Mutant Gliomas Research Articles

EPCO-23. EFFECTS OF H3.3G34V MUTATION ON GENOMIC H3K36 AND H3K27 METHYLATION PATTERNS IN ISOGENIC PEDIATRIC GLIOMA CELLS

Abstract BACKGROUND Histone H3.3 mutation (H3F3A) occurs in 50% of cortical pediatric high-grade gliomas (pHGGs). This mutation replaces glycine 34 with arginine or valine (G34R/V), impairing SETD2 activity (H3K36-specific trimethyltransferase). Consequently, reduced H3K36me3 is observed on H3.3G34V nucleosomes relative to wild-type, contributing to genomic instability and driving a distinct gene expression signature associated with tumorigenesis. However, it is not known if this differential H3K36me3 enrichment is due to H3.3G34V mutant protein alone. Therefore, we set to elucidate the effect of H3.3G34V mutant protein in pediatric glioma on H3K36me3, H3K27me3 and H3.3 enrichment in vitro. METHODS Experiments were conducted using an established H3.3G34V mutant patient-derived pediatric high-grade glioma cell line (KNS42), and Histone H3.3 wild-type human astrocytes (NHAs). Lentiviral delivery of a doxycycline-inducible RNAi vector targeted against H3F3A was transduced to KNS42 cells to knock down H3.3G34V protein expression. Cell lines were then compared via ChIP-Seq, Western Blot, proliferation and viability studies. RESULTS We found that the doxycycline-inducible shRNA knockdown of mutant H3F3A encoding the H3.3G34V protein resulted in loss of H3.3G34V enrichment and increased H3K36me3 enrichment throughout the genome. After knockdown, H3.3G34V enrichment was preserved at loci observed to have the greatest H3.3G34V and H3K36me3 enrichment prior to knockdown. Induced expression of mutant H3.3G34V protein in vitro was insufficient to induce genomic H3K36me3 enrichment patterns observed in H3.3G34V mutant glioma cells. We also observed strong co-enrichment of H3.3G34V and wild-type H3.3 protein, as well as greater H3K27me3 enrichment, in cells expressing H3.3G34V. CONCLUSION Our study demonstrates the effects of H3.3G34V mutant protein on genomic H3K36me3, H3K27me3 and H3.3 enrichment patterns in isogenic cell lines.

BIOM-09. OPPOSING EFFECTS OF LINKED CX3CR1 GERMLINE VARIANTS ON SURVIVAL AND IMMUNE ACTIVATION IN GLIOMA

Abstract Linkage-disequilibrium connects CX3CR1 V249I and T280M single-nucleotide polymorphisms; V249I occurs as a single-variant (SV) and T280M exclusively as a double-variant (DV) concurrent with V249I. In low-grade gliomas (LGG), we found improved median overall-survival (mOS) in SV (n = 14, mOS 21.03yrs), compared to WT (n = 45, mOS 9.41yrs) and DV (n = 31, mOS 10.49yrs; p = 0.037 Log-rank). In TCGA glioblastoma (GBM), SV (n = 69, mOS 15.65mo) improved survival over DV (n = 106, mOS 11.74mo; p = 0.044 Log-rank). In a separate GBM cohort, bulk RNA-sequencing revealed MMP9 over-expression in DV (n = 15) compared to SV (n = 10; log2FC 2.87, p-adj 0.079). On a single-gene basis, MMP9 expression was decreased in SV compared to DV and WT (n = 33; p = 0.044, Kruskal-Wallis). MMP9 promotes tumor angiogenesis and invasion. SV showed SPP1 over-expression compared to DV (log2FC -1.75, p-adj 0.026), with decreased SPP1 expression in both WT and DV compared to SV (p = 0.004, Kruskal-Wallis). SPP1 is a pro-inflammatory cytokine enriched in glioma-associated macrophages. Gene Ontology revealed enrichment of biological processes: neutrophil migration, granulocyte chemotaxis, and neutrophil chemotaxis in SV (p< 0.05). We single-cell RNA-sequenced IDH mutant gliomas (51777 cells). We subset and re-clustered myeloid-derived cells (microglia/macrophages, 8944 cells), identifying thirteen clusters. Myeloid-derived cells clustered by CX3CR1 status, with 50.8% of SV cells in cluster five and 85.3% of DV cells in clusters six or seven. Cluster five over-expressed IL1B (average-log2FC 3.23), a tumor-suppressive M1 macrophage marker. Clusters six and seven were characterized by IFIT genes – immunomodulatory in myeloid-derived cells (IFIT3 average-log2FC 2.65 and 2.06). Assessing CX3CR1 status in normal population – 74.2% are WT, 10.0% SV, and 15.7% DV. In glioma patients, we identify 59.9% WT, 14.9% SV, and 25.2% DV – an increase of 49% in SV and 61% in DV (p = 1.95e-6, Chi-square). Our findings demonstrate opposing effects of linked CX3CR1 germline variants on glioma survival and indicate enhanced immune activation in SV.

