Background: Recurrent glioblastoma portends a poor prognosis and the role of repeat surgery in improving survival remains uncertain. Therefore, we undertook a systematic review and meta-analysis in order to determine if repeat surgical resection provides a meaningful survival benefit for patients with recurrent glioblastoma. Methods: Two independent reviewers searched for articles that reported on overall-survival of patients with recurrent glioblastoma using MEDLINE, Embase, Google Scholar, and Cochrane from January 2000 to 2018. Studies that compared overall survival of patients treated with single surgery compared to repeat surgery in the temozolomide era were included for analysis.Primary outcomes were odds ratio for survival at 6, 12, and 24 months from date of initial diagnosis. Secondary outcomes were ratio odds ratio for survival at 6, 12, and 24 months from date of repeat surgery. The proportions of patients who had the outcomes of interest were pooled using random-effects model. Quality assessment was performed using the Newcastle Ottawa Scale. Heterogeneity across trials was quantified by the I2 statistic. Publication bias was evaluated visually using funnel plots and quantified by the Egger regression. Results: Fourteen articles reporting on 3048 patients were included for analysis. The majority of articles were deemed to be of high quality with Newcastle Ottawa scale greater than 7 points. Pooled analysis showed improved overall survival following repeat surgery at 6- (OR 1.73, 95% CI 1.23-2.45, p<0.05), 12- (OR 1.71, 95% CI 1.20-2.45, p<0.05), and 24-months (OR 2.24,95% CI 1.01-4.95, p<0.05) and from date of initial diagnosis at 6- (OR 8.22, 95% CI 5.23-12.93, p<0.01), 12- (OR 4.16, 95% CI 3.25-5.36, p<0.01), and 24- (2.35, 95% CI 1.77-3.11, p<0.05) months. Conclusions: Repeat surgery for recurrent glioblastoma is associated with a significant survival advantage independent of other salvage therapies that include chemotherapy, radiation, and other antineoplastic regimens.

Malignant gliomas (MG) are highly infiltrative tumours with a poor prognosis. Nuclear factor I (NFI) is a family of 4 transcription factors (NFIA, B, C and X) implicated in the regulation of genes involved in MG cell migration and infiltration, particularly the neural stem cell marker, brain fatty acid binding protein (B-FABP). NFI activity is regulated by its phosphorylation status, with hypophosphorylated NFI being the active form. Our results indicate that the phosphatase calcineurin is able to dephosphorylate NFI. In turn, calcineurin is cleaved and activated by calpain proteases. We have identified CAST, a gene that encodes calpain inhibitor, calpastatin, as a putative target of NFI based on chromatin immunoprecipitation. Putative NFI binding elements are located in intron 3 of the CAST gene. To determine whether there is a bona fide alternative promoter within intron 3 of CAST, we carried out gel shifts as well as luciferase reporter gene assays using both the canonical and alternative promoters of CAST. These assays confirmed CAST alternative promoter usage in MG cells. Knockdown of individual NFIs revealed a role for NFIC and NFIX in the repression of CAST gene expression, specifically in cells expressing the hypophosphorylated (active) form of NFI. NFI depletion also altered the subcellular localization of both calpain and calineurin protein. Our results suggest a feedback loop for the NFI –calcineurin – calpain – calpastatin pathway in MG cells which may regulate cell migration.

Glioblastoma is inherently resistant to radiation and drug treatments. This is mediated by the most common forms of cell death are often actively inhibited. Identifying and exploiting alternative cell death pathways are essential to overcoming or bypassing drug resistance. Ferroptosis, a newly described, morphologically and biochemically distinct, cell death mechanism is characterized by iron-dependent cellular accumulation of reactive oxygen species. The combination of siramesine, a lysosome disruptor, and lapatinib, a dual tyrosine kinase inhibitor (TKI), synergistically induced death in glioma cancer cells. This cell death had characteristics of ferroptosis: it was blocked by the ferroptosis inhibitor ferrostatin-1 and the iron chelator deferoxamine. In addition, the amount of ROS and lipid peroxidation were increased in glioma cells. Iron transport protein remained unchanged but reactive iron levels increased. One target for kinase inhibitors is protein bisulfate isomerase (PDI). Knockdown of PDI in combination with siramesine increased cell death that was blocked by ferrostatin-1. Taken together, drug combinations that alter reactive iron and ROS levels might induce ferroptosis and overcome drug resistance in glioma cells.

