activation of immunosuppressive signaling pathways in these patients. This pattern aligns with the enrichment of M2 macrophages and other suppressive immune subsets, reinforcing the link between the high-risk signature and a dysfunctional, immunosuppressive microenvironment. DOK3 as a Key Regulator of M2 Macrophage Polarization We next examined the three key genes– CTSB , LITAF , and DOK3 –that constitute the macrophage-based risk model. In the TCGA dataset, high expression of all three genes was associated with worse prognosis (Fig. 7 A). In the CGGA dataset, elevated expression of CTSB and DOK3 similarly correlated with poor survival, whereas LITAF expression showed no significant prognostic association (Fig. 7 B). Given the significant positive correlation between the risk score, M2 macrophage infiltration, and CD163 expression (Fig. 5 J-K), we further investigated the cellular localization of these genes using single-cell RNA sequencing (scRNA-seq) data from the GSE131928 and CGGA datasets. Clustering of tumor and microenvironmental cell populations–annotated canonical markers including CD68 (macrophages), TMEM119 (microglia), and CD163 (M2 macrophages)–revealed that DOK3 expression was predominantly restricted in microglia and macrophages (Fig. 7 C-D). In contrast, CTSB and LITAF exhibited broad, non-specific expression across multiple cell types, including malignant cells, stromal cells, lymphocytes, and oligodendrocytes ( Fig. S4A-D ) suggesting that their functional relevance may not be limited to macrophage functions. These findings highlight DOK3 as the most cell-type-specific gene with the risk signature and suggest that it may act as a regulator of M2 macrophage polarization. This polarization likely facilitates the establishment of an immunosuppressive tumor microenvironment, promoting immune evasion and supporting the malignant progression of gliomas. DOK3 Knockdown Inhibits M2 Macrophage Polarization and Suppressed Glioma Cell Migration To experimentally validate the role of DOK3 in macrophage polarization and tumor progression, we performed cell-based functional assays using THP-1 cells, a well-established model for human monocyte-macrophage biology. Differentiation into M0 macrophages was first induced with PMA treatment, followed by IL-4 and IL-13 stimulation for 48 hours to drive M2 polarization, which was confirmed by the morphological transition from round to spindle-shaped cells (Fig. 8 A). Gene silencing of DOK3 in M2-polarized macrophages led to a marked reduction in DOK3 expression and a significant decrease in the M2 marker CD163 , as confirmed by Western blot analysis (Fig. 8 B). Immunofluorescence staining corroborated these findings, showing diminished expression of both DOK3 and CD163 in the knockdown group compared with controls (Fig. 8 C-E). These results indicate that DOK3 is essential for the maintenance of the M2-polarized phenotype. To assess the functional relevance of DOK3 -mediated polarization in tumor biology, we co-cultured LN229 glioma cells with conditioned media derived from DOK3 -silenced M2 macrophages. Compared to controls, glioma cells invasiveness was significantly attenuated in the knockdown group (Fig. 8 F-G), demonstrating that DOK3 -mediated M2 polarization facilitates a tumor-promoting microenvironment that enhances glioma cell migration. Collectively, these findings provide direct functional evidence that DOK3 promotes M2 macrophage polarization and contributes to glioma aggressiveness. Targeting DOK3 in macrophages may therefore represent a promising therapeutic strategy to reprogram the tumor immune microenvironment and restain glioma progression. Discussion Glioblastoma (GBM) remains one of the most lethal malignancies of the central nervous system, characterized by extreme heterogeneity, aggressive invasiveness, and resistance to standard therapies [ 31 ]. Increasing evidence underscores the pivotal role of the TME in driving tumor progression and therapeutic resistance. Among the immune components, TAMs–particularly those polarized toward the M2 phenotype–are recognized as key facilitators of immune evasion, tumor growth, and treatment failure [ 32 – 34 ]. In this study, we systematically dissected the immunological and molecular landscape of M2 macrophages in GBM and identified DOK3 as a potential regulator of M2 polarization and glioma progression. Using xCell-based immune deconvolution across TCGA and CGGA cohorts, we demonstrated that elevated M2 macrophage infiltration is associated with poor prognosis and reduced tumor purity, consistent with their recognized role in shaping an immunosuppressive microenvironment. Unsupervised clustering of 299 M2-related genes identified three immune subtypes (C1-C3), among which the C1 subtype exhibited the poorest prognosis and was enriched for the mesenchymal molecular subtype. This C1 subtype displayed a higher infiltrated but profoundly immunosuppressive microenvironment, characterized by enrichment of M2 macrophages, monocytes, and resting NK cells, along with depletion of cytotoxic immune populations such as Th1-polarized CD4 + T cells and activated NK cell. Although CD8 + T cells were numerically abundant in C1 tumors, their functional exhaustion likely rendered them ineffective, reflecting an “immune-high but functionally suppressed” state that has been increasingly recognized as a hallmark of immune dysfunction in GBM [ 30 , 35 – 37 ]. This dysfunctional immune architecture may drive immune escape, malignant progression, and resistance to therapy. Through integrative machine learning combining XGBoost and LASSO modeling, we developed a macrophage-related risk score that stratifies patients by prognosis and immune contexture. Among the three key genes identified ( CTSB , LITAF , and DOK3 ), DOK3 emerged as the most cell-type–specific marker, with strong enrichment in microglia and macrophages and a robust correlation with the M2 marker CD163 . Single-cell transcriptomic analysis confirmed its selective expression in myeloid-derived populations, supporting a macrophage-specific role. Functional experiments provided direct evidence for this: DOK3 knockdown in THP-1–derived macrophages inhibited M2 polarization, reduced CD163 expression, and suppressed the pro-invasive phenotype of glioma cells exposed to conditioned medium from M2 macrophages. Collectively, these findings demonstrate that DOK3 promotes M2 polarization and contribute to the establishment of an immunosuppressive, tumor-promoting microenvironment in GBM. These findings carry important translational implications. DOK3 may serve not only as a prognostic biomarker but also a potential therapeutic target to reprogram the TME in GBM. Therapeutic strategies aimed at inhibiting DOK3 or modulating macrophage polarization could restore antitumor immune activity and potentially enhance the efficacy of existing immunotherapeutic approaches, including checkpoint inhibitors and myeloid-targeted therapies, in patients with GBM. Despite the strengths of this study, including integrative multi-omics analyses and functional validation, several limitations should be acknowledged. The absence of in vivo validation limits our ability to fully assess the therapeutic potential and mechanistic pathways of DOK3 in the complex TME. Moreover, while our data support a critical role of DOK3 in macrophage polarization, the downstream signaling pathways and regulatory networks remain to be elucidated. Finally, the profound heterogeneity of GBM underscores the need for larger, multi-center studies and the integration of spatial single-cell multi-omics to better characterize immune interactions and uncover context-specific therapeutic vulnerabilities. Conclusions This study comprehensively clarifies the characteristics of the immunosuppressive C1 subtype driven by M2 macrophages in glioblastoma multiforme (GBM), and confirms that DOK3 can promote the polarization of macrophages toward the M2 pro-tumor phenotype, thereby participating in the construction of the immunosuppressive microenvironment of GBM and driving the malignant progression of gliomas (Fig. 9 ). Additionally, the risk model constructed based on macrophages provides a reliable basis for prognostic stratification and clinical therapeutic decision-making in GBM patients. These findings fully highlight the potential of DOK3 and macrophage reprogramming as therapeutic targets, which are expected to remodel the tumor immune microenvironment in GBM patients and further improve the clinical efficacy of immunotherapeutic strategies. Declarations Ethics approval and consent to participate All datasets were accessed in accordance with their respective usage policies and ethical guidelines (IRB ID: KY2024-171-02). The study was conducted in compliance with the principles of the Declaration of Helsinki. Consent for publication Not applicable Availability of data and material All data are included in this published article and Additional files. Competing interest The authors declare that they have no competing financial interests. Funding This work was supported by the Noncommunicable Chronic Diseases–National Science and Technology Major Project (2024ZD0525300), the National Natural Science Foundation of China (NSFC) fund (No. 82192894 and 82472841), the Beijing Hospitals Authority Youth Programme (QML20230507), the Natural Science Foundation of Beijing in China (No. 7254342), the Beijing Municipal Science & Technology Commission and Administrative Commission of Zhongguancun Science Park (Z241100009024044), and the Beijing Municipal Health Commission Fund (11000023T000002044300‐5). Authors’ contributions C.R. performed data collection, processing and analysis, conducted experimental validation, contributed to the biological interpretation of the results, and drafted the manuscript. J.S. conducted experimental validation. C.Y. performed data collection, processing, and analysis. J.Z. performed data collection, processing, and analysis. W.T. performed data collection, processing, and analysis. H.Z. contributed to the biological interpretation of the results. P.R. contributed to the biological interpretation of the results. J.C. contributed to the biological interpretation of the results. W.F. conceived and designed the study and contributed to the biological interpretation of the results. Y.Z. conceived and designed the study. Z.Z. conceived and designed the study and drafted the manuscript. All authors critically reviewed, revised, and approved the final manuscript. Acknowledgements Not applicable References Koshy, M., et al., Improved survival time trends for glioblastoma using the SEER 17 population-based registries. J Neurooncol, 2012. 107 (1): p. 207-12. Ostrom, Q.T., et al., CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2015-2019. Neuro Oncol, 2022. 