NIMG-51. RADIOGRAPHIC IMPROVEMENT IN A PATIENT WITH OLIGODENDROGLIOMA UNDERGOING TREATMENT FOR ACUTE MYELOID LEUKEMIA

Abstract INTRODUCTION: The standard chemotherapy for treating oligodendroglioma consists of a combination of procarbazine, lomustine, and vincristine (PCV). The combination of hypomethylating agents, like azacitidine, and BCL2 inhibitors like venetoclax have not been formally studied in the treatment of glial tumors. The combination of these two drugs is commonly used to treat acute myeloid leukemia, with IDH mutant disease being a particularly sensitive subtype. CASE REPORT: A 48-year-old man was diagnosed with a left frontal, grade II oligodendroglioma following multiple episodes of speech arrest and right-hand numbness. He was subsequently treated with six cycles of PCV with a minor response. After slow progression over eight years, the tumor started displaying contrast enhancement. While planning for radiation therapy, he presented to the Emergency Department with severe fatigue and shortness of breath. Workup led to a diagnosis of pure erythroid leukemia, with molecular and cytogenetic studies suggesting a neoplasm related to prior exposure to cytotoxic chemotherapy. He was treated with three cycles of venetoclax and azacitidine, but unfortunately, the erythroleukemia progressed after a transient response to this treatment. On the other hand, however, a re-staging MRI brain showed a remarkable reduction of his tumor size of approximately 40% per RANO criteria. The patient was off steroids and remained neurologically asymptomatic during this time. Discussion: The use of azacitidine for the treatment of IDH mutant gliomas has been reported in the literature, with mixed results that might suggest at least some benefit in a subtype of patients. Other studies have also suggested that the BCL2 gene is associated with treatment resistance and tumor recurrence in gliomas. Our case presents an interesting observation that combining a hypomethylating agent with a BCL2 inhibitor might convey clinical and radiographic benefits for patients with relapsed IDH mutant gliomas.

MODL-46. RECAPITULATING CELLULAR HETEROGENEITY AND TREATMENT-SPECIFIC EFFECTS THROUGH SINGLE-CELL ANALYSIS OF PATIENT-DERIVED TUMOR SLICE CULTURE MODELS

Abstract INTRODUCTION Malignant brain tumors are almost universally fatal despite standard-of-care (SOC) therapy with radiation (XRT) and temozolomide (TMZ). Low Grade Glioma (LGG) is a subclass of malignant glioma that can harbor an R132H mutation in Isocitrate Dehydrogenase 1 (IDH1). Extensive clinical trials have explored the clinical utility of IDH1 R132H mutation small molecule inhibitors such as ivosidenib and vorasidenib; however, the direct intracellular and intercellular effects of IDH1 R132H inhibition relative to SOC remain largely unknown. Objective This study examines these issues by developing a novel patient-specific tumor slice culture model to study the effects of both SOC and IDH1 R132H inhibition on IDH1 mutant glioma. Method En-bloc tumor samples from a patient with IDH mutant LGG undergoing surgery were sectioned into 400 um thick slices using a vibratome and cultured on porous microwell inserts for up to 14 days. Tumor slices were then either left untreated or treated with SOC (XRT/TMZ) or ivosidenib and harvested at various time points. Samples were hashed, and single-cell RNA sequencing was performed and analyzed. Result: Patient-derived tumor slice cultures were able to grow and were observed to contain heterogeneous cell populations, including tumor cells, astrocytes, neurons, oligodendrocytes, and microglia. When slice cultures were treated with SOC and ivosidenib treatment, separate tumor subpopulations were shown to be sensitive to either treatment. Furthermore, scRNA analysis revealed unique subpopulations that were insensitive to either treatment. We further demonstrate that slice cultures treated with Ivosidenib enriched for gene signatures related to protein translation and antigen processing. Intriguingly, we find that tumor cultures treated with ivosidenib abrogated the Notch signaling pathways between tumor-tumor and tumor-immune clusters. CONCLUSION Herein, we demonstrate that patient-derived tumor slice cultures are a viable strategy for understanding drug effects and tumor sensitivity to different therapies.

BIOS-03. COPY NUMBER VARIATION (CNV) AND CLINICAL OUTCOMES IN IDH-WILDTYPE VERSUS IDH-MUTANT GLIOMA

Abstract Genomic classification of gliomas has the potential for improved prognostication and treatment guidance. However, much work remains to elucidate the full potential of genomics in glioma characterization. IDH mutant gliomas generally carry an improved survival compared to their wild-type counterparts, but their clinical course is significantly more variable and challenging to predict. The goal of this study is to characterize tumor genetic/genomic parameters associated with progression-free survival (PFS) and overall survival (OS) of both IDH-wildtype and IDH-mutant gliomas and compare the two subtypes for further stratification. We collected data for 203 patients diagnosed with IDH-wildtype gliomas (WHO Grade IV) and 88 patients with IDH-mutant gliomas (WHO Grade II-IV) from the Emory Healthcare system. Microarray single nucleotide polymorphism-copy number analysis was performed to detect total CNVs which are associated with tumor mutational burden. We analyzed total CNV based on the median, quartiles, and optimal cut points via the ideal cutpoint calculator, Cutoff Finder. We then estimated the correlation between % total CNV, PFS, and OS using the Pearson correlation coefficient. For IDH-wildtype gliomas, the mean CNV was 18.69 (standard deviation (SD): 9.62), and the median was 16.00 (IQR: 1-57). The optimal cutoff was defined as 39 for PFS analysis and 20.5 for OS analysis. For IDH-mutant gliomas, the mean CNV was 12.8 (SD: 11.36), and the median was 8.5 (IQR: 5 – 17.5). The optimal cutoff was defined as 10.5 for OS analysis and 7.5 for PFS analysis. Overall, higher CNV was associated with lower PFS and OS in both groups. However, the results were only statistically significant for OS at the optimal cutoff (20.5) for IDH-wildtype gliomas (p-value = 0.0379), and for PFS at the optimal cutoff (7.5) for IDH-mutant gliomas (p-value = 0.0420). These findings suggest that CNV may be useful as a potential prognosis factor for both tumor subtypes in different contexts.

