Introduction

Gliomas are the most prevalent and malignant tumors in the central nervous system. They exhibit rapid growth and invade surrounding tissues1,2,3. Despite recent advances in medical technology, the treatment of gliomas continues to pose a challenge2,4,5. Conventional treatment approaches, including surgery, radiation therapy, and chemotherapy, offer limited advantages in enhancing patient survival rates and quality of life3,6,7. Hence, the high recurrence rate and low five-year survival rate among gliomas necessitate an urgent investigation into innovative treatment strategies8.

Macrophages are crucial in the tumor microenvironment9,10,11. Tumor-associated macrophages (TAMs) constitute a population of immune cells within tumor tissues, exerting profound effects on tumor development and progression12,13,14. Based on their function and phenotype, TAMs could be categorized into two types: pro-inflammatory M1 and anti-inflammatory M210,15,16. M1 macrophages typically demonstrate anti-tumor activity, whereas M2 macrophages facilitate tumor growth, invasion, and metastasis11,17. Multiple studies have consistently demonstrated a substantial increase in the proportion of M2-type TAMs in gliomas, which is strongly correlated with a negative disease prognosis10,16,18,19.

In recent years, immunotherapy, specifically antibody therapy directed towards immune checkpoints, has displayed promising effectiveness in different cancer types20,21,22. For instance, antibodies that target PD-1 and PD-L1 can reactivate T cells that tumors have suppressed, resulting in an effective attack against the tumors23,24,25. Nevertheless, immunotherapy for gliomas continues to encounter several challenges, including the immunosuppressive tumor microenvironment and the physical barrier imposed by the blood-brain barrier26,27,28. In this context, gaining a comprehensive understanding of key regulatory factors is highly significant, particularly about the function and regulation of TAMs in the tumor microenvironment to advance novel treatment strategies for Glioma. To achieve this, we utilized the orthotopic glioma model in C57BL/6 mice, a well-established system that closely mimics the invasive growth patterns and immunosuppressive microenvironment observed in human glioblastoma.

This study uses single-cell transcriptome sequencing technology to perform an in-depth analysis of cells in the glioma microenvironment, specifically focusing on the functions and regulations of TAMs. Our study identified a novel regulatory factor, LAPTM4A, which exhibits upregulation in gliomas and is associated with a poor disease prognosis. We conducted additional investigations to explore the impact of LAPTM4A on the M2 polarization of TAMs, anti-tumor immunity, and tumor progression. We anticipate that the findings of this study will offer a novel theoretical foundation and identify molecular targets for the immunotherapy of glioblastoma, thereby enhancing treatment options and overall quality of life for patients.

Materials and methods

Ethics statement

This study adheres strictly to the applicable ethical principles and regulations governing animal experimentation. Our university’s Institutional Animal Care and Use Committee (IACUC) has granted approval for all experimental procedures (review number: 2023PS1086K). All animals are kept and cared for in accordance with humane principles, and experiments are conducted to minimize pain to the greatest extent. After the experiment, all mice were euthanized humanely with ether anesthesia. All animal experiments were conducted in accordance with the ARRIVE guidelines. The maximal tumor volume permitted by our Institutional Animal Care and Use Committee (IACUC) was 1500 mm3, and in none of the experiments were these limits exceeded. We have complied with all relevant ethical regulations for animal use.

In situ transplantation tumor model of glioma in mice

Six-week-old male C57BL/6 mice (strain: 219, obtained from Beijing Vitonlihua Experimental Animal Technology Co., Ltd., Beijing, China) were randomly assigned into two groups: normal mice and model mice. The mice were housed in conventional pathogen-free conditions. The model mouse was anesthetized and secured in a stereotaxic frame, followed by exposing the skull through an incision. A small hole was drilled on the outer side of the prefrontal cortex, approximately 2.0 mm away from the skull. Subsequently, a 10 μL Hamilton syringe was used for stereotaxic injection of 1 μL of a double-labeled cell suspension, GL261-LUC-GFP-Puro, containing 10,000 cells. The injection was targeted at the left striatum of the mouse, with a flow rate of 0.2 μL/min. The tumor was implanted in the anterior lateral thorax at the following coordinates (in mm): AP: +0.5, ML: 2.0, DV: -2.529. Tumor volume was observed after 7 days of intracranial implantation using real-time fluorescence imaging. Intraperitoneal injection of a 15 mg/mL D-potassium fluorescein solution (SY0232, Beijing Bioeasy Biotechnology Co., Ltd., Beijing, China) was administered to each mouse. After 10 min, the mice were introduced into the live animal fluorescence imaging system (NirVivo-MIX, Beijing RuiGuang Science and Technology Co., Ltd., Beijing, China) for in situ observation of tumor bioluminescence30. When mice exhibit signs of neurological dysfunction, such as the inability to walk, excessive sleepiness, hunchback posture, or weight loss exceeding 20% (considered a humane endpoint), euthanasia is carried out31,32. All animal experiments were conducted in accordance with the ARRIVE guidelines.

Acquisition of single-cell RNA-seq data

The single-cell sequencing dataset GSE136001 was downloaded from the gene expression omnibus database (http://www.ncbi.nlm.nih.gov/geo/), containing brain tissue samples from 4 normal mice and 4 glioma model mice33.

Single-cell RNA-seq analysis

Two normal mice and two glioma model mice were randomly selected, and the Seurat package in R software was used34. The glioblastoma-associated scRNA-seq dataset GSE84465 was analyzed. Initially, low-quality cells with fewer than 1000 genes (nFeature) and more than 10% of mitochondrial gene proportion (percent.mt) in glioma cells were eliminated. In order to remove potential doublets, single cells that detected more than 6000 genes were also excluded35. Next, the LogNormalize method was applied to perform normalization, followed by the utilization of the “FindVariableFeatures” function to identify the top 1000 highly variable genes. Subsequently, Principal Component Analysis was utilized to reduce the dimensionality of the scRNA-seq data. The “FindClusters” function, with the “resolution” parameter set to 0.4, was utilized to identify cell clusters. These clusters were then visualized through a t-distributed stochastic neighbor embedding (tSNE) plot. The marker genes for each cell cluster were identified by utilizing the FindAllMarkers function and the cells were annotated using the CellMarker database36.

Cell culture

The T98G (CRL-1690) and LN229 (CRL-2611) glioblastoma cell lines were obtained from the American Type Culture Collection (ATCC) and authenticated using short tandem repeat (STR) profiling. All cell lines were tested for mycoplasma contamination using PCR and confirmed to be mycoplasma-free.

They were cultured in DMEM (Dulbecco’s Modified Eagle Medium) supplemented with 10% fetal bovine serum and 1% antibiotics (penicillin/streptomycin) at 37 °C in a 5% CO2 environment. The DMEM was obtained from Gibco, Thermo Fisher Scientific (China) Co., Ltd., Shanghai. The GL261-LUC-GFP-Puro double-labeled mouse glioma cells (SAc0135LG, Shanghai Aoyin Biotechnology Co., Ltd.) were cultured in a GL261 cell-specific culture medium (BNCC360453, Beijing Beina Chuanglian Biotechnology Research Institute) consisting of 88% DMEM-H, 10% FBS, 1% L-glutamine, and 1% HEPES 1 M Buffer solution.