24 (Suppl 5): p. v1-v95. Ostrom, Q.T., et al., National-level overall survival patterns for molecularly-defined diffuse glioma types in the United States. Neuro Oncol, 2023. 25 (4): p. 799-807. Gangoso, E., et al., Glioblastomas acquire myeloid-affiliated transcriptional programs via epigenetic immunoediting to elicit immune evasion. Cell, 2021. 184 (9): p. 2454-2470.e26. Mei, Y., et al., Siglec-9 acts as an immune-checkpoint molecule on macrophages in glioblastoma, restricting T-cell priming and immunotherapy response. Nat Cancer, 2023. 4 (9): p. 1273-1291. Zhao, R., et al., Blocking ITGA5 potentiates the efficacy of anti-PD-1 therapy on glioblastoma by remodeling tumor-associated macrophages. Cancer Commun (Lond), 2025. 45 (6): p. 677-701. Wu, L., et al., Natural Coevolution of Tumor and Immunoenvironment in Glioblastoma. Cancer Discov, 2022. 12 (12): p. 2820-2837. Ge, W. and W. Wu, [Influencing Factors and Significance of Tumor-associated Macrophage Polarization in Tumor Microenvironment]. Zhongguo Fei Ai Za Zhi, 2023. 26 (3): p. 228-237. Cao, C., et al., Intra-Operative Definition of Glioma Infiltrative Margins by Visualizing Immunosuppressive Tumor-Associated Macrophages. Adv Sci (Weinh), 2023. 10 (28): p. e2304020. Kureshi, C.T. and S.K. Dougan, Cytokines in cancer. Cancer Cell, 2025. 43 (1): p. 15-35. Zheng, C., et al., IFNα-induced BST2(+) tumor-associated macrophages facilitate immunosuppression and tumor growth in pancreatic cancer by ERK-CXCL7 signaling. Cell Rep, 2024. 43 (4): p. 114088. Li, J., et al., Tumor-Associated Macrophage-Derived Exosomal LINC01232 Induces the Immune Escape in Glioma by Decreasing Surface MHC-I Expression. Adv Sci (Weinh), 2023. 10 (17): p. e2207067. Krneta, T., et al., M2-polarized and tumor-associated macrophages alter NK cell phenotype and function in a contact-dependent manner. J Leukoc Biol, 2017. 101 (1): p. 285-295. Ghiringhelli, F., et al., CD4+CD25+ regulatory T cells inhibit natural killer cell functions in a transforming growth factor-beta-dependent manner. J Exp Med, 2005. 202 (8): p. 1075-85. Groh, V., et al., Tumour-derived soluble MIC ligands impair expression of NKG2D and T-cell activation. Nature, 2002. 419 (6908): p. 734-8. Batlle, E. and J. Massagué, Transforming Growth Factor-β Signaling in Immunity and Cancer. Immunity, 2019. 50 (4): p. 924-940. Pyonteck, S.M., et al., CSF-1R inhibition alters macrophage polarization and blocks glioma progression. Nat Med, 2013. 19 (10): p. 1264-72. Zhao, Z., et al., Chinese Glioma Genome Atlas (CGGA): A Comprehensive Resource with Functional Genomic Data from Chinese Glioma Patients. Genomics Proteomics Bioinformatics, 2021. 19 (1): p. 1-12. Zhao, Z., et al., Comprehensive RNA-seq transcriptomic profiling in the malignant progression of gliomas. Sci Data, 2017. 4 : p. 170024. Puchalski, R.B., et al., An anatomic transcriptional atlas of human glioblastoma. Science, 2018. 360 (6389): p. 660-663. Edgar, R., M. Domrachev, and A.E. Lash, Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res, 2002. 30 (1): p. 207-10. Li, T., et al., TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res, 2020. 48 (W1): p. W509-w514. Aran, D., Z. Hu, and A.J. Butte, xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol, 2017. 18 (1): p. 220. Butler, A., et al., Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol, 2018. 36 (5): p. 411-420. Daigneault, M., et al., The identification of markers of macrophage differentiation in PMA-stimulated THP-1 cells and monocyte-derived macrophages. PLoS One, 2010. 5 (1): p. e8668. Martinez, F.O., et al., Transcriptional profiling of the human monocyte-to-macrophage differentiation and polarization: new molecules and patterns of gene expression. J Immunol, 2006. 177 (10): p. 7303-11. Gu, Z. and D. Hubschmann, simplifyEnrichment: A Bioconductor Package for Clustering and Visualizing Functional Enrichment Results. Genomics Proteomics Bioinformatics, 2023. 21 (1): p. 190-202. Shalapour, S., et al., Immunosuppressive plasma cells impede T-cell-dependent immunogenic chemotherapy. Nature, 2015. 521 (7550): p. 94-8. Kalluri, R., The biology and function of fibroblasts in cancer. Nat Rev Cancer, 2016. 16 (9): p. 582-98. Kennedy, R. and E. Celis, Multiple roles for CD4+ T cells in anti-tumor immune responses. Immunol Rev, 2008. 222 : p. 129-44. Lim, M., et al., Current state of immunotherapy for glioblastoma. Nat Rev Clin Oncol, 2018. 15 (7): p. 422-442. Quail, D.F. and J.A. Joyce, The Microenvironmental Landscape of Brain Tumors. Cancer Cell, 2017. 31 (3): p. 326-341. Hambardzumyan, D., D.H. Gutmann, and H. Kettenmann, The role of microglia and macrophages in glioma maintenance and progression. Nat Neurosci, 2016. 19 (1): p. 20-7. Chen, Y., et al., Tumor-associated monocytes promote mesenchymal transformation through EGFR signaling in glioma. Cell Rep Med, 2023. 4 (9): p. 101177. Vivier, E., et al., Innate or adaptive immunity? The example of natural killer cells. Science, 2011. 331 (6013): p. 44-9. Wherry, E.J., T cell exhaustion. Nat Immunol, 2011. 12 (6): p. 492-9. Woroniecka, K., et al., T-Cell Exhaustion Signatures Vary with Tumor Type and Are Severe in Glioblastoma. Clin Cancer Res, 2018. 24 (17): p. 4175-4186. Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigureslegends.docx SF.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7534936","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":511019928,"identity":"b18ef636-2b72-450b-896d-d850ec7139dd","order_by":0,"name":"Chang-Yuan Ren","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Chang-Yuan","middleName":"","lastName":"Ren","suffix":""},{"id":511019929,"identity":"0a5d609b-9d16-4e2c-93f8-ba45989f4903","order_by":1,"name":"Ji Shi","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ji","middleName":"","lastName":"Shi","suffix":""},{"id":511019930,"identity":"47af1ae9-d8ba-4444-a0aa-8018c25ff06b","order_by":2,"name":"Chang-Lin Yang","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Chang-Lin","middleName":"","lastName":"Yang","suffix":""},{"id":511019931,"identity":"4c0c0ee8-8d35-4a01-9cf8-e5400ec5a238","order_by":3,"name":"Jin-Hao Zhang","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jin-Hao","middleName":"","lastName":"Zhang","suffix":""},{"id":511019932,"identity":"228fa359-2aae-476d-8c47-32f02eacdd65","order_by":4,"name":"Wen-Lu Tan","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Wen-Lu","middleName":"","lastName":"Tan","suffix":""},{"id":511019933,"identity":"56898cb2-f7f8-48f9-80e2-ac7e37e7db02","order_by":5,"name":"Han-Xiao Zhou","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Han-Xiao","middleName":"","lastName":"Zhou","suffix":""},{"id":511019934,"identity":"533173db-9da6-42cc-a114-c0d5d8fa4efb","order_by":6,"name":"Pan-Yu Ren","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Pan-Yu","middleName":"","lastName":"Ren","suffix":""},{"id":511019935,"identity":"91a6aebb-1326-470a-b2e0-72251e9a496e","order_by":7,"name":"Josip Cvitković","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Josip","middleName":"","lastName":"Cvitković","suffix":""},{"id":511019936,"identity":"477df866-e1fb-40e9-b042-91d20995566d","order_by":8,"name":"Wen-Hua Fan","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Wen-Hua","middleName":"","lastName":"Fan","suffix":""},{"id":511019937,"identity":"d4991140-c1b5-4351-8200-9ca97b52b5f2","order_by":9,"name":"Ying Zhang","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ying","middleName":"","lastName":"Zhang","suffix":""},{"id":511019938,"identity":"9646458f-9db5-44a4-8077-101b2648da7a","order_by":10,"name":"Zheng Zhao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABDUlEQVRIiWNgGAWjYDACCSBmbAASzPwPDn6ogIryEKWFvYfxscQZkrTwnGE24G0jQov87OaHD3/usMmTj8g9JiE5z05ed0YC44O3bQzy5ji0MM45ZmwgeSat2PBGXppE4bZkw203EpgN57YxGO5swK6FWSLBTMKw7XDixhlAhuS2AwlmNxLYpIEuTDA4gF0Lm0T6N4nEtv8QLbxzwFrYf+PTwiORYyZxsO1A4nyeM8YGvA0QW5jxaZGQyCk2bGxLTtzA3pb4WOIY0C9nHjZLzjknYbgBhxb5GekbH/5ss0uc38x84OCHGjt5s+PJBz+8KbORx2ULHCApAEeTBAH1IOsaCKsZBaNgFIyCEQoAgBJfOUR44toAAAAASUVORK5CYII=","orcid":"","institution":"Capital Medical University","correspondingAuthor":true,"prefix":"","firstName":"Zheng","middleName":"","lastName":"Zhao","suffix":""}],"badges":[],"createdAt":"2025-09-04 10:08:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7534936/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7534936/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90895005,"identity":"1bf13630-40ba-481f-9df8-1d29c71b4340","added_by":"auto","created_at":"2025-09-09 11:31:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":605535,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHigh M2 macrophage infiltration is associated with reduced tumor purity, spatial regions, complement, and poor prognosis in GBM.\u003c/strong\u003e \u003cstrong\u003eA,\u003c/strong\u003eScatter plot showing negative correlations between M2 macrophage scores (xCell) and tumor purity estimated by ESTIMATE in the TCGA cohorts. \u003cstrong\u003eB,\u003c/strong\u003e M2 macrophage scores in different GBM histological regions, including Cellular Tumor (CT), Infiltrating Tumor (IT), Leading Edge (LE), MicroVascular Proliferation (MVP), and Pseu- dopalisading cells Around Necrosis (PAN). \u003cstrong\u003eC,\u003c/strong\u003eCorrelation analysis showing that M2 macrophage scores (xCell) are positively associated with the complement hallmark pathway estimated by ssGSEA in the TCGA cohort. \u003cstrong\u003eD, \u003c/strong\u003eCorrelation analysis showing that M2 macrophage scores (xCell) are positively associated with the expression of key complement hallmark pathway genes, including C1QA, C1QB, and C3AR1. \u003cstrong\u003eE,\u003c/strong\u003e Kaplan–Meier survival curves comparing overall survival (left) and progression-free survival (right) between high and low M2 macrophage score groups in TCGA cohorts. Survival differences assessed by log-rank test (p values indicated).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/08b95997e5dd35d23a2bdc6e.png"},{"id":90895006,"identity":"983ee6a3-efb7-4901-a0c2-ac1650b59061","added_by":"auto","created_at":"2025-09-09 11:31:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1198532,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification and characterization of immune subtypes based on M2 macrophage–associated genes in GBM. A,\u003c/strong\u003e Venn diagram showing 299 genes that were highly positively correlated with Macrophages_M2 scores (\u003cem\u003eR\u003c/em\u003e \u0026gt; 0.5 \u0026amp; \u003cem\u003ep\u003c/em\u003e\u0026lt; 0.05) in both TCGA and CGGA cohorts. \u003cstrong\u003eB,\u003c/strong\u003e Unsupervised consensus clustering based on the 299 M2-related genes identified three immune subtypes (C1, C2, and C3) in the TCGA cohort. \u003cstrong\u003eC,\u003c/strong\u003e A random forest classifier trained on TCGA data was applied to classify CGGA samples into the same three immune subtypes. \u003cstrong\u003eD-E, \u003c/strong\u003ePrincipal component analysis (PCA) of gene expression profiles showing clear separation among the three immune subtypes in both TCGA and CGGA cohorts. \u003cstrong\u003eF-G, \u003c/strong\u003eKaplan–Meier survival curves comparing overall survival (OS) among the immune subtypes in TCGA and CGGA cohorts. Patients in the C1 group exhibited the worst prognosis. \u003cstrong\u003eH-I, \u003c/strong\u003eDistribution of molecular subtypes (Classical, Mesenchymal, Proneural) across the immune subtypes, with the mesenchymal subtype significantly enriched in the C1 group.