Hypoxic stress dysregulates functions of glioma-associated myeloid cells through epigenomic and transcriptional programs

Abstract

Hypoxia rapidly alters gene expression to allow cellular adaptation to challenging conditions and support tumour growth. Hypoxia also affects the chromatin structure by modifications of histones and DNA methylation. Glioblastoma (GBM) is an aggressive, deadly primary brain tumour for which there is no effective treatment. The tumour microenvironment of GBM is highly heterogeneous, with infiltration of glioma-associated microglia and macrophages (GAMs) and the presence of necrotic, hypoxic regions which significantly impair effectiveness of therapies. The mechanisms through which hypoxia alters the tumour microenvironment and regulates functions of infiltrating immune cells remain poorly understood. Here, we show that hypoxia modulates the expression of myeloid markers in distinct ways: upregulates the expression of monocytic marker Lgals3 and downregulates the microglial markers P2ry12 and Tmem119 in microglial and monocytic GAMs in vitro and in vivo. Underlying genome-wide hypoxia-dependent transcriptomic changes in microglial cells were identified using microglia-glioma co-cultures and validated in human and mouse GBM single- cell transcriptomics datasets. Numerous GAM subtype markers are dysregulated in response to hypoxic stress due to associated changes in chromatin accessibility, as determined using ATACseq. While hypoxia alone drives a decrease of the overall chromatin accessibility at gene promoters, the exposure to glioma cells under hypoxic conditions leads to both increases and decreases of chromatin accessibility at promoter regions in microglial cells. Hypoxia downregulates the chromatin accessibility at the regions enriched in motifs for transcription factors known as master regulators of microglial cell identity and function, including SPI1 or IRF8. Overall, our results highlight the importance of hypoxic stress as a strong intratumoral regulator of myeloid cell functions, which adds a new dimension to the characterisation of particular GAM subpopulations.