DDDR-09. FUNCTIONAL GENOMICS AND CHEMOTHERAPY RESISTANCE SCREENING IDENTIFIES BIOMARKERS AND TARGETABLE MECHANISMS OF CCNU RESISTANCE IN GLIOMAS

Abstract The bifunctional alkylator lomustine (CCNU) remains widely but heterogeneously used as one of the few effective treatments in glioblastoma (GBM) and IDH mutant gliomas. Despite its widespread use, mechanisms of response and resistance to this agent are controversial and patients who benefit are poorly understood. To address this with next-generation patient models and functional genomics approaches, we generated and characterized CCNU response in a large cohort of 71 patient-derived glioma cell lines (PDCLs, spheroids/organoids). Models had fully defined clinical and molecular characteristics (WES, RNAseq, and patient and model MGMT). Patient models, including those derived from patients who acquired resistance to CCNU, exhibited a surprisingly wide range of differential response to CCNU (median IC50: 21.66uM; range: 1.52uM-159uM) using metabolic (CTG) as well as cell death readouts, with no lines showing complete resistance (EC100 < 291uM). MGMT promoter methylation status but not mismatch repair (MMR) trended, but did not reach, statistical significance as a predictor of CCNU sensitivity (p=0.053), in striking contrast to TMZ (p=0.01). Chronic CCNU exposure of two PDCLs led to acquired resistance to CCNU (IC50 increase from 5uM to 50uM on average). RNA sequencing showed 194 significantly differentially expressed genes while no mutations acquired in resistant models were identified in DNA sequencing. PRISM pooled screening of over 900 barcoded cancer cell lines and correlation with DepMap whole genome CRISPR-knockout data showed unique patterns of CCNU response and strong upregulation of DNA repair targets in the resistant state, which differed from TMZ. Shared targets were also identified in a CRISPR-activator screen performed on a CCNU sensitive cell line in the presence and absence of CCNU and converged on multiple mTOR pathway genes as top enriched targets and biomarkers of resistance in CCNU-resistant models. These druggable targets are currently being investigated for combination therapies with CCNU and potential to overcome resistance.

Differential metabolic alterations in IDH1 mutant vs. wildtype glioma cells promote epileptogenesis through distinctive mechanisms.

Glioma-related epilepsy (GRE) is a hallmark clinical presentation of gliomas with significant impacts on patient quality of life. The current standard of care for seizure management is comprised of anti-seizure medications (ASMs) and surgical resection. Seizures in glioma patients are often drug-resistant and can often recur after surgery despite total tumor resection. Therefore, current research is focused on the pro-epileptic pathological changes occurring in tumor cells and the peritumoral environment. One important contribution to seizures in GRE patients is metabolic reprogramming in tumor and surrounding cells. This is most evident by the significantly heightened seizure rate in patients with isocitrate dehydrogenase mutated (IDHmut) tumors compared to patients with IDH wildtype (IDHwt) gliomas. To gain further insight into glioma metabolism in epileptogenesis, this review compares the metabolic changes inherent to IDHmut vs. IDHwt tumors and describes the pro-epileptic effects these changes have on both the tumor cells and the peritumoral environment. Understanding alterations in glioma metabolism can help to uncover novel therapeutic interventions for seizure management in GRE patients.

CircIQGAP1 regulates RCAN1 and RCAN2 through the mechanism of ceRNA and promotes the growth of malignant glioma

Circular RNA (circRNA) is one of non-coding RNA with specific circular structure, which has been found to be involved in regulating a series of malignant biological behaviors in many malignant tumors. In this study, based on the IDH1 molecular typing of gliomas, we identified a significant downregulation of circRNA (circIQGAP1) expression in IDH1 mutant gliomas by high-throughput sequencing. In 79 tissue samples, we confirmed that circIQGAP1 expression was significantly downregulated in IDH1 mutant gliomas, and that low circIQGAP1 expression was positively associated with better prognosis. Knockdown of circIQGAP1 in glioma cell lines inhibited glioma cell malignancy and conversely overexpression of circIQGAP1 promoted glioma malignancy. circIQGAP1 regulated glioma cell migration, proliferation, invasion and apoptosis through miR-1256/RCAN1/Bax/Bcl-2/Caspase3 and miR-622/RCAN2/Bax/Bcl-2/Caspase3 axes. These results suggest that circIQGAP1 plays an important role in glioma development, promotes tumor growth, and is a potential therapeutic target for glioma.