Human THP-1 monocytes (TIB-20, ATCC, Manassas, VA, USA) and mouse RAW264.7 macrophages (TIB-71, ATCC, Manassas, VA, USA) were cultured in RPMI-1640 medium (11875119, Thermo Fisher Scientific (China) Co., Ltd., Shanghai), supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Phorbol 12-myristate 13-acetate (PMA) at 150 nM (P8139, Sigma) initiated macrophage differentiation. The cells were then collected after two days37.

CRISPR-Cas9 Knockouts, shRNA gene knockdowns, and lentiviral infections

The LAPTM4A CRISPR/Cas9 KO plasmid (m)(sc-421745, Santa Cruz Biotechnology, Inc, USA) was transfected along with the packaging plasmid (psPAX2, 5 mg, #12260, AddGene, USA) and the envelope plasmid (pVSVg, 5 mg, #8454, AddGene, USA) into 293 T cells (CL-0469, Promega, China) to generate lentivirus and knock out the LAPTM4A gene.

The lentiviral overexpression vector pCDH-CMV-MCS-EF1α-copGFP (Lv-, CD511B-1, System Biosciences, USA) and the lentiviral interference vector pGreenPuro (CMV) shRNA Lentivector (sh-, SI505A-1, System Biosciences, USA) were utilized in the construction of lentiviral vectors for LAPTM4A overexpression or silencing. The lentiviral vector was transfected into 293 T cells (CRL-3216, ATCC, USA) using Lipofectamine 3000 reagent (L3000015, Invitrogen, New York, California, USA).

The method of lentiviral infection was established based on the corresponding cell lines, including GL261 (LAPTM4A-KO or oe-LAPTM4A) with LAPTM4A knockout or overexpression, THP-1 or RAW264.7 cells interfered with sh-LAPTM4A, and their control cell lines (WT, oe-NC, or sh-NC). Inoculate 5 × 105 cells into a 6-well plate and incubate until the cell confluence reaches 70–90%. Next, introduce a suitable quantity of packaging lentivirus (MOI = 10, with a working titer of approximately 5 × 106 TU/mL) and 5 μg/mL polybrene (TR-1003, Merck, USA) to the culture medium for transfection purposes. Following transfection for 4 h, the polybrene was diluted by adding equal amounts of culture medium. The culture medium was then replaced after 24 h of transfection. After 48 h, the transfection efficiency was assessed using a luciferase reporter gene, and stable cell lines were obtained by resistance screening using 60 μg/mL of ampicillin (A100339, Sengong Biotech, Shanghai, China). Select the variant with high infection efficiency determined by Western blot analysis for further experiments. Each experiment is repeated three times. The shRNA sequence can be found in Supplementary Table 138,39,40,41.

Co-Culture Of M0 macrophages and T98g/Ln229 in vitro

T98G/LN229 cells were cultured in RPMI 1640 medium and treated with conditioned medium for 2 days to generate TAMs. THP-1 cells were then transfected with either sh-NC or sh-LAPTM4A, and subsequently incubated with the same conditioned medium for 2 days. Next, the TAM supernatant was added to the T98G/LN229 cell culture medium for co-cultivation.

The Millicell co-culture system (PIHP03050, Millicell, USA) was employed for the co-culture experiment, where macrophages were co-cultured with T98G/LN229 cells. In this experiment, a Transwell insert was utilized as a barrier to segregate the THP-1 macrophage differentiation product, which was seeded on the upper surface from the T98G/LN229 cells seeded on the lower surface. Several specific groups were included in the study: the control group, consisting of only PMA-induced THP-1 cells; the co-culture group which involves PMA-induced THP-1 cells co-cultured with T98G/LN229 culture supernatant; the co-sh-NC group comprising PMA-induced THP-1 cells transfected with sh-NC and co-cultured with T98G/LN229 culture supernatant; and the co-sh-LAPTM4A group, which consists of PMA-induced THP-1 cells transfected with sh-LAPTM4A and co-cultured with T98G/LN229 culture supernatant. Twenty-four hours later, continue the subsequent analysis42.

A situ transplantation tumor model to group intracranial glioma mice

C57BL/6 mice (6 weeks old) were randomly allocated into the wild-type (WT) group and the LAPTM4A-KO group, with 10 mice per group. Both GFP-GL261 or GL261-luc and 1 × 106 WT or LAPTM4A knockout (LAPTM4A-KO) TAM (induced by mouse RAW264.7) were co-injected into the left striatum of mice using stereotaxic positioning. Orthotopic transplantation tumor model of Glioma in other mice.

The CSF-1R inhibitor BLZ945 (M00188, Beijing Bioalloy Biotechnology Co., Ltd.) depletes M2 macrophages. The GL261-luc-induced glioma mice were randomly divided into four groups (n = 10/group): oe-NC+Vehicle (injection of oe-NC-transfected TAM and Captisol vehicle), oe-NC + BLZ945 group (injection of oe-NC-transfected TAM and oral administration of BLZ945), oe-LAPTM4A+Vehicle (injection of oe-LAPTM4A-transfected TAM and Captisol vehicle), and oe-LAPTM4A + BLZ945 group (injection of oe-LAPTM4A-transfected TAM and oral administration of BLZ945). The control group in the Vehicle group was administered the carrier Captisol (FY22772, Nantong Feiyu Biotechnology Co., Ltd., China) via injection. The mice treated with BLZ945 commenced treatment on the 7th day following tumor implantation. BLZ945 is administered orally as a capsule, with 200 mg/kg daily in 20% ethanol, for 10 treatment cycles.

The α-PD1 group or LAPTM4A-KO+α-PD1 group received intraperitoneal injections of 200 μg α-PD1 (BE0146, Bioxcell, USA) for 7 days, followed by twice-weekly injections. Observe tumor changes through bioluminescence43,44.

All surgical procedures were performed under general anesthesia using isoflurane, and mice were housed in SPF conditions. To reduce pain and distress, post-operative monitoring was conducted twice daily during the first 3 days and once daily thereafter. Humane endpoints were predefined and included neurological dysfunction (e.g., inability to walk, lethargy, hunchback posture) or >20% body weight loss, upon which humane euthanasia with ether was performed. No unexpected adverse events, such as infection or early death unrelated to tumor burden, were observed.

Cck-8

Cell proliferation status was determined using the CCK-8 assay kit (40203ES60, Yeasen, Shanghai, China). During the CCK-8 assay, glioma cells were treated with TAM supernatant containing sh-NC or sh-LAPTM4A. Next, the cancer cells were inoculated into 96 wells with 100 μL of cell culture medium per well. The cells were then incubated in a cell incubator for 0, 24, 48, and 72 h. The old medium should promptly be replaced with a fresh culture medium. Then, 10 μL of CCK-8 solution was added to each well, followed by an incubation at 37 °C for 2 h. After the incubation period, the absorbance (A) was measured at a wavelength of 450 nm using a Multiskan FC microplate reader (51119080, purchased from Thermo Fisher Scientific, USA)45. Three parallel holes were set in each group, and the average was taken. Furthermore, the experiment was repeated three times.

