Single-cell sequencing analysis reveals the promotion of ELR CXCL genes on angiogenesis and the influence on malignant progression in glioblastoma

Abstract

Glioblastoma (GBM) is a highly aggressive cancer that significantly impacts human health. Myeloid cells and bone marrow-derived macrophages (BMDMs) are major components of tumor microenvironment; however, their roles in GBM tumorigenesis, progression, and treatment response remain unclear due to considerable heterogeneity. Single-cell sequencing data for GBM were collected from the Gene Expression Omnibus (GEO) database to analyze the heterogeneity of GBM cell subpopulations. Trends in myeloid cell subpopulations were examined, identifying cancer-associated BMDMs and key genes that were specifically overexpressed in this subgroup. Validation was performed through experiments in vitro, including quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) to measure mRNA expressions of chemokine (C-X-C motif) ligand 3 (CXCL3), vascular endothelial growth factor A (VEGF-A), and matrix metallopeptidase 9 (MMP9); angiogenesis assays to observe endothelial cell tube formation; Cell Counting Kit-8 (CCK-8) assays to assess cell viability; colony formation assays for cell proliferation; Transwell assays to evaluate cell migration and invasion; and flow cytometry to measure apoptosis. Eight distinct cell types were identified in GBM, with a notable proportion of myeloid cells. Cancer-related BMDMs were characterized within the myeloid cell population, revealing specific overexpression of Glu-Leu-Arg (ELR) CXCL genes associated with angiogenesis and tumor development. Experiments in vitro confirmed that the ELR CXCL-related gene CXCL3 derived from BMDMs promoted angiogenesis and influenced GBM malignancy. The ELR CXCL + BMDM subgroup represented a critical cell type associated with the rapid growth of GBM, as the secreted CXCL3 enhanced angiogenesis and impacted GBM malignant progression.

Introduction

Glioblastoma (GBM) is the most lethal intracranial tumor, characterized by highly infiltrative and aggressive growth, as well as abundant angiogenesis.^(1,2) Current treatment options for GBM include surgery, chemotherapy, radiotherapy, and targeted therapies. While these approaches have improved survival and prognosis to some extent, overall outcomes remain unsatisfactory.⁽³⁾ Given the critical importance of a robust blood supply for tumor growth and metastasis, anti-angiogenic therapies have become vital components of GBM treatment, aiming to inhibit new vessel formation and thereby slow tumor progression.⁽⁴⁾ However, these therapies face several limitations, such as issues with resistance, lack of biomarkers, and unclear therapeutic windows.^(5,6) These challenges highlight the need for further research and exploration in anti-angiogenic treatments to enhance therapeutic efficacy and improve patient outcomes.

GBM induces pathological changes in angiogenesis, increasing tumor aggressiveness and complicating treatment.⁽⁷⁾ The abnormal vascularization observed in GBM is a result of various molecular regulatory mechanisms and cellular communications mediated by different cell types.⁽⁸⁾ For instance, tumor cells can secrete multiple pro-angiogenic factors, such as vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF), which directly promote neovascularization, thereby supporting tumor growth and invasion.⁽⁹⁾ Polarization of tumor-associated macrophages (TAMs) within GBM microenvironment can enrich the vascular extracellular matrix, potentially enhancing angiogenesis.⁽¹⁰⁾ The multiple pathways involved in angiogenesis create various opportunities for therapeutic targeting. Therefore, it is essential to explore the tumor heterogeneity in GBM and identify key drivers of angiogenesis to provide new targets and strategies for improving anti-angiogenic therapies.