Estimation of tumor recurrence in meningioma patients is one of the important clinical challenges. The prognostic impact of immune modulatorymolecule PD-L1 in several malignancies has been demonstrated. We studied the association of PD-L1 expression in meningioma with tumor recurrence and the underlying mechanism of its activation. Immunohistochemical staining(IHC) was performed for detection of PD-L1and NFKB2 on whole sections of meningiomas diagnosed between1998-2016 at TorontoWesternHospital. The biologic role of hypoxia in activation of PD-L1 in meningioma was investigated using gene set enrichment analysis(GSEA)-based on RNAseq data in validation cohorts. We analyzed a total of 93 meningioma cases: F/M ratio58/35;WHOgrade I(41),II(43),III(9),42(47%)cases with tumor recurrence and median follow up was 6.97yrs. PD-L1 expression on tumor cells(PD-L1TC) in 33(35%) cases was identified with distinctive patchy distribution. Univariate analysis indicated expression of PD-L1TC as a prognostic factor for tumor recurrence(p<0.0001). Multivariate analysis showed that PD-L1TC expression is an independent prognostic factor for tumor recurrence after adjusting for extent of resection(EOR),WHO grade and maximum tumor diameter(p<0.0001). Analysis of RNAseq data of two GEOmeningioma studies demonstrated prominent expression of NFKB activation associated with PD-L1expression. IHC analysis confirmed increased expression of NFKB2 protein in 26(30%)cases,which correlated with PD-L1TC expression(p=0.02). Furthermore,GSEAon a RNAseq data of 88 sporadic meningiomas using Hypoxia gene signature of HUVEC cells we found that the hypoxic sporadic meningiomas have significantly elevated PD-L1 expression(p<0.001). Our data strongly suggest that PD-L1TCexpression serves as a significant prognostic marker for tumor recurrence and. We found that hypoxia andNFKB2 activation are potential underlying mechanisms. These results also provide a rationale for potential adjuvant therapeutic role for PD-L1 inhibitors in meningioma.

Background: Glioblastoma is the most common adult malignant glioma, with poor prognosis and adverse neurological sequelae. Physical activity improves outcomes in patients with other cancers, but has not been evaluated in GBM. This prospective, single-arm intervention trial examines feasibility and preliminary efficacy of exercise on PFS, cognition and QOL in newly diagnosed GBM patients. Method: Participants are English-speaking GBM patients scheduled for concurrent chemoradiation at PMH, 18-65 years old, ECOG ≤ 2. The 3-month home-based exercise program includes aerobic and resistance training, tailored to prior fitness level, current physical status, and individual interests. Assessments of physical and neurocognitive functions, mood, fatigue, sleep, and QOL, occur within 2 weeks of starting chemoradiation, and approximately 3, 6, 12, and 18 months later, or until tumor progression. Feasibility will be assessed by accrual, retention, and adherence rates. Outcomes include PFS (RANO criteria), change in cognition (reliable change index method), physical activity and sleep (actigraphy, self-report questionnaires). Time-to-event outcomes will be estimated (Kaplan-Meier), and mixed modelling will explore individual and disease variables that contribute to outcomes. Results: During the first five months of recruitment, 13 of 19 eligible patients consented. Nine completed the exercise program. One patient died after the intervention and none of the others progressed. No exercise-related serious adverse events occurred. Preliminary results will be presented at the meeting. Discussion: Exercise appears feasible for GBM patients. Effects on survival, performance status, cognition, sleep, mood, and QOL are ongoing. Results may guide physical activity recommendations in GBM and generate avenues for translational research.