Use, access, and initial outcomes of off-label ivosidenib in patients with IDH1 mutant glioma.

Isocitrate dehydrogenase (IDH) is commonly mutated (mIDH) in gliomas, and this mutant enzyme produces the oncometabolite 2-hydroxyglutarate (2HG). 2HG promotes gliomagenesis and is implicated in epileptogenesis. Ivosidenib (IVO), a small molecule oral mIDH1 inhibitor, is FDA-approved for mIDH1 newly diagnosed and relapsed/refractory acute myeloid leukemia. Moreover, IVO has efficacy in clinical trials for recurrent mIDH1 gliomas. Given the lack of targeted treatments for gliomas, we initiated off-label IVO for mIDH glioma patients in October 2020. Retrospectively, we sought to assess early outcomes in our patients and describe their experience on IVO from October 2020 through February 2022. Our objective was to report on the following variables of off-label use of IVO: radiographic response, seizure control, tolerability, and access to the medication. All patients initially received single-agent IVO dosed at 500 mg orally once daily. The cohort age range was 21-74 years. Tumor types included astrocytoma (n = 14) and oligodendroglioma (n = 16), with most being grade 2 (n = 21). The best radiographic response in nonenhancing disease (n = 22) was 12 stable diseases, 5 minor responses, 3 partial responses, and 2 progressive diseases. Seizure frequency was stable to improved for most patients (70%, n = 21). IVO was well-tolerated, with the most common toxicities being diarrhea, elevated creatine kinase, and QTc interval prolongation. Most patients (66.7%, n = 20) received drugs via the patient assistance program, with insurance initially covering a third of patients and with ongoing use, later covering 60%. Targeted therapies like IVO are options for mIDH glioma patients and can provide positive oncologic and neurological outcomes.

H3K27M mutant glioma: Disease definition and biological underpinnings.

High-grade glioma (HGG) is the most common cause of cancer death in children and the most common primary central nervous system tumor in adults. While pediatric HGG was once thought to be biologically similar to the adult form of disease, research has shown these malignancies to be significantly molecularly distinct, necessitating distinct approaches to their clinical management. However, emerging data have shown shared molecular events in pediatric and adult HGG including the histone H3K27M mutation. This somatic missense mutation occurs in genes encoding one of two isoforms of the Histone H3 protein, H3F3A (H3.3), or HIST1H3B (H3.1), and is detected in up to 80% of pediatric diffuse midline gliomas and in up to 60% of adult diffuse gliomas. Importantly, the H3K27M mutation is associated with poorer overall survival and response to therapy compared to patients with H3 wild-type tumors. Here, we review the clinical features and biological underpinnings of pediatric and adult H3K27M mutant glioma, offering a groundwork for understanding current research and clinical approaches for the care of patients suffering with this challenging disease.

Proteomic landscape of primary and metastatic brain tumors for heterogeneity discovery.

Despite recent advancements in our understanding of driver gene mutations and heterogeneity within brain tumors, whether primary or metastatic (also known as secondary), our comprehension of proteomic changes remains inadequate. The aim of this study is to provide an informative source for brain tumor researches, and distinguish primary brain tumors and secondary brain tumors from extracranial origins based on proteomic analysis. We assembled the most frequent brain tumors as follows: gliomas from WHO grade 2 to 4, with IDH1 mutations and wildtypes; brain metastases (BrMs) originating from lung cancer (LC), breast cancer (BC), ovarian cancer (OC), and colorectal cancer (CC). A total of 29 tissue samples were analyzed by label free quantitative mass spectrometry-based proteomics. In total, 8165 protein groups were quantified, of which 4383 proteins were filtered at 50% valid intensity values for downstream analysis. Proteomic analysis of BrMs reveals conserved features shared among multiple origins. While proteomic heterogeneities were found for discriminating different grades of gliomas, as well as IDH1 mutant and wildtype gliomas. In addition, notable distinctions were observed at the pathway level between BrMs and gliomas. Specifically, BrMs exhibited characteristic pathways focused on proliferation and immunomodulation after colonizing the brain, whereas gliomas primarily engaged in invasion processes. We characterized an extensive proteomic landscape of BrMs and gliomas. These findings have promising implications for the development of targeted therapies for BrMs and gliomas.

Efficient diagnosis of IDH-mutant gliomas: 1p/19qNET assesses 1p/19q codeletion status using weakly-supervised learning

Accurate identification of molecular alterations in gliomas is crucial for their diagnosis and treatment. Although, fluorescence in situ hybridization (FISH) allows for the observation of diverse and heterogeneous alterations, it is inherently time-consuming and challenging due to the limitations of the molecular method. Here, we report the development of 1p/19qNET, an advanced deep-learning network designed to predict fold change values of 1p and 19q chromosomes and classify isocitrate dehydrogenase (IDH)-mutant gliomas from whole-slide images. We trained 1p/19qNET on next-generation sequencing data from a discovery set (DS) of 288 patients and utilized a weakly-supervised approach with slide-level labels to reduce bias and workload. We then performed validation on an independent validation set (IVS) comprising 385 samples from The Cancer Genome Atlas, a comprehensive cancer genomics resource. 1p/19qNET outperformed traditional FISH, achieving R2 values of 0.589 and 0.547 for the 1p and 19q arms, respectively. As an IDH-mutant glioma classifier, 1p/19qNET attained AUCs of 0.930 and 0.837 in the DS and IVS, respectively. The weakly-supervised nature of 1p/19qNET provides explainable heatmaps for the results. This study demonstrates the successful use of deep learning for precise determination of 1p/19q codeletion status and classification of IDH-mutant gliomas as astrocytoma or oligodendroglioma. 1p/19qNET offers comparable results to FISH and provides informative spatial information. This approach has broader applications in tumor classification.