CXC chemokines are a class of small secreted peptides that play critical roles in leukocyte recruitment and regulation of tumor microenvironment, influencing tumor cell proliferation, metastasis, invasion, angiogenesis, and treatment resistance.⁽¹¹⁾ ELR CXC chemokine family, characterized by specific amino acid sequence Glu-Leu-Arg (ELR), includes CXCL1, CXCL2, CXCL3, CXCL5, CXCL6, CXCL7, CXCL8, and CXCL17.⁽¹²⁾ These chemokines exert their effects by activating CXCR1 and CXCR2 receptors, with one of their primary functions being the regulation of angiogenesis.⁽¹³⁾ Increasing evidence indicates that ELR CXCL genes are closely associated with various cancers, including breast cancer, colorectal cancer, and hepatocellular carcinoma,^(14–16) showing significant upregulation in cancer and a positive correlation with cancer proliferation, invasion, metastasis, and chemotherapy resistance. However, the specific mechanisms of ELR CXCL genes in the pathogenesis and progression of GBM remain poorly understood, and their application and clinical research in GBM treatment have not been extensively explored.

In this study, we analyzed and constructed a single-cell atlas of GBM, identifying eight distinct cell populations and demonstrating the heterogeneity of myeloid cells. We characterized a subgroup of bone marrow-derived macrophages (BMDMs) associated with GBM development, revealing specific upregulation of ELR CXCL genes in BMDMs. Further cell experiments confirmed that CXCL3, encoded by the ELR CXCL gene, promotes angiogenesis and malignant progression in GBM, providing new insights and potential therapeutic directions for the development of anti-angiogenic therapies in GBM patients.

Materials and Methods

Single-cell sequencing (scRNA-seq) data source and quality control (QC)

GBM-related scRNA-seq data were obtained from Gene Expression Omnibus (GEO, GSE182109), consisting of tumor data from four low-grade glioma (LGG) cases and four primary GBM (ndGBM) cases. We performed QC on integrated single-cell data, setting the following criteria: mitochondrial gene percentage <15%, nCount <50,000, and nFeature <7,000. Gene expression was normalized using SeuratNorm. 38,078 single cells passed QC standards and were used for analysis.⁽¹⁷⁾

Single-cell landscape construction and cell cluster annotation

Cells were clustered using Seurat (ver. 5.5.0) with default parameters, and dimensionality reduction and visualization were conducted using Uniform Manifold Approximation and Projection (UMAP). The FindAllMarkers function in Seurat R package was utilized to identify highly expressed marker genes in each cell cluster. Cell clusters were classified based on expressions of classical cell-type marker genes from cellMarkers2.0 and literature.

Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses

R package clusterProfiler was employed (p<0.05) to perform GO and KEGG enrichment analyses on differentially expressed genes between ELR CXCL + BMDMs and ELR CXCL-BMDMs, revealing their potential functions and molecular pathways.

Pseudotime trajectory analysis

Pseudotime analysis was conducted using R package Monocle3 (1.0.0) to investigate cell trajectories andexpression patterns of specific genes over pseudotime.

Cell culture

Human GBM cells LN229 (CL-0578; Procell, Wuhan, China) and Human BMDMs (CP-H186; Procell) were cultured in dulbecco’s modified eagle medium (DMEM) complete medium (Thermo Fisher, Waltham, MA), respectively, both containing 1% penicillin-streptomycin (P/S), with 5% fetal bovine serum (FBS) for the former and 10% FBS for the latter. Human umbilical vein endothelial cells (HUVEC) (BNCC378266; BNCC, Beijing, China) were maintained in endothelial cell medium (ECM) (Thermo Fisher) supplemented with 5% FBS, 1% endothelial cell growth supplement (ECGS) (Thermo Fisher), and 1% P/S. A 5% CO₂ environment was used to incubate all three cell types at 37°C.

LN229 cells were sown in the lower chamber of a Transwell, and BMDMs were seeded in the top chamber for co-culture studies. After incubating at 37°C for 24 ‍h, the upper chamber was removed, and LN229 cells from the lower chamber were collected for further research.

Cell transfection

Before being transfected, cells were cultivated in a medium devoid of antibiotics until they achieved a confluence of around 70–90%. GenePharma (Shanghai, China) generated the CXCL3 overexpression plasmid (oe-CXCL3) and overexpressing negative control (oe-NC) constructs, which were then transfected into BMDMs using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Forty-eight hours after transfection, the effectiveness of the process was evaluated.

Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR)

TRIzol reagent (Invitrogen) was used to extract total RNA from cells, and a reverse transcription kit (Invitrogen) was used to convert the RNA into cDNA. Utilizing TB Green Premix Ex Taq II (Takara Bio, Kusatsu, Japan), amplification was carried out. The 2^−ΔΔCt technique was used to calculate the relative transcript levels of the target genes, with GAPDH acting as an internal reference. The following were the primers: CXCL3: F: CGCCCAAACCGAAGTCATAG; R: GCTCCCCTTGTTCAGTATCTTTT; VEGF-A: F: AGGGCAGAATCATCACGAAGT; R: AGGGTCTCGATTGGATGGCA; MMP9: F: TTCTCCAGAAGCAACTGTCC; R: TAGGTGATGTTGTGGTGGTG; GAPDH: F: ACAAITTTGGTATCGTGGAAGG; R: GCCATCACGCCACAGTTTC.

Human umbilical vein endothelial cells (HUVEC) tube formation assay

HUVECs were cultured using the culture media from GBM cells. Matrigel (BD Biosciences, San Jose, CA)-coated 96-well plates (5 × 10⁴ cells/well) were used for HUVEC seeding. Following 6 h of incubation at 37°C, tube formation was noted, and pictures were taken using an inverted microscope (Olympus, Hachioji, Japan). Using ImageJ software, the quantity of branching points was measured.

Cell counting kit-8 (CCK-8) assay

In 96-well plates, cells that grew logarithmically (2,000 cells/ml) were planted. 100 ‍μl of cell culture media was added to each well, and the wells were incubated for 0, 24, 48, 72, and 96 ‍h. 10 ‍μl of CCK-8 solution (Beyotime, Shanghai, China) was added to each well after the supernatant was promptly disposed away. A microplate reader (Thermo Fisher) was used to measure the optical density (OD) at 450 ‍nm after the plates had been incubated for 4 ‍h at 37°C.

Colony formation assay

After being digested with 0.25% trypsin, cells in the logarithmic growth phase were suspended in DMEM that included 10% FBS. In 6-well plates, 100 cells were planted, and they were cultured for a week at 37°C with 5% CO₂. The culture was terminated and the supernatant disposed of as soon as colonies were evident. After being fixed with 4% paraformaldehyde, cells were stained for 20 ‍min with crystal violet. The plates were air-dried after being cleaned under running water to get rid of extra dye. Ultimately, pictures were taken, and a quantitative analysis was done.

Transwell assay

Cells were introduced to the upper chamber of a Transwell with an uncoated membrane to perform migration tests. Cells were positioned in the top chamber of a Transwell covered with Matrigel for invasion tests. The upper insert received 200 ‍μl of serum-free media having 2 × 10⁴ cells suspended in it, whereas the bottom chamber received 500 ‍μl of DMEM containing 10% FBS. Cells that invaded or migrated to the bottom of the membrane were fixed with 4% paraformaldehyde and stained with crystal violet following a 24-h incubation period at 37°C. We used an optical microscope (Olympus) to take pictures and count the cells.

Flow cytometry

Apoptosis was measured using the Annexin V-Fluorescein Isothiocyanate (FITC)-Propidium Iodide (PI) apoptosis detection kit (Sigma, St. Louis, MO). After being cultivated for 24 ‍h at a concentration of 1 × 10⁶/ml, cells were resuspended in Annexin V binding buffer. For 20 ‍min, cells were stained using a solution containing PI and Annexin V/FITC. A flow cytometer and FlowJo software (BD Biosciences) were used to conduct flow cytometric analysis.

Statistical analysis

SPSS ver. 21.0 was used for data analysis, and GraphPad Prism 8.0 (GraphPad, La Jolla, CA) was used for data visualization. The data are shown as mean ± SD, and all experiments were conducted three times. Student’s t test was used to compare differences between two groups. Statistical significance was defined as a p<0.05.

