P04.69 Differential microglia - glioma cell interaction during tumor progression

Abstract

Background

Microglia cells resemble the innate immune system of the brain, which makes them prime candidates for interactions with glioma cells, and thus potential targets for therapeutic interventions. Their role in glioma invasion and progression is contradictory, implying heterogeneous actions of this cell type. An impact on the formation and function of the recently discovered progression- and resistance-promoting long cellular processes of glioma cells, tumor microtubes (TMs), has not been assessed.

Material and Methods

Patient-derived glioblastoma stem-like cell (GBMSC) lines with an invading growth pattern were implanted under a cranial window in CX3Cr1-EGFP knock-in mice. Further, a syngeneic model (GL261) was used to validate the findings in an immunocompetent glioma mouse model. Repetitive in vivo two-photon laser scanning microscopy was used to gain dynamic insights into how microglia and glioma cells interact during brain colonization.

Results

Microglia cells form a dense network of inter-connected cells resting in the brain parenchyma as well as in the perivascular space that is invaded by glioma cells after cortical injection. Microglia density was stable during the process of glioma cell infiltration (8666 cells/mm³ before infiltration vs. 8590 cells/mm³ densely infiltrated region). A diverse pattern of microglia - glioma cell interactions was observed, with direct cell-to-cell connections, as well as co-localization of cell processes of microglia cells and TMs of glioma cells. Direct somatal microglia - glioma cell interactions were more frequently observed in the perivascular niche compared to the brain parenchyma (64.2% vs. 37.0%; p=0.05, t-test). Inhibiting the function of microglia cells by using CX3Cr1-EGFP ^+/+ mice with a functional loss of the chemokine receptor lead to a higher invasion speed of the glioma cells in contact with microglia cells in the perivascular space vs. ^+/- mice (6.0 µm/h vs. 3.5 µm/h; p=0.036, Mann-Whitney U test) but not in the other brain parenchyma. Inhibition of the CSF-1 pathway by PLX3397 profoundly reduced microglia cell function, again with a pro-invasive effect on glioma cells in the perivascular space (10.2 µm/h vs. 3.7µm/h, treatment vs. control; p=0.021, Mann-Whitney U test). In all experimental paradigms, there was no apparent impact of microglia interactions on TM formation and dynamics, and TM-dependent tumor cell invasion and proliferation outside the perivascular niche. Tracing invading tumor cells in a syngeneic GL261 tumor model confirmed the specific importance of microglia interactions for perivascular niche invasion.

Conclusion

Functional microglia cells inhibit the perivascular route of glioma cell invasion in vivo, but have no apparent impact on TM formation and function. In consequence, microglia inhibition appears counterproductive when the invasive nature of the disease is the therapeutic target.