P02.18.B INTEGRATIVE OMICS ANALYSIS OF IDH1 MUTANT GLIOMA PATIENTS REVEALS ALTERATIONS IN BUTYRATE METABOLISM

Abstract BACKGROUND Mutations in isocitrate dehydrogenase (IDH) 1 or 2 define glioma classification and determine the biology of these tumors. Although IDH is an essential enzyme in the cellular metabolism, powerful genome-wide analyses focus mostly at the genetic and epigenetic levels, while differences between distinct glioma entities at the proteomics, metabolomics, and functional levels are still poorly understood. Here, we provide an integrative multi-omics analysis of glioma patients to reveal metabolomic vulnerabilities of gliomas stratified by IDH mutation. MATERIAL AND METHODS We performed a multi-omics analysis of two cohorts of glioma patients (n=28) stratified based on IDH mutation status. Molecular assessment included LC-MS analysis, EPIC DNA methylation arrays, RNA-seq/HTA array-based transcriptomics and label-free quantitative proteomics. In 13 patients 13C-glucose was injected prior to surgery and tumor and adjacent brain tissue was collected. Gut microbial profiling using 16S ribosomal RNA (rRNA) sequencing was performed in a separate cohort of 128 glioma patients and 24 healthy individuals. Additional validation was performed in isogenic glioma stem-like cultures with CRISPR knock-in of IDH mutation. RESULTS We show that several molecular layers collectively contribute to biological differences between IDH mutant and wild-type gliomas, with strongest differences observed at the DNA methylation level. IDH mutant gliomas present specific molecular features linked to tumor metabolism. We confirm previous findings on 2-HG, cysteine and de novo glutathione synthesis pathways, which are specifically altered in IDH mutant gliomas. We further identify butyrate metabolism as a novel pathway highly active in IDH mutant gliomas. Butyrate is a bacterial metabolite with multiple systemic effects that is normally synthesized in the gut. This is supported by our gut microbiota metagenomics analysis of glioma patients. Yet the detection of enzymes associated with butyrate metabolism in tumor tissue and cell lines suggests a differential/enhanced metabolic rewiring of the butyrate signaling in IDH mutant gliomas CONCLUSION Our data shows the importance of multiple molecular layers in shaping specific features of IDH mutant gliomas. We unravel an activated butyrate metabolism in IDH mutant gliomas that may underly a novel gut-brain axis in glioma patients.

P14.05.B 2-HYDROXYGLUTARATE IN CEREBROSPINAL FLUID AS A MONITORING BIOMARKER FOR IDH-MUTANT GLIOMAS

Abstract BACKGROUND D-2-hydroxyglutarate (D-2-HG) is a well-established oncometabolite of isocitrate dehydrogenase (IDH) mutant gliomas. While prior studies have demonstrated that D-2-HG is elevated in the cerebrospinal fluid (CSF) of patients with IDH-mutant gliomas, no study has evaluated the ability of CSF D-2-HG to quantitatively monitor tumor burden and therapeutic response. Toward this goal, we are routinely obtaining CSF samples from patients prior to, during, and after treatment via intra-operative collection and CSF-access devices. MATERIAL AND METHODS Patients were enrolled into our Ommaya reservoir study (NCT04692337), which enables longitudinal access to CSF. The catheter for the implanted reservoir was placed within the surgical resection cavity following a standard-of-care resection. Intra-operatively collected CSF samples were also biobanked through Mayo Clinic’s Neuro-Oncology biorepository. CSF samples were also acquired via clinically indicated ventriculoperitoneal shunts (VP shunts) or external ventricular drains via our ongoing CSF biomarkers study (NCT04692324). D-and-L-2-HG were assayed from each CSF sample via gas chromatography-mass spectrometry (GC-MS) at Mayo Clinic’s Biochemical Genetics Laboratory. RESULTS CSF was collected prior to and after resection in patients via Ommaya reservoirs or VP shunts, all of whom underwent near gross total resections (>95% of tumor volume). D-2-HG and the D/L ratio fell on average by 57.9% (range: 15.1%-98.6%) and 68.8% (range: 49.3%-98.9%), respectively, following resection. In one patient with a largely intraventricular grade 4 IDH-mutant astrocytoma, D-and-L-2-HG fell by 99.1% and 99.3% by post-operative day 2, consistent with the impact of ventricular contact on baseline CSF D-2-HG levels. In patients for whom CSF was collected throughout chemoradiation, with or without immunotherapy, D-2-HG and the D/L ratio continued to decrease over time, despite concern for pseudoprogression versus progression on imaging. CONCLUSION Our early experience with this strategy demonstrates the potential for intracranial CSF D-2-HG as a monitoring biomarker for IDH-mutant gliomas. Work is ongoing to evaluate the reproducibility of these findings in a larger cohort of patients with low or high-grade gliomas, in addition to the impact of tumor recurrence on 2-HG. Future studies will evaluate the impact of IDH inhibitors and other experimental therapies on CSF 2-HG levels for pharmacodynamic monitoring of treatment response.