BrdU cell proliferation assay

Glioma cells in the logarithmic growth phase were harvested, and a single-cell suspension was prepared. The cells were then inoculated into a 24-well plate with a density of 1 × 105 cells per well. Afterward, the BrdU staining reagent (ST1056, Beyotime, Shanghai, China) was added. After a 12 h incubation period, the cells were fixed using 4% paraformaldehyde. They were then incubated with the anti-BrdU antibody (ab308341, 1:2000, Abcam, USA). Next, the cell nuclei were labeled with a DAPI staining solution (C1006, Beyotime, Shanghai, China). The BrdU-positive cells and DAPI-positive cells were quantified in three fields of view using a fluorescence microscope (Olympus, Tokyo, Japan). The cell proliferation rate is calculated by dividing the number of BrdU-positive cells by the number of DAPI-positive cells46.

Transwell assays

The experiment used a 24-well plate and 8 μm Transwell chambers (3422, Corning, USA) to conduct migration and invasion assays. For invasion experiments, each well was coated with 100 μL of matrix gel before the experiment and incubated at 37 °C for 2 h. Digest the cells, wash them twice with PBS, resuspend them in a serum-free culture medium, and adjust the cell density to 3 × 105 cells/mL. Each group consists of three chambers, with 200 μL of cell suspension in each chamber. A volume of 700 μL of complete culture medium is added to the bottom of each chamber. The chambers are placed in a 37 °C, 5% CO2 incubator for 48 h. After fixation in methanol for 30 min, the chambers are stained with 0.05% crystal violet (G1062, Solarbio, Beijing, China) for 5 min. The internal cells of the chambers are removed using a cotton swab and subsequently observed and photographed under an inverted microscope (IX73, OLYMPUS, Japan). Image processing and quantification are performed using ImageJ 1.8.0 software. In migration experiments, there is no need to coat the transwell inserts with a basement membrane matrix. The remaining steps of the experiment remain the same as those in invasion experiments. The experiment is repeated three times.

ELISA detection

The expression levels of TNF-α (human: PT518, mouse: PT512) and IL-10 (human: PI528, mouse: PI522) in the cell culture medium or mouse serum were detected based on the instructions provided in the ELISA kit (Beyotime, Shanghai, China). The absorbance values at 450 nm of each well were determined in less than 3 min using the versatile microplate reader Synergy 2 (BioTek, USA). The regression equation of the standard curve is calculated by utilizing the standard concentration as the x-axis and the absorbance value (A) as the y-axis. To determine the concentration of the target protein in the sample, the absorbance value of the sample is substituted into the equation42.

Immunofluorescence

The cells to be tested were pre-plated in a 12-well cell culture plate to promote cell adhesion. Once the cells adhered to the plate, the culture medium was removed. Subsequently, the cells were washed thrice with PBS in each well, lasting 5 min. Following the washes, the cells were fixed with 4% paraformaldehyde (PFA), permeabilized with 0.1% Triton X-100, and blocked with 10% normal donkey serum before slicing. Glioma tissue samples were initially fixed in 4% paraformaldehyde, dehydrated, cleared, embedded in paraffin, and subsequently sectioned for histological examination. During the immunofluorescence staining process, sections were dewaxed and rehydrated before being blocked with 2% BSA before subsequent experimental procedures. Following the addition of the antibody, incubate the mixture overnight at a temperature of 4 °C. Dilute all antibodies in accordance with the product instructions, which can be found in Supplementary Table 2. After 24 h, the slices were washed with PBS. Then, incubate them with goat anti-mouse IgG (A10551, Thermofisher, 1:200, USA) or goat anti-rabbit IgG (A-11008, Thermofisher, 1:500, USA) secondary antibody at a 1:200 or 1:500 dilution, respectively, for 1 h at room temperature. Following the incubation, the slices were washed again with PBS. To stain the nuclei, use a staining solution of DAPI (C1002, Beyotime, China) for 5 min and then remove the excess DAPI by washing with PBS. The cell coverslip was carefully removed from the cell culture plate using a curved fine needle and small forceps. Subsequently, it was placed on a glass slide with added anti-fluorescence quenching mounting medium, ensuring that the cell side was facing down. Observations and images were captured using a fluorescent microscope (FV-1000/ES, Olympus, Japan). The quantification method measures the fluorescence coverage area within a fixed field of view under a 40x objective. Six fields of view were recorded per group, and the values were averaged47.

Flow cytometry analysis

The resected mouse tumor was separated and digested in RPMI culture medium containing 20 µg/mL of DNase I (11284932001, Merck KGaA, Darmstadt, Germany) and 5 µg/mL of Collagenase IV (C4-BIOC, Merck KGaA, Darmstadt, Germany) for 1 h at 37 °C. After digestion, it was filtered through a 70 µm Corning® Cell Strainer (CLS431751, Merck KGaA, Darmstadt). Subsequently, the single-cell suspension was treated with red blood cell lysis buffer (00-4333-57, Invitrogen), followed by washing and resuspension in PBS. After removing dead cells, resuspend the sample using Percoll (P1644, Merck KGaA, Darmstadt, Germany), following the guidelines provided by the manufacturer. The concentration was adjusted to 1 × 107 cells/mL and then resuspended in 100 µL of PBS for each sample. Cells were collected from the interface of 35–45% Percoll and subjected to flow cytometry (FCM) analysis. The macrophage phenotype was assessed using antibodies against F4/80, CD11b, CD163, CD206, and CD86. The T cell phenotype was evaluated using antibodies against CD45, IFNγ, TNFα, and CD8. The samples were fixed overnight using a 1% formaldehyde solution, and the test was conducted on the following day. The detection of cells was performed using the LSR Fortessa™ cell analyzer (BD Bioscience, USA), and the samples were analyzed using Flowjo software (Ashland, OR, USA). Detailed information about the antibodies used in flow cytometry can be found in Supplementary Table 248,49.

RT-qPCR

Total RNA was extracted using a Trizol reagent (15596026, Invitrogen, USA) per the manufacturer’s instructions. The extracted RNA underwent reverse transcription into cDNA using the PrimeScript RT reagent Kit (RR047A, Takara, Japan). Subsequently, RT-qPCR analysis was conducted on the synthesized cDNA utilizing the Fast SYBR Green PCR Kit (11736059, Thermo Fisher Scientific (China) Co., Ltd., Shanghai, China). Each well had three replicates. The reference gene used was β-actin. The 2-ΔΔCt calculation for relative expression. The experiment was repeated three times. The primer sequences utilized for RT-qPCR in this investigation are presented in Supplementary Table 3. The primers were synthesized by Takara50,51,52.

Western blot

The protein samples extracted from the cellular components and whole cell lysate were quantified using the Pierce BCA Protein Assay Kit (23227, Thermo Fisher, USA). Extract tissues and cell proteins using RIPA buffer. Load 20 μg of protein from each sample onto a sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel and transfer it onto a nitrocellulose membrane. After sealing the skim milk powder at a concentration of 5%, it was incubated for one hour before antibody hatching. All antibodies were diluted according to the instructions manual. On the following day, Peroxidase-conjugated AffiniPure Goat Anti-Rabbit IgG (H + L) (#111035003, Jackson ImmunoResearch, USA) or Peroxidase-conjugated AffiniPure Goat Anti-rat IgG (H + L) (#115035003, Jackson ImmunoResearch, USA) was employed for 1 h. Subsequently, the immunoreactive bands were visualized using the enhanced chemiluminescence reagent (WP20005, thermofisher, USA), and imaging was conducted using the ChemiDoc XRS Plus luminescent image analyzer (Bio-Rad). Band intensities of each group in Western blot images were quantified using ImageJ analysis software, with β-actin as the internal reference (Supplementary Table 2)53. Repeat three times. All the original, uncropped Western Blot (WB) images have been provided in the supplementary files.

