Glioblastomas (GBMs) account for nearly half of all primary malignant brain tumours, and current therapies are often only marginally effective. Our understanding of the underlying biology of these tumours and the development of new therapies have been complicated in part by widespread inter- and intratumoural heterogeneity. To characterize this heterogeneity, we performed regional subsampling of primary glioblastomas and derived organoids from these tissue samples. We then performed single-cell RNA-sequencing (scRNA-seq) on these primary regional subsamples and 1-3 matched organoids per sample. We have profiled samples from six tumour sets to date and have obtained sequencing data for 21,234 primary tissue cells and 14,742 organoid cells. While the most apparent differences in gene expression appear to be between individual tumours, we were also able to identify similar cellular subpopulations across tissue samples and across organoids. Importantly, organoids derived from the same tissue sample appeared to be composed of similar cellular subpopulations and were highly comparable to each other, indicating that replicate organoids faithfully represent the original tumour tissue. Overall, our scRNA-seq approach will help evaluate the utility of tumour-derived organoids as model systems for GBM and will aid in identifying cellular subpopulations defined by gene expression patterns, both in primary GBM regional subsamples and their associated organoids. These analyses will allow for the characterization of clonal or subclonal populations that are likely to respond to different therapeutic approaches and may also uncover novel therapeutic targets previously unrevealed through bulk analyses.

Background: Gliomas are the most common and fatal adult brain tumor with distinct genomic subgroups defined by isocitrate dehydrogenase (IDH) mutation status. Mutations in IDH result in overproduction of the oncometabolite 2-hydroxyglutarate (2HG). The landscape of metabolic changes that define gliomas has not previously been explored. Methods: We performed liquid chromotography-mass spectrometry (LC-MS) to examine over 700 metabolites on 90 fresh-frozen glioma samples (30 IDH-wildtype, 30 IDH-mutant 1p/19q codeleted, 30 IDH-mutant 1p/19q non-codeleted) from our institutional biobank. R and S enantiomers of 2HG were quantified using high pressure liquid chromatography tandem mass spectrometry coupled with a CHIROBIOTIC R column. Genome wide DNA methylation was performed on all tumors using Illumina 850k EPIC array. Unsupervised consensus clustering of differentially expressed metabolites and methylated post-processed probes was performed. Copy number variations were determined based on intensity values of the methylation array. Survival of unsupervised cluster groups was determined using the Kaplan-Meier Estimate. Results: Unsupervised clustering of 689 metabolites revealed 2 distinct subgroups of gliomas associated with recurrence-free survival (RFS, P = 0.021). IDH mutant tumours were found in both cluster groups where as IDH-wildtype tumors were found only in Group 2. Group 2 IDH-mutant tumors had unfavourable PFS, higher R/S-2HG levels, and higher proportion of copy number alterations (4q, 9p, 13q, 17q) compared to group 1 IDH-mutant tumors (P=0.048, P=0.0194, P<0.0001 respectively) compared to group 1 IDH-mutant tumors. (P= p=0.048). . Conclusions: Metabolic profiling of gliomas reveals 2 distinct subtypes of IDH-mutant independent of 1p/19q codeletion status with differing survival patterns and large scale chromosomal alterations that may be driven by varying levels of R/S-2HG.

BACKGROUND: Metabolomics technology has the potential to revolutionize how we screen, diagnose, and treat cancer, as well as improve upon existing cancer molecular tests that may not sufficiently capture the complexity of most malignancies. In this study, we explore the clinical potential of metabolomics analysis in the diagnosis and risk-stratification of brain tumors. METHODS: To test the hypothesis that brain tumor type and survival could be predicted with metabolomics, we analyzed the pre-operative serum and urine samples of patients with glioblastoma (GBM), oligoastrocytoma (OA2), meningioma (M1) and compared them to healthy controls. (HC). Sera from immune-deficient NOD-SCID mice xenografted with human GBM brain tumor initiating cells were also studied. RESULTS: Metabolomics analysis of patient samples was able to accurately differentiate GBM, OA2, M1 and HC (p = 2.3 x 10-26). Subsequently, a prediction model developed and validated internally was able to diagnose GBM with a sensitivity of 86.7% and specificity of 93.8%, and distinguish whether a GBM patient possess O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation (p = 7.4 x 10-10). Within the MGMT methylated group, the model was able to predict longevity (p = 3.25 x 10-4). The model was also able to predict survival irrespective of MGMT methylation status (p = 2.9 x 10-6). CONCLUSIONS: In this study, we demonstrate that metabolomic analysis of patient biofluids can identify brain tumors, distinguish brain tumor subtypes, and independently predict MGMT status as well as longevity among GBM patients. Metabolomics analysis may facilitate non-invasive diagnosis of aggressive brain tumours.