P10.17.A GLUTAMINE METABOLISM AS A PROGNOSTIC FACTOR IN GBM AND OLIGODENDROGLIOMA

Abstract BACKGROUND Increased uptake of principal nutrients is one of the attributes of cancer cells. A recent study has shown that glutamine, but not glucose, consumption differed markedly between gliomas and adjacent healthy tissue. However, the prognostic role of glutaminolysis or glutaminolysis-related genes (GRGs) in gliomas remains unclear. Here we sought to determine if the GRG signature correlates with glioma patient survival and if it could identify patients with different outcomes. MATERIAL AND METHODS The RNA-seq and clinical data for The Cancer Genome Atlas (TCGA) was obtained from the GlioVis webpage. We compared the expression of the GRGs (GLUD1, GLUL, GLS1, GLS2, GOT1, GOT2, GPT, GPT2, SLC1A5, and SLC7A5) between 4 non-tumor, 220 IDH wild-type (IDHwt) and 389 IDH mutant (IDHmut) gliomas, of which 154 harbored 1p/19q codeletion. We further performed unsupervised hierarchical clustering for all glioma patients and compared survival between clusters. RESULTS Our results revealed that most GRGs were differentially expressed between non-tumor, IDHwt, and IDHmut gliomas. We found that gliomas grouped into 3 clusters based on the varying GRGs expression. Eighty-eight percent of glioblastomas (GBM, IDHwt) and 19% of all IDHmut tumors were characterized by low GRGs expression. These were denoted as cluster 1, and the remaining samples belonged to clusters 2 or 3, corresponding to intermediate and high GRGs expression groups, respectively. We performed survival analysis separately for IDHwt and IDHmut gliomas and found that cluster 1 patients had the worst prognosis, regardless of IDH status. We further separated IDHmut gliomas according to 1p/19q codeletion status. Oligodendroglioma (1p/19q codeleted) patients had a median survival of 33.2 months, 134.3 months, and 169.8 months in clusters 1, 2, and 3, respectively. No cluster-dependent differences in survival were seen for astrocytoma. For IDHwt (GBM) we found that patients from clusters 1 and 2 had survival of 14 months and 25 months, respectively. Due to the small number of deceased GBM patients in cluster 3, we could not estimate the median survival. The univariate analysis confirmed that cluster membership was an independent predictor of survival in oligodendroglioma (cluster 2 and 3 vs. 1, HR 0.14; 95% CI, 0.05-0.4; p=0.0005) and GBM (cluster 1 vs. 2 and 3, HR 3.05; 95% CI, 1.5-5.9; p=0.0012). CONCLUSION The GRG signature could define metabolic groups with distinct survivals in oligodendroglioma and GBM and constitute a new prognostic factor. We are planning further investigations to confirm our findings.

P19.04.B EXAMINATION OF THE ROLES OF KDM6A AND KDM6B HISTONE DEMETHYLASES IN ISOCITRATE DEHYDROGENASE (IDH) MUTANT GLIOMA

Abstract Glioma, comprising 80% percent of all malignant brain tumors, is one of the most aggressive and difficult to treat tumors. Isocitrate dehydrogenase (IDH) is an enzyme that catalyzes the oxidative decarboxylation of isocitrate, producing α-ketoglutarate. Most IDH-mutant gliomas have a heterozygous point mutation in IDH1 that causes an arginine-to-histidine substitution at amino acid 132. Mutant IDH enzyme leads to 2-hydroxyglutarate accumulation in cells, inhibiting DNA and histone demethylases and ultimately generating a hypermethylated epigenetic phenotype. Considering that the mutant IDH enzyme generates a discrete epigenetic profile, we aimed to develop an epigenetic-based therapy approach for IDH mutant glioma. To this end, we performed chemical library screens consisting of inhibitors of important epigenetic enzymes and identified GSK-J4 as a chemical selectively targeting IDH1 mutant cells.GSK-J4 is a well-known inhibitor for histone demethylases KDM6A and KDM6B; therefore, we aimed to phenocopy the effects of the chemical approach with the genetic approach by depleting KDM6A and/or KDM6B using CRISPR/Cas9. We successfully established IDH-mutant primary glioma cell lines deficient in KDM6A and/or KDM6B expression, as revealed by qRT-PCR and Western Blots. Cells with double knockout were markedly slower in growth compared to controls which had received nontargeting gRNA or single gRNAs targeting each enzyme. In the future, we will examine changes in cell cycle and apoptosis pathways upon KDM6A and KDM6B knockouts and test the response of these cells to mutant IDH inhibitors, TMZ, or irradiation. In parallel, genome-wide changes in gene expression and cellular pathways, histone modifications, and chromatin accessibility will be analyzed, via RNA-seq, ChIP-seq, and ATAC-seq techniques, respectively.Even though KDM6A and KDM6B have been shown to have important but sometimes conflicting roles in various tumor types, no study showed their roles in IDH mutant glioma. Therefore, revealing the functional mechanisms of KDM6A and KDM6B in IDH-mutant gliomas will serve as an important step that leads to the development of novel epigenetic-based therapeutic approaches for IDH-mutant gliomas.This project is supported by TUBITAK with the given grant number: 219S882.