Data source and data processing

Retrieve RNA sequencing data from the cancer genome atlas database (https://portal.gdc.cancer.gov/) for the TCGA-GBM and lower grade glioblastoma (LGG) datasets. These datasets comprise 5 normal tissue samples and 704 cancer tissue samples obtained from patients with glioblastoma. The remaining samples of normal brain tissue were obtained from the GTEx database, which consists of 1147 cases (https://www.gtexportal.org). Differential analysis was conducted using the “limma” package in the R language. A significance threshold of P < 0.05 was employed to filter differentially expressed genes54. The Gene Expression Profiling Interactive Analysis (GEPIA) online database contains RNA sequencing expression data from 9736 tumor samples and 8587 normal samples. The data is sourced from TCGA and GTEx data. GEPIA enables differential expression analysis between tumor and normal samples and gene survival analysis55.

Immune infiltration analysis

The R package “GSEAbase” was used to conduct single-sample gene set enrichment analysis (ssGSEA), which was based on the expression levels of 29 immune-related markers. Enrichment scores were calculated for 29 immune-related features in patients with TCGA-GBM and LGG using the “GSVA” package in R. The proportions of infiltrating immune cells in the samples were visualized using the R packages “Vioplot” and “Pheatmap”, and the immune cell content was obtained for each sample56,57.

WGCNA and LASSO algorithms identify disease-specific genes

Constructing an immune co-expression network using the “WGCNA” package in R software entails constructing a gene co-expression network, identifying modules, analyzing module relationships, and identifying highly correlated genes. We set the soft thresholding parameter β to 3 and the scale-free R2 to 0.90. We then select the key module with the highest gene differential expression and extract the genes in this module for further analysis34,58,59.

The LASSO algorithm penalizes least squares regression by employing the L1 penalty function as a special instance. The glmnet R package is utilized for conducting LASSO logistic regression to ascertain the optimal number of features. It refers to the minimum number at which lambda is attained60.

Statistics and reproducibility

The data were obtained from a minimum of three independent experiments, and the results are presented as the mean ± standard deviation (Mean ± SD). Cell experiments were repeated three times to ensure reproducibility. To compare two groups, an independent samples t-test was performed. For comparisons among three or more groups, analysis of variance (ANOVA) was used, followed by Tukey’s HSD post-hoc test if significant differences were detected. For data that did not meet normality or homogeneity of variance assumptions, nonparametric tests were applied: the Mann-Whitney U test for two-group comparisons and the Kruskal-Wallis H test for multiple-group comparisons. All statistical analyses were conducted using GraphPad Prism 9 (GraphPad Software, Inc.) and R software. A two-sided p-value of less than 0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Results

Single-cell transcriptomic analysis reveals diverse immune cell populations and predominance of TAMs in glioblastoma

To investigate the characteristics of immune cells in glioblastoma, we used a mouse orthotopic GL261 glioblastoma model61 for single-cell sequencing analysis. We selected single-cell transcriptome sequencing data from the brain tissues of 2 normal mice and 2 GL261 glioma model mice. After quality control, filtering, and PCA dimensionality reduction, we classified individual cells (Fig. 1A). Quality control and normalization of the single-cell RNA-seq data resulted in the selection of the top 2000 highly variable genes for downstream analysis (Supplementary Fig. 1A). As shown in Supplementary Fig. 1B, the top 4 principal components were selectively displayed. We further performed principal component analysis using the RunPCA function in the Seurat package and visualized the results. As shown in Supplementary Fig. 1C, we selectively displayed the major upregulated and downregulated genes in the first 2 principal components. Using the ElbowPlot function, we quantitatively determined the principal components to select, and the top 17 components were chosen for subsequent analysis (Supplementary Fig. 1D).

Fig. 1: Single-cell Landscape Analysis of Glioblastoma based on Single-cell Sequencing.
figure 1

A Workflow of single-cell analysis in GL261 tumor cells to construct glioblastoma mouse models; B t-SNE clustering plot after batch effect removal; C t-SNE clustering plot with annotated cell types after batch effect removal. Cluster annotations: microglia, monocytes, macrophages, dendritic cells, natural killer cells, neutrophils, T cells; D Bubble heatmap showing expression of cell type markers in cell clusters. Bubble size indicates the percentage of cells expressing the markers, and their relative expression levels are color-coded; E Distribution of different cell types in normal and tumor samples; F Bar plot showing the overall cell composition in normal and tumor samples, color-coded by cell type; G t-SNE clustering plot showing the proportions of individual cell types in normal and tumor samples; H Circular plots showing the number or strength of cell communications, with line thickness representing the number of pathways; I Heatmap showing differential cell communications regarding the number or strength. The top color bar represents the sum of column values (incoming signals) displayed in the heatmap, and the right color bar represents the sum of row values (outgoing signals). In the color bar, red (or blue) represents increased (or decreased) signaling in the glioblastoma group compared to the HC group; J Circular plots showing cell communications in normal and glioblastoma samples, with line thickness representing the number of pathways; K Comparison of intercellular communication networks between normal and glioblastoma samples; L associated interactions with TAMs in normal and glioblastoma samples.

Full size image

Using the tSNE method, we further corrected and clustered the data into 15 cell subgroups (Fig. 1B; Supplementary Fig. 1E, F). As shown in Fig. 1C, D, we annotated the cell subgroups based on known cell marker genes reported in the singlerR package, ultimately identifying seven cell types in the tumor microenvironment: microglia, monocytes, macrophages, dendritic cells, natural killer cells, neutrophils, and T cells. Additionally, we found that TAMs were the predominant immune-related cells within the tumor (Fig. 1E). Compared to the normal controls, microglia cells were notably decreased, while immune cells such as T cells, DCs, and macrophages were increased in glioma samples (Fig. 1F, G). These results indicate that glioma samples are enriched with TAMs.

To determine the mechanisms driving glioblastoma cell evolution, we investigated cell-cell communication mediated by ligand-receptor interactions across all cellular components. Using the “CellChat” R package, we inferred the probability of communication at the signaling pathway level by calculating the communication probability of all ligand-receptor (L-R) interactions associated with each pathway, to study interactions between identified glioma cell phenotypes. Figure 1H, I show the number and strength of cell-cell communications among the seven cell subgroups, with macrophages displaying a high frequency and intensity of interactions. Further comparisons of cell communication in normal controls and glioma samples revealed an increase in macrophage interactions in glioma samples compared to controls (Fig. 1J). We also examined cell interactions within the macrophage migration inhibitory factor (MIF) pathway, observing enhanced intercellular interactions, especially in macrophages, in the glioma sample group compared to the normal group (Fig. 1K). Notably, in tumor samples, the Ccl5/Ccr5 interaction was prominent among all pairs, primarily mediating crosstalk between TAM and microglia-like cells, achieving the highest score, followed by crosstalk with T-cells (Fig. 1L).

