Preview

PDF, English - main document Download (44MB) | Terms of use

Abstract

Meningiomas are thought to arise from the arachnoid cells of the leptomeninx and make up the most common primary intracranial tumor in adults. They are usually benign, however in about 20 % of cases, tumors present with an aggressive phenotype and higher risk of recurrence. Risk stratification thereby remains challenging especially for NF2-mutated meningiomas, which make up about two thirds of all cases, as they can occur at the full spectrum of WHO grades in meningioma from 1 to 3. Recently, molecular profiling has gained importance for prognosis in meningioma with several classification systems that have been established mostly based on the DNA methylation of the tumors. However, the DNA methylation-based classification has not been extensively linked to phenotypic traits of the tumor. Nor have meningiomas been investigated regarding intratumoral subpopulations that may exist in parallel and may have different characteristics, especially regarding the stage of progression and ability to recur. In addition, the role of the immune microenvironment in meningiomas is poorly understood, despite the identification of an immune-enriched meningioma subgroup with beneficial outcome in two independent DNA methylation classification systems. In this dissertation, I investigated the consistency of subgroups initially defined on epigenomic level across molecular levels by comparison to transcriptomic and proteomic data. Further, I leveraged single nuclei transcriptomic profiling to dissect intertumoral differences in the expression profile specific to the tumor cell population depending on the tumor subgroup, and to investigate the abundance and phenotype of intratumoral tumor cell subpopulations across samples. Similarly, I analyzed the single nuclei transcriptomic data to characterize tumor-infiltrating immune cells with respect to their abundance and activation status. I furthermore correlated the differences in immune infiltration with the progression-free survival of patients by deconvoluting DNA methylation array data according to their cellular composition. These analyses underlined the coherence of epigenomic meningioma subgroups across transcriptome and proteome. Moreover, I identified six tumor cell subpopulations that were defined by distinct expression profiles und could be identified across samples at varying abundancies depending on the stage of progression. Similarly, I observed profound differences in infiltrating immune cells between tumor subgroups, with a significant enrichment of tumor-associated macrophages in a benign subgroup of NF2-mutated meningiomas as compared to more progressed tumors. In parallel to their abundancy, macrophages changed in activation between benign and malignant cases from an anti- to a pro-tumorigenic phenotype. The evaluation of progression-free survival data revealed a positive correlation to the proportion of infiltrating immune cells as estimated from epigenomic profiles. Altogether, these results highlight the role of multi-level molecular profiling for tumor grading in a paradigmatic, epidemiologically relevant tumor type. They further indicate an important role of tumor-infiltrating macrophages during meningioma progression with possible consequences for risk prediction as well as therapeutic targets in meningioma.