Search

Diffuse Intrinsic Pontine Glioma (DIPG) is a universally fatal paediatric brainstem tumour with a median overall survival of 9-11 months. The unique genetic, anatomical and developmental identity of the tumour means that there are currently no effective chemotherapies or targeted therapies. This thesis aimed to identify and characterise novel targets for the treatment of DIPG. Potential therapeutic opportunities arise from the frequent mutations found in histone H3 (H3K27M) and kinase ACVR1 found in DIPG tumours. A screen of epigenetic protein inhibitors revealed that the viability of the patient-derived DIPG cell line HSJD-DIPG-007 was highly sensitive to arginine methyltransferase (PRMT) inhibitors, particularly in ACVR1 mutant lines. The PRMT5 inhibitor LLY-283 was identified as the most promising lead for its potent inhibition of DIPG viability and the recognised role of PRMT5 in the regulation of the cell cycle and stemness. RNA-sequencing and phenotypic analyses revealed that LLY-283 did not induce cell death, but reduced stemness and invasiveness, altered metabolism and restored H3K27me3 levels in DIPG cells. Through collaboration with Ángel Montero Carcaboso, LLY-283 was tested in an orthotopic murine xenograft model of DIPG. LLY-283 reduced infiltration of DIPG cells into the murine forebrain, but did not prolong survival, perhaps reflecting inadequate drug exposure. In parallel, the CRISPR/Cas9 system was used to correct the causative G328V ACVR1 mutation in DIPG cells. The resultant isogenic DIPG cells demonstrated slowed proliferation after correction of ACVR1 mutations, rather than a complete growth arrest as observed with published shRNA mediated ACVR1 knockdown. Surprisingly, the isogenic cell lines also showed comparable sensitivity (GI50) to PRMT5 and ACVR1 inhibition, suggesting other genetic differences contribute to the variation in treatment sensitivity between DIPG cell lines. My findings highlight the value of isogenic DIPG cell lines for interrogating the effects of individual tumorigenic mutations. It remains incredibly challenging to develop new strategies to treat DIPG. This thesis supports the future design of combination treatments for DIPG and provides new experimental tools to aid their development.

Glioma is the commonest intrinsic brain tumor, and its high-grade form has a devastating prognosis. These tumors also arise in the spinal cord, carrying significant morbidity in children; however the genetics of these spinal gliomas is poorly understood. EGFRvIII is a common driver mutation in brain gliomas; it is unclear when this is acquired during glioma evolution and what its cooperative genetic drivers are. Here, we show that EGFRvIII initiates gliomagenesis in vivo; EGFRvIII leads to glioma precursors in the subventricular zone and brain surface, and later glioma formation in the brain and spinal cord. The long latency for tumor formation implies the need for additional mutations to drive gliomagenesis. In these tumors, we detected further genetic alterations including amplification of EGFRvIII, mutations of Trp53 and Tead2, and Cdkn2a deletion, through whole-exome sequencing. To shed further light on EGFR-cooperative genes for glioma progression, we conducted a genome-wide piggyBac transposon mutagenesis screen in vivo, which identified known glioma drivers (including Cdkn2a, Pten and Nf1) and novel putative partners, including genes that regulate neuronal differentiation such as Sox6 and Tcf12, and a novel regulator of the Ras pathway Spred1. RNA-sequencing confirmed the presence of fusion transcripts (transposon mediated effects) for these genes. We demonstrate the clinical relevance of these cooperative genes through comparison with large human glioma databases, demonstrating recurrent genetic alterations of these genes are in patient tumors implicating them as putative drivers, and we highlight that expression levels of Sox6 and Tcf12 correlate with patient prognosis. We show that there are shared and distinct mutated genes in brain and spinal gliomas. Although Pten is a well-known tumor suppressor for brain gliomas, it was not previously known whether Pten drives spinal gliomagenesis. Given recurrent transposon insertions in Pten were found in both brain and spinal gliomas, we generated conditional mice with EGFRvIII and Pten loss, demonstrating Pten accelerates spinal glioma formation. Our work elucidates the genetic evolutionary processes behind EGFRvIII-driven gliomas, provides a detailed genomic comparison between brain and spinal gliomas, and provides functional genomic datasets to help decipher complex human glioma genomes.