Results

QC of GBM scRNA-Seq data and cell subgroup identification

To analyze cellular characteristics within GBM tumors, we collected scRNA-Seq data from four LGG cases and four ndGBM cases from GSE182109. After QC and batch effect removal (Fig. 1A and B), we obtained 38,078 high-quality single cells. Using UMAP for visualization, these cells were categorized into 34 primary clusters (Fig. 1C). Based on known surface marker gene expression (Fig. 1D), these clusters were identified as eight distinct cell types, including B cells, endothelial cells, red blood cells, glioma cells, mast cells, myeloid cells, oligodendrocytes, and T cells (Fig. 1E). Additionally, the distribution of each cell subtype is illustrated in Fig. 1F, showing a relatively high proportion of myeloid cells and glioma cells. In summary, we constructed a human GBM single-cell atlas with precise cell type annotations, revealing a significant presence of myeloid cells.

Fig. 1.

QC of GBM scRNA-Seq data and cell subgroup identification. (A) QC plots for single-cell UMI counts, gene counts, and mitochondrial proportions; (B) Principal component analysis results; (C) UMAP visualization of cell clustering; (D) Expression levels of various cell marker genes; (E) UMAP visualization of cell annotation results; (F) Distribution of different cell types.

Heterogeneity among myeloid cell subgroups in GBM

Myeloid cells constitute the largest stromal component in gliomas and are key immune elements within the glioma microenvironment. Understanding the interaction mechanisms between myeloid cells and tumor cells, as well as the formation mechanisms of the intrinsic heterogeneity of tumor cells, is crucial for exploring new targeted therapy strategies.⁽¹⁸⁾ After isolating myeloid cells, we re-clustered them (Fig. 2A and B), applying UMAP algorithm to classify myeloid cells into three subgroups: BMDMs, microglia, and dendritic cells (DC) (Fig. 2C). By analyzing proportions of myeloid cell subgroups in LGG and GBM, we found that microglia predominated in LGG, while the proportion of BMDMs significantly increased when LGG progressed to GBM. This notable change in the BMDM subgroup was significantly associated with the transition from low-grade to high-grade glioma (Fig. 2D and E). Thus, we hypothesized that BMDMs played a particularly important role in GBM development.

Fig. 2.

Heterogeneity among myeloid cell subgroups in GBM. (A) Principal component analysis of myeloid cell subgroups; (B) UMAP visualization of myeloid cell clustering; (C) UMAP visualization of myeloid cell subgroup annotation; (D) Visualization of myeloid cell subgroups in LGG and GBM; (E) Proportions of LGG and GBM subgroups within myeloid cells.

Upregulation of ELR CXCL genes in ELR CXCL + BMDMs

We performed subgroup analysis of BMDMs (Fig. 3A and B) and identified three subgroups—GRID2 + BMDMs, MCM10 + BMDMs, and ELR CXCL + BMDMs—based on the actual data and marker genes from cellmarker2.0 database (Fig. 3C). GO enrichment analysis for ELR CXCL + BMDMs revealed significant enrichment in pathways such as TNF and IL-17 (Fig. 3D). KEGG enrichment analysis indicated that genes associated with ELR CXCL + BMDMs were primarily enriched in chemokine signaling pathways (Fig. 3E). A UMAP visualization of gene expression showed that differentially expressed ELR CXCL genes were predominantly found in ELR CXCL + BMDMs subgroup (Fig. 3F). Pseudotime analysis indicated that ELR CXCL + BMDMs represent a post-differentiation type of cell, first differentiating from GRID2 + BMDMs to MCM10 + BMDMs, and subsequently from MCM10 + BMDMs to ELR CXCL + BMDMs. Notably, most ELR CXCL genes were upregulated in ELR CXCL + BMDMs (Fig. 3G and H). Collectively, our analysis suggested that the ELR CXCL + BMDMs subgroup was a critical population associated with rapid GBM growth, and the highly expressed ELR CXCL genes in this subgroup could serve as important targets for hampering GBM progression.