P12.05.B THE ROLE OF ONCO-METABOLITE IN THE IDH MUTANT GLIOBLASTOMA MICROENVIRONMENT

Abstract BACKGROUND Recently, we found that the reactive astrocytes in the IDH wild type glioblastoma contribute to anti-tumor immunity and support pro-oncogenic signaling. Here, we characterized the transcriptomic signature of IDH1/2-mutant glioma associated astrocytes and determined a unique inflammatory transformation, profoundly different to astrocytes in IDH wildtype glioma patients due to an onco-metabolite R-2-hydroxyglutarate using Next generation sequencing and human ex-vivo slice model. MATERIAL AND METHODS We purified and transcriptionally profiled astrocytes from 9 patients with confirmed IDH1-R132H mutation, by means of RNA-sequencing and the data were analyzed using the established pipelines. We also used spatial transcriptomics to evidently show the spatial distribution of astrocytes in IDH-mutated/wildtype glioma samples. We validated our findings using human organotypic slice model inoculated with IDH-mutant cell line or treated with oncometabolite 2-hydroxy glutarate. Additionally, LC-MS was further used to give us a chart of neurotransmitters due to altered microenvironment. RESULTS Our results from RNA sequencing showed a transcriptional transformation of reactive astrocytes within the microenvironment of IDH-mutated tumors compared to wildtype glioma by means of RNA-sequencing of purified astrocytes. And, using our established human neocortical GBM model inoculated with IDH mutant tumor and R-2HG treatment, we showed that we were able to activate inflammatory transcriptional programs in astrocytes, mediated by the presence of microglia. Further, by spatially mapping the transcriptomic profiles of purified microglia, we were able to confirm that microglia also demonstrate inflammatory activation in IDH mutated glioma. This inflammatory astrocyte transformation is associated with a loss of neurotransmitter homeostasis (disrupted levels of glutamate) in the treated sections, as has been previously reported in IDH mutated tumors. Additionally, R-H2G increased neuronal spiking rate in, pointing to potential excitotoxicity. CONCLUSION Our work provides a crucial contribution towards understanding the role of R-2HG in the IDH mutant glioma microenvironment and sheds light on the significant microenvironmental differences to IDH wild-type glioma.

P14.04.B CIRCULATING TUMOR CELLS AS A NON-INVASIVE GLIOBLASTOMA DIAGNOSTIC AND PROGNOSTIC TOOL

Abstract BACKGROUND Circulating tumor cells (CTCs) detection in the blood of glioblastoma (GBM) patients can facilitate the assessment of tumoral information without the need for invasive approaches. Our aim is to use GBM CTCs to obtain knowledge about the tumor and improve the current GBM diagnosis and prognosis. MATERIAL AND METHODS We collected primary, matched recurrent GBM and healthy controls blood samples. CTCs were enriched from whole blood controls with Parsortix Cell Separation System and analysed on DEPArray system. CTCs were recognized by immunofluorescence with GFAP.The CTCs have been counted for all the samples. Copy Number Aberrations (CNAs) analysis was performed on CTCs. To compare CTCs genetic background with the same patient’s primary and recurrence tissues, whole exome sequencing (WES) was realized on tumor tissues. Recurrence of GBM was diagnosed according to the RANO criteria +/-path proven second surgery +/- FET PET imaging proven. We assessed tumor volume of contrast enhanced (CE) (as a surrogate marker of BBB damage) and nCE at the time of the CTCs collection by seminautomatic segmentation on ITKsnap software. RESULTS We found CTCs in GBM patients newly diagnosed at the time of first surgery (86%) but also at recurrence (100 %). Cell counting demonstrated in both primary and recurrent GBM patients’ blood a statistically significantly higher number of CTCs compared to healthy controls (average of cells number 15.8 vs 2.3). Interestingly, recurrent GBM patients showed more CTCs than primary GBM patients (average 9 vs 22.7). CNAs analysis confirmed the tumor status of the selected CTCs. Some of GBM classical mutations (in HTERT promoter, BRAF, PTEN, PI3KCA and EGFR genes) have been searched both in the tumor tissues and in the CTCs. HTERT was found mutated in CTCs of one patient’s tumor carrying mutant HTERT and wild type IDH1/2. We also included three IDH1 mutant glioma patients and one BRAFV600E mutant GBM patient in order to track these in CTCs. We did not find BRAF mutations in CTCs isolated from one recurrent BRAFV600E GBM patient with a very low allele frequency in the recurrent tumor (<1%). CONCLUSION We found CTCs in GBM patients newly diagnosed and at recurrence. Several factors may affect CTCs release such as tumor volume, CE and the CTCs genetic BACKGROUND , which seems to be heterogeneous. CTCs detection by our methodology could be helpful in the clinical setting for diagnostic, prognostic and theranostic purposes and for patients monitoring. Prospective study is ongoing in order to assess CTCs variations on and after therapy. This study, first investigated the genetic heterogeneity of CTCs released in the blood flow from the parental tumor.

