Unveiling novel cell clusters and biomarkers in glioblastoma and its peritumoral microenvironment at the single-cell perspective

Abstract

Background: Glioblastoma (GBM) is a highly heterogeneous, recurrent and aggressively invasive primary malignant brain tumor. The heterogeneity of GBM results in poor targeted therapy. Therefore, the aim of this study is to depict the cellular landscape of GBM and its peritumor from a single-cell perspective. Discovering new cell subtypes and biomarkers, and providing a theoretical basis for precision therapy.

Methods: We collected 8 tissue samples from 4 GBM patients to perform 10 × single-cell transcriptome sequencing. Quality control and filtering of data by Seurat package for clustering. Inferring copy number variations to identify malignant cells via the infercnv package. Functional enrichment analysis was performed by GSVA and clusterProfiler packages. STRING database and Cytoscape software were used to construct protein interaction networks. Inferring transcription factors by pySCENIC. Building cell differentiation trajectories via the monocle package. To infer intercellular communication networks by CellPhoneDB software.

Results: We observed that the tumor microenvironment (TME) varies among different locations and different GBM patients. We identified a proliferative cluster of oligodendrocytes with high expression of mitochondrial genes. We also identified two clusters of myeloid cells, one primarily located in the peritumor exhibiting an M1 phenotype with elevated TNFAIP8L3 expression, and another in the tumor and peritumor showing a proliferative tendency towards an M2 phenotype with increased DTL expression. We identified XIST, KCNH7, SYT1 and DIAPH3 as potential factors associated with the proliferation of malignant cells in GBM.

Conclusions: These biomarkers and cell clusters we discovered may serve as targets for treatment. Targeted drugs developed against these biomarkers and cell clusters may enhance treatment efficacy, optimize immune therapy strategies, and improve the response rates of GBM patients to immunotherapy. Our findings provide a theoretical basis for the development of individualized treatment and precision medicine for GBM, which may be used to improve the survival of GBM patients.

Keywords: Biomarkers; Glioblastoma; Single-cell sequencing; Tumor heterogeneity; Tumor microenvironment.

Fig. 1

Preliminary exploration of the GBM microenvironment. A Flowchart. B Magnetic Resonance Imaging in Patient 1. C Biological markers for each cell cluster. D The uniform manifold approximation and projection (UMAP) plots of cell clusters in GBM, 24 clusters (top left), patient origin (top right), site of origin (bottom left), annotated cell subgroups (bottom right). Each point represents a cell, with different colors indicating different cell clusters, patients, and origins. E inferCNV, where red represents copy number gain, and blue represents copy number loss. F Percentage distribution of cell clusters after annotation, with different colors representing different cell clusters. G Differences in the content of cell clusters between the tumor and its surrounding area

Fig. 2

The role of subtypes of oligodendrocytes in the GBM microenvironment. A UMAP plots of oligodendrocyte cell clusters in the GBM microenvironment, 8 clusters (top left), patient origin (top right), site of origin (bottom left), annotated cell subgroups (bottom right). Each point represents a cell, with different colors indicating different cell clusters, patients, and origins. B One gene from the top 20 genes in each cell cluster is selected as a biomarker for that specific cell cluster. C and D Percentage distribution of cell clusters after annotation, with different colors representing different cell clusters. E Gene set variation analysis (GSVA) in oligodendrocytes (tumor versus peritumor). F GSVA, a heatmap showing pathway enrichment for each cell cluster. G SCENIC, a transcription factor heatmap for each cell cluster. H and I Top 5 transcription factors for both tumor and peritumor. J The expression of proliferative genes in each cell cluster. (K) Expression levels of STMN1 and MT-CO2 in OL_C1_MT-CO2 in the tumor and peritumor. L Differential analysis in OL_C1_MT-CO2 using Wilcoxon test, with avg_log2FC > 0.4 and p_val_adj < 0.05. Red indicates high expression, while blue indicates low expression

Fig. 3

Functional enrichment and cell trajectory inference in oligodendrocytes. A GO functional enrichment of OL_C1_MT-CO2 in the tumor and peritumor. B KEGG functional enrichment of OL_C1_MT-CO2 in the tumor and peritumor. C Protein–protein interaction network of OL_C1_MT-CO2 in the tumor and peritumor. D Cell trajectory inference, with each point representing a cell. The gradient from deep blue to light blue indicates time progression from early to late (left), and different colors represent different cell clusters (right). E Dynamic expression of top 5 genes in oligodendrocyte cell clusters, with different colors representing different clusters. F Dynamic expression of top 100 genes in oligodendrocyte cell clusters, with a gradient from blue to red indicating expression levels from low to high