Fig. 3.

Upregulation of ELR CXCL genes in ELR CXCL + BMDMs. (A) Principal component analysis of BMDM subgroups; (B) UMAP visualization of BMDM clustering; (C) UMAP visualization of BMDM subgroup annotation; (D) GO enrichment analysis of ELR CXCL + BMDMs; (E) KEGG enrichment analysis of ELR CXCL + BMDMs genes; (F) Expression of ELR CXCL genes in myeloid cells (top) and BMDMs (bottom); (G) Trajectory plot of BMDMs; (H) Curves showing the expression of select genes over pseudotime.

CXCL3 promotes angiogenesis and affects malignant progression of GBM

To investigate the role of CXCL3, one of the proteins encoded by the ELR CXCL genes, in GBM malignant progression and angiogenesis, we designed experiments in vitro. BMDMs were transfected to overexpress CXCL3, and qRT-PCR was performed to assess transfection efficiency (Fig. 4A) before co-culturing with GBM cells. The levels of pro-angiogenic mediators were measured by qRT-PCR, indicating that CXCL3 overexpression significantly enhanced expressions of pro-angiogenic factors VEGF-A and MMP9 in GBM cells (Fig. 4B). Conditioned media from GBM cells were collected for HUVEC culture, and it was found that CXCL3 overexpression markedly increased tube formation ability of HUVECs (Fig. 4C). This further confirmed that CXCL3 derived from BMDMs facilitated angiogenesis in GBM. Subsequently, a series of experiments were done to explore effects of CXCL3 derived from BMDMs on biological behavior of GBM cells. CCK-8 assay demonstrated that elevated CXCL3 levels enhanced GBM cell viability (Fig. 4D), and similar results were observed in colony formation assays (Fig. 4E). Transwell assays showed that CXCL3 overexpression increased migratory and invasive capabilities of GBM cells (Fig. 4F). Finally, flow cytometric analysis of apoptosis revealed that the proportion of apoptotic cells in the oe-CXCL3 group was significantly lower than in the control group (Fig. 4G). In summary, CXCL3 derived from BMDMs not only promoted angiogenesis but also enhanced GBM cell viability, proliferation, migration, and invasion.

Fig. 4.

CXCL3 derived from BMDMs promotes angiogenesis and affects malignant progression of GBM. (A) qRT-PCR analysis of CXCL3 mRNA expression; (B) qRT-PCR analysis of VEGF-A and MMP9 mRNA expression; (C) Angiogenesis assay assessing HUVEC tube formation; (D) CCK-8 assay measuring cell viability; (E) Colony formation assay measuring cell proliferation; (F) Transwell assay assessing cell migration and invasion; (G) Flow cytometric analysis of apoptosis. *p<0.05, indicating a significant difference.

Discussion

The median survival period for GBM, the most prevalent and malignant primary brain tumor in adults, is around 15 months.⁽¹⁹⁾ Its unique tumor and immune microenvironment, along with its biological characteristics, present significant challenges for the development of new therapies.⁽²⁰⁾ Therefore, there is an urgent need for innovative experimental designs that incorporate biomarker enrichment strategies to ultimately improve prognosis of GBM patients. In this study, we utilized scRNA-seq technology to comprehensively identify heterogeneous cell populations in GBM and demonstrated the critical role of ELR CXCL genes in promoting angiogenesis and tumor progression. This provides new molecular targets and therapeutic strategies for developing targeted treatments for GBM, potentially leading to breakthrough advancements in this field.