Atypical Teratoid Rhabdoid Tumors (AT/RTs) often affect children under the age of 3 and are the most common malignant CNS tumors in children younger than 6 months. It is very rare to see these tumors in patients older than 6 years of age. We discuss the case of a 14 year old male with AT/RT of the right insula. He had a prior diagnosis of Dysembryoplastic Neuroectodermal Tumor (DNET) at the age of 10, after two years of intermittent headaches, nausea, and seizures, which was treated with conformal radiation and chemotherapy for a year. Following the diagnosis of AT/RT, he underwent radiotherapy, multiple lines of chemotherapy, and two additional debulking surgeries of the left temporal lobe due to continuing progression. He was then treated with Alisertib (an Aurora-A kinase inhibitor) with good response on sequential MRIs after the first three cycles. He progressed after nine cycles of Alisertib and required further debulking surgery. Six years after his AT/RT diagnosis (and 10 years after his DNET diagnosis), the patient expired at the age of 20 due to ongoing progression. To our knowledge, this is only the second reported case of Alisertib use in a non-pediatric AT/RT case. We also performed a literature review of all reported cases of AT/RT in adults between the years 2000 – 2017 and discuss treatment options, patient demographics, and survival.

BACKGROUND: The use of high-dose chemotherapy (HDC) and autologous hematopoietic stem cell transplant (HSCT) has been used in certain pediatric patients with brain tumours to delay/spare radiotherapy. We aimed to study factors predicting a successful stem cell collection (SCC) and correlate stem cell dose infused with HSCT outcomes. METHODS: A retrospective chart review was undertaken for pediatric patients with brain tumours treated at our centre with HDC/HSCT between 2004-2016. RESULTS: Fifty-five patients were identified (32 male) with median age of 6.3 years at time of SCC (range 0.4-18.7). Patients' diagnoses were medulloblastoma (62%), ATRT (20%), and PNET (18%). Most patients (82%) underwent a single/1-day SCC, while the remaining required 2 SCC procedures. Peripheral blood stem cells were the source in most collections (95%). Successful SCC (CD34 collected greater-than-or-equal-to 2 x10^6/kg/transplant) and ideal SCC (greater-than-or-equal-to 5 x10^6/kg/transplant) was achieved in 85% and 45% of patients, respectively. Use of mobilizing chemotherapy with G-CSF was the only factor associated with achieving an ideal collection, while gender, age, stem cell source, and pre-apheresis peripheral blood CD34 count were not significant. Higher CD34/kg infused was associated with faster neutrophil engraftment in the first 3 courses of HDC/HSCT and platelet engraftment in the first course. CONCLUSIONS: The majority of SCC for autologous HSCT can be successfully completed with a single apheresis session. Mobilization with both chemotherapy and G-CSF yields higher CD34 compared to G-CSF alone. Higher dose of CD34/kg infused was associated with faster neutrophil and to a more limited scale platelet recovery post-HSCT.

Background: Brain tumors present unique challenges to patient and family quality of life (QOL). Cognitive dysfunction is common and functionally limiting, with no established treatments. These studies evaluate feasibility and preliminary efficacy of behavioral interventions developed for neuro-oncology patients. Study 1: A randomized controlled trial (N=25 primary brain tumor patients) compared an adapted version of Goal Management Training (GMT, a neuroscience-based integration of mindfulness and strategy training) and a newly-designed supportive psychoeducational intervention (Brain Health Program, BHP) to standard of care. Each intervention comprised 8 individual sessions and at-home practice between sessions. GMT patients’ executive functions improved immediately (p=.077, d=1.13), with maintenance at 4-month follow-up (p=.046, d=1.09). Both intervention groups reported improvements in everyday cognitive functioning immediately (p=.049; d’s GMT=0.43, BHP=0.79) and at follow-up (p=.001; d’s GMT=0.22, BHP=1.01). BHP patients also reported improved mood (p’s=.026 & .012, d’s=0.61 & 0.62). Study 2: Following a needs assessment about cognitive concerns and QOL in brain metastases patients (N=109) and caregivers (N=31), we developed a novel, brief (3 sessions + homework) Cognitive Support Program to provide education and strategy-training in key areas of concern: executive functions, memory, and communication. Options include caregiver co-training, and in-person or web-based delivery. Preliminary data from a pilot trial in progress demonstrate objective and subjective improvements. Conclusions: Cognitive rehabilitation may be a feasible and effective option for primary or metastatic brain tumor patients, addressing a need that is largely unmet in standard cancer care. Further development and larger trials appear warranted, with capacity for remote delivery recommended.

