Single-cell transcriptomics reveals immunosuppressive microenvironment and highlights tumor-promoting macrophage cells in Glioblastoma

Abstract

Glioblastoma (GBM) is the most prevalent and aggressive primary brain malignancy in adults. Nevertheless, the cellular heterogeneity and complexity within the GBM microenvironment (TME) are still not fully understood, posing a significant obstacle in the advancement of more efficient immunotherapies for GBM. In this study, we conducted an integrated analysis of 48 tumor fragments from 24 GBM patients at the single-cell level, uncovering substantial molecular diversity within immune infiltrates. We characterized molecular signatures for five distinct tumor-associated macrophages (TAMs) subtypes. Notably, the TAM_MRC1 subtype displayed a pronounced M2 polarization signature. Additionally, we identified a subtype of natural killer (NK) cells, designated CD56dim_DNAJB1. This subtype is characterized by an exhausted phenotype, evidenced by an elevated stress signature and enrichment in the PD-L1/PD-1 checkpoint pathway. Our findings also highlight significant cell-cell interactions among malignant glioma cells, TAM, and NK cells within the TME. Overall, this research sheds light on the functional heterogeneity of glioma and immune cells in the TME, providing potential targets for therapeutic intervention in this immunologically cold cancer.

Fig 1. Dissection of the tumor microenvironment in human GBM with scRNA-seq.

(A) UMAP plots of cells from the 48 tumor fragments profiled in this study, with each cell color coded to indicate the associated cell types. (B) The violin plot displayed the classic markers used for cell type annotation. (C) FeaturePlot displayed the classic markers used for cell type annotation. (D) UMAP plot of cells from the 48 tumor fragments profiled in this study, with each cell color coded to dataset. (E) Major cell type composition of each tumor fragment and dataset. (F) UMAP plot of malignant glioma cells, with each cell color coded to fragments. (G) Heatmap of CNA profiles inferred from scRNA-seq of tumor cells. Red indicated genomic amplifications, and blue indicated genomic deletions. (H) Heatmap of CNA profiles inferred from scRNA-seq of tumor from ndGBM_04 (left) and rGBM_O4_1 (right). T_NK and Myeloid cells were used as reference.

Fig 2. Cellular composition of myeloid cells.

(A) UMAP plots of myeloid cells, with each cell color coded to indicate the associated cell types. (B) Heatmap displayed the classic markers used for myeloid cell type annotation. (C) UMAP plots of five TAM subtypes, with each cell color coded to TAM subsets. (D) FeaturePlot displayed the classic markers used for TAM subtype annotation. (E) M1-like, M2-like macrophage, and pro_inflammatory scores of five TAM cell subtyps. (F, G) Key transcription factors across five TAM subsets and a cis-regulatory analysis using pySCENIC. (H) Different enrichment of ‘Hallmark’ pathways between TAM_MRC1 and TAM_ISG15 by GSEA analysis.

Fig 3. Cellular composition of T and NK cells.

(A) UMAP plots of T and NK cells, with each cell color coded to indicate the associated cell types. (B) Heatmap displayed the classic markers used for T and NK cell type annotation. (C) T_NK cell type composition of each tumor fragment and dataset. (D) UMAP plots of CD4 + T and CD8 + T cells, with each cell color coded to indicate the associated cell types. (E) FeaturePlot displayed the classic markers used for CD4 + T and CD8 + T subtype annotation. (F) T regulatory scores of six T cell subtypes. (G) FeaturePlot displayed the T regulatory scores in six T cell subtypes. (H) T exhaustion and cytotoxicity scores of six T cell subtypes.

Fig 4. Cellular composition of NK cells.

(A) UMAP plots of CD56bright NK cells, with each cell color coded to indicate the associated subtypes. (B) FeaturePlot displayed the classic markers used for CD56bright NK subtype annotation. (C) UMAP plots of CD56dim NK cells, with each cell color coded to indicate the associated subtypes. (D) FeaturePlot displayed the classic markers used for CD56dim NK subtype annotation. (E) NK cytotoxicity and stress scores of two CD56dim NK cell subtypes. (F) Different enrichment of KEGG and ‘Hallmark’ pathways between CD56dim_DNAJB1 and CD56dim_CCL3 by GSEA analysis.

Fig 5. Cell-cell communication among glioma and immunity cell types.

(A) Capacity for intercellular communication between malignant cells and immune cells. Each line color indicates the ligands expressed by the cell population represented in the same color (labeled). The lines connect to the cell types that express the cognate receptors. (B) Heatmap displayed the communication strength between two cell types. (C) Dot plot showing the average expression of highlighted ligands and receptors across cell types used for CellphoneDB analysis. (D) Cell-cell communication analysis using CellphoneDB. Depicted are the dot plots of ligand-receptor pairs for glioma-immunity and immunity-glioma signaling.