Fig. 4

The function of myeloid clusters in the GBM microenvironment. A UMAP plots of myeloid cell clusters in the GBM microenvironment, 9 clusters (top left), patient origin (top right), site of origin (bottom left), annotated cell subgroups (bottom right). Each point represents a cell, with different colors indicating different cell clusters, patients, and origins. B One gene from the top 20 genes in each cell cluster is selected as a biomarker for that specific cell cluster. C and D Percentage distribution of cell clusters after annotation, with different colors representing different cell clusters. E GSVA, a heatmap showing pathway enrichment for each cell cluster. F and H Top 5 transcription factors for both tumor and peritumor. G SCENIC, a transcription factor heatmap for each cell cluster. I UMAP plots of M1 and M2 scores, with each point representing a cell. The color gradient from deep purple to yellow signifies higher scores. J Differences of M1 and M2 scores in different myeloid cell clusters. K The expression of proliferative genes in each cell cluster. L Expression levels of TNFAIP8L3 in Myeloid_C2_TNFAIP8L3 in the tumor and peritumor. Expression levels of DTL, CDK1, STMN1 and MKI67 in Myeloid_C9_DTL in the tumor and peritumor

Fig. 5

Functions and developmental trajectories of myeloid. A GO functional enrichment of Myeloid_C2_TNFAIP8L3 in the peritumor. B KEGG functional enrichment of Myeloid_C2_TNFAIP8L3 in the peritumor. C Protein–protein interaction network of Myeloid_C2_TNFAIP8L3 in the peritumor. D and E Protein–protein interaction networks of Myeloid_C9_DTL in the tumor and peritumor. F Cell trajectory inference, with each point representing a cell. The gradient from deep blue to light blue indicates time progression from early to late (left), and different colors represent different cell clusters (right). G The cell trajectories for each myeloid cell cluster, where different colors represent different cell clusters. H Dynamic expression of top 100 genes in myeloid cell clusters, with a gradient from blue to red indicating expression levels from low to high

Fig. 6

The function of neuronal clusters in the GBM microenvironment. A UMAP plots of neuronal cell clusters in the GBM microenvironment, 14 clusters (top left), patient origin (top right), site of origin (bottom left), annotated cell subgroups (bottom right). Each point represents a cell, with different colors indicating different cell clusters, patients, and origins. B One gene from the top 20 genes in each cell cluster is selected as a biomarker for that specific cell cluster. C Markers for excitatory and inhibitory neurons. D and E Percentage distribution of cell clusters after annotation, with different colors representing different cell clusters. F GSVA, a heatmap showing pathway enrichment for each cell cluster. G and H Top 5 transcription factors for both tumor and peritumor. I Cell trajectory inference, with each point representing a cell. The gradient from deep blue to light blue indicates time progression from early to late (left), and different colors represent different cell clusters (right). J Dynamic expression of top 5 genes in neuronal cell clusters, with different colors representing different clusters

Fig. 7

The role of malignant cell clusters in the GBM microenvironment. A UMAP plots of malignant cell clusters in the GBM microenvironment, 14 clusters (top left), patient origin (top right), site of origin (bottom left), annotated cell subgroups (bottom right). Each point represents a cell, with different colors indicating different cell clusters, patients and origins. B One gene from the top 20 genes in each cell cluster is selected as a biomarker for that specific cell cluster. C and D Percentage distribution of cell clusters after annotation, with different colors representing different cell clusters. E Top 5 transcription factors for both tumor and peritumor. F The expression of proliferative genes in each cell cluster. G SCENIC, a transcription factor heatmap for each cell cluster. H RNA velocity, where each point represents a cell, different colors indicate distinct cell clusters, arrow direction represents the potential differentiation trajectory of cells, and arrow size indicates the strength of differentiation ability. I Differential analysis in Malignant_C14_SYT1 using Wilcoxon test, with avg_log2FC > 0.4 and p_val_adj < 0.05. Red indicates high expression, while blue indicates low expression. J GSVA in Malignant_C2_DIAPH3 (Malignant_C2_DIAPH3 versus other cell clusters). K Expression levels of SYT1 and STMN1 in Malignant_C14_SYT1 in the tumor and peritumor. (L and M) GO functional enrichment of Malignant_C14_SYT1 in the tumor and peritumor. N and O Protein–protein interaction networks of Malignant_C14_SYT1 in the tumor and peritumor

Fig. 8

Intercellular communication. A Interactions heatmap between cell clusters, where the color gradient from blue to red represents an increasing number of interactions. B and D Interactions between two cell clusters via ligand-receptor pairs, where larger points indicate smaller p-values. The color gradient from dark gray to red represents the average expression level of ligand-receptor pairs from low to high. C and E Nodes represent cell clusters, nodes of the same color indicate the starting point, and reaching another node represents the endpoint. The numbers on the edges represent the number of ligand-receptor pairs, with thicker lines indicating more pairs (top left). The outer circle represents cell clusters, and the inner circle represents ligands or receptors. Arrows indicate direction, with the thickness of the lines representing the expression levels of the originating genes. The arrow size represents the expression levels of the recipient genes. Light green indicates the originating direction, while dark green represents the receiving direction (bottom right)