In our research, we analyzed scRNA-seq data from GEO database and found that myeloid cells comprised the largest proportion of cells in GBM. As GBM progresses from low-grade to high-grade, the proportion of BMDMs within myeloid cells increases. The infiltration of BMDMs into GBM is a hallmark of the disease and plays a key role in tumor progression.⁽²¹⁾ Consistent with our findings, Li et al.⁽²²⁾ reported that BMDMs from high-grade GBM had a higher representation among immune infiltrating cells. This highlights the critical role of BMDMs in GBM progression. BMDMs influence tumor progression through various mechanisms, such as suppressing T cell activation and proliferation or promoting the polarization of M2 macrophages.⁽²³⁾ BMDMs directly impact angiogenesis by secreting factors like VEGF and bFGF.⁽²⁴⁾ Notably, Xu’s research⁽²⁵⁾ confirmed that BMDMs might drive tumor growth via IL-6/gp130/STAT3 signaling pathway, providing likely intervention targets for tumor therapy. Therefore, more investigations into specific mechanisms of BMDMs in GBM are crucial for advancing our understanding of tumor biology.

Compared to LGG, the rapid growth of tumor cells in GBM results in significantly shorter survival times for patients. Angiogenesis plays a vital role in occurrence and progression of gliomas and is strongly related to disease severity.⁽²⁶⁾ We identified a subset of ELR CXCL + BMDMs within myeloid cells, which specifically express multiple ELR CXC chemokines. ELR CXC chemokine genes are typically associated with angiogenesis and are important in tumor growth and development.⁽²⁷⁾ By analyzing expressions of select ELR CXCL genes and their pseudotime expression curves, we found that CXCL3, CXCL5, and CXCL8 were relatively highly expressed. Increased CXCL8 secretion activates NF-κB signaling, fostering angiogenesis in gastric cancer.⁽²⁸⁾ Research by Mao et al.⁽²⁹⁾ indicated that CXCL5 drives tumorigenesis and angiogenesis in GBM via JAK-STAT/NF-kappaB pathway. However, the role of CXCL3 in GBM angiogenesis has been less explored. Our study addresses this gap, confirming the pro-tumorigenic effects of CXCL3 derived from BMDMs in GBM. CXCL3 mediates tumor growth, migration, angiogenesis, and leukocyte recruitment.^(12,30) CXCL3 is overexpressed in several cancers, including cervical, colorectal, and head and neck squamous cell carcinoma, and is implicated in dismal prognosis.^(31–33) However, its role in GBM remains under-researched. Our findings established that CXCL3 from BMDMs significantly contributed to angiogenesis and tumor progression in GBM, providing scientific evidence for targeting CXCL3 in GBM therapy, thereby improving patient outcomes.

In summary, this study analyzed scRNA-seq data to reveal heterogeneity of GBM cell subpopulations and identified the important ELR CXCL + BMDMs associated with rapid GBM growth. The CXCL3 secreted by this subgroup influenced malignant progression by driving angiogenesis and manipulating biological behaviors of GBM cells. This discovery offers new directions for anti-angiogenic therapy in GBM and underscores the importance of exploring roles of BMDMs in cancer progression and treatment. Future research could employ multi-omics analysis, integrating gene expression, proteomics, and metabolomics data to comprehensively understand the mechanisms of BMDMs in GBM.

Author Contributions

Project administration: FG; Data curation: HW; Formal Analysis: YH and Junjun Zhang; Investigation: Junjun Zhang; Methodology: YH and Jianfei Zhang; Resources: FG; Software: HW and Jianfei Zhang; Supervision: YH; Validation: HW and Junjun Zhang; Visualization: HW; Writing – original draft: FG; Writing – review & editing: YH and Jianfei Zhang.

Data Availability Statement

The data and materials in the current study are available from the corresponding author on reasonable request.

Funding

This study was supported by grants from Ningbo Top Medical and Health Research Program (No. 2022020304).

Ethics Approval and Consent to Participate

Ethical approval is not required for this study in accordance with local or national guidelines.

Consent to Participate Statement

Patient consent was not required in accordance with local or national guidelines.

Conflict of Interest

No potential conflicts of interest were disclosed.