P19.06.B T2/FLAIR MISMATCH, SPECIFIC RADIOGENOMIC MARKER FOR A DIFFUSE LOW GRADE ASTROCYTOMA

Abstract A 36-year-old woman presented with recurring episodes of generalized tonic clonic seizures. An MRI from an outside institution demonstrated an intra-axial brain mass suspicious for a primary brain tumor. Repeat conventional and advanced MRI imaging revealed a primary brain mass centred on the left postcentral gyrus involving cortex and subcortical white matter. The lesion was homogenously hyperintense on T2-weighted imaging, and centrally iso-to hypointense on FLAIR with a rim of hyperintense signal on the FLAIR sequence. There was no significant peritumoral edema or minimal effacement of the central sulcus. No abnormal enhancement was seen post-contrast. The lesion demonstrated no reduced diffusivity on diffusion-weighted imaging. MR perfusion showed no elevation in regional cerebral blood volume or flow. Findings are most compatible with a low grade glioma. Patient underwent total surgical resection and biopsy for histopathology. Genetic sequencing revealed an IDH mutant, non 1P,19q codeleted grade II astrocytoma. The diagnosis of adult infiltrating gliomas is predicted using imaging, and confirmed by histopathology, and molecular diagnostic methods. Regardless of grade the 1st phase in glioma diagnosis and management is IDH status. A multidisciplinary approach to the diagnosis of low grade and high grade glioma including genetic and molecular analysis is essential, as different genetic and molecular subtypes have different management and prognostic implications. Conventional MRI still represents a mandatory and extremely important step in the diagnostic workflow of neuro oncological patients for preoperative diagnosis, treatment planning, and subsequent follow up. The T2/FLAIR mismatch sign describes the MRI appearance of an IDH-mutant astrocytoma that is non 1p19q codeleted, and it is considered a preoperative novel specific MRI marker for this type of diffuse astrocytoma. In this case, the MRI findings were used preoperatively as a noninvasive tool to diagnose the patient`s low grade glioma which was confirmed by subsequent histological and genetic analysis. TEACHING POINTS1. T2/FLAIR mismatch allows for pre-operative diagnosis of IDH-mutant astrocytomas with a high level of specificity.2. It is particularly helpful in distinguishing diffuse astrocytoma from an oligodendroglioma3. Conventional MR sequences can suggest the diagnosis of IDH mutant gliomas which can assist neurosurgeons in presurgical planning and can help neuro-oncologists with treatment planning including potential clinical trials.

HDAC1 and HDAC6 are essential for driving growth in IDH1 mutant glioma

Low-grade and secondary high-grade gliomas frequently contain mutations in the IDH1 or IDH2 metabolic enzymes that are hypothesized to drive tumorigenesis by inhibiting many of the chromatin-regulating enzymes that regulate DNA structure. Histone deacetylase inhibitors are promising anti-cancer agents and have already been used in clinical trials. However, a clear understanding of their mechanism or gene targets is lacking. In this study, the authors genetically dissect patient-derived IDH1 mutant cultures to determine which HDAC enzymes drive growth in IDH1 mutant gliomas. A panel of patient-derived gliomasphere cell lines (2 IDH1 mutant lines, 3 IDH1 wildtype lines) were subjected to a drug-screen of epigenetic modifying drugs from different epigenetic classes. The effect of LBH (panobinostat) on gene expression and chromatin structure was tested on patient-derived IDH1 mutant lines. The role of each of the highly expressed HDAC enzymes was molecularly dissected using lentiviral RNA interference knock-down vectors and a patient-derived IDH1 mutant in vitro model of glioblastoma (HK252). These results were then confirmed in an in vivo xenotransplant model (BT-142). The IDH1 mutation leads to gene down-regulation, DNA hypermethylation, increased DNA accessibility and H3K27 hypo-acetylation in two distinct IDH1 mutant over-expression models. The drug screen identified histone deacetylase inhibitors (HDACi) and panobinostat (LBH) more specifically as the most selective compounds to inhibit growth in IDH1 mutant glioma lines. Of the eleven annotated HDAC enzymes (HDAC1-11) only six are expressed in IDH1 mutant glioma tissue samples and patient-derived gliomasphere lines (HDAC1-4, HDAC6, and HDAC9). Lentiviral knock-down experiments revealed that HDAC1 and HDAC6 are the most consistently essential for growth both in vitro and in vivo and target very different gene modules. Knock-down of HDAC1 or HDAC6 in vivo led to a more circumscribed less invasive tumor. The gene dysregulation induced by the IDH1 mutation is wide-spread and only partially reversible by direct IDH1 inhibition. This study identifies HDAC1 and HDAC6 as important and drug-targetable enzymes that are necessary for growth and invasiveness in IDH1 mutant gliomas.