Neuro-Oncology

Abstract BACKGROUND/OBJECTIVES MYC family proteins are implicated in many human cancers, but their therapeutic targeting has proven challenging. MYCN and c-MYC amplification in childhood neuroblastoma (NB) and medulloblastoma (MB) are associated with aggressive disease and high mortality. Novel and effective therapeutic strategies are therefore urgently needed for these tumors. MYC-driven oncogenic transformation impairs cell survival under nutrient deprivation (ND), a characteristic stress condition within the tumor microenvironment. We recently identified eukaryotic Elongation Factor 2 Kinase (eEF2K) as a pivotal mediator of the adaptive response of tumor cells to ND. We therefore hypothesized that eEF2K facilitates the adaptation of MYCN/MYC amplified NB/MB to ND, and that inhibiting this pathway can impair tumor progression. METHODS To test our hypothesis, we first analyzed publicly available genomic databases and tissue microarrays for eEF2K expression in NB and MB, and for links between eEF2K, MYCN/MYC, and clinical outcome. Effects of eEF2K inhibition were evaluated on survival of MYCN/MYC amplified versus non-amplified NB/MB cell lines under ND. Finally, NB xenograft mouse models were used to confirm in vitro observations. RESULTS Our results indicate that high eEF2K expression and activity are strongly predictive of poor outcome in NB and MB (p<0.001), and correlate significantly with MYCN/MYC amplification (p<0.001). Inhibition of eEF2K significantly decreases survival of MYCN/MYC amplified NB/MB cell lines in vitro under ND. Knockdown of eEF2K under caloric restriction determines a twofold growth decrease of MYCN amplified NB xenografts. CONCLUSIONS eEF2K represents a critical mediator for the adaptive response of MYCN/MYC amplified NB and MB to acute metabolic stress, and is therefore a promising therapeutic target. Future therapeutic studies will aim to combine eEF2K pharmacological inhibition with caloric restriction mimetics such as metformin or glycolysis inhibitors, as eEF2K activity appears to be critical under metabolic stress conditions. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

doi: 10.1093/neuonc/nox168.822pmid: N/A

Abstract BACKGROUND In an international randomized controlled phase II study of temozolomide (TMZ) vs TMZ in combination with Bevacizumab (BEV) in locally diagnosed non-1p/19q co-deleted grade II or III glioma with a first and contrast-enhancing recurrence after initial radiotherapy, overall survival at 12 months (OS12) was not significantly different (61% in the TMZ arm and 55% in the TMZ+BEV arm). OBJECTIVES Health-Related Quality of Life (HRQoL) was a key secondary endpoint in this trial, and the main objective was to determine the impact of addition of BEV to TMZ on HRQoL. METHODS HRQoL was assessed using the EORTC QLQ-C30 (version 3) and QLQ-BN20 at baseline, and then every 12 weeks until progression of disease. The primary HRQoL endpoint was the QLQ-C30 global health scale, with cognitive functioning and pain selected as secondary HRQoL issues. Analysis was done using linear mixed modeling and complimented with sensitivity analyses using summary statistics and imputation of missing data. A clinical difference of ≥10 points on the 100-point QLQ-C30 scale between the two arms was considered as clinically relevant. RESULTS Baseline compliance was high at 94%, and remained above 60% until 72 weeks, limiting the analysis to 60 weeks. Compliance was similar in both arms. We found no statically significant or clinically significant differences between the primary HRQoL endpoints of both treatment arms (P=0.2642). The sensitivity analyses confirmed this finding. We observed no statistical or clinical significant differences regarding the selected secondary endpoint scales of cognitive functioning and pain. Among the other scales, social functioning (P=0.0254), physical functioning (P=0.0388) and itchy skin (P=0.0364) were significantly in favor of the TMZ arm, while motor dysfunction (P=0.0285) scores were better in the TMZ+BEV arm. INTERPRETATION: The addition of BEV to TMZ in this patient group neither improves nor negatively impacts HRQoL. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