Despite decades of trials, the prognosis for diffuse intrinsic pontine gliomas (DIPG) remains dismal. DIPG is inoperable and standard treatment is radiation alone, as the addition of chemotherapeutic agents, such as temozolomide, have not improved survival. In addition to inherent chemoresistance, treatment of DIPG is impeded by an intact blood-brain barrier (BBB). VAL-083 is a structurally unique bi-functional DNA-targeting agent that readily crosses the BBB. VAL-083 forms interstrand DNA crosslinks at N7-guanine, resulting in DNA double-strand breaks (DSB), S/G2-phase cell-cycle arrest, and ultimately cancer cell death. We have previously demonstrated that VAL-083 is able to overcome temozolomide-resistance in vitro and in vivo, and that its cytotoxicity is independent of the DNA-repair enzyme O6-methylguanine DNA-methyltransferase (MGMT). MGMT is almost universally expressed in DIPG and its expression is strongly correlated with temozolomide-resistance. VAL-083’s distinct mechanism-of-action suggests the potential for combination with inhibitors of DNA DSB repair or S/G2 cell-cycle progression (e.g. Wee1 inhibitor AZD1775). Here, we investigated the effects of VAL-083 in combination with radiation, AZD1775 or irinotecan (topoisomerase inhibitor) in three DIPG cell-lines: SF10693 (H3.1), SF8628 (H3.3) and NEM157 (H3.3). VAL-083 showed activity at low uM-concentration in all three cell-lines. In addition, VAL-083 showed synergy with AZD1775 in all three cell-lines. Combined with its ability to cross the BBB, accumulate in brain tumor tissue and overcome MGMT-related chemoresistance, these results suggest VAL-083 as a potentially attractive treatment option for DIPG as single agent or in combination with AZD1775. Combination studies with radiation are ongoing and will be presented at the meeting.

Current standard-of-care for glioblastoma (GBM) includes surgery, radiation and temozolomide. Most tumors recur within a year from diagnosis and median survival for recurrent GBM (rGBM) is 3-9 months. Unmethylated promoter status for O6-methylguanine-DNA-methyltransferase (MGMT) is a validated biomarker for temozolomide-resistance, exhibited by most GBM patients. VAL-083 is a DNA-targeting agent with a mechanism-of-action that is independent of MGMT. VAL-083 overcomes temozolomide-resistance in GBM cell-lines, cancer stem cells, and in vivo models. VAL-083 readily crosses the blood-brain barrier and accumulates in brain-tumor tissue. We recently completed a VAL-083 dose-escalation trial in temozolomide- and bevacizumab-refractory rGBM and determined that 40mg/m2/day given intravenously on days 1,2,3 of a 21-day cycle is generally well-tolerated. This dosing regimen was selected for subsequent GBM trials, including an ongoing single-arm, biomarker-driven Phase 2 trial (N=48) in temolozomide-refractory, bevacizumab-naïve rGBM , MGMT-unmethylated (Clinicaltrials.gov:NCT02717962). The primary objective of this study is to determine if VAL-083 improves OS compared to a historical control of 7.15 months for MGMT-unmethylated rGBM patients treated with lomustine (EORTC26101). In addition, another single-arm, biomarker-driven, Phase 2 study (N=25) of VAL-083 in combination with radiotherapy in newly diagnosed GBM, MGMT-unmethylated is ongoing (Clinicaltrials.gov:NCT03050736). This trial aims to determine a dose for further study of VAL-083 in combination with radiotherapy and explore if VAL-083 improves PFS and OS compared to historical results in newly diagnosed GBM. Enrollment and safety data updates will be provided at the meeting. The results of these studies, if successful, may support VAL-083 as part of a new chemotherapeutic treatment paradigm for GBM.

