FUNCTIONAL DISSECTION OF HISTONE H3 LYSINE 27 METHYLATION IN GLIOMAGENESIS

Gliomas represent 80% of the central nervous system tumors. Historically, they have been classified by the World Health Organization according to their supposed cell of origin and to their malignancy grade. Among gliomas, astrocytomas are the most common and are classified histologically with a four grade system where grade IV corresponds to the most malignant form, also known as glioblastoma multiforme (GBM). Furthermore, in recent years a molecular characterization has been defined for GBM. Different subtypes have been identified, each of them over expressing a specific group of genes. For decades, it has been assumed that cancer was caused only by genetic alterations. Now the view is changed and cancer is considered both a genetic and epigenetic disease. It was demonstrated indeed that epigenetic silencing of genes through DNA or histone methylation at their promoters can be an alternative way to achieve their loss of function; in addition, DNA demethylation of constitutive heterochromatin can promote genome instability. In this context, we refer to epigenetics as the sum of heritable changes in phenotype and/or gene expression without altering the primary DNA sequence. Methylation of lysine 27 on histone H3 is a post-translational modification that is mediated by the histone methyltransferase complex known as Polycomb Repressive Complex 2 (PRC2) through its active subunit, Enhancer of Zeste Homolog 2 (EZH2). This facultative heterochromatin mark promotes the recruitment of Polycomb Group (PcG) proteins to achieve gene silencing. PcG proteins have been shown to play a major role in embryonic development and adult somatic cell differentiation. Initial studies on embryonic stem cells (ESCs) showed that Polycomb Complexes are required to maintain stem cell identity. However further investigation showed that this process is much more elaborated and the current model proposes that PcG proteins function dynamically during development and differentiation to lock off the expression of alternative cell fate regulators in any particular lineage. Moreover EZH2 and other Polycomb members have been found to be dysregulated in a variety of cancer types. EZH2 is overexpressed in tumors like prostate, breast and bladder, and BMI1, a member of Polycomb Repressive Complex 1 (PRC1), is over expressed in GBM, causing aberrant expression of neural stem cell (NSC) markers and preventing apoptosis. It was also demonstrated that genes that are directly regulated by Polycomb in ESCs are up to 12-fold more likely to have cancer specific DNA hypermethylation at their promoters than other genes. In my project I have pursued two main questions. First, I wanted to assess whether EZH2 is required for glioma initiation. Moreover, since tumor-initiating cells have been found also in GBM, I wanted to check if EZH2 is required to maintain this stem cell pool. In order to address those questions, I chose a well established animal model of GBM which relies on the loss of Ink4a/Arf together with the over expression of EGFRvIII. A conditional knockout allele for Ezh2 was introduced in order to remove this protein at different stages of the disease. With this work I could demonstrate that while EZH2 is required for the establishment of the tumor, it can be dispensable for its maintenance. The Polycomb axis acts early on during tumor formation, causing the relocation of H3K27me3 in an instructive manner and this process seems to be fundamental in order to achieve the full transformation of the cells. I was able to show that EZH2 is dispensable for glioma maintenance, pointing to a unique window of Polycomb sensitivity that characterizes the primary phase of gliomagenesis prior to the establishment of the glioma propagating cell (GPC) compartment that is able to reconstitute tumors.