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/6a2adb7306ff77fd9f169eb0.png"},{"id":90895004,"identity":"73838b7b-d4e0-47e9-94cb-07d75534b14c","added_by":"auto","created_at":"2025-09-09 11:31:42","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":686054,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmunological characterization of the C1–C3 subtypes in GBM. A-C,\u003c/strong\u003e ESTIMATE analysis showing ImmuneScore, StromalScore, and tumor purity across the three immune subtypes in TCGA and CGGA cohorts. The C1 subtype exhibited the highest immune and stromal scores and the lowest tumor purity, whereas C2 displayed the opposite trend. \u003cstrong\u003eD-F,\u003c/strong\u003e Comparison of Macrophages_M2 scores and expression levels of canonical M2 markers (CD163 and IL10) among the three subtypes, showing highest M2 infiltration and marker expression in the C1 subtype.\u003cstrong\u003e G-H,\u003c/strong\u003eInfiltration of selected immune cell types across subtypes, estimated by xCell and CIBERSORT. The C1 subtype was enriched in immunosuppressive cells (monocytes, neutrophils, resting NK cells, exhausted CD8+ T cells) and depleted in antitumor effector cells (Th1 CD4+ T cells, activated NK cells).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/9ed45ee24486d5ca9ba463b1.png"},{"id":90897708,"identity":"aea2326d-abe5-4ab9-b9ef-16094bbda05f","added_by":"auto","created_at":"2025-09-09 11:47:42","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":568349,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunctional enrichment analysis of genes upregulated in the C1 subtype. A-B, \u003c/strong\u003eDAVID functional enrichment analysis of C1-upregulated genes (logFC \u0026gt; 1, \u003cem\u003ep\u003c/em\u003e\u0026lt; 0.05), highlighting immune-related biological processes. \u003cstrong\u003eC-D, \u003c/strong\u003eGene set enrichment analysis (GSEA) based on CGGA and TCGA datasets showing significant enrichment of hallmark immune-related pathways in the C1 subtype, including Allograft Rejection, IL6_JAK_STAT3 Signaling, Inflammatory Response, and TNFA Signaling via NF-κB.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/df6350712cbaca9b24af9cae.png"},{"id":90897213,"identity":"ae9c5005-93d2-48a0-9aaf-7ab0578e9790","added_by":"auto","created_at":"2025-09-09 11:39:42","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":782595,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction of a macrophage-related risk score model to identify the C1 immunosuppressive subtype.\u003c/strong\u003e \u003cstrong\u003eA-B,\u003c/strong\u003e Feature importance analysis by XGBoost in the TCGA and CGGA cohorts, identifying 200 and 143 M2-related genes with non-zero gain values, respectively. \u003cstrong\u003eC,\u003c/strong\u003e Venn diagram showing 102 overlapping genes with non-zero gain across both datasets. \u003cstrong\u003eD,\u003c/strong\u003e LASSO regression analysis of the 102 genes identified three key genes (\u003cem\u003eCTSB\u003c/em\u003e, \u003cem\u003eLITAF\u003c/em\u003e, and \u003cem\u003eDOK3\u003c/em\u003e) used to construct the risk score model. \u003cstrong\u003eE,\u003c/strong\u003e Using the three candidate genes to group the risk scores, it was observed that the high-risk group (orange) had a poorer prognosis compared to the low-risk group (cyan). \u003cstrong\u003eF-G,\u003c/strong\u003e Kaplan-Meier survival analysis demonstrated that higher risk scores were associated with shorter OS and PFS. \u003cstrong\u003eH,\u003c/strong\u003e Comparison of risk scores among the three immune subtypes (C1-C3).\u003cstrong\u003e I,\u003c/strong\u003e Comparison of risk scores across GBM molecular subtypes, with the highest scores observed in the mesenchymal subtype. \u003cstrong\u003eJ-K,\u003c/strong\u003e Correlation of the risk score with M2 macrophage infiltration (J) and \u003cem\u003eCD163\u003c/em\u003e expression (K).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/68ad189e8c7d6c034501e24d.png"},{"id":90895019,"identity":"d143a409-5f48-44ce-ac7a-fab9b77a9d36","added_by":"auto","created_at":"2025-09-09 11:31:42","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":511529,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe association between the risk score and immune microenvironment characteristics in the TCGA cohort.\u003c/strong\u003e \u003cstrong\u003eA, \u003c/strong\u003eImmune cell infiltration levels in high- and low-risk groups were assessed using the xCell algorithm. The high-risk group showed significantly elevated levels of several immunosuppressive cell types, including M2 macrophages, B cells, monocytes, cancer-associated fibroblasts (CAFs), and plasmacytoid dendritic cells (pDCs), while CD4\u003csup\u003e+\u003c/sup\u003e Th1 cells were more abundant in the low-risk group.\u003cstrong\u003e B,\u003c/strong\u003e Correlation between risk score and ESTIMATE-derived stromal score, immune score, and tumor purity. \u003cstrong\u003eC,\u003c/strong\u003e Expression levels of immune checkpoint molecules (\u003cem\u003eTIMD3\u003c/em\u003e, \u003cem\u003eCSF1R\u003c/em\u003e, \u003cem\u003ePD-1\u003c/em\u003e) are positively correlated with the risk score.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/bef037a63ae4d6bb6de05681.png"},{"id":90895012,"identity":"c9e5ddea-21d5-411d-925c-eadb3dbfd920","added_by":"auto","created_at":"2025-09-09 11:31:42","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1010823,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExpression patterns and prognostic relevance of CTSB, LITAF, and DOK3, with emphasis on DOK3 in myeloid populations.\u003c/strong\u003e \u003cstrong\u003eA-B, \u003c/strong\u003eKaplan-Meier survival analysis of \u003cem\u003eCTSB\u003c/em\u003e, \u003cem\u003eLITAF\u003c/em\u003e, and \u003cem\u003eDOK3\u003c/em\u003e in TCGA (A) and CGGA (B) cohorts. High expression of \u003cem\u003eCTSB\u003c/em\u003e and \u003cem\u003eDOK3\u003c/em\u003e was associated with poor prognosis in both datasets, while LITAF showed prognostic value only in TCGA. \u003cstrong\u003eC-D,\u003c/strong\u003e Single-cell RNA-seq analysis of glioma samples from the CGGA scRNA-seq dataset (C) and GSE131928 scRNA-seq dataset (D). UMAP plots depict cell clustering and gene expression of canonical myeloid markers (\u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eTMEM119\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e) and DOK3, which is predominantly enriched in macrophages and microglia.\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/e916cd40c918a72a1f706864.png"},{"id":90897711,"identity":"5d10e3bd-6d32-4705-8987-1cbcac190e58","added_by":"auto","created_at":"2025-09-09 11:47:42","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":2202044,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eValidation of DOK3’s role in promoting M2 macrophage polarization and glioma cell migration.\u003c/strong\u003e \u003cstrong\u003eA,\u003c/strong\u003eTHP-1 cells changed from round to spindle shape after IL-4 and IL-13 treatment, indicating M2 polarization. \u003cstrong\u003eB, \u003c/strong\u003eWestern blot showed decreased \u003cem\u003eDOK3\u003c/em\u003eand \u003cem\u003eCD163\u003c/em\u003e protein levels after \u003cem\u003eDOK3\u003c/em\u003e knockdown. \u003cstrong\u003eC-E,\u003c/strong\u003e Immunofluorescence staining confirmed reduced \u003cem\u003eDOK3\u003c/em\u003e and \u003cem\u003eCD163\u003c/em\u003e expression in the knockdown group. \u003cstrong\u003eF-G,\u003c/strong\u003e Glioma cells cultured with conditioned medium from \u003cem\u003eDOK3\u003c/em\u003e-silenced M2 macrophages showed reduced migration ability.\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/b7657e3623cdf853df0c278e.png"},{"id":90897709,"identity":"bf991a5c-8f93-4de9-b923-beb37c7836a4","added_by":"auto","created_at":"2025-09-09 11:47:42","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":219725,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDOK3 can promote the polarization of macrophages toward the M2 pro-tumor phenotype, thereby participating in the construction of the immunosuppressive microenvironment of glioblastoma multiforme (GBM) and driving the malignant progression of gliomas.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/bb695a15879a109c692f96c0.png"},{"id":93465402,"identity":"39ff641e-8040-45c4-8464-0126b96ac544","added_by":"auto","created_at":"2025-10-14 07:17:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9336028,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/a80e26f3-1306-4410-bd1f-352be8dff833.pdf"},{"id":90895002,"identity":"bbd648dc-53ba-45c6-993c-cd9326c2cc5d","added_by":"auto","created_at":"2025-09-09 11:31:42","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":15165,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigureslegends.docx","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/2bb6b45b4baacd4f584b6dd6.docx"},{"id":90897211,"identity":"3cd7ad12-4add-4a92-94f8-3e54963bb236","added_by":"auto","created_at":"2025-09-09 11:39:42","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1995962,"visible":true,"origin":"","legend":"","description":"","filename":"SF.docx","url":"https://assets-eu.researchsquare.com/files/rs-7534936/v1/93b22290983181d2775cdacf.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"M2 Macrophage-Based Classification Identifies DOK3 as a Driver of Pro-tumoral Polarization and Migration in Glioblastoma","fulltext":[{"header":"Introduction","content":"\u003cp\u003eGlioblastoma (GBM) is the most common and highly aggressive primary malignant tumor of the central nervous system (CNS), accounting for approximately 80% of primary malignant CNS tumors [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Despite advances in multimodal treatment\u0026ndash;including maximal surgical resection, radiotherapy, and temozolomide-based chemotherapy\u0026ndash;the prognosis remains dismal, with a median overall survival (OS) of only 14\u0026ndash;18 months and a 5-year survival rate below 10% [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The highly infiltrative nature of GBM hampers completed surgical resection, contributing to inevitable recurrence and resistance to conventional therapies [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eRecent molecular and immunological profiling has revealed that GBM harbors a profoundly immunosuppressive tumor microenvironment (TME), enriched with regulatory immune cells and inhibitory soluble factors that dampen the activity of cytotoxic T cells and natural killer (NK) cells, thereby promoting immune evasion and disease progression [\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. This immunosuppressive milieu not only accelerates tumor growth but also limits the efficacy of emerging immunotherapeutic strategies, highlighting the urgent need for precise prognostic stratification systems and robust molecular target to guide personalized treatment strategies.