INTRODUCTION: Gangliogliomas (GGs) are neuroepithelial tumours of the central nervous system (CNS) composed of mature ganglion cells or a mixed population of ganglion and glial cells. Microarray data of low grade gliomas (LGG) including GGs revealed overexpression of the Dlx2 gene, a homeobox gene essential for interneuron migration and differentiation. We hypothesized that GGs are arrested in development, and began to explore the role of the Dlx2 gene. BRAF rearrangements and BRAF V600E point mutations have been reported in pediatric LGG. METHODS: DLX2 expression was examined in GGs using immunofluorescence (IF) and immunohistochemistry (IHC) labelling of formalin fixed paraffin embedded (FFPE) tissue sections, along with staining of glial and neuronal markers. BRAF mutations were detected using a commercial antibody and/or sequence verification of the DNA extracted from the FFPE blocks. RESULTS: In the Discovery cohort 10/30 were DLX2+ (33.3%) and in the Validation cohort 15/40 were DLX2+ (37.5%). Of these 15 cases, 15 were GFAP+ (100%), 15 were synaptophysin and/or NeuN+ (100%), and 13 were OLIG2+ (86.7%); 6 had a BRAF V600E mutation (40.0%). For the Validation cohort of 40 GGs, 28 were OLIG2+ (70.0%); 13/28 co-expressed DLX2 (46.4%). 18/40 cases had a BRAF V600 mutation(17 V600E, 1 V600G; 45.0%) and 6/18 were DLX2+ (33.3%). CONCLUSIONS: DLX2 is expressed in GGs in both neuronal and glial marker expressing tumour cells. Our results support that GGs arise from CNS progenitors arrested at the neuronal-glial cell fate “decision” point.

Glioblastoma is the most common and malignant brain tumor with a median overall survival of 20.5 months. There is an urgent need to develop novel therapeutic strategies. Using a glioblastoma TCGA dataset, we have determined that high NSUN5 mRNA expression is strongly associated with poor survival in glioblastoma patients. NSUN5 is a ribosomal RNA (rRNA) cytosine methyltransferase. Human NSUN5 is located in chromosome 7 and is completely deleted in the Williams-Beurren syndrome, a complex neurodevelopmental disorder. However, RNA targets of NSUN5 in mammals and its role in cancer are unknown. The objective of this project is to determine whether elevated NSUN5 changes rRNA methylation pattern and thereby leads to pro-tumorigenic translational reprogramming and pro-tumorigenic phenotypes in glioblastoma. Western blotting showed that NSUN5 is expressed in 7 out of 9 established glioblastoma cell lines and in 8 out of 12 primary patient-derived glioblastoma cell lines. Bisulfite sequencing confirmed that NSUN5 methylates C3782 of human 28S rRNA in glioblastoma cells. Functionally, overexpression of NSUN5 increases, whereas NSUN5 knockout decreases global protein synthesis and sphere formation in glioblastoma cells. More importantly, mice bearing intracranial NSUN5-expressing U87 tumors survived for a shorter time than mice bearing tumors derived from U87 control cells. Our results suggest that NSUN5 methylates 28S rRNA and may enhance cancer stem cell phenotypes and tumor formation and/or progression in glioblastoma. Experiments are ongoing to determine whether NSUN5 promotes tumor formation and/or progression through translational reprogramming in glioblastoma. This study may help identify novel therapeutic targets for glioblastoma.

