Development of brain tumors from neural stem/progenitor cells

Brain tumors are neoplasms of the central nervous system (CNS) and are classified according to their histopathology. Current therapies rely on this classification, but the exact factors specifying these tumor types are not known. Identifying the cellular origin and molecular basis of brain tumor development will contribute to the advancement of new targeted therapies. This thesis investigated the connection between normal neural stem/progenitor cells and brain tumors with respect to the cellular origin of brain tumors and molecular events contributing to their development. The first part of this work focused on surface markers of neural stem/progenitor cells in the lateral ventricle wall (LVW) and spinal cord (SC), which allowed the prospective isolation and investigation of such cells. CD133 was discussed as a “neural stem cell marker” and a phenotypic hint toward a possible lineage relationship between neural stem cells and CD133-positive brain tumor stem cells. A thorough investigation of neural stem cell types carrying this surface marker was however missing at the beginning of this thesis. In this work, CD133 was identified on two cell types in the postnatal LVW region: post-mitotic ependymal cells – the vast majority of CD133-positive LVW cells – and a subpopulation of ventricle-contacting astrocytic stem cells. In addition, CD133-positive ependymal cells in the adult SC were investigated. In contrast to ependymal cells in the LVW, SC ependymal cells showed in vitro self-renewal and multipotency. Comparative gene expression analysis of both cell types revealed that SC ependymal cells express certain genes which likely contribute to their stem cell properties. Furthermore, several genes were found upregulated in SC ependymal cells which have previously been reported as signature genes in ependymomas – i.e. tumors of the SC, which may originate from SC ependymal cells. The observed molecular and functional similarities between self-renewing stem cell types and tumor cells indicate a possible derivation of brain tumors from a normal stem/progenitor cell type. In the second part of this work, two main questions were addressed: 1) Which oncogenic factors are sufficient to induce brain tumor development from normal neural stem/progenitor cells of the LVW, and 2) Do genetic events direct brain tumor phenotypes? The combination and order of HRAS and MYC over-expression in Trp53-deficient neural stem/progenitor cells was found to instruct the development of gliomas, CNS PNETs or atypical teratoid/rhabdoid (AT/RT)-like tumors. AT/RT-like tumors histologically resembled human AT/RTs and the gene expression profile of both of these tumors indicated an activation of the unfolded protein response (UPR). This cellular pathway is induced upon stress conditions like hypoxia or nutrient-deprivation and has an essential and supportive role for rapidly growing tumors like AT/RTs. AT/RTs are characterized by the loss of function of the tumor suppressor SMARCB1 and investigations in this thesis demonstrate that this loss leads to an increased sensitivity toward eIF2alpha phosphorylation, which is a central UPR component. Based on these findings, an interference with the UPR is suggested as a novel strategy for the treatment of SMARCB1-deficient tumors. In contrast to MYC and HRAS, the over-expression of other candidate oncogenes, like Ezh2, FoxM1 or Bmi1 failed to induce tumor development. This work demonstrates that Bmi1 over-expression leads to an increased neurosphere cell self-renewal and proliferation and a decrease in cell death. These BMI1 effects are relevant for the maintenance of tumor cells. How BMI1 exerts its functions remains only partly understood, and in this thesis, four novel BMI1 target genes are identified, which – based on their known functions - could contribute to the described BMI1 effects.