These results indicate that TAMs are not only significantly increased in number within the glioma microenvironment but also play a key role in cell-cell communication, particularly through the Ccl5/Ccr5 interaction, mediating crosstalk between TAMs and other immune cells.

Polarization of TAMs into distinct M1 and M2 subtypes in the tumor microenvironment: functional implications and associated gene expression patterns

To investigate the functionality of TAM in a preliminary analysis, we conducted a tSNE clustering analysis by separately extracting macrophages. Our findings reveal that TAM could be categorized into two subtypes: Ccl5+ macrophages and Pf4+ macrophages (Fig. 2A–C). They both highly express TAM-related genes (Fig. 2D). To investigate the functionality of TAMs, we analyzed the expression of M1 and M2 markers, phagocytosis-related genes, and cytokine interaction genes specific to each TAM subtype. Our findings revealed that Ccl5+ macrophages exhibited enhanced expression of genes associated with M1-like polarization (Il-1b, Socs3, and Cd86), indicating heightened inflammatory functionality. They also showed increased expression of phagocytosis-related genes (Cxcl16 and Lgals3bp). On the other hand, Pf4+ macrophages demonstrated elevated levels of immunosuppressive M2-like genes (Mrc1, Cd163, and Lyve1) and cytokine interaction genes (Cx3cr1, Csf1r, and Pf4) (Fig. 2E). We performed Gene Ontology analysis on the enriched marker genes within our subgroups and discovered their involvement in several biological processes. Specifically, these genes were associated with the cytokine-mediated signaling pathway and negative regulation of the immune system processes in BP. In terms of cellular components, they were found to be associated with the phagocytic vesicle and MHC class I peptide loading complex in CC. Additionally, in terms of molecular functions, they were found to be associated with chemokine receptor binding in MF (Fig. 2F). This observation demonstrates that within the tumor microenvironment (TME), TAMs could undergo polarization, giving rise to distinct M1 or M2 subtypes, thereby mediating an immunomodulatory microenvironment.

Fig. 2: Secondary t-SNE clustering analysis of macrophages.
figure 2

A Heatmap of gene expression in two subclusters of macrophages after secondary clustering; B t-SNE clustering plot of the two macrophage subclusters; C t-SNE plot showing the expression of Ccl5 and Pf4 in different subclusters; D t-SNE plot showing the expression of TAM-associated genes C1qa, C1qc, Apoe, and Trem2 in different subclusters; E Bubble plot showing the expression of genes involved in M1, M2, phagocytosis, and cytokine interaction in each TAM subtype; F GO enrichment analysis of marker genes enriched in the two TAM subclusters.

Full size image

Identification of key regulators in TAM functionality: LAPTM4A emerges as a potential driver of M2 polarization in glioblastoma

To identify the key genes that regulate the function of TAMs, we extracted differentially expressed genes from the scRNA sequencing results of the macrophage subpopulation (Fig. 3A). Furthermore, we conducted WGCNA analysis by combining the differential gene analysis results from the GTEx and TCGA databases. From this analysis, we chose the module genes with the highest correlation (MEgery module genes) as glioma-specific genes for subsequent analysis (Fig. 3B, C). Furthermore, we employed the LASSO machine learning algorithm to analyze and identify the crucial genes among the differentially expressed genes in the GTEx and TCGA databases (Fig. 3D). Combining the genetic screenings of LAPTM4A, PCBP1, PDCD6, and GABARAP (Fig. 3E) yielded four intersecting genes. The scRNAseq results showed high expression of all four genes in the tumor samples (Fig. 3F).

Fig. 3: Key Genes mediating glioblastoma tumor microenvironment macrophage polarization identified by bioinformatics analysis.
figure 3

A Volcano plot of differentially expressed genes in macrophage subclusters in scRNA analysis, with low expression on the left and high expression on the right, Ctrol = 2, Tumor = 2; B WGCNA analysis of differentially expressed genes in GTEx and TCGA databases, heatmap showing the correlation between modules and Glioma in the dataset, each cell contains the corresponding correlation and P-value, Normal = 1152 (GTEx = 1147; TCGA = 5), Glioma = 704 (GBM = 170; LGG = 534); C Analysis of the correlation between module membership and gene importance; D LASSO analysis results, the x-axis represents log(λ) values, the y-axis represents Binomial Deviance, the top Supplementary Figure hows the number of cells retained when calculating with the corresponding log(λ) value, the dashed line represents the log(λ) value corresponding to the optimal Binomial Deviance and the number of cells retained; E Venn diagram of candidate genes from three datasets; F Differential expression of 4 core genes in single-cell RNA sequencing data between normal and tumor groups. The size of the dots represents the percentage of gene expression in each cell cluster (25%, 50%, 75%), while the shade of blue indicates the level of gene expression, with darker blue indicating higher expression; G Correlation analysis between the four core genes and 29 immune components; H Correlation analysis between the four core genes and the M2-type macrophage marker gene CD163.

Full size image

Additionally, the expression of these genes in GBM and LGG samples was analyzed using GEPIA data. Compared to the control group, LAPTM4A and GABARAP were upregulated in GBM and LGG samples, while PCBP1 was only upregulated in GBM samples. There was no difference in the upregulation of PDCD6 (Supplementary Fig. 2A). Furthermore, the analysis of clinical prognosis data revealed that glioma patients with low expression of LAPTM4A, PCBP1, and PDCD6 had a better prognosis, while the opposite was observed for GABARAP (Supplementary Fig. 2B). Furthermore, the immune cell infiltration in gliomas was investigated using the TCGA database. The levels of 29 immune components, comprising immune cells, immune factors, and immune pathways, were assessed using the ssGSEA method. The results demonstrated the enrichment of multiple immune cells, immune factors, and immune pathways, including macrophages, in gliomas compared to normal controls (Supplementary Fig. 3A, B). We conducted a comprehensive correlation analysis between these four core genes and immune components. The findings revealed that LAPTM4A and PCBP1 exhibited predominantly positive correlations, whereas PDCD6 and GABARAP showed mainly negative correlations (Fig. 3G).

Furthermore, we compared the results of the correlation analysis for these four genes with the M2 macrophage polarization-related gene CD163. The comparison revealed that only LAPTM4A positively correlated with CD163 (Fig. 3H). Additionally, we validated the expression of LAPTM4A in the scRNA-seq data from the GSE136001 dataset, which includes samples from 4 normal mice and 4 glioma model mice. The 18 cell clusters were annotated as seven cell types: microglia, monocytes, macrophages, dendritic cells, natural killer cells, neutrophils, and T cells (Supplementary Fig. 3C–E). The changes in TAM and microglia cell proportions were consistent with our sequencing results (Supplementary Fig. 3F, G). LAPTM4A expression in tumor tissues from this dataset was also higher than in normal tissues, particularly in macrophages and microglia cells (Supplementary Fig. 3H). These results suggest that LAPTM4A could play a crucial role as an immune gene in the M2 polarization of macrophages in glioblastoma.