BACKGROUND: Neurofibromatosis type 2 (NF2) is a rare genetic condition caused by mutations in the Merlin gene on chromosome 22. It results in acoustic neuromas (schwannomas) and other CNS tumors including meningiomas and ependymomas. Most patients develop hearing loss as a result of neuroma-driven destruction of auditory nerves. Surgery and radiation therapy remain the two most commonly recommended treatment options. However, there is a risk of further hearing loss with these procedures. There is emerging evidence that bevacizumab, a monoclonal antibody against VEGF-A, can shrink acoustic neuromas and mitigate hearing loss. CASE PRESENTATION: A 34-year-old female with bilateral acoustic neuromas from NF2 suffers partial hearing loss in the left ear and total hearing loss in the right ear after removal of the right-sided neuroma. Baseline MRI showed a left-sided acoustic neuroma (15 x 13 mm) and recurrence of the right-sided neuroma (18 x 14 mm). Bevacizumab was initiated at 5 mg/kg IV every 14 days. After 8 cycles, the patient reported marked improvement in hearing. At lower frequencies (< 1,000 Hz, the range of human voice), auditory thresholds improved by up to 60% of baseline, while at higher frequencies, improvements of up to 46% were seen. Repeat imaging showed no disease progression. CONCLUSIONS: Bevacizumab led to hearing improvement and prevention of disease progression after 8 cycles of therapy. This treatment should be considered in patients with NF2 and acoustic neuromas who wish to pursue a less-invasive treatment option with the potential of delaying progression and mitigating hearing loss.

Toca 511 (vocimagene amiretrorepvec) is an investigational retroviral replicating vector that selectively infects dividing cancer cells, integrates into the genome and replicates due to immune defects in tumors. Toca 511 spreads through tumors and stably delivers the gene encoding an optimized yeast cytosine deaminase that converts the prodrug Toca FC (investigational, extended-release of 5-fluorocytosine) into 5-fluorouracil. In preclinical models, 5-fluorouracil kills infected dividing cancer cells, myeloid derived suppressor cells and tumor associated macrophages, enabling immune activation against the tumor. In this dose ascending Ph1 trial (NCT01470794), Toca 511 was injected into the resection cavity wall of patients with rHGG, followed by courses of oral Toca FC. Additional cohorts included combination with bevacizumab or lomustine. Across the Ph1 program, the safety profile remains favorable. Objective responses (ORs) were assessed by IRR using MRI scans prior to Toca FC treatment as baseline. ORs occurred 6-19 months after Toca 511 administration, suggesting an immunologic mechanism. The ORs were observed in 4 patients with IDH1-wildtype and 2 patients with IDH1-mutant tumors, including 5 complete responses (CRs) with the investigational therapy alone, and 1 CR in combination with bevacizumab. The median duration of response (mDoR) was 35.1+ months. As of AUG2017, all responders were CR and remain alive. In a 23-patient subgroup who received high doses of Toca 511 and met Ph3 trial criteria, mOS was 14.4 months, 3-year survival rate was 26.1%, and mDoR was 35.7+ months with a durable response rate of 21.7%. Data suggest a positive association of durable response with OS.

Few studies investigated the associations between intervention modalities, timing, and survival in glioblastoma (GBM) patients. A total of 20511 eligible GBM patients underwent biopsy and craniotomy surgeries followed by adjuvant treatments (2005-2014) were derived from the National Cancer Database (NCDB). The time intervals (days) from the date of diagnosis to the initiation date of adjuvant treatment [radiation therapy only (RT), chemotherapy only, concurrent chemoradiation (CRT), or non-concurrent RT and chemotherapy] were categorized into quartiles (Q1-Q4). Kaplan-Meier method and Cox proportional hazards regression were applied for survival analysis. Multivariate logistic regression was performed to compare differences in treatment timing, intervention modalities, and secondary outcomes. The patients underwent biopsy obtained significant survival benefit by having delayed adjuvant treatment [comparing to Q1, Q2: HR (hazard ratio), 0.88, Q3: HR, 0.86]. For patients underwent resection, the prolonged waiting time of adjuvant treatment had 5-6% reduced risk of death [comparing to Q1, Q2: HR, 0.95; Q3: HR, 0.94]. Patients received more RT fractions [comparing to 10-29 fractions, 30-33 fractions: HR: 0.62 (biopsy), 0.62 (resection); ≥34 fractions: HR: 0.53 (biopsy), 0.62 (resection)] and high-dose RT [comparing to 34-46 Gy, 50-60 Gy: HR: 0.91 (biopsy), 0.95 (resection); ≥ 60 Gy: HR: 0.77 (biopsy), 0.88 (resection)] experienced significantly superior survival in both biopsy and resection groups. The impact of timing to adjuvant treatment on GBM survival varied by surgery procedures. Having adjuvant treatment initiated within 21 days for both biopsy and craniotomy groups may not guarantee a significant survival benefit. More RT fractions and high-dose RT are associated with better GBM survival.

