Glioblastoma (GBM) is the deadliest primary brain tumour in adults with a median survival of
14-20 months from initial diagnosis. Despite aggressive treatment involving maximal surgical
debulking followed by radiation therapy and chemotherapy, GBM remains incurable owing
to a high rate of fatal recurrence. Understanding why unresected tumour cells survive
treatment is necessary to better treat this disease. GBM tumours recur due to the presence
of treatment-resistant cells, and this resistance phenotype is posited to be mediated by
several epigenetic mechanisms including DNA methylation, histone modifications and
chromatin remodelling. Work within the Glioma Genomics group in Leeds has specifically
highlighted a potential role for histones, so this study focuses on understanding the role of
histone modifications in GBM treatment resistance. More specifically, this group has recently
proposed, for the first time, that Jumonji and AT- Rich Interacting Domain 2 (JARID2) plays a
role in tumour recurrence via chromatin remodelling in GBM. JARID2 is an accessory protein
of Polycomb Repressive Complex 2 (PRC2) which is the sole complex responsible for
trimethylation on lysine 27 of histone H3 (H3K27me3). JARID2 promotes PRC2 recruitment to
chromatin and regulates its enzymatic activity. PRC2 and JARID2 have a fundamental role in
neurodevelopment but the role in GBM treatment resistance needs to be investigated. This
can be achieved by generating genome-wide profiles of histone modifications associated with
JARID2, and the binding of JARID2 itself, in paired primary (untreated) and recurrent (posttreatment) samples. Despite the tremendous work in generating global genome-wide
profiling of histone modifications in various types of cancers, few genome-wide maps for
H3K27me3 are available for GBM, therefore, current interest is placed on generating and
comparing genome-wide mapping of histone modifications to locate and identify key
epigenetic changes that are associated with GBM development and progression. Another
histone mark (H3K4me3, which is a transcriptional activator) is known to work in concert with
H3K27me3 during cell lineage determination in the brain, so it was also deemed necessary to
prolife this mark. Thus, this study aimed to establish a workflow for generating and
comparing the global distribution patterns of two histone modifications, along with binding
patterns of the catalytic component of PRC2 (Enhancer of zeste homolog 2: EZH2) and JARID2,
in paired primary and recurrent GBM samples. My hypothesis is that histone remodeling is
driving the changes in the gene expression observed in GBM through treatment. I established
iv
a workflow and then generated a genome- wide chromatin landscape for H3K27me3,
H3K4me3 and EZH2 binding from matched fresh frozen pair primary and recurrent GBM
samples of our in-house dataset. Then, I performed an integrative analysis on histone marks
along with EZH2 by correlating their modifications with the changes in gene expression. The
analysis was performed on a subset of genes that have been found to be dysregulated in
GBM’s patient following standard treatment due to the epigenetic remodeling of their
promoters. The findings revealed that these genes are significantly found in the bivalent state,
but the balance of histone marks is altered by therapy. Also, it revealed that histone
modifications are driving gene expression in those genes more than the others This leads to
the hypothesis that this bivalency is what causes the tumours to be able to adapt to
treatment. I concluded that JARID2 genes facilitate tumour recurrence through
transcriptional reprogramming in patients following standard therapy. Additionally, it implies
that bivalent areas promote GBM tumorigenicity and are linked to chemo-resistance.