Purpose: Diffuse gliomas can be divided on the basis of presence or absence of mutation in IDH genes. IDH-mutant diffuse gliomas represent a wide range of clinical outcome, which is not accounted for by current clinical and pathologic parameters. We aim to identify clinically and biologically relevant subgroups within IDH-mutant gliomas to gain a deeper insight into finer sub-classification. Methods: We used 412 IDH-mutant glioma samples that were profiled by The Cancer Genome Atlas (TCGA) Research Network, utilising methylation/mRNA datasets to identify subtypes with unique molecular signatures. We applied a Similarity Network Fusion (SNF) on individual platforms and their integrations. Results: SNF approach split glioma into four groups. The integrated RNA/methylation subtype produced a highly prognostic groups that predict survival (p-value=0.003) compared to mRNA and methylation alone. We observed a high degree of correlation between integrative subtypes and somatic mutations. Groups 1&4 had higher TERT promoter mutations (35% and 16%, respectively) compared to groups 2&3. Groups 1&4 showed increased TERT expression (34% and 14% respectively), and high percentage of TP53 and ATRX mutations. Multivariate analysis after adjusting for confounding factors including grade and age showed prognostic factors associated with survival (HR=3.2, p-value=0.001) in group 4 versus others. Conclusions: The results indicate that clinically relevant alterations exist within IDH-mutant gliomas that could stratify patients for treatment. Interestingly, group 4 showed high expression of HOX genes (18/18) (p-value=0.01) and higher methylation of Hox genes (21) (p-value=0.01) compared to others. Higher expression of specific Hox genes were associated with worse survival.

Standard-of-care for glioblastoma (GBM) includes surgery, radiation and temozolomide. Nearly all tumors recur and 5-year survival is less than 3%. Unmethylated promoter status O6-methylguanine-DNA-methyltransferase (MGMT) is a validated biomarker for temozolomide-resistance. Second-line treatment with bevacizumab has not only failed to improve survival, but has also been shown to induce intratumor hypoxia and increased chemoresistance. VAL-083 is a bi-functional DNA-targeting agent that readily crosses the blood-brain barrier. VAL-083 targets N7-guanine, causing DNA double-strand breaks and cancer cell-death in GBM cancer stem cells (CSCs) and non-CSCs, independent of MGMT. To investigate the in vivo anti-tumor effect of VAL-083+bevacizumab, we used an orthotopic GBM T16 PDX model. All mice carried MGMT-unmethylated, temozolomide-resistant recurrent GBM tumors detected by MRI 35 days post-implantation. Tumor progression was measured by MRI on days 49 and 56, and was calculated for the entire study (day 35 vs. 56) and for the last 7 days (day 49 vs. 56). Mice were grouped into control, bevacizumab, VAL-083, and VAL-083+bevacizmab. VAL-083 treatment started 3 days after bevacizumab treatment to ensure induction of hypoxia. Results: Tumors were significantly smaller in VAL-083-treated mice both compared to control (-83%, p<0.001) and compared to bevacizumab-treated (-75%, p<0.001) mice. Additionally, analysis of tumor growth in-time showed significantly reduced tumor progression for VAL-083+bevacizumab compared to VAL-083 alone (p<0.01). Conclusions: These results show strong in vivo anti-tumor efficacy of VAL-083 against MGMT-unmethylated, recurrent GBM. This effect was further augmented in combination with bevacizumab, providing rationale of clinical investigation of VAL-083+bevacizumab in GBM.

Brain tumor behavior is driven by aberrations in the genome and epigenome. Many of these changes, such as IDH mutations in diffuse low-grade glioma (DLGG), are common amongst the same class of tumour and can be incorporated into the diagnostic criteria. However, any given tumor may have other, less common genomic aberrations that are essential for its biological behavior and may inform on underlying aberrant cellular pathways, and potential therapeutic agents. Precision oncology is a genomics-based approach which profiles these alterations to better manage cancer patients and has established itself within the practice of oncology and is slowly making its way into neuro-oncology. The BC Cancer’s Personalized OncoGenomics (POG) program has profiled 16 adult tumours originating from the central nervous system using whole genome and transcriptome analysis (WGTA), for the first time, within a meaningful clinical timeframe/setting. As expected, primary genomic drivers were consistent with their respective diagnoses, though secondary drivers were found to be unique to each tumour. Although these analyses did not result in altered clinical management for these patients, primarily due to availability of drug or clinical trials, they highlight the heterogeneity of secondary drivers in cancers and provide clinicians with meaningful biological information. Lastly, the data generated by POG has highlighted the frequency and complexity of novel driver fusions which are predicted to behave similarly to canonical driver events in their respective tumours. The information available to clinicians through POG has provided paramount knowledge into the biology of each unique tumour.