Glioblastomas are the most frequent and aggressive primary brain tumor in adults and despite recent therapeutic advances, they are resistant to treatment. Increasing malignancy of gliomas correlates with an increase in cellularity and a poorly organized tumor vasculature, leading to insufficient blood supply, hypoxic areas, and ultimately to the formation of necrosis. Hypoxia induces direct or indirect changes in the biology of solid tumor and their microenvironment through the activation of HIF transcription factors, leading to increased aggressiveness and tumor resistance to therapy. Not much is known about the epigenetic alterations induced by hypoxia and how they could alter tumor biology. In the present study, we have utilized PIMO as a specific marker of hypoxia in glioblastoma patients, treated with PIMO preoperatively. We have estimated PIMO positivity in each tumor (5-45%) and determined that it positively correlates with the hypoxia marker CA IX (r=0.57). In addition, 10 surgical PIMO cases were dissociated, immune labeled using PIMO antibody, followed by DNA isolation and methylation profiling. Our analysis of differentially top 4000 differentially methylated probes suggests that PIMO-positive (hypoxic) cells are differentially methylated compared to the PIMO-negative cells and these changes are associated with genes involved in hypoxic cellular response. We will validate these findings in additional glioblastoma cases and assess the mechanism of these epigenetic alterations in vitro in glioma stem cell culture conditions and upon exposure of the cells hypoxic conditions.

Molecular characterization of pediatric low-grade glioma (pLGG) over the last decade has identified recurrent alterations, most commonly involving BRAF, and less frequently other pathways including MYB and MYBL1. Many of these molecular markers have been exploited clinically to aid in diagnosis and treatment decisions. However, their frequency and prognostic significance remain unknown. Further, a significant portion of cases do not have any of these alterations and what underlies these cases remains unknown. To address this we compiled a cohort of 562 patients diagnosed at SickKids from 1990-2017. We identified molecular alterations in 454 cases (81% of the cohort). The most frequent events were those involving BRAF; either as fusions (most commonly with KIAA1549 (30%)) or V600E mutations (17%) and NF-1 (22%). Less frequently, we identified recurrent FGFR1 fusions and mutations (3%), MYB/MYBL alterations (2%), H3F3AK27M (2%) or IDH1R132H (0.5%) mutations, as well as other novel rare events. Survival analysis revealed significantly better progression-free survival (PFS) and overall survival (OS) of KIAA1549-BRAF fused patients compared to BRAFV600E with 10-year OS 97.7% (95%, CI 95.5-100) and 83.9% (95%, CI 72.5-95.6), respectively. In addition to survival, molecular alterations predicted differences in response to conventional therapeutics; BRAF fused patients showed a 46% response-rate, versus only 14% in V600E patients. pLGGs harboring H3F3AK27M progressed early with median PFS of 11 months. In patients with MYB/MYBL1, FGFR1/FGFR2 alterations, we observed only one death (FGFR1N546K case). The work here represents the largest cohort of pLGGs with molecular profiling and their impact on the clinical behaviour of the disease.