Recent advances in collecting sequencing data from tumors is promising for both immediate individual patient treatment and investigation of cancer mechanisms. A resultant central goal is identifying changes in tumors that are impactful towards these ends. Here, we develop two tools to identify impactful changes at different levels. We develop both methods in the context of gliomas, a common form of brain cancer. Firstly, we develop and assess a tool for assessing the impact of fusion genes, a type of common mutation using RNA-sequencing data. We validate the tool by working with collaborators in The Cancer Genome Atlas. Secondly, we develop a tool for an overall assessment of patient outcome by integrating data from diverse sequencing platforms. We validate this tool using simulation, data from consortiums, and collaborators at UCSF.

Glioblastoma (GBM), a genetically heterogeneous disease, has a significant burden on our society. Currently, the standard treatment for newly diagnosed GBM patients consists of surgery followed by concomitant radiotherapy and temozolomide chemotherapy, resulting in a median survival of 12-15 months. Targeted therapies are being developed to inhibit oncogenes based upon GBM molecular profiling, though have not been as successful as expected. Our understanding of DNA and RNA alterations in GBM has grown considerably over the past few years. However, our understanding of the lipid biology, specifically sphingolipids, in GBM is lagging and may prove useful in the arsenal of targeted therapies. The sphingolipid pathway contains lipid signalling molecules, which modulate cellular survival through the balance of ceramide, a pro-apoptotic metabolite, and Sphingosine-1-Phosphate (S1P), a pro-survival metabolite. Herein, I characterise for the first time the sphingolipid profile of normal grey matter (NGM), diffuse astrocytomas (AII), anaplastic astrocytomas (AIII), and GBM using liquid chromatography tandem mass spectrometry. The lipid profile is supported by an enzyme expression profile favouring ceramide catabolism and S1P formation, including upregulation of acid ceramidase (ASAH1) and sphingosine kinase 1 (SPHK1), and a down regulation of S1P phosphatase 2 (SGPP2). Significantly, C18 ceramide was reduced 5-fold in GBM compared to NGM, while S1P was increased in GBM by approximately 9-fold compared to NGM. Based on the sphingolipid profiles, ASAH1 and SPHK1 were assessed for functional relevance in vitro. Using gene silencing and pharmacological inhibition, I found SPHK1 to be critical for U87MG-induced angiogenesis through S1P paracrine signalling, which was independent of VEGF levels. EGFR mutations were associated with increased C16 and C22 ceramide levels. For the first time, I measured sphingolipid metabolites in plasma extracted from GBM patients and healthy controls. Elevated levels of S1P were found in GBM plasma and together with tumour S1P levels, were associated with a poor survival outcome. In contrast low S1P levels in tissue combined with MGMT methylation was associated with a good survival outcome. Overall, the data presented in this thesis reaffirm the importance of sphingolipid metabolism in GBM biology, reflected by a shift in the ceramide-S1P balance, favouring the pro-angiogenic S1P. Additionally, sphingolipid interactions with altered genetic pathways and potential biomarker capacity are novel findings that require further validation, with the hope of informing and monitoring therapeutic responses for GBM patients.

Brain tumors are highly lethal, most often resulting in a terminal prognosis. Today's standard-of-care has not significantly altered patient survival time. The status quo indicates the need for a more targeted, multi-functional therapeutic approach. Nanotechnology has been implicated as a potential magic bullet drug carrier solution to cancer, with magnetic iron oxide nanoparticles (MNPs) representing a unique, magnetically responsive type of nanomaterial. MNPs have been targeted to brain tumors by applying an external magnetic field to lesions. The short circulation times of preferred MNPs, though, limits their availability for targeting, and, thus, tumor delivery. If plasma residence were enhanced, the question then arises: would targeting efficiency improve? This dissertation aimed to develop long-circulating MNPs suitable for enhanced magnetic brain tumor targeting. Aminosilane (A) and cross-linked starch coated, polyethylene glycol (PEG) conjugated MNP candidates were synthesized using N-hydroxysuccinimide chemistry and characterized. Candidates were similar in size (140--190 nm) and PEG content, (up to 1.5 mol %) and deemed magnetically suitable for targeting. Candidates were subjected to in vitro reticuloendothelial system simulations to predict long-circulating behavior in vivo. Unlike the A-platform, cross-linked starch PEG-MNPs (D5/D20) performed well in vitro showing sustained size stability in cell culture medium and up to 10-fold reduced uptake in RAW 264.7 macrophages, when compared to previously studied starch-coated MNPs (D). Simulations predicted improved circulation behavior, with intravenously administered D5/D20 (12 mg Fe/kg) showing up to an estimated 150-fold increase in tumor exposure over D in 9L-glioma bearing rats. Validated electron spin resonance methodology was used to study biodistribution of D5/D20 in tissues of elimination. MNP concentrations in the spleen were high, and, unexpectedly, appeared attenuated in the liver. Improved exposure of MNPs to rat 9L-glioma lesions resulted in enhanced magnetic brain tumor targeting (t=1 hr) with up to 1.0% injected dose (ID)/g tissue (200 g animal) delivery achieved, a 15-fold improvement over targeted D (0.07% ID/g tissue). Magnetic resonance imaging confirmed targeting enhancements. Results achieve proof-of-concept that long-circulating MNPs permit enhanced magnetic targeting of nanoparticles to brain tumors, warranting further development of the materials and strategy studied.