\u003c/p\u003e\u003cp\u003eTumor-associated macrophages (TAMs) are the dominant immune cell component within the GBM microenvironment, with the M2-polarized phenotype constituting the predominant and functionally pro-tumoral subset [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. High densities of M2 macrophages have been strongly associated with higher WHO grades, enhanced invasiveness, and poorer clinical outcomes [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Mechanistically, M2 macrophages exert potent immunosuppressive effects by secreting IL10, TGF-β, Arg-1, and VEGF, upregulating PD-L1 expression [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], promoting T-cell exhaustion, and expanding regulatory T-cells (Treg) populations [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Enrichment of M2 macrophages in glioma tissues has been correlated with reduced IFN-γ expression, indicative of impaired T-cell activity in vivo [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Furthermore, M2 macrophages suppress NK cells cytotoxicity through TGF-β\u0026ndash;mediated downregulation of the activating receptor NKG2D [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], a key mechanism of tumor immune escape [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Notably, blocking TGF-β signaling in preclinical glioma models restores NKG2D expression and enhances NK cell-mediated tumor killing [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. These findings collectively highlight the central role of M2 macrophages as orchestrators of GBM immunosuppression and tumor progression, positioning them as compelling targets for therapeutic intervention. Preclinical studies using CSF1R inhibitors, for example, have demonstrated the potential to reprogram M2 macrophages toward a pro-inflammatory M1 phenotype, thereby enhancing antitumor immunity [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn this study, we combined integrative multi-omics analyses with functional experiments to investigate the molecular drivers of M2 macrophage polarization in GBM. We identified \u003cem\u003eDOK3\u003c/em\u003e as a key regulator of pro-tumoral polarization and glioma cell migration, providing mechanistic insights into macrophage-mediated tumor progression and highlighting \u003cem\u003eDOK3\u003c/em\u003e as a promising immunotherapeutic target for patients with IDH-wildtype GBM.\u003c/p\u003e"},{"header":"Methods and materials","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eData Acquisition\u003c/h2\u003e\u003cp\u003eMultiple glioblastoma (GBM) datasets were obtained from publicly available, authoritative databases. RNA sequencing data and corresponding clinical information from 205 GBM patients were retrieved from The Cancer Genome Atlas (TCGA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://cancergenome.nih.gov/\u003c/span\u003e\u003cspan address=\"http://cancergenome.nih.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] and used as the training cohort for feature selection and model construction. An additional 74 GBM samples from the Chinese Glioma Genome Atlas (CGGA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://cgga.org.cn/\u003c/span\u003e\u003cspan address=\"http://cgga.org.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] served as an independent external validation cohort to evaluate the robustness and generalizability of the model. Clinical variables included, but were not limited to, age, sex, IDH mutation status, MGMT promoter methylation status, chemoradiotherapy, and survival outcomes such as overall survival (OS) and progression-free survival (PFS).\u003c/p\u003e\u003cp\u003eTo investigate the spatial distribution of M2 macrophages in different histological regions, we analyzed data from the Ivy Glioblastoma Atlas Project (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://glioblastoma.alleninstitute.org/\u003c/span\u003e\u003cspan address=\"http://glioblastoma.alleninstitute.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. This dataset comprises gene expression profiles from five anatomically defined regions: (1) cellular tumor (CT), (2) infiltrating tumor (IT), (3) leading edge (LE), (4) microvascular proliferation (MP), and (5) pseudopalisading cells (PC).\u003c/p\u003e\u003cp\u003eTo further exploring the immune microenvironment at single-cell resolution, we incorporated single-cell RNA sequencing data (scRNA-seq) data from both the CGGA database and the GSE131928 dataset in the Gene Expression Omnibus (GEO) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. These datasets include multiple primary and recurrent GBM samples, enabling high resolution characterization of immune cell heterogeneity and dynamics within the tumor microenvironment.\u003c/p\u003e\u003cp\u003eAll datasets were accessed in accordance with their respective usage policies and ethical guidelines. The study was conducted in compliance with the principles of the Declaration of Helsinki.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eEstimation of Immune Infiltrates\u003c/h3\u003e\n\u003cp\u003eImmune cell infiltration was quantified using the TIMER2.0 platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://timer.cistrome.org/\u003c/span\u003e\u003cspan address=\"http://timer.cistrome.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] and xCell algorithm [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] based on bulk RNA-seq data from both the TCGA, CGGA, and IVY cohorts. TIMER2.0 integrates multiple state-of-the-art deconvolution algorithms, including TIMER, CIBERSORT, quanTIseq, and MCP-counter, enabling robust and comprehensive estimates of immune cell abundance across tumor samples. These estimates serves as the basis for subsequent immunological analyses and correlation studies. To identify genes associated with M2 macrophage activity, Pearson correlation analysis was performed between gene expression levels and the M2 macrophages score in each dataset, with genes showing strong positive correlation (r\u0026thinsp;\u0026gt;\u0026thinsp;0.5) in both cohorts retained for downstream analyses.\u003c/p\u003e\n\u003ch3\u003eUnsupervised Clustering and Subtype Identification\u003c/h3\u003e\n\u003cp\u003eGenes significantly correlated with M2 macrophage infiltration (\u003cem\u003eR\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.5 and \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in both the TCGA and CGGA datasets were subjected to unsupervised clustering using the \u003cem\u003eConsensusClusterPlus\u003c/em\u003e R package. To ensure the stability and predictive utility of the identified subtypes, a random forest classifier was trained using the TCGA cohorts and subsequently applied to the CGGA cohort for external validation. The robustness and separability of the resulting clusters were further assessed using principal component analysis (PCA).\u003c/p\u003e\n\u003ch3\u003eRisk Model Construction\u003c/h3\u003e\n\u003cp\u003eTo identify key genes predictive of the immunosuppressive subtype, we applied Extreme Gradient Boosting (XGBoost) and least absolute shrinkage and selection operator (LASSO) regression to the TCGA cohort. Three candidate genes\u0026ndash;CTSB, LITAF, and DOK3\u0026ndash;were selected to construct a macrophage-associated risk score model. The prognostic value of this model was evaluated using Kaplan-Meier survival analysis and univariate and multivariate Cox regression. In addition, correlations between the risk score and CD163 expression, immune checkpoint markers, and immune infiltration patterns were systematically analyzed to elucidate the immunological and clinical relevance of the model.\u003c/p\u003e\n\u003ch3\u003eSingle-cell Transcriptomic Analysis\u003c/h3\u003e\n\u003cp\u003eSingle-cell RNA sequencing (scRNA-seq) data were processed using the Seurat R package (version 5.0.0) [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Low-quality cells and lowly expressed genes were filtered out based on standard quality control metrics, followed by normalization and batch-effect correlation. Dimensionality reduction and clustering were performed using the Uniform Manifold Approximation and Projection (UMAP) algorithm. The expression patterns of key marker genes (CD68, TMEM119, CD163, and DOK3) was analyzed across different immune and non-immune cell clusters to characterize their cellular distribution and functional relevance.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eTHP-1 Cell Culture and Polarization Induction\u003c/h2\u003e\u003cp\u003eTHP-1 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FPS), 1% penicillin-streptomycin, and 0.1 mM β-mercaptoethanol at 37\u0026deg;C in a humidified atmosphere containing 5% CO₂. To induce differentiation into M0 macrophages, cells were treated with 185 ng/mL phorbol 12-myristate 13-acetate (PMA, dissolved in DMSO) for 12 hours, resulting in an adherent phenotype [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. For M2 polarization, adherent cells were co-incubated with 20 ng/mL IL-4 and 20 ng/mL IL-13 in the continued presence of PMA for 48 hours [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. For gene silencing experiments, siRNA negative control (siRNA-NC) or siRNA targeting DOK3 (siRNA-DOK3) was transfected into the respective groups, followed by 48 hours of incubation.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eWestern-blot Assay\u003c/h3\u003e\n\u003cp\u003eAfter polarization to the M2 phenotype, THP-1 cells were transfected with siRNA targeting DOK3. Total protein was extracted using RIPA buffer, and protein concentrations were measured by BCA assay. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% skim milk for 1 hour at room temperature, followed by overnight incubation at 4\u0026deg;C with primary antibodies: DOK3 (Immunoway, YT1397, 1:1000), CD163 (Abcam, ab156769, 1:1000), and GAPDH (Immunoway, YM3029, 1:20000). After washing, membranes were incubated with HRP-conjugated secondary antibodies for 1 hour at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) and imaged with a Bio-Rad imaging system.