The extensive heterogeneity both between and within the medulloblastoma (MB) subgroups underscores a critical need for variant-specific biomarkers and therapeutic strategies. We previously identified a role for the CD271/p75 neurotrophin receptor (p75NTR) in regulating stem/progenitor cells in the SHH MB subgroup. Here, we demonstrate the utility of CD271 as a novel diagnostic and prognostic marker for SHH MB using immunohistochemical analysis as well as transcriptome data across 763 primary tumors. Characterization of CD271+ and CD271- cells by RNA sequencing revealed that these two subpopulations are molecularly distinct, co-existing cellular subsets both in vitro and in vivo. MAPK/ERK signaling is upregulated in the CD271+ population and inhibiting this pathway reduced CD271 levels, stem/progenitor cell proliferation and cell survival as well as cell migration in vitro. Importantly, the MEK inhibitor selumetinib extends survival and reduces CD271 levels in vivo. Our study demonstrates the clinical utility of CD271 as both a diagnostic and prognostic tool for SHH MB tumors and reveals a novel role for MEK inhibitors in targeting CD271+ SHH MB cells.

BACKGROUND: IGTS is a rare phenomenon of paradoxical germ cell tumor (GCT) growth during or following treatment despite normalization of tumor markers. We sought to evaluate the frequency, clinical characteristics and outcome of IGTS in patients in 21 North-American and Australian institutions. METHODS: Patients with IGTS diagnosed from 2000-2017 were retrospectively evaluated. RESULTS: Out of 739 GCT diagnoses, IGTS was identified in 33 patients (4.5%). IGTS occurred in 9/191 (4.7%) mixed-malignant GCTs, 4/22 (18.2%) immature teratomas (ITs), 3/472 (0.6%) germinomas/germinomas with mature teratoma, and in 17 secreting non-biopsied tumours. Median age at GCT diagnosis was 10.9 years (range 1.8-19.4). Male gender (84%) and pineal location (88%) predominated. Of 27 patients with elevated markers, median serum AFP and Beta-HCG were 70 ng/mL (range 9.2-932) and 44 IU/L (range 4.2-493), respectively. IGTS occurred at a median time of 2 months (range 0.5-32) from diagnosis, during chemotherapy in 85%, radiation in 3%, and after treatment completion in 12%. Surgical resection was attempted in all, leading to gross total resection in 76%. Most patients (79%) resumed GCT chemotherapy/radiation after surgery. At a median follow-up of 5.3 years (range 0.3-12), all but 2 patients are alive (1 succumbed to progressive disease, 1 to malignant transformation of GCT). CONCLUSION: IGTS occurred in less than 5% of patients with GCT and most commonly after initiation of chemotherapy. IGTS was more common in patients with IT-only on biopsy than with mixed-malignant GCT. Surgical resection is a principal treatment modality. Survival outcomes for patients who developed IGTS are favourable.

Medulloblastoma (MB) is the most common malignant pediatric brain tumor. Current clinical trials for recurrent MB patients based on genomic profiles of primary, treatment-naïve tumors, provide limited clinical benefit since recurrent metastatic MBs are highly genetically divergent from their primary tumors. By adapting the existing Children’s Oncology Group treatment protocol for children with newly diagnosed high-risk MB for treatment of mice intracranially engrafted with human MB cells, we have characterized the rare treatment-refractory cell population in Group 3 MBs. MB cell populations recovered separately from brains and spines during the course of tumor development and therapy were comprehensively profiled for gene expression analysis, stem cell and molecular features to generate a global, comparative profile of MB cells through therapy to relapse. One of the most intriguing observations from our gene expression data was consistent over-expression of proteins belonging to Inhibitor of DNA-binding/differentiation (ID) family and a longevity associated protein bactericidal/permeability-increasing fold-containing-family-B-member-4 (BPIFB4) in our refractory population. The persistent upregulation of genes preserving undifferentiated state and cellular longevity further strengthens the hypothesis of stem-cell like cells driving tumor relapse in MB. Targeting BPIFB4 using both knockdown (KD) and knockout (KO) strategies have resulted in decreased proliferation and self-renewal of both primary and recurrent MB cells, further highlighting its potential as a novel therapeutic target in MB. Our differential genomic and gene expression profiles of the “treatment-responsive” tumors against those that fail therapy have successfully contributed to discovery and characterization of novel therapeutic targets for the most aggressive subgroup of MB.