Inhibition of LAPTM4A Modulates TAM polarization: a shift from M2 to M1 phenotype in glioblastoma co-culture experiments

To further investigate the involvement of LAPTM4A in the polarization of TAMs, we conducted a co-culture experiment using T98G or LN229 glioma cells and THP-1 cells infected or uninfected with sh-NC/shLAPTM4A to induce TAM differentiation (Fig. 4A). RT-qPCR and Western blot analysis confirmed the successful downregulation of LAPTM4A in TAMs (sh-NC vs. sh-LAPTM4A#1 P < 0.0001; sh-NC vs. sh-LAPTM4A#2 P < 0.0001), and the most effective interference was chosen for subsequent experiments (Supplementary Fig. 4A,B). First, RT-qPCR and ELISA were used to analyze the expression of TAM-related genes. The results showed that, compared to the Control group, macrophages in the co-culture group exhibited a decrease in M1 phenotype markers (CD86: Control vs. co-culture P < 0.0001; iNOS: Control vs. co-culture P < 0.0001; TNF-α: Control vs. co-culture P < 0.0001) and an increase in M2 phenotype markers (CD163: Control vs. co-culture P < 0.0001; IL-10: Control vs. co-culture P < 0.0001; Arg-1: Control vs. co-culture P < 0.0001). Conversely, compared to the co-sh-NC group, the co-sh-LAPTM4A group exhibited increased M1 markers (CD86: co-sh-NC vs. co-sh-LAPTM4A P < 0.0001; iNOS: co-sh-NC vs. co-sh-LAPTM4A P = 0.0008; TNF-α: co-sh-NC vs. co-sh-LAPTM4A P = 0.0013) and decreased M2 markers (CD163: co-sh-NC vs. co-sh-LAPTM4A P < 0.0001; IL-10: co-sh-NC vs. co-sh-LAPTM4A P < 0.0001; Arg-1: co-sh-NC vs. co-sh-LAPTM4A P < 0.0001) (Fig. 4B, C; Supplementary Fig. 4C,D).

Fig. 4: Interference of LAPTM4A Affects M1/M2 Phenotype Switch in Glioblastoma TAMs.
figure 4

A Induction of TAMs by co-culturing T98G or LN229 glioblastoma cells with THP-1 cells infected with or without sh-NC/shLAPTM4A; B RT-qPCR analysis of the expression levels of M2 (CD163, IL-10, Arg-1) and M1 (CD86, iNOS, TNF-α) macrophage markers in macrophages co-cultured with T98G glioblastoma cells; C Detection of TNF-α and IL-10 production by ELISA; D, E Immunofluorescence staining of CD86 and CD206 expression in macrophages, with green fluorescence labeling CD86 or CD206, blue fluorescence labeling DAPI (Scale Bar = 25 μm); F Western blot analysis of CD86 and MRC-1 expression. *P < 0.05, cell experiments repeated at least three times.

Full size image

Immunofluorescence (MRC-1: Control vs. co-culture P < 0.0001; co-sh-NC vs. co-sh-LAPTM4A P < 0.0001; CD86: Control vs. co-culture P < 0.0001; co-sh-NC vs. co-sh-LAPTM4A P < 0.0001) and Western blot (MRC-1: Control vs. co-culture P < 0.0001; co-sh-NC vs. co-sh-LAPTM4A P < 0.0001; CD86: Control vs. co-culture P < 0.0001; co-sh-NC vs. co-sh-LAPTM4A P < 0.0001) results further confirmed that, compared to the Control group, the co-culture group exhibited lower CD86 expression in the M1 phenotype (Control vs. co-culture P < 0.0001) and higher MRC-1 expression in the M2 phenotype. Conversely, compared to the co-sh-NC group, the co-sh-LAPTM4A group exhibited higher CD86 expression (co-sh-NC vs. co-sh-LAPTM4A P < 0.0001) and lower MRC-1 expression (co-sh-NC vs. co-sh-LAPTM4A P < 0.0001) (Fig. 4D-F; Supplementary Fig. 4E-G). The findings suggest that inhibiting LAPTM4A could switch the M2 phenotype of TAM in gliomas to the M1 phenotype.

Inhibition of LAPTM4A suppresses glioblastoma progression: in vitro and in vivo evidence highlighting reduced tumor growth, metastasis, and enhanced survival

To further verify if LAPTM4A regulates the effect of TAM on glioblastoma in vitro. Glioma cell proliferation, invasion, and migration were induced using the conditioned medium of tamoxifen (TAM)-infected cells with sh-LAPTM4A. The results from CCK-8, BrdU, and Transwell experiments indicated a reduction in the proliferation (sh-NC vs sh-LAPTM4A P < 0.0001), invasion (sh-NC vs sh-LAPTM4A P < 0.0001), and migration (sh-NC vs sh-LAPTM4A P < 0.0001) of glioma cells in the sh-LAPTM4A group compared to the sh-NC group (Fig. 5A–E). Furthermore, immunofluorescence experiments demonstrated that TAM in the sh-LAPTM4A group could upregulate E-cadherin expression and downregulate N-cadherin’s expression compared to the sh-NC group (Fig. 5F). The results above suggest that inhibiting LAPTM4A expression in TAMs could suppress the growth and metastasis of glioblastoma.

Fig. 5: Impact of interfering with LAPTM4A on the progression of Glioma.
figure 5

A Diagram of the process of stimulating TAM formation by infecting macrophages with sh-LAPTM4A using the supernatant of tumor cell stimulation; B BrdU detection of glioma cell proliferation capability (Scale Bar = 25 μm); C CCK-8 detection of cell viability; D, E Transwell detection of cell migration and invasion capability (Scale Bar = 50 μm); F Immunofluorescence detection of epithelial-to-mesenchymal transition-associated gene expression, N-cadherin, and E-cadherin. N-cadherin or E-cadherin is labeled with green fluorescence, and DAPI is labeled with blue fluorescence (Scale Bar = 25 μm). *P < 0.05, cell experiments were repeated at least three times.

Full size image

To investigate the potential impact of LAPTM4A on tumors in vivo, we injected WT and LAPTM4A-KO GL261 glioma cells, which expressed the fluorescent marker GFP, separately into C57BL/6 mice. After 21 days of inoculation, when the mice were euthanized, it was observed that the tumor volume (WT vs LAPTM4A-KO P < 0.0001) and weight (WT vs LAPTM4A-KO P < 0.0001) of LAPTM4A-KO mice were smaller than those of WT C57BL/6 mice (Fig. 6A–D). Subsequently, Ki67 immunofluorescence staining was utilized to examine glioma cell proliferation within the tumor microenvironment. The findings revealed fewer Ki67-positive cells (WT vs LAPTM4A-KO P < 0.0001) in LAPTM4A-KO mice than in wild-type (WT) mice (Fig. 6E). Angiogenesis plays a crucial role in the growth and progression of malignant gliomas62. We assessed the density of CD31+ vessels in tumor tissues and observed a decrease (WT vs LAPTM4A-KO P < 0.0001) in CD31+ vessel density in LAPTM4A-KO mice compared to the WT group (Fig. 6F). We conducted additional experiments to investigate the impact of LAPTM4A on the survival time of male mice inoculated with tumors. LAPTM4A-KO mice exhibited a longer survival period than WT mice (WT vs LAPTM4A-KO P = 0.0111) when observed over 90 days (Fig. 6G). The results above suggest that defects in LAPTM4A could impede the in vivo growth of gliomas.