Gliomas are the most common type of malignant brain tumor. Even with the recent progress in conventional therapies, the prognosis remains poor and development of effective immunotherapy is needed. Tumor vaccines using CpG and tumor lysate have been demonstrated as effective in glioma therapies. We hypothesized that tumor lysate grown in a physiologic 5 % O2 condition would increase immunogenicity compared to atmospheric 20 % O2 condition since glioma in situ has been demonstrated hypoxic; and hypoxic conditions can activate the danger signal to induce antitumor immunogenicity in addition to CpG. In this study, we characterized the immunogenicity of tumor lysate derived from 5 % O2 condition for the first time in comparison with 20 % O2 condition. Vaccination with lysate from 5 % O2 condition increased the numbers of several lymphocyte subsets at draining lymph nodes compared to the tumor lysate from 20 % O2 condition. Vaccination with lysate from 5 % O2 condition did not change the cytokine levels in the sera compared to the vaccination with lysate from 20 % O2. Tumor reactive antibody levels were increased with the amount of lysate used for vaccines; however, there was not a significant difference in antibody levels with 5 % O2 in comparison with 20 % O2. Lastly, using the tumor lysate from 5 % O2 condition had superior efficacy in inducing cytotoxicity against glioma compared to the tumor lysate from 20 % O2 condition. CD8+ T depletion showed that there were other cells that play a role in this cytotoxicity. Together, these data show that 5 % oxygen tumor lysate has distinct effects on immunogenicity compared to 20 % O2 lysate. These findings indicate a potential application in cancer treatment.

Human glioblastoma multiforme (GBM) is the most common malignant glioma in adults. GBM constitutes a major problem in clinical management due to its high infiltrative ability, resistance to the current cancer therapy that account for its bad prognosis. Specific target molecules produced in the cells are required to control the biology of malignant glioma cells and facilitate the treatment strategies. Many reports demonstrated that exogenous addition of the disialoganglioside GD3 induces apoptosis in various cell types, acting mainly through the mitochondrion. The overall aim of this work is to elucidate the biological roles of endogenously synthesized GD3 on U-1242MG glioblastoma cell line. We conducted experiments to elucidate the biological effects of endogenous GD3 expression on U-1242MG cells, and the results of these showed that the expression of GD3 was associated with apoptosis. After establishing that GD3 causes apoptosis, we investigated some of the mechanisms leading to apoptosis that are induced by three death ligands (FasL, TNF-α and TRAIL). We found that of the three death ligands only TRAIL-induced apoptosis in U-1242MG cells. We also found that three death receptors (DR5, Fas and TNF-R1) involved in the early stage of apoptosis are located in the caveolin-1 enriched fractions. Furthermore, TRAIL initiates the subcellular relocalization of several death receptors and other apoptotic molecules. We found that endogenously expressed GD3 is localized in the caveolae of U42-GRD2 cells, and that GD3 expression induced translocation of the death receptors Fas, TNF-R1, DR5 and other apoptotic molecules such as caspase-8 into the caveolae as a part of the molecular mechanism of GD3-induced apoptosis. Moreover, in contrast to caspase-8, the subcellular localization of cytochrome c was not affected by endogenous GD3 expression indicating that the mitochondrial pathway leading to apoptosis was probably not involved. From the above we conclude that the endogenous expression of GD3 induces apoptosis in U-1242MG glioma cells mediated through the extrinsic, not the intrinsic pathway.