Background: CNS innate immune cells, microglia and macrophages (MMs), are the largest component of the inflammatory infiltrate in glioblastoma (GBM). They initially participate in tumor surveillance, but are subverted by GBM. Immunotherapies have proven incredibly successful in cancers such as melanoma, but not against GBM in part because GBM-associated MMs are not well understood. We hypothesized the content and inflammatory phenotype of MMs in GBM is variable between patients. We suspect MMs in IDH-wildtype and –mutant GBMs display divergent inflammatory phenotypes that helps explain the latter's better prognosis. Understanding GBM-associated MM heterogeneity will allow for better immunotherapy development and selection. Methods: MMs were isolated from untreated human IDH-wildtype and -mutant GBMs using flow cytometry and cultured for collection of conditioned media and analysis of secretory products. Automated segmentation with a high-content analysis system was used to quantitate MM content and inflammatory phenotype in frozen sections. New bioinformatics techniques allowed the comparison of MM profiles in publicly available single-cell RNA-sequencing databases with IDH-wildtype and -mutant GBMs. Results: Surprisingly marked variation in MM content exists between GBMs ranging from ~0-70%. A mixture of pro- and anti-inflammatory MMs are found in each GBM. Interestingly, IDH-mutant GBM-associated MMs were more activated than MMs in IDH-wildtype GBMs. Conclusions: Taken together, the highly variable MM content and phenotype of GBMs suggests the success of immunotherapies hinges on taking a precision medicine approach. MM-rich GBMs would benefit more from therapies that target them. MM activation in IDH-mutant GBMs may contribute to better patient prognoses.

Group-3 medulloblastoma (MBL) is highly resistant to radiation (IR) and chemotherapy and has the worst prognosis. Hence, there is an urgent need to elucidate targets that sensitize these tumors to chemotherapy and IR. Employing standard assays for viability and sensitization to IR, we identified PRDX1 as a therapeutic target in Group-3 MBL. Specifically, targeting PRDX1 by RNAi or inhibition by Adenanthin led to specific killing and sensitization to IR of Group-3 MBL cells. We rescued sensitization of Daoy and UW228 cells by hypermorphic expression of PRDX1. PRDX1 knockdown caused oxidative DNA damage and induced apoptosis. We correlated PRDX1 expression to patient outcomes in a validated MBL tumor-microarray. Whole genome sequencing identified pathways/genes that were dysregulated with PRDX1 inhibition or silencing. Our in vivo studies in mice employing flank/orthotopic tumors from patient derived xenografts/Group-3 MBL cells confirmed in vitro observations. Animals with tumors in which PRDX1 was targeted by RNAi or Adenanthin (using mini osmotic pumps) showed decreased tumor burden and increased survival when compared to controls. Since, Adenanthin does not cross the blood brain barrier (BBB) we used HAV6 peptide to transiently disrupt the BBB and deliver Adenanthin to the tumor. Immunohistochemistry confirmed that targeting PRDX1 resulted in increased oxidative DNA damage, apoptosis and decreased proliferation. In summary, we have validated PRDX1 as a therapeutic target in group-3 MBL, identified Adenanthin as a potent chemical inhibitor of PRDX1 and confirmed the role of HAV peptide (in the transient modulation of BBB permeability) in an orthotopic model of group-3 MBL.