Fig. 6: Influence of LAPTM4A deficiency on tumor growth in GL261 glioma mice.
figure 6

A Representative images of whole-brain tumors in WT and LAPTM4A KO mice. Scale Bar = 1 cm; B Representative immunofluorescence images of coronal brain sections of WT and LAPTM4A KO mice after stereotactic injection of GL261-GFP for 21 days. GL261-GFP (green), DAPI (blue) (Scale Bar = 1 mm); C Quantification of glioma volume based on GFP signal; D Glioma tumor weight measurement; E Immunofluorescence detection of Ki67-positive expression. Ki67 staining image (left), GFP fluorescence image of GL261 glioma cells (middle), merged image of WT (top) and LAPTM4A KO mice (bottom) (right) (Scale Bar = 100 μm); F Immunofluorescence detection of CD31+ vessel density. CD31 red staining image (left), GFP fluorescence image of GL261 glioma cells (middle), merged image of WT (top) and LAPTM4A KO mice (bottom) (right) (Scale Bar = 100 μm); G Kaplan-Meier curve representing the cumulative survival rate of WT and LAPTM4A KO mice injected with GL261-GFP cells. Except for the survival rate detection with n = 15–18 mice in each group, the remaining experiments had n = 10. *P < 0.05.

Full size image

LAPTM4A modulates macrophage polarization in glioma: inhibiting M2 phenotype and supporting M1 dominance to impede tumor progression

To further verify whether LAPTM4A influences macrophage polarization within the glioma immune microenvironment in vivo, we performed RT-qPCR and ELISA analyses. The results showed that, compared to the WT group, LAPTM4A-KO macrophages exhibited a significant increase in M1 phenotype markers (CD86: WT vs. LAPTM4A-KO P < 0.0001; iNOS: WT vs. LAPTM4A-KO P < 0.0001; TNF-α: WT vs. LAPTM4A-KO P < 0.0001) and a decrease in M2 phenotype markers (CD163: WT vs. LAPTM4A-KO P < 0.0001; IL-10: WT vs. LAPTM4A-KO P < 0.0001; Arg-1: WT vs. LAPTM4A-KO P < 0.0001) (Fig. 7A,B).

Fig. 7: Influence of LAPTM4A on the polarization of M2 macrophages in Glioma.
figure 7

A RT-qPCR detection of M1 (CD86, iNOS, and TNF-α) and M2 (CD163, IL-10, Arg-1) macrophage marker expression levels in glioma model mice; B ELISA detection of TNF-α and IL-10 production; C Flow cytometry detection of CD86 and CD206 expression; D Western blot detection of CD86 and MRC-1 expression; E Representative images of Iba1 (red), LAPTM4A (white), and GL261-GFP (green) in glioma models, as well as merged images of WT (top) and LAPTM4A-KO mice (bottom) (Scale Bar = 25 μm); F Quantitative determination of the percentage of Iba1+ coverage area in Glioma (GL261-GFP); G Quantitative determination of the LAPTM4A+ coverage area inside and outside Glioma (normalized to Iba1+ coverage area). n = 10, *P < 0.05.

Full size image

Flow cytometry and Western blot analyses further confirmed that, compared to the WT group, LAPTM4A deletion significantly reduced CD206 (MRC-1) expression (M2 macrophage marker, WT vs. LAPTM4A-KO P < 0.0001) while increasing CD86 expression (M1 macrophage marker, WT vs. LAPTM4A-KO P < 0.0001) (Fig. 7C, D).

Additionally, we examined LAPTM4A expression in LAPTM4A-KO tumor tissues, revealing a significant reduction compared to the WT group (Fig. 7E–G) (WT vs. LAPTM4A-KO P = 0.635; WT vs. LAPTM4A-KO P < 0.0001). These findings suggest that LAPTM4A deletion inhibits M2 macrophage polarization in gliomas.

Furthermore, to deplete M2 macrophages, glioma-bearing mice were treated with BLZ94543. Flow cytometry confirmed that BLZ945 treatment significantly reduced the proportion of M2 macrophages (Supplementary Fig. 5A) (oe-NC+Vehicle vs. oe-NC + BLZ945 P < 0.0001; oe-NC + BLZ945 vs. oe-LAPTM4A + BLZ945 P < 0.0001; oe-LAPTM4A+Vehicle vs. oe-LAPTM4A + BLZ945 P < 0.0001) and downregulated Arg-1 expression (Supplementary Fig. 5B) (oe-NC+Vehicle vs. oe-NC + BLZ945 P < 0.0001; oe-NC + BLZ945 vs. oe-LAPTM4A + BLZ945 P < 0.0001; oe-LAPTM4A+Vehicle vs. oe-LAPTM4A + BLZ945 P < 0.0001).

Additionally, BLZ945 significantly inhibited tumor growth (Supplementary Fig. 5C) (oe-NC+Vehicle vs. oe-NC + BLZ945 P < 0.0001; oe-NC + BLZ945 vs. oe-LAPTM4A + BLZ945 P < 0.0001; oe-LAPTM4A+Vehicle vs. oe-LAPTM4A + BLZ945 P < 0.0001). In contrast, compared to oe-NC+Vehicle, the oe-LAPTM4A+Vehicle group exhibited an increased proportion of M2 macrophages and Arg-1 expression, promoting tumor growth. Notably, LAPTM4A overexpression reversed the effects of BLZ945 (Supplementary Fig. 5A–C), further confirming that LAPTM4A deficiency blocks M2 macrophage polarization and inhibits tumor progression.

LAPTM4A deficiency augments immune activation in glioma: amplifying CD8+ T cell response and enhancing sensitivity to anti-PD1 therapy