\u003c/p\u003e\n\u003ch3\u003eImmunofluorescence\u003c/h3\u003e\n\u003cp\u003eCells were fixed with 4% paraformaldehyde, permeabilized with Triton X-100, and blocked with goat serum. Cells were incubated overnight with primary antibodies against DOK3 and CD163, followed by incubation with Alexa Fluor-conjugated secondary antibodies. Nuclei were counterstained with DAPI, and fluorescence imaging was performed using a confocal microscope.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eTranswell Migration Assay\u003c/h2\u003e\u003cp\u003eTranswell assays were performed using 8-\u0026micro;m pore size polycarbonate membrane inserts. LN229 glioma cells (4 \u0026times; 10\u003csup\u003e4\u003c/sup\u003e cells in 200 \u0026micro;L serum-free medium) were seeded into the upper chamber, while M2-polarized macrophages (1 \u0026times; 10\u003csup\u003e5\u003c/sup\u003e cells in 200 \u0026micro;L medium containing 5% FBS) were placed in the lower chamber. After 12 hours of incubation, non-invading cells on the upper surface of the membrane were carefully removed and invading cells on the lower surface were fixed and stained with crystal violet.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eAll statistical analyses were conducted using R software (version 4.2.3) and IBM SPSS Statistics (version 27.0.1). R packages included ConsensusClusterPlus, simplifyEnrichment [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e], ggplot2, pheatmap, GSVA, ggpubr, survival, survminer, dplyr, Seurat, patchwork, and pROC. Pearson correlation, Student\u0026rsquo;s t-test, and one-way ANOVA were applied to assess group differences. Cox regression analysis was performed for survival analysis. \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were considered statistically significant.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eM2 Macrophages Are Spatially Enriched and Associated with Poor Prognosis in GBM\u003c/h2\u003e\u003cp\u003eTo clarify the clinical and biological significance of M2 macrophage in glioblastoma (GBM), we quantified their infiltration using xCell-based deconvolution in two independent transcriptomic cohorts: TCGA (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;205) and CGGA (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;74). In both cohorts, higher M2 macrophage scores were strongly correlated with reduced tumor purity (\u003cem\u003eR\u003c/em\u003e = -0.72, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001; CGGA: \u003cem\u003eR\u003c/em\u003e = -0.73, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, \u003cb\u003eFig. S1A\u003c/b\u003e), indicating preferential accumulation of these cells within the non-neoplastic tumor microenvironment (TME) at the expense of tumor cell proportion.\u003c/p\u003e\u003cp\u003eTo explore their spatial context, we analyzed anatomically annotated transcriptional profiles from the Ivy Glioblastoma Atlas Project (Ivy GAP, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;270). M2 macrophage scores were markedly elevated in microvascular proliferation (MVP) and pseudopalisading cells around necrosis (PAN), two histological niches strongly linked to angiogenesis, hypoxia and aggressive tumor behavior (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Consistent with these findings, these regions were also exhibited in anti-inflammatory and pro-angiogenic pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC, \u003cb\u003eFig. S1B\u003c/b\u003e), suggesting that M2 macrophages may be recruited to these specialized niches to promote vascular remodeling and immune evasion.\u003c/p\u003e\u003cp\u003eFunctionally, gene set enrichment analysis (GSEA) further revealed that M2 macrophage infiltration positively correlated with pathways involved in complement activation (e.g., \u003cem\u003eC1QA\u003c/em\u003e, \u003cem\u003eC1QB\u003c/em\u003e, \u003cem\u003eC3AR1\u003c/em\u003e) and immune modulation, including \u003cem\u003einflammatory response\u003c/em\u003e, \u003cem\u003einterferon gamma response\u003c/em\u003e, and \u003cem\u003eTNFα signaling via NF-κB\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD, \u003cb\u003eFig. S1C-E\u003c/b\u003e). These results indicate that beyond supporting angiogenesis, M2 macrophages actively shape an immunosuppressive microenvironment that suppresses anti-tumor immunity and facilitating tumor immune evasion.\u003c/p\u003e\u003cp\u003eSurvival analyses underscored the clinical importance of these observations: across both cohorts, patients with high M2 macrophage infiltration exhibited significantly shorter overall survival (OS) and progression-free survival (PFS) compared with those with lower infiltration (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE, \u003cb\u003eFig. S1F\u003c/b\u003e).\u003c/p\u003e\u003cp\u003eCollectively, these findings demonstrate that M2 macrophages are spatially enriched in aggressive tumor niches, where they drive an immunosuppressive and pro-tumoral microenvironment, ultimately contributing to poor clinical outcomes in patients with GBM.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eM2 Macrophage\u0026ndash;Related Gene Signatures Define Three Prognostically Relevant Immune Subtypes in GBM\u003c/h2\u003e\u003cp\u003eTo characterize the immune heterogeneity of GBM, we first performed Pearson correlation analysis to identify genes strongly correlated with the M2 macrophages infiltration (\u003cem\u003eR\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.5 and \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in both TCGA and CGGA cohorts. This analysis yielded 299 overlapping M2 macrophage-related genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Using these genes, unsupervised consensus clustering of the TCGA cohort stratified patients into three robust immune subtypes (C1, C2, and C3) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). A random forest classifier trained on the TCGA cohort was subsequently applied to the CGGA cohort, confirming the reproducibility of the three-subtype classification (TCGA: C1, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;90; C2, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;55; C3, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;60; CGGA: C1, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;32; C2, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;29; C3, \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;13) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC).\u003c/p\u003e\u003cp\u003ePrincipal component analysis (PCA) demonstrated clear separation among the three immune subtypes in both datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD-E), supporting the robustness of the clustering. Survival analysis revealed that patients in the C1 subtype exhibited significantly short overall survival compared with those in the C2 and C3 subtypes (TCGA: \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.042; CGGA: \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.021) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF-G), indicating that the C1 immune subtype is associated with the poorest prognosis.\u003c/p\u003e\u003cp\u003eFurther molecular profiling revealed that mesenchymal transcriptional subtype was enriched within the C1 immune subtype (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eH-I), suggesting that the high M2 macrophage activity is linked to mesenchymal transition and aggressive tumor biology.\u003c/p\u003e\u003cp\u003eCollectively, these findings demonstrate that M2 macrophage-related gene signature define three biologically and clinically distinct immune subtypes of GBM, providing a framework for understanding immune heterogeneity and informing the development of subtype-specific therapeutic strategies.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eC1 Subtype Exhibits an Immune-Infiltrated but Immunosuppressive Microenvironment\u003c/h2\u003e\u003cp\u003eTo better characterize the tumor microenvironment (TME) across M2 macrophage-base subtype, we applied the ESTIMATE algorithm to calculate immune score, stromal score, and tumor purity in both the TCGA and CGGA cohorts. In the TCGA cohort, the C1 subtype\u0026ndash;which showed the poorest prognosis\u0026ndash;exhibited the highest immune and stromal scores and markedly reduced tumor purity. Conversely, the C2 subtype, linked to the most favorable outcomes, displayed the lowest immune and stromal scores and the highest tumor purity. These patterns were consistently reproduced in the CGGA cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-C), confirming the robustness of subtype-specific microenvironmental features.\u003c/p\u003e\u003cp\u003eImportantly, the elevated immune and stromal scores in C1 did not indicate effective antitumor immunity. Instead, this subtype reflected a profoundly immunosuppressive TME dominated by M2-like tumor-associated macrophages (TAMs). Quantitative analysis revealed that the C1 subtype harbored the highest levels of M2 macrophage infiltration and elevated expression of canonical M2 markers, including \u003cem\u003eCD163\u003c/em\u003e and \u003cem\u003eIL10\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD-F).\u003c/p\u003e\u003cp\u003eImmune cell composition profiling further highlighted the tumor-supportive and dysfunctional nature of the C1 immune landscape. This subtype was enriched with monocytes, neutrophiles, resting NK cells, and a population of CD8\u0026thinsp;+\u0026thinsp;T cells with an exhausted phenotype (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG-H). In contrast, effector cell populations with recognized antitumor activity\u0026ndash;such as Th1-polarized CD4\u0026thinsp;+\u0026thinsp;T cells and activated NK cells\u0026ndash;were depleted in C1 and relatively enriched in the C2 subtype.\u003c/p\u003e\u003cp\u003eCollectively, these findings indicate that the C1 subtype represents an immune-infiltrated but immunosuppressive TME, where high immune cell content is driven by dysfunctional or tumor-supportive subsets rather than effective antitumor effectors. This immunological architecture, likely orchestrated by M2-polarized macrophages, may underlie the aggressive clinical behavior of C1 tumors and highlight the need for therapeutic strategies targeting immunosuppressive myeloid populations.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003eEnrichment Analysis Reveals Immune Response-Associated but Potentially Immunosuppressive Signatures in the C1 Subtype\u003c/h2\u003e\u003cp\u003eTo explore the molecular underpinnings of the distinct tumor microenvironment observed in the C1 subtype, we performed functional enrichment analysis on genes significantly upregulated in C1 compared to C2 and C3 (logFC\u0026thinsp;\u0026gt;\u0026thinsp;1, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). This analysis revealed significant enrichment of immune-related biological processes, such as \u003cem\u003einflammatory response\u003c/em\u003e and \u003cem\u003eimmune response\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA-B), indicating that the C1 subtype is characterized by a strong immune response\u0026ndash;associated signature.