Abstract BACKGROUND Although it is known that upfront complete surgical resection is associated with improved survival in glioblastoma, studies are lacking to confirm whether this holds true for the unfavorable patient population of O-6-methylguanine-DNA methyltransferase (MGMT) non-methylated glioblastoma. Here, we investigated the association of survival with the extent of resection in the GLARIUS study, a randomized phase II trial of bevacizumab/irinotecan (BEV/IRI) versus standard temozolomide (TMZ) in MGMT non-methylated glioblastoma. METHODS Patients included in the modified intention-to-treat (ITT, n=170) population were stratified by extent of resection (partial vs. complete) as determined by early (>72h) postoperative contrast-enhanced MRI. In a Kaplan-Meier analysis, we compared overall survival (OS) between the groups in each treatment arm. A Cox regression analysis was used to detect whether complete resection was of prognostic value and independent of canonical prognostic markers, including age and Karnofsky performance score (KPS). RESULTS In the BEV/IRI arm, 58 patients (50%) each underwent partial (PR) and complete resection (CR). In the TMZ arm, 29 patients (55%) underwent partial and 24 (45%) complete resection. Prognostic factors, including age and KPS, were balanced between patients with PR and CR. Neither in the BEV/IRI arm (CR, median OS [mOS], 17.3 [95% CI, 15.5-21.4] versus PR, mOS, 16.5 [95% CI, 14.9-17.9]; Hazard Ratio [HR], 0.79 [95% CI, 0.53-1.20]; p=0.28) nor in the TMZ arm (CR, mOS, 18.0 [95% CI, 17.1-21.3] versus PR, mOS, 15.3 [95% CI, 10.6-21.1]; HR, 0.77 [95% CI, 0.43-1.40]; p=0.40) patients with CR derived a significant OS benefit compared to those with PR. In a multivariable Cox regression analysis, this effect did not change after accounting for canonical prognostic markers. CONCLUSIONS In the GLARIUS trial, patients with complete resection derived no relevant survival benefit as compared with partially resected patients. However, the small sample size, in the TMZ arm in particular, limits our analysis. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

doi: 10.1093/neuonc/nox168.329pmid: N/A

Abstract Glioblastomas are lethal cancers defined by angiogenesis and pseudopalisading necrosis. Here, we demonstrate that these features are associated with distinct transcriptional programs, with vascular regions showing a proneural profile and hypoxic regions a mesenchymal pattern. As these regions harbor glioma stem cells (GSCs), we investigated the epigenetic regulation of these two niches. Proneural, perivascular GSCs activated EZH2, whereas mesenchymal GSCs in hypoxic regions expressed BMI1 protein, which promoted cellular survival under stress. Using both genetic and pharmacologic inhibition, we found that proneural GSCs are selectively sensitive to EZH2 disruption, whereas mesenchymal GSCs are sensitive to BMI1 inhibition. Given that glioblastomas contain both proneural and mesenchymal GSCs, combined EZH2 and BMI1 targeting proved more effective than either agent alone both in culture and in vivo, suggesting that strategies that simultaneously target multiple epigenetic regulators within glioblastomas may be necessary to overcome resistance to therapies caused by intratumoral heterogeneity. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

Abstract INTRODUCTION The EORTC22033-26033 clinical trial investigated whether initial temozolomide (TMZ) chemotherapy confers a survival advantage compared to radiotherapy (RT) in high-risk low-grade glioma patients. The aim of this study was to identify markers associated with survival and treatment response in the tumors’ transcriptome. METHODS Gene expression profiling (n=195/477) was used to assign tumors to one of six intrinsic glioma subtypes (IGS; molecularly similar tumors predefined by unsupervised gene expression analysis) and to extract in silico the cellular composition of immune infiltrates. DNA copy number changes were determined on samples assigned to IGS-16. RESULTS We confirm that IGS-subtypes are prognostic in EORTC22033-26033 clinical trial samples. Specific genetic changes segregate in distinct IGS subtypes: most samples assigned to IGS-9 have IDH-mutations combined with 1p19q codeletion, while samples assigned to IGS-17 have IDH-mutations without 1p19q codeletion (non-codeleted) and samples assigned to other intrinsic subtypes often are IDH-wildtype. A trend towards longer PFS in the RT arm as compared to TMZ was observed for samples assigned to IGS-9 (HR for TMZ is 1.90, 95% CI [0.95, 3.80], P=0.065), but not for samples assigned to IGS-17 (HR for TMZ vs RT is 0.87, 95% CI[0.50, 1.51], P=0.62). We did not identify genes significantly associated with progression free survival within intrinsic subtypes. Follow-up time is however limited. We also show that LGGs and GBMs differ in their immune-infiltrate with LGGs having higher suppressor (TREG, MDSC) and lower effector cell (CD4, CD8) populations compared to GBMs. This suggests that LGGs are less amenable to checkpoint inhibitor type immune therapies than GBMs. Copy number analysis of samples assigned to IGS-16 (a subtype to which pilocytic astrocytomas (PAs) are assigned) confirmed the typical genomic hallmark of PAs, a tandem duplication on 7q34, in one patient. CONCLUSION Intrinsic glioma expression subtypes are prognostic for PFS in EORTC22033-26033 clinical trial samples. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