The above results demonstrate the mechanism of LAPTM4A in TAM. Consequently, further investigation is necessary to explore why LAPTM4A fails to have an overall effect in counteracting tumor immunity in TAM. Immunofluorescence results demonstrated that the knockout of LAPTM4A resulted in an upregulation of CD8 (WT vs LAPTM4A-KO P < 0.0001) (Fig. 8A), suggesting that the absence of LAPTM4A activates the immune response in Glioma. Additionally, we co-injected TAMs and GL261 cells infected with either WT or LAPTM4A-KO into C57BL/6 mice, followed by α-PD1 injections for 7 days to evaluate their anti-tumor efficacy. We then administered α-PD1 injection for 7 days to evaluate their anti-tumor efficacy. The results demonstrate a reduction in tumor signals in the LAPTM4A-KO+IgG and LAPTM4A-KO+α-PD1 groups compared to the WT+IgG group (WT+IgG vs WT+anti-PD1 P < 0.0001; WT+IgG vs LAPTM4A-KO+IgG P < 0.0001; WT vs LAPTM4A-KO+anti-PD1 P < 0.0001; LAPTM4A-KO+IgG vs LAPTM4A-KO+anti-PD1 P = 0.0011) (Fig. 8B). We hypothesize that the absence of LAPTM4A triggers an intrinsic immune response in tumor cells, leading to the recruitment of CD8+ T cells and improving the effectiveness of PD1 blockade in vivo. To verify this hypothesis, we conducted additional research on the synergistic impact of LAPTM4A knockout and α-PD1. The findings demonstrated that the LAPTM4A knockout group exhibited an increase in the proportion of CD8+ T cells (WT+IgG vs WT+anti-PD1 P = 0.5372; WT+IgG vs LAPTM4A-KO+IgG P < 0.0001; WT vs LAPTM4A-KO+anti-PD1 P < 0.0001; LAPTM4A-KO+IgG vs LAPTM4A-KO+anti-PD1 P < 0.0001) and CD4+ T cells within tumor cells (WT+IgG vs WT+anti-PD1 P = 0.4665; WT+IgG vsLAPTM4A-KO+IgG P < 0.0001; WT vs LAPTM4A-KO+anti-PD1 P < 0.0001; LAPTM4A-KO+IgG vs LAPTM4A-KO+anti-PD1 P < 0.0001). Furthermore, the activation status of CD8 + T cells was significantly elevated, as evidenced by increased IFN-γ levels. Compared to the WT+IgG group, no significant difference was observed in the WT+anti-PD1 group (P = 0.3941), while a significant increase was detected in the LAPTM4A-KO+IgG group (P < 0.0001). Additionally, IFN-γ levels were significantly higher in the LAPTM4A-KO+anti-PD1 group compared to the WT+IgG group (P < 0.0001) and the LAPTM4A-KO+IgG group (P = 0.0152). Similarly, TNF-α levels showed no significant difference between the WT+IgG and WT+anti-PD1 groups (P = 0.9091). However, a significant increase was observed in the LAPTM4A-KO+IgG group compared to the WT+IgG group (P < 0.0001). Moreover, TNF-α levels were significantly elevated in the LAPTM4A-KO+anti-PD1 group compared to both the WT+IgG group (P < 0.0001) and the LAPTM4A-KO+IgG group (P < 0.0001) (Fig. 8C, D). The results above suggest that the absence of LAPTM4A in Glioma’s immune microenvironment enhances immune activation and increases sensitivity to anti-PD1 therapy.

Fig. 8: Lack of LAPTM4A mediates anti-tumor immunity in Glioma.
figure 8

A Immunofluorescence detection of CD8α expression in glioma tissues, Scale Bar = 25 µm; B Fluorescence images of tumor size in mice treated with α-PD1; C Flow cytometry detection of CD4+ and CD8+ T cell content in glioma tissues; D Flow cytometry analysis of IFN-γ or TNF-α expression in CD8+ T cells in glioma tissues. n = 10, *P < 0.05, ns indicates no difference.

Full size image

Discussion

Based on the above results, we could preliminarily draw the following conclusions: the absence of LAPTM4A inhibits M2-polarized macrophages associated with Glioma, promotes immune activation, enhances the effectiveness of anti-PD1 therapy, and ultimately suppresses glioma development.

Glioma, a highly malignant tumor of the central nervous system, has long been a focal point of research in neuro-oncology63. In recent years, there has been a growing recognition of the role of the immune microenvironment in different types of tumors, including gliomas3. Initially, studies predominantly emphasized the biological characteristics of glioma cells. However, in recent years, there has been an increasing trend in research focusing on immune cells within the glioma microenvironment64,65. In particular, TAMs play a crucial role in the development and progression of glioblastoma66. In comparison, our research enhances the comprehension of the role of TAMs, particularly M2-type TAMs, in gliomas. Our study utilized single-cell transcriptome sequencing technology to identify two primary types of TAMs, specifically M1 and M2, within the glioma microenvironment. This finding is in line with previous research, which suggests that within the tumor microenvironment, macrophages could be categorized as either pro-inflammatory M1 type or anti-inflammatory M2 type67. Earlier studies have demonstrated that M2-type TAMs in gliomas are associated with a negative disease prognosis68. This particular subtype of macrophages is believed to facilitate tumor growth, invasion, and metastasis. However, the precise regulatory mechanisms remain unclear16. Numerous researchers have investigated the origin and function of M2-type TAMs from various perspectives. However, regulating their function in Glioma remains enigmatic18,69. In contrast to these studies, we have presented a novel approach to understanding the functional regulation of M2-type TAMs by examining the role of LAPTM4A.

LAPTM4A is a novel discovery in our research, and its expression in gliomas is correlated with a negative disease prognosis. Although some previous studies have mentioned the association between LAPTM4A and certain diseases, the research on this topic in gliomas is still relatively new70. Our findings demonstrate that LAPTM4A impacts both the M2 polarization of TAMs and the biological behavior of glioma cells. This perspective offers us a fresh vantage point to investigate the cellular interactions within the microenvironment of gliomas. In future studies, we will further explore the interactions between macrophages and tumor cells and investigate the molecular mechanisms by which LAPTM4A regulates these interactions.

In glioma research, numerous genes are associated with the development and prognosis of the disease71. Mutations or abnormal expressions of genes, such as IDH1, EGFR, and P53, are associated with gliomas’ growth, invasion, and chemotherapy resistance72. The mechanism by which LAPTM4A acts in gliomas differs from that of these genes70. It does not directly affect the glioma cells; it influences the tumor’s behavior by regulating TAMs in the microenvironment73. This indirect regulatory mechanism offers an innovative strategy for the treatment of gliomas.

Our research findings indicate that LAPTM4A has potential as a therapeutic target for treating Glioma. Modulating the expression or function of LAPTM4A can potentially influence the M2 polarization of TAMs, suppressing glioma growth, invasion, and metastasis. Furthermore, LAPTM4A may exhibit synergistic effects when combined with other therapeutic strategies, including immune checkpoint inhibitors. This study offers a novel approach to comprehensively treating gliomas.

In general, our research offers novel insights into the immune microenvironment of gliomas, specifically focusing on the role and regulatory mechanisms of TAMs. LAPTM4A, being a novel key regulatory factor, warrants further investigation into its function in gliomas and its potential as a therapeutic target. Our research findings can potentially contribute to developing novel therapeutic strategies by modulating the expression or activity of LAPTM4A, thereby influencing tumor growth and immune response.

Nevertheless, several limitations should be acknowledged. First, although in vitro co-culture systems and in vivo murine experiments were utilized to explore the function of LAPTM4A, these models do not fully recapitulate the complexity of human glioma immunobiology. Second, the GL261 orthotopic mouse model, while widely used and reflective of several glioblastoma features, may not capture species-specific immune responses or genetic heterogeneity present in human tumors. Third, although LAPTM4A was shown to influence M2 polarization and immune activation, its downstream signaling pathways and molecular interactors remain incompletely defined. Lastly, our findings were validated in a single mouse strain and tumor model; additional validation in other glioma models, humanized mouse systems, and clinical patient samples will be necessary to assess translational relevance.

Despite these limitations, the conserved nature of macrophage polarization mechanisms and the upregulation of LAPTM4A in human glioma datasets indicate that the findings of this study may have relevance beyond the specific experimental system. With further mechanistic and translational validation, LAPTM4A holds potential as a broadly applicable therapeutic target in glioblastoma immunotherapy.