\u003c/p\u003e\u003cp\u003eConsistently, Gene Set Enrichment Analysis (GSEA) in both the TCGA and CCGA cohorts demonstrated that C1-upregulated genes were significantly enriched in hallmark pathways, including \u003cem\u003eallograft rejection\u003c/em\u003e, \u003cem\u003eIL6-JAK-STAT3 signaling\u003c/em\u003e, \u003cem\u003einflammatory response\u003c/em\u003e, and \u003cem\u003eTNFα signaling via NF-κB\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC-D). These pathways are well known to drive chronic inflammation, cytokine signaling, and immune cell recruitment\u0026ndash;processes that often foster immune dysregulation rather that effective antitumor immunity.\u003c/p\u003e\u003cp\u003eTaken together, these findings suggest that although the C1 subtype exhibits a transcriptomic signature heightened immune responses, this activity primarily reflects an inflammation-driven, immunosuppressive microenvironment. This is consistent with the high abundance of M2 macrophages and dysfunctional effector immune subsets identified in our cellular infiltration analyses, underscoring the immune-evasive nature of the C1 subtype.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003eXGBoost-LASSO\u0026ndash;Derived Macrophage Signature Identified a High-Risk, Immunosuppressive C1 Subtype in Glioma\u003c/h2\u003e\u003cp\u003eTo pinpoint genes most closely associated with the C1 subtype and macrophage-driven immunosuppression, we leveraged the 299 macrophage-related genes previously identified as strongly correlated with M2 macrophage infiltration. Using these candidate features, we constructed a XGBoost-based classifier in both the TCGA and CGGA cohorts. The models identified 200 genes with non-zero gain values in TCGA and 143 genes in CGGA (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). In the XGBoost framework, gain value represents the average performance improvement contributed by a feature when splitting decision nodes; thus, genes with non-zero gain values were considered important contributors to identifying the C1 subtype classification. Notably, 102 overlapping genes were consistently identified across both datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), underscoring their stable predictive utility in independent cohorts.\u003c/p\u003e\u003cp\u003eTo further refine these candidates, we performed LASSO regression analysis to the 102 overlapping genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC), which yielded a concise three-gene macrophage signature composed of \u003cem\u003eCTSB\u003c/em\u003e, \u003cem\u003eLITAF\u003c/em\u003e, and \u003cem\u003eDOK3\u003c/em\u003e, with regression coefficients of 0.18895, 0.04699, and 0.00799, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). Based on these coefficients, we calculated a composite macrophage-related risk model for each patient and stratified individuals within the C1 subtype in both cohorts into high- and low-risk groups according to the median score. Patients in the high-risk group exhibited significantly higher mortality than those in the low-risk group \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE\u003cb\u003e).\u003c/b\u003e\u003c/p\u003e\u003cp\u003eKaplan-Meier survival analyses confirmed that higher risk scores were significantly shorter overall survival (OS) and progression-free survival (PFS) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF-G\u003cb\u003e)\u003c/b\u003e. The risk score was markedly elevated in the C1 subtype compared with C2 and C3 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e and was also increased in the mesenchymal transcriptional subtype \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eI\u003cb\u003e)\u003c/b\u003e, indicating a strong association with more aggressive tumor phenotypes.\u003c/p\u003e\u003cp\u003eCorrelation analyses further revealed that the macrophage-based risk score was positively associated with M2 macrophage infiltration (\u003cem\u003eR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.72, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ\u003cb\u003e)\u003c/b\u003e and strongly correlated with the canonical M2 marker \u003cem\u003eCD163\u003c/em\u003e (\u003cem\u003eR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.75, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001;) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eK\u003cb\u003e)\u003c/b\u003e. These findings indicate that the risk score captures not only macrophage abundance but also the immunosuppressive activity of the tumor microenvironment.\u003c/p\u003e\u003cp\u003eImportantly, these results were consistently validated in the CGGA cohort (\u003cb\u003eFig. S2A-G\u003c/b\u003e), demonstrating the robustness and reproducibility of this macrophage-associated risk model across independent datasets.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eHigh Risk Score Reflects an Immune-Infiltrated but Immunosuppressive Tumor Microenvironment\u003c/h2\u003e\u003cp\u003eTo elucidate the biological significance of the macrophage-based risk score within the TME, we analyzed immune cell composition using the xCell algorithm in both the TCGA and CGGA cohorts. The high-risk group exhibited significantly elevated infiltration of B cells, cancer-associated fibroblasts, M2 macrophages, monocytes, and plasmacytoid dendritic cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA, \u003cb\u003eFig. S3A\u003c/b\u003e)\u0026ndash;cell populations widely implicated in immune suppression and tumor progression [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. In contrast, the low-risk group exhibited higher levels of CD4\u003csup\u003e+\u003c/sup\u003e Th1 T cells, a subset known for potent antitumor activity [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e], suggesting that these cells may contribute to more active immune response in this subgroup.\u003c/p\u003e\u003cp\u003eWe next assessed the overall immune and stromal contexture using the ESTIMATE algorithm. Both stromal and immune score were positively correlated with the risk score, whereas tumor purity showed a negative correlation (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, \u003cb\u003eFig. S3B\u003c/b\u003e). These findings indicate that tumors with higher risk scores harbor greater infiltration of non-tumor components, consistent with a highly infiltrated yet immunologically suppressive TME.\u003c/p\u003e\u003cp\u003eTo further explore potential mechanisms of immune suppression, we examined the relationship between the risk score and the expression of classical immune checkpoint molecules. Expression levels of \u003cem\u003eTIMD3\u003c/em\u003e, \u003cem\u003eCSF1R\u003c/em\u003e, and \u003cem\u003ePD-1\u003c/em\u003e were significantly elevated in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC, \u003cb\u003eFig. S3C\u003c/b\u003e), suggesting enhanced activation of immunosuppressive signaling pathways in these patients. This pattern aligns with the enrichment of M2 macrophages and other suppressive immune subsets, reinforcing the link between the high-risk signature and a dysfunctional, immunosuppressive microenvironment.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003eDOK3 as a Key Regulator of M2 Macrophage Polarization\u003c/h2\u003e\u003cp\u003eWe next examined the three key genes\u0026ndash;\u003cem\u003eCTSB\u003c/em\u003e, \u003cem\u003eLITAF\u003c/em\u003e, and \u003cem\u003eDOK3\u003c/em\u003e\u0026ndash;that constitute the macrophage-based risk model. In the TCGA dataset, high expression of all three genes was associated with worse prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA). In the CGGA dataset, elevated expression of \u003cem\u003eCTSB\u003c/em\u003e and \u003cem\u003eDOK3\u003c/em\u003e similarly correlated with poor survival, whereas \u003cem\u003eLITAF\u003c/em\u003e expression showed no significant prognostic association (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB).\u003c/p\u003e\u003cp\u003eGiven the significant positive correlation between the risk score, M2 macrophage infiltration, and \u003cem\u003eCD163\u003c/em\u003e expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ-K), we further investigated the cellular localization of these genes using single-cell RNA sequencing (scRNA-seq) data from the GSE131928 and CGGA datasets. Clustering of tumor and microenvironmental cell populations\u0026ndash;annotated canonical markers including \u003cem\u003eCD68\u003c/em\u003e (macrophages), \u003cem\u003eTMEM119\u003c/em\u003e (microglia), and \u003cem\u003eCD163\u003c/em\u003e (M2 macrophages)\u0026ndash;revealed that \u003cem\u003eDOK3\u003c/em\u003e expression was predominantly restricted in microglia and macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC-D). In contrast, \u003cem\u003eCTSB\u003c/em\u003e and \u003cem\u003eLITAF\u003c/em\u003e exhibited broad, non-specific expression across multiple cell types, including malignant cells, stromal cells, lymphocytes, and oligodendrocytes (\u003cb\u003eFig. S4A-D\u003c/b\u003e) suggesting that their functional relevance may not be limited to macrophage functions.\u003c/p\u003e\u003cp\u003eThese findings highlight \u003cem\u003eDOK3\u003c/em\u003e as the most cell-type-specific gene with the risk signature and suggest that it may act as a regulator of M2 macrophage polarization. This polarization likely facilitates the establishment of an immunosuppressive tumor microenvironment, promoting immune evasion and supporting the malignant progression of gliomas.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e\u003ch2\u003eDOK3 Knockdown Inhibits M2 Macrophage Polarization and Suppressed Glioma Cell Migration\u003c/h2\u003e\u003cp\u003eTo experimentally validate the role of \u003cem\u003eDOK3\u003c/em\u003e in macrophage polarization and tumor progression, we performed cell-based functional assays using THP-1 cells, a well-established model for human monocyte-macrophage biology. Differentiation into M0 macrophages was first induced with PMA treatment, followed by IL-4 and IL-13 stimulation for 48 hours to drive M2 polarization, which was confirmed by the morphological transition from round to spindle-shaped cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA).\u003c/p\u003e\u003cp\u003eGene silencing of \u003cem\u003eDOK3\u003c/em\u003e in M2-polarized macrophages led to a marked reduction in \u003cem\u003eDOK3\u003c/em\u003e expression and a significant decrease in the M2 marker \u003cem\u003eCD163\u003c/em\u003e, as confirmed by Western blot analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB). Immunofluorescence staining corroborated these findings, showing diminished expression of both \u003cem\u003eDOK3\u003c/em\u003e and \u003cem\u003eCD163\u003c/em\u003e in the knockdown group compared with controls (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC-E). These results indicate that \u003cem\u003eDOK3\u003c/em\u003e is essential for the maintenance of the M2-polarized phenotype.