Introduction: Diffuse intrinsic pontine glioma (DIPG) is refractory to therapy. The identification of histone H3.1/H3.3 K27M mutations in most DIPG has provided new insights. The DLX homeobox genes are expressed in the developing forebrain. The Dlx1/Dlx2 double knockout (DKO) mouse loses tangential GABAergic interneuron migration to the neocortex. We have identified genes that encode glutamic acid decarboxylase (GAD) enzymes as direct targets of DLX1/DLX2. In DIPG patients with H3.3 K27M mutations there is decreased Dlx2 and increased expression of the myelin transcription factor, Myt1. Methods and Results: We used bioinformatics approaches and chromatin immunoprecipitation (ChIP) assays to identify Olig2, Nkx2.2 and Myt1 promoter sequences as candidate DLX2 targets in vivo. DNA binding specificity was confirmed. The functional consequences of Dlx2 co-expression with reporter constructs of ChIP-isolated promoter fragments of Olig2 and Nkx2.2 demonstrated repression of gene targets in vitro. qPCR showed increased Olig2 and Nkx2.2 expression in the DKO forebrain. Stable transfection of a murine DIPG cell line with Dlx2 resulted in increased Gad1 and Gad2 and decreased Olig2 and Nkx2.2 expression. Of significance, we demonstrated decreased expression of H3.3 K27M and restoration of H3.3 K27 tri-methylation (me3). Conclusions: DLX transcription factors promote GABAergic interneuron and concomitant inhibition of oligodendroglial differentiation in neural progenitors by repression of a suite of genes including Olig2 and Nkx2.2. Restoration of H3 K27me3 expression in DIPG provides a promising lead towards exploration of differentiation as a therapeutic strategy for DIPG.

The Canadian Brain Tumour Registry (CBTR) project was established in 2016 with the aim of enhancing infrastructure for surveillance and clinical research to improve health outcomes for brain tumour patients in Canada. We present a national surveillance report on malignant primary brain and central nervous system (CNS) tumours diagnosed in the Canadian population from 2009-2013. Patients were identified through the Canadian Cancer Registry (CCR); an administrative dataset that includes cancer incidence data from all provinces/territories in Canada. Cancer diagnoses are coded using the ICD-O3 system. Tumour types were classified by site and histology using The Central Brain Tumour Registry of the United States definitions. Incidence rates (IR) and 95% confidence intervals (CI) were calculated per 100,000 person-years and standardized to the 2011 census population age-distribution. Overall, 12,115 malignant brain and CNS tumours were diagnosed in the Canadian population from 2009-2013 (IR:8.43;95%CI:8.28,8.58). Of these, 6,845 were diagnosed in males (IR:9.72;95%CI:9.49,9.95) and 5,270 in females (IR:7.20;95%CI:7.00,7.39). The most common histology overall was glioblastoma (IR:4.06;95%CI:3.95,4.16). Among those aged 0-19 years, 1,130 malignant brain and CNS tumours were diagnosed from 2009-2013 (IR:3.36;95%CI:3.16,3.56). Of these, 625 were diagnosed in males (IR:3.32;95%CI:3.34,3.92) and 505 in females (IR:3.08;95%CI:2.81,3.36). The most common histology among the paediatric population was pilocytic astrocytoma (IR:0.73;95%CI:0.64,0.83). The presentation will include: IRs for other histologies, the geographic distribution of cases and a comparison between Canada and the United States.

Introduction: Radiation-induced meningiomas (RIMs) are increasing in prevalence as cancer patients live longer. Our laboratory has demonstrated that RIMs have a unique genomic landscape compared to sporadic meningiomas. Notably, a subset of RIMs harbor genomic rearrangement resulting in NF2 gene fusion with a nonrecurrent reciprocal gene. We aimed to compare the gene expression of NF2-Fusion and NF2-Wild Type (NF2-WT) RIMs. Methods: RNA sequencing using Illumina HiSeq was performed on 7 NF2-Fusion and 12 NF2-WT RIMs. Short read sequences obtained from sequencing were mapped to reference human genome(hg19). We performed differential expression analysis using edgeR statistical packages. Pathway analysis was performed using Gene Set Enrichment Analysis (GSEA). Immunohistochemistry was performed to validate findings. Results: Principal component analysis revealed that 5/7 of NF2-Fusion RIMs had similar gene expression profiles. One outlier had no chromosome 1p loss like the other NF2-Fusion RIMs. Pathway analysis demonstrated there was an upregulation of immune pathways in NF2-Fusion RIMs. Immunohistochemistry of PD-L1 revealed that 0/7 and 7/7 of NF2-Fusion tumors had positive expression in tumoral and inflammatory cells, respectively. In comparison, 6/12 of RIMs had tumoral and inflammatory cell expression of PD-L1. In addition, there was a higher CD3 lymphocyte infiltration in NF2-WT (42.2 vs. 12.4 number of cells per HPF). Discussion: Preliminary data in our lab demonstrates that NF2-Fusion tumors have a distinct immune microenvironment compared to NF2-WT tumors. Although pathway analysis indicates that NF2-Fusion RIMs have overexpression of immune pathways, immunohistochemistry reveals that inflammatory cells have positive PD-L1 expression, suggestive of immune burnout.