Abstract Preoperative embolization of radiographically suspected meningiomas is often performed to facilitate subsequent tumor resection. Its effects on the post-operative course have not been studied in detail and randomized trials are lacking. To gain insight into the role of tumor embolization, we conducted a retrospective study in all patients undergoing resection of an intracranial meningioma at the University Hospital Zurich 2000-2013 with a post-operative follow-up of at least 3 months (N=741). Patients were reviewed for the inclusion of pre-operative embolization in the management strategy. Annotations included demographics, radiographic, surgical, histological and hematological parameters, cardiovascular risk factors, pre- and postoperative neurological function and genome-wide methylation-based classification. Binary regression and Cox proportional hazards models were applied to determine factors associated with outcome. Pre-operative embolization was performed in 337 patients (42%). The decision for embolization was associated with larger tumor size and presence of peritumoral edema. On multivariate analyses controlling for established prognostic factors, pre-operative embolization was associated with inferior post-operative neurological outcome (odds ratio [OR] 1.85, 95% confidence interval [CI] 1.30-2.63), whereas gross total resection was associated with better neurological outcome (OR 0.59, 95% CI 0.39-0.87). Cardiovascular events after surgery comprised mostly deep vein thrombosis (N=39) and pulmonary embolisms (N=64). There were also associations with embolization (OR 2.38, 95% CI 1.37-4.00) and female gender (OR 2.18, 95% CI 1.17-4.08). Recurrence-free survival (RFS) of embolized patients was less favorable among patients with WHO grade II or grade III meningiomas (median RFS: 4.3 versus 7.0 years, multivariate hazard ratio [HR] 1.92, 95% CI 0.97-3.79, p=0.062) or in patients with intermediate or malignant gene methylation subtype meningiomas (median RFS: 2.0 versus 8.2 years, multivariate HR 8.66, 95% CI 3.00-25.05, p<0.001). Pre-operative meningioma embolization may cause adverse surgical outcomes. Randomized trials to determine benefit-risk ratios are warranted to clarify the role of pre-operative embolization for meningioma surgery. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

Abstract WHO criteria now classify adult glioma into groups using combination of tumor grade and two tumor markers: 1p/19q codeletion and IDH1/2 mutation. TERT promoter mutation has also been shown to further classify tumors into clinically-meaningful groups. The Cancer Genome Atlas (TCGA) recently identified six methylation subtypes of glioma. Thus, as part of a clinical transformation process, we evaluated sensitivity and specificity of the Illumina methylation array to classify adult gliomas into appropriate WHO groups. We molecularly profiled 86 specimens from our institution via Illumina EPIC methylation array: 76 adult gliomas, seven other brain tumors (anaplastic ependymoma, glial neoplasm, pilocytic astrocytoma) and three gliosis samples. The 76 adult gliomas also have Affymetrix OncoScan array, IDH1/2 and TERT promoter mutation sequencing data. Using only methylation data, the EPIC array correctly classified 70 (92%) of 76 gliomas into appropriate WHO groups. When EPIC array copy number data were used to verify codeletion status, 71 (93%) were correctly classified. Similarly, using only Illumina methylation data, 560 (90%) of 622 TCGA gliomas were correctly classified into WHO groups. When methylation array copy number data were used to verify codeletion status 611 (98%) were correctly classified. While the OncoScan and EPIC methylation array had 100% concordant codeletion status, there were discrepancies for other genomic regions; especially in gliomas that carry IDH1/2, ATRX and TP53 mutations. All seven (100%) of the other brain tumors were classified as LGm6. With respect to TERT mutation, of 44 TCGA IDH wildtype, TERT wild-type and non-codeleted gliomas, 64% were classified as LGm6; thus, supporting the importance of TERT for further classifying gliomas. Overall, when methylation and 1p/19q copy number data are used, the Illumina methylation array correctly classified most adult gliomas into appropriate WHO groups. However, caution should be used when evaluating genome-wide copy number alterations solely using the methylation array. This content is only available as a PDF. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected]. © The Author(s) 2017. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].