\u003c/p\u003e\u003cp\u003eTo assess the functional relevance of \u003cem\u003eDOK3\u003c/em\u003e-mediated polarization in tumor biology, we co-cultured LN229 glioma cells with conditioned media derived from \u003cem\u003eDOK3\u003c/em\u003e-silenced M2 macrophages. Compared to controls, glioma cells invasiveness was significantly attenuated in the knockdown group (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eF-G), demonstrating that \u003cem\u003eDOK3\u003c/em\u003e-mediated M2 polarization facilitates a tumor-promoting microenvironment that enhances glioma cell migration.\u003c/p\u003e\u003cp\u003eCollectively, these findings provide direct functional evidence that DOK3 promotes M2 macrophage polarization and contributes to glioma aggressiveness. Targeting \u003cem\u003eDOK3\u003c/em\u003e in macrophages may therefore represent a promising therapeutic strategy to reprogram the tumor immune microenvironment and restain glioma progression.\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eGlioblastoma (GBM) remains one of the most lethal malignancies of the central nervous system, characterized by extreme heterogeneity, aggressive invasiveness, and resistance to standard therapies [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Increasing evidence underscores the pivotal role of the TME in driving tumor progression and therapeutic resistance. Among the immune components, TAMs\u0026ndash;particularly those polarized toward the M2 phenotype\u0026ndash;are recognized as key facilitators of immune evasion, tumor growth, and treatment failure [\u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. In this study, we systematically dissected the immunological and molecular landscape of M2 macrophages in GBM and identified \u003cem\u003eDOK3\u003c/em\u003e as a potential regulator of M2 polarization and glioma progression.\u003c/p\u003e\u003cp\u003eUsing xCell-based immune deconvolution across TCGA and CGGA cohorts, we demonstrated that elevated M2 macrophage infiltration is associated with poor prognosis and reduced tumor purity, consistent with their recognized role in shaping an immunosuppressive microenvironment. Unsupervised clustering of 299 M2-related genes identified three immune subtypes (C1-C3), among which the C1 subtype exhibited the poorest prognosis and was enriched for the mesenchymal molecular subtype. This C1 subtype displayed a higher infiltrated but profoundly immunosuppressive microenvironment, characterized by enrichment of M2 macrophages, monocytes, and resting NK cells, along with depletion of cytotoxic immune populations such as Th1-polarized CD4\u003csup\u003e+\u003c/sup\u003e T cells and activated NK cell. Although CD8\u003csup\u003e+\u003c/sup\u003e T cells were numerically abundant in C1 tumors, their functional exhaustion likely rendered them ineffective, reflecting an \u0026ldquo;immune-high but functionally suppressed\u0026rdquo; state that has been increasingly recognized as a hallmark of immune dysfunction in GBM [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan additionalcitationids=\"CR36\" citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. This dysfunctional immune architecture may drive immune escape, malignant progression, and resistance to therapy.\u003c/p\u003e\u003cp\u003eThrough integrative machine learning combining XGBoost and LASSO modeling, we developed a macrophage-related risk score that stratifies patients by prognosis and immune contexture. Among the three key genes identified (\u003cem\u003eCTSB\u003c/em\u003e, \u003cem\u003eLITAF\u003c/em\u003e, and \u003cem\u003eDOK3\u003c/em\u003e), \u003cem\u003eDOK3\u003c/em\u003e emerged as the most cell-type\u0026ndash;specific marker, with strong enrichment in microglia and macrophages and a robust correlation with the M2 marker \u003cem\u003eCD163\u003c/em\u003e. Single-cell transcriptomic analysis confirmed its selective expression in myeloid-derived populations, supporting a macrophage-specific role. Functional experiments provided direct evidence for this: \u003cem\u003eDOK3\u003c/em\u003e knockdown in THP-1\u0026ndash;derived macrophages inhibited M2 polarization, reduced \u003cem\u003eCD163\u003c/em\u003e expression, and suppressed the pro-invasive phenotype of glioma cells exposed to conditioned medium from M2 macrophages. Collectively, these findings demonstrate that \u003cem\u003eDOK3\u003c/em\u003e promotes M2 polarization and contribute to the establishment of an immunosuppressive, tumor-promoting microenvironment in GBM.\u003c/p\u003e\u003cp\u003eThese findings carry important translational implications. \u003cem\u003eDOK3\u003c/em\u003e may serve not only as a prognostic biomarker but also a potential therapeutic target to reprogram the TME in GBM. Therapeutic strategies aimed at inhibiting \u003cem\u003eDOK3\u003c/em\u003e or modulating macrophage polarization could restore antitumor immune activity and potentially enhance the efficacy of existing immunotherapeutic approaches, including checkpoint inhibitors and myeloid-targeted therapies, in patients with GBM.\u003c/p\u003e\u003cp\u003eDespite the strengths of this study, including integrative multi-omics analyses and functional validation, several limitations should be acknowledged. The absence of in vivo validation limits our ability to fully assess the therapeutic potential and mechanistic pathways of \u003cem\u003eDOK3\u003c/em\u003e in the complex TME. Moreover, while our data support a critical role of \u003cem\u003eDOK3\u003c/em\u003e in macrophage polarization, the downstream signaling pathways and regulatory networks remain to be elucidated. Finally, the profound heterogeneity of GBM underscores the need for larger, multi-center studies and the integration of spatial single-cell multi-omics to better characterize immune interactions and uncover context-specific therapeutic vulnerabilities.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study comprehensively clarifies the characteristics of the immunosuppressive C1 subtype driven by M2 macrophages in glioblastoma multiforme (GBM), and confirms that DOK3 can promote the polarization of macrophages toward the M2 pro-tumor phenotype, thereby participating in the construction of the immunosuppressive microenvironment of GBM and driving the malignant progression of gliomas (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e). Additionally, the risk model constructed based on macrophages provides a reliable basis for prognostic stratification and clinical therapeutic decision-making in GBM patients. These findings fully highlight the potential of DOK3 and macrophage reprogramming as therapeutic targets, which are expected to remodel the tumor immune microenvironment in GBM patients and further improve the clinical efficacy of immunotherapeutic strategies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll datasets were accessed in accordance with their respective usage policies and ethical guidelines (IRB ID: KY2024-171-02). The study was conducted in compliance with the principles of the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data are included in this published article and Additional files.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing financial interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by\u0026nbsp;the Noncommunicable Chronic Diseases\u0026ndash;National Science and Technology Major Project (2024ZD0525300), the\u0026nbsp;National Natural Science Foundation of China (NSFC) fund (No. 82192894 and 82472841), the Beijing Hospitals Authority Youth Programme (QML20230507), the Natural Science Foundation of Beijing in China (No. 7254342), the Beijing Municipal Science \u0026amp; Technology Commission and Administrative Commission of Zhongguancun Science Park (Z241100009024044), and the Beijing Municipal Health Commission Fund (11000023T000002044300‐5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eC.R. performed data collection, processing and analysis, conducted experimental validation, contributed to the biological interpretation of the results, and drafted the manuscript. J.S. conducted experimental validation. C.Y. performed data collection, processing, and analysis. J.Z. performed data collection, processing, and analysis. W.T. performed data collection, processing, and analysis. H.Z. contributed to the biological interpretation of the results. P.R. contributed to the biological interpretation of the results. J.C. contributed to the biological interpretation of the results. W.F. conceived and designed the study and contributed to the biological interpretation of the results. Y.Z. conceived and designed the study. Z.Z. conceived and designed the study and drafted the manuscript. All authors critically reviewed, revised, and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eKoshy, M., et al., \u003cem\u003eImproved survival time trends for glioblastoma using the SEER 17 population-based registries.\u003c/em\u003e J Neurooncol, 2012. \u003cstrong\u003e107\u003c/strong\u003e(1): p. 207-12.\u003c/li\u003e\n\u003cli\u003eOstrom, Q.T., et al., \u003cem\u003eCBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2015-2019.\u003c/em\u003e Neuro Oncol, 2022. \u003cstrong\u003e24\u003c/strong\u003e(Suppl 5): p. v1-v95.\u003c/li\u003e\n\u003cli\u003eOstrom, Q.T., et al., \u003cem\u003eNational-level overall survival patterns for molecularly-defined diffuse glioma types in the United States.\u003c/em\u003e Neuro Oncol, 2023. \u003cstrong\u003e25\u003c/strong\u003e(4): p. 799-807.\u003c/li\u003e\n\u003cli\u003eGangoso, E., et al., \u003cem\u003eGlioblastomas acquire myeloid-affiliated transcriptional programs via epigenetic immunoediting to elicit immune evasion.\u003c/em\u003e Cell, 2021. \u003cstrong\u003e184\u003c/strong\u003e(9): p. 2454-2470.e26.\u003c/li\u003e\n\u003cli\u003eMei, Y., et al., \u003cem\u003eSiglec-9 acts as an immune-checkpoint molecule on macrophages in glioblastoma, restricting T-cell priming and immunotherapy response.\u003c/em\u003e Nat Cancer, 2023. \u003cstrong\u003e4\u003c/strong\u003e(9): p. 1273-1291.\u003c/li\u003e\n\u003cli\u003eZhao, R., et al., \u003cem\u003eBlocking ITGA5 potentiates the efficacy of anti-PD-1 therapy on glioblastoma by remodeling tumor-associated macrophages.\u003c/em\u003e Cancer Commun (Lond), 2025. \u003cstrong\u003e45\u003c/strong\u003e(6): p. 677-701.\u003c/li\u003e\n\u003cli\u003eWu, L., et al., \u003cem\u003eNatural Coevolution of Tumor and Immunoenvironment in Glioblastoma.\u003c/em\u003e Cancer Discov, 2022. \u003cstrong\u003e12\u003c/strong\u003e(12): p. 2820-2837.