Dataset Information

Deciphering hypoxia related glioblastoma-associated macrophages and predicting prognostic markers by single-cell, spatially-resolved transcriptomics and experimental analysis

ABSTRACT: Glioblastomas (GBM), the most common and lethal type of primary brain tumor in adult, currently lack effective treatment and prognostic indicators. Hypoxia，a pathology promoting tumor progression in most solid tumors, has been reported to exhibit a negative correlation with GBM prognosis. Hypoxia related tumor-associated macrophages (hTAMs), the major cellular component of hypoxic regions, have not being systematically studied. We characterized the transcriptomic identity of hTAMs at the single-cell level, simultaneously discovering signatures indicative of poor outcomes. Immunofluorescent staining and spatially resolved transcriptomics revealed pronounced disparities regarding the morphology and distribution of TAMs between hypoxic and non-hypoxic regions of GBMs. We identified ARL4C and HSPA5 as prognostic indicators, exhibiting spatial patterns congruent with hypoxic regions and predicting worse outcomes, verified by qPCR and cell culture. Intercellular communication analysis using single-cell-omics revealed that hypoxia related GBM tumor cells (hGBM cells) exploit the TIMP1/LRP1 ligand-receptor axis to modulate hTAMs, resulting in a worse prognosis. Our study has offered clinically prognostic indicators, along with potentially therapeutic targets for GBMs.

ORGANISM(S): Homo sapiens

PROVIDER: GSE251908 | GEO | 2024/01/02

REPOSITORIES: GEO

ACCESS DATA

Similar Datasets

Project description:Tumor hypoxia is linked to poor outcome for many cancers, but the underlying mechanisms and the environmental factors that initiate tumor hypoxia are poorly understood. Here, we tracked tumor hypoxia in glioblastoma (GBM), a highly malignant brain cancer, in immunocompetent mice with a sensitive fluorescent reporter combined with single cell transcriptomics. We found that hypoxic GBM cells are quiescent and display a mesenchymal transition, both linked to malignant potency. We also captured in vivo hypoxia gene signature, which is more represented in recurrent GBM and predicts worse outcome. Interestingly, hypoxic GBM cells is a diverse population, consisted of four subclusters, and enriched for immune pathways. Concordantly, our reporter highlighted a distinct geographic pattern of immune cells in hypoxic regions, with phagocytic tumor-associated macrophages (TAMs) and CD8+ cytotoxic T cells (CTLs) congregated in hypoxic cores confined by hypoxic GBM cells in pseudopalisading patterns. Mechanistically, this is a dynamic temporospatial process, requiring cytokine CCL8. Remarkably, the sequestered TAMs also experience hypoxia, and they are reprogrammed to an immunotolerant state by factors released from hypoxic GBM cells. Contrary to the conventional viewpoint that hypoxia arises from rapid tumor expansion outstripping vascular supply, we discovered anticancer immunity as an important driving force of tumor hypoxia. Attenuating immune responses by implanting GBM in host mice with immunodeficiency or IL1β deletion significantly decreased GBM hypoxia in well-established tumors. Unexpectedly, radiation therapy (XRT) greatly reduced tumor hypoxia, and when combined with evofosfamide, a prodrug activated by hypoxia, achieved enhanced efficacy in tumor shrinkage and eradication of hypoxic GBM population. Analyses of human patient GBM samples highlighted a connection of mesenchymal subtype, immune response, and tumor hypoxia, all contributing to poor survival. Altogether, our study revealed a reciprocal influence of anti-tumor immunity and tumor hypoxia, which has important ramification for prognosis and immunotherapy for GBM.

Project description:Oxygen fluctuation during tissue remodeling imposes a major metabolic challenge in human tumors. Stem-like tumor cells in glioblastomas are believed to possess extraordinary metabolic flexibility, enabling them to initiate growth even under non-permissive conditions. To identify adaptive response mechanism, we compared the effects of acute and chronic hypoxia versus oxygenation on stem-like glioblastoma (GS) cells. GS cell lines were established and propagated either under normoxia (21% O2) or hypoxia (1% O2) and subsequently exposed to acute normoxia or hypoxia, respectively. Gene expression profiling revealed that acute hypoxia predominantly induced metabolic pathways and cell cycle arrest, whereas chronic hypoxia activated neurodevelopmental processes. In particular, we found increased expression of glycolytic enzymes, especially of the preparatory phase of glycolysis, under acute hypoxia, whereas pentose phosphate pathway (PPP) enzymes were downregulated. The opposite was found for acute oxygenation of hypoxic GS cells. Findings were confirmed by qPCR and immunoblot analyses. Despite downregulation by hypoxia, expression of PPP enzymes is increased in GBMs compared to normal brain, whereas expression of hypoxia-inducible enzymes of the parallel preparatory phase of glycolysis is decreased. Immunohistochemistry revealed strong staining for PPP enzymes in the bulk of GBM tissue, especially in highly proliferative areas, but not in pseudopalisading (hypoxic) cells. Glycolytic enzymes displayed an inverse pattern. Mass spectrometric analysis using [1,2-13C2]-D-glucose showed reduced glucose flux through the PPP under hypoxia in favor of flux through glycolysis. Acute and chronic hypoxia increased cell migration but reduced proliferation, whereas normoxia had opposite effects. Our findings extend Warburg’s observations by showing that in most tumor cells the PPP, which supplies metabolites for biomass production, is favored over the parallel preparatory phase of glycolysis, but is suppressed under acute severe hypoxia, causing a switch to direct glycolysis to protect against hypoxic stress.

Dataset Information

Deciphering hypoxia related glioblastoma-associated macrophages and predicting prognostic markers by single-cell, spatially-resolved transcriptomics and experimental analysis

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure