Tumour‐associated microglia/macrophages predict poor prognosis in high‐grade gliomas and correlate with an aggressive tumour subtype
In removed GBM tissue, tumour‐associated microglia/macrophages (TAMs) have been reported to constitute up to 30% 7. TAMs are recruited from the brain and bone marrow to the site of the tumour through the action of signalling molecules released by glioma cells and other cells in the microenvironment 8. Reportedly, glioma cells suppress the immune surveillance of TAMs by skewing them towards the pro‐tumourigenic M2 phenotype while inhibiting development of the anti‐tumourigenic M1 phenotype 8. However, recent studies indicate that glioma cells induce a mixed population of TAMs expressing both M1‐ and M2‐related molecules 11 possibly resembling a more undifferentiated M0 phenotype 21. In return, TAMs support tumour growth and progression 8 by stimulating proliferation 22, migration 23, invasion 12 and angiogenesis 8, overall indicating a complex bidirectional communication between glioma cells and TAMs.
Using immunohistochemistry, the number of TAMs was shown to increase with malignancy grade in gliomas 15. In addition, high amounts of TAMs expressing M2‐related markers, e.g. cluster of differentiation 204 (CD204) and 163 (CD163), have been associated with increasing WHO grade and poorer prognosis 15. However, these studies neither investigated the influence of M2 TAMs on survival in separate WHO grades nor performed multivariate analysis to examine the independent prognostic value. Based on these unresolved issues, our aim was to explore the prognostic impact of TAMs in a large population‐based glioma cohort by quantifying the expression of the microglial/macrophage markers ionized calcium‐binding adaptor molecule‐1 (IBA‐1) and CD204 using double immunofluorescence. IBA‐1 is considered a specific TAM marker 31 and CD204 a marker of M2‐like TAMs 9. We used an automated quantitative fluorescence approach enabling continuous measurements of area and intensity reducing some of the inter‐ and intra‐observer variability seen in conventional semi‐quantitative pathologist‐based scoring 32. Early on, we discovered that the distribution of TAMs in especially GBMs was highly heterogeneous. We, therefore, investigated the correlation between TAMs, gemistocytic tumour cells and GBM subtype as tumour aggressiveness may depend on both the specific molecular subtype and the level of gemistocytic tumour cells. In addition, we characterized the phenotype of CD204+ TAMs by double immunofluorescence using a panel of eight markers related to immune activation and tumour aggressiveness.