\u003c/li\u003e\n\u003cli\u003eGe, W. and W. Wu, \u003cem\u003e[Influencing Factors and Significance of Tumor-associated Macrophage Polarization \u003c/em\u003e\u003cem\u003e\u2029\u003c/em\u003e\u003cem\u003ein Tumor Microenvironment].\u003c/em\u003e Zhongguo Fei Ai Za Zhi, 2023. \u003cstrong\u003e26\u003c/strong\u003e(3): p. 228-237.\u003c/li\u003e\n\u003cli\u003eCao, C., et al., \u003cem\u003eIntra-Operative Definition of Glioma Infiltrative Margins by Visualizing Immunosuppressive Tumor-Associated Macrophages.\u003c/em\u003e Adv Sci (Weinh), 2023. \u003cstrong\u003e10\u003c/strong\u003e(28): p. e2304020.\u003c/li\u003e\n\u003cli\u003eKureshi, C.T. and S.K. Dougan, \u003cem\u003eCytokines in cancer.\u003c/em\u003e Cancer Cell, 2025. \u003cstrong\u003e43\u003c/strong\u003e(1): p. 15-35.\u003c/li\u003e\n\u003cli\u003eZheng, C., et al., \u003cem\u003eIFN\u0026alpha;-induced BST2(+) tumor-associated macrophages facilitate immunosuppression and tumor growth in pancreatic cancer by ERK-CXCL7 signaling.\u003c/em\u003e Cell Rep, 2024. \u003cstrong\u003e43\u003c/strong\u003e(4): p. 114088.\u003c/li\u003e\n\u003cli\u003eLi, J., et al., \u003cem\u003eTumor-Associated Macrophage-Derived Exosomal LINC01232 Induces the Immune Escape in Glioma by Decreasing Surface MHC-I Expression.\u003c/em\u003e Adv Sci (Weinh), 2023. \u003cstrong\u003e10\u003c/strong\u003e(17): p. e2207067.\u003c/li\u003e\n\u003cli\u003eKrneta, T., et al., \u003cem\u003eM2-polarized and tumor-associated macrophages alter NK cell phenotype and function in a contact-dependent manner.\u003c/em\u003e J Leukoc Biol, 2017. \u003cstrong\u003e101\u003c/strong\u003e(1): p. 285-295.\u003c/li\u003e\n\u003cli\u003eGhiringhelli, F., et al., \u003cem\u003eCD4+CD25+ regulatory T cells inhibit natural killer cell functions in a transforming growth factor-beta-dependent manner.\u003c/em\u003e J Exp Med, 2005. \u003cstrong\u003e202\u003c/strong\u003e(8): p. 1075-85.\u003c/li\u003e\n\u003cli\u003eGroh, V., et al., \u003cem\u003eTumour-derived soluble MIC ligands impair expression of NKG2D and T-cell activation.\u003c/em\u003e Nature, 2002. \u003cstrong\u003e419\u003c/strong\u003e(6908): p. 734-8.\u003c/li\u003e\n\u003cli\u003eBatlle, E. and J. Massagu\u0026eacute;, \u003cem\u003eTransforming Growth Factor-\u0026beta; Signaling in Immunity and Cancer.\u003c/em\u003e Immunity, 2019. \u003cstrong\u003e50\u003c/strong\u003e(4): p. 924-940.\u003c/li\u003e\n\u003cli\u003ePyonteck, S.M., et al., \u003cem\u003eCSF-1R inhibition alters macrophage polarization and blocks glioma progression.\u003c/em\u003e Nat Med, 2013. \u003cstrong\u003e19\u003c/strong\u003e(10): p. 1264-72.\u003c/li\u003e\n\u003cli\u003eZhao, Z., et al., \u003cem\u003eChinese Glioma Genome Atlas (CGGA): A Comprehensive Resource with Functional Genomic Data from Chinese Glioma Patients.\u003c/em\u003e Genomics Proteomics Bioinformatics, 2021. \u003cstrong\u003e19\u003c/strong\u003e(1): p. 1-12.\u003c/li\u003e\n\u003cli\u003eZhao, Z., et al., \u003cem\u003eComprehensive RNA-seq transcriptomic profiling in the malignant progression of gliomas.\u003c/em\u003e Sci Data, 2017. \u003cstrong\u003e4\u003c/strong\u003e: p. 170024.\u003c/li\u003e\n\u003cli\u003ePuchalski, R.B., et al., \u003cem\u003eAn anatomic transcriptional atlas of human glioblastoma.\u003c/em\u003e Science, 2018. \u003cstrong\u003e360\u003c/strong\u003e(6389): p. 660-663.\u003c/li\u003e\n\u003cli\u003eEdgar, R., M. Domrachev, and A.E. Lash, \u003cem\u003eGene Expression Omnibus: NCBI gene expression and hybridization array data repository.\u003c/em\u003e Nucleic Acids Res, 2002. \u003cstrong\u003e30\u003c/strong\u003e(1): p. 207-10.\u003c/li\u003e\n\u003cli\u003eLi, T., et al., \u003cem\u003eTIMER2.0 for analysis of tumor-infiltrating immune cells.\u003c/em\u003e Nucleic Acids Res, 2020. \u003cstrong\u003e48\u003c/strong\u003e(W1): p. W509-w514.\u003c/li\u003e\n\u003cli\u003eAran, D., Z. Hu, and A.J. Butte, \u003cem\u003exCell: digitally portraying the tissue cellular heterogeneity landscape.\u003c/em\u003e Genome Biol, 2017. \u003cstrong\u003e18\u003c/strong\u003e(1): p. 220.\u003c/li\u003e\n\u003cli\u003eButler, A., et al., \u003cem\u003eIntegrating single-cell transcriptomic data across different conditions, technologies, and species.\u003c/em\u003e Nat Biotechnol, 2018. \u003cstrong\u003e36\u003c/strong\u003e(5): p. 411-420.\u003c/li\u003e\n\u003cli\u003eDaigneault, M., et al., \u003cem\u003eThe identification of markers of macrophage differentiation in PMA-stimulated THP-1 cells and monocyte-derived macrophages.\u003c/em\u003e PLoS One, 2010. \u003cstrong\u003e5\u003c/strong\u003e(1): p. e8668.\u003c/li\u003e\n\u003cli\u003eMartinez, F.O., et al., \u003cem\u003eTranscriptional profiling of the human monocyte-to-macrophage differentiation and polarization: new molecules and patterns of gene expression.\u003c/em\u003e J Immunol, 2006. \u003cstrong\u003e177\u003c/strong\u003e(10): p. 7303-11.\u003c/li\u003e\n\u003cli\u003eGu, Z. and D. Hubschmann, \u003cem\u003esimplifyEnrichment: A Bioconductor Package for Clustering and Visualizing Functional Enrichment Results.\u003c/em\u003e Genomics Proteomics Bioinformatics, 2023. \u003cstrong\u003e21\u003c/strong\u003e(1): p. 190-202.\u003c/li\u003e\n\u003cli\u003eShalapour, S., et al., \u003cem\u003eImmunosuppressive plasma cells impede T-cell-dependent immunogenic chemotherapy.\u003c/em\u003e Nature, 2015. \u003cstrong\u003e521\u003c/strong\u003e(7550): p. 94-8.\u003c/li\u003e\n\u003cli\u003eKalluri, R., \u003cem\u003eThe biology and function of fibroblasts in cancer.\u003c/em\u003e Nat Rev Cancer, 2016. \u003cstrong\u003e16\u003c/strong\u003e(9): p. 582-98.\u003c/li\u003e\n\u003cli\u003eKennedy, R. and E. Celis, \u003cem\u003eMultiple roles for CD4+ T cells in anti-tumor immune responses.\u003c/em\u003e Immunol Rev, 2008. \u003cstrong\u003e222\u003c/strong\u003e: p. 129-44.\u003c/li\u003e\n\u003cli\u003eLim, M., et al., \u003cem\u003eCurrent state of immunotherapy for glioblastoma.\u003c/em\u003e Nat Rev Clin Oncol, 2018. \u003cstrong\u003e15\u003c/strong\u003e(7): p. 422-442.\u003c/li\u003e\n\u003cli\u003eQuail, D.F. and J.A. Joyce, \u003cem\u003eThe Microenvironmental Landscape of Brain Tumors.\u003c/em\u003e Cancer Cell, 2017. \u003cstrong\u003e31\u003c/strong\u003e(3): p. 326-341.\u003c/li\u003e\n\u003cli\u003eHambardzumyan, D., D.H. Gutmann, and H. Kettenmann, \u003cem\u003eThe role of microglia and macrophages in glioma maintenance and progression.\u003c/em\u003e Nat Neurosci, 2016. \u003cstrong\u003e19\u003c/strong\u003e(1): p. 20-7.\u003c/li\u003e\n\u003cli\u003eChen, Y., et al., \u003cem\u003eTumor-associated monocytes promote mesenchymal transformation through EGFR signaling in glioma.\u003c/em\u003e Cell Rep Med, 2023. \u003cstrong\u003e4\u003c/strong\u003e(9): p. 101177.\u003c/li\u003e\n\u003cli\u003eVivier, E., et al., \u003cem\u003eInnate or adaptive immunity? The example of natural killer cells.\u003c/em\u003e Science, 2011. \u003cstrong\u003e331\u003c/strong\u003e(6013): p. 44-9.\u003c/li\u003e\n\u003cli\u003eWherry, E.J., \u003cem\u003eT cell exhaustion.\u003c/em\u003e Nat Immunol, 2011. \u003cstrong\u003e12\u003c/strong\u003e(6): p. 492-9.\u003c/li\u003e\n\u003cli\u003eWoroniecka, K., et al., \u003cem\u003eT-Cell Exhaustion Signatures Vary with Tumor Type and Are Severe in Glioblastoma.\u003c/em\u003e Clin Cancer Res, 2018. \u003cstrong\u003e24\u003c/strong\u003e(17): p. 4175-4186.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"glioblastoma, DOK3, M2 macrophage, pro-tumoral polarization, tumor microenvironment","lastPublishedDoi":"10.21203/rs.3.rs-7534936/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7534936/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eIDH-wildtype glioblastoma (GBM) is the most aggressive primary brain tumor, characterized by limited therapeutic options and dismal prognosis. Among the components of the immunosuppressive tumor microenvironment (TME), M2-polarized macrophages are pivotal mediators of tumor progression, yet the molecular mechanisms underlying their polarization and pro-tumoral functions remain inadequately understood.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eWe integrated bulk and single-cell RNA sequencing datasets from CGGA, TCGA, and GSE131928. M2 macrophage infiltration was quantified using the xCell algorithm, and unsupervised clustering of M2-associated genes defined immune subtypes, further refined by XGBoost and LASSO modeling. A macrophage-associated risk score based on \u003cem\u003eCTSB\u003c/em\u003e, \u003cem\u003eLITAF\u003c/em\u003e, and \u003cem\u003eDOK3\u003c/em\u003e was constructed and validated across independent cohorts for prognostic and immune relevance. Functional validation was performed by silencing DOK3 in THP-1–derived macrophages, followed by co-culture with glioma cells to assess macrophage polarization and tumor cell behavior.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults \u003c/strong\u003eElevated M2 macrophage infiltration correlated with reduced tumor purity, spatial heterogeneity, and worse survival. Three immune subtypes (C1-C3) were identified; notably, the C1 subtype exhibited the highest M2 infiltration, strongest immunosuppressive features, and poorest prognosis. The macrophage-based risk score robustly stratified patient survival and correlated with \u003cem\u003eCD163\u003c/em\u003e expression and immune checkpoint activation. Single-cell analysis revealed predominant \u003cem\u003eDOK3 \u003c/em\u003eexpression in macrophages and microglia. Functional assays demonstrated that \u003cem\u003eDOK3 \u003c/em\u003eknockdown reduced \u003cem\u003eCD163\u003c/em\u003e expression and attenuated glioma cell invasiveness, supporting its role in promoting M2 polarization and tumor aggressiveness.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003eThis integrative analysis identifies \u003cem\u003eDOK3\u003c/em\u003e as a pivotal regulator of M2 macrophage polarization and a driver of glioblastoma progression. The macrophage-based risk score provides a practical tool for prognostic stratification, and targeting \u003cem\u003eDOK3\u003c/em\u003e offers a promising therapeutic strategy to reprogram the TME and improve clinical outcomes in patients with IDH-wildtype GBM.\u003c/p\u003e","manuscriptTitle":"M2 Macrophage-Based Classification Identifies DOK3 as a Driver of Pro-tumoral Polarization and Migration in Glioblastoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-09 11:31:37","doi":"10.21203/rs.3.rs-7534936/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"eca7f33f-9337-4383-beef-76b2b5355096","owner":[],"postedDate":"September 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-10-14T07:09:03+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-09 11:31:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7534936","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7534936","identity":"rs-7534936","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}