Glioblastoma Hijacks Microglial Gene Expression to Support Tumor Growth

Background Glioblastomas are the most common and lethal primary brain tumors. Microglia, the resident immune cells of the brain survey their environment and respond to pathogens, toxins, and tumors. Glioblastoma cells communicate with microglia, in part by releasing extracellular vesicles (EVs). Despite the presence of large numbers of microglia in glioblastoma, the tumors continue to grow, and these neuroimmune cells appear incapable of keeping the tumor in check. To understand this process, we analyzed gene expression in microglia interacting with glioblastoma cells. Methods We used RNASeq to analyze the expression patterns of genes involved in key microglial function of microglia in mice with glioblastoma. We focused on microglia that had taken up tumor-derived EVs and therefore were within and immediately adjacent to the tumor. Results We show that these microglia have downregulated expression of genes involved in sensing tumor cells and tumor-derived danger signals, as well as genes used for tumor killing. In contrast, expression of genes involved in facilitating tumor spread was upregulated. These changes appear to be in part EV-mediated, since intracranial injection of EVs in normal mice led to similar transcriptional changes in microglia. We observed a similar microglial transcriptomic signature when we analyzed datasets from human patients with glioblastoma. Conclusion Our data define a MicrogliaGlioblastoma specific phenotype, whereby glioblastomas have hijacked gene expression in the neuroimmune system to favor avoiding tumor sensing, suppressing the immune response, clearing a path for invasion and enhancing tumor propagation. For further exploration we developed an interactive online tool at www.glioma-microglia.com with all expression data and additional functional and pathway information for each gene. mRNA expression of microglia isolated from control and tumor-bearing brain were analyzed by RNAseq, using Illumina Nextseq and MiniSeq.