Tumor-associated macrophages in meningiomas: a novel biomarker for poor survival outperforming the benefits of T cells

Abstract

Tumor-associated macrophages (TAMs) represent the main immune cell population in various brain malignancies, but there is rare knowledge on the functional and, in particular, the prognostic role of TAMs in the meningioma (MGM) microenvironment. Here, we investigated TAM frequencies, activation state, survival-associated changes, and their association with tumor-infiltrating T lymphocytes (TILs) in two independent study samples comprising altogether 680 MGMs. To this end, we performed tissue cytometry analyses, quantified tissue cytokine levels, and integrated previously published TIL infiltration and microarray datasets in the discovery cohort comprising n = 195 clinically well-annotated cases. This was complemented by a DNA methylation-based deconvolution approach to predict TAM and TIL infiltration rates using immune cell-specific CpG sites as well as survival associations in an independent validation cohort of n = 485 MGMs. Our findings revealed substantial but heterogeneous TAM infiltration in newly diagnosed MGMs, with increased numbers of pro-tumoral TAMs in clinically aggressive tumors. Additional cytokine and transcriptome analyses corroborated the presence of an immunosuppressive niche in TAM-enriched MGMs. Importantly, a high frequency of pro-tumoral TAMs was associated with poor patient outcome, and high TAM infiltration was further identified as an independent prognostic factor for inferior survival, counteracting the beneficial prognostic effect of TILs. Moreover, methylation-based deconvolution analyses confirmed the opposing prognostic roles of TAMs and TILs in the validation cohort. Altogether, higher numbers of TAMs appear to be a hallmark of clinically aggressive behavior in newly diagnosed and recurrent MGMs. Unlike TILs, immunosuppressive TAMs seem to play a dominant role in the immunological landscape of MGMs with a significant negative impact on patient outcome, highlighting pro-tumoral TAMs to be an attractive treatment target in MGMs. Furthermore, our deconvolution approach presents a pipeline to computationally determine TAM and TIL infiltrates in the MGM microenvironment, which might be highly valuable for patient stratification for future immunotherapeutic treatments.