Development of a rabbit human glioblastoma model for testing of endovascular selective intra-arterial infusion (ESIA) of novel stem cell-based therapeutics

BACKGROUND

Endovascular selective intra-arterial (ESIA) infusion of cellular oncotherapeutics is a rapidly evolving strategy for treating glioblastoma. Evaluation of ESIA infusion requires a unique animal model. Our goal was to create a rabbit human GBM model to test IA infusions of cellular therapies and to test its usefulness by employing clinical-grade microcatheters and infusion methods to deliver mesenchymal stem cells loaded with an oncolytic adenovirus, Delta-24-RGD (MSC-D24).

METHODS

Rabbits were immunosuppressed with mycophenolate mofetil, dexamethasone, and tacrolimus. They underwent stereotactic xenoimplantation of human GBM cell lines (U87, MDA-GSC-17, and MDA-GSC-8-11) into the right frontal lobe. Tumor formation was confirmed on magnetic resonance imaging, histologic, and immunohistochemistry analysis. Selective microcatheter infusion of MSC-D24 was performed via the ipsilateral internal carotid artery to assess model utility and the efficacy and safety of this approach.

RESULTS

Twenty-five rabbits were implanted (18 with U87, 2 MDA-GSC-17, and 5 MDA-GSC-8-11). Tumors formed in 68% of rabbits (77.8% for U87, 50.0% for MDA-GSC-17, and 40.0% for MDA-GSC-8-11). On MRI, the tumors were hyperintense on T2-weighted image with variable enhancement (evidence of blood brain barrier breakdown). Histologically, tumors showed phenotypic traits of human GBM including varying levels of vascularity. ESIA infusion into the distal internal carotid artery of 2 ml of MSCs-D24 (107 cells) was safe in the model. Examination of post infusion specimens documented that MSCs-D24 homed to the implanted tumor at 24 hours.

CONCLUSIONS

The intracranial immunosuppressed rabbit human GBM model allows testing of ESIA infusion of novel therapeutics (eg, MSC-D24) in a clinically relevant fashion.

Investigative needle core biopsies for multi-omics in Glioblastoma

Glioblastoma (GBM) is a primary brain cancer with an abysmal prognosis and few effective therapies. The ability to investigate the tumor microenvironment before and during treatment would greatly enhance both understanding of disease response and progression, as well as the delivery and impact of therapeutics. Stereotactic biopsies are a routine surgical procedure performed primarily for diagnostic histopathologic purposes. The role of investigative biopsies - tissue sampling for the purpose of understanding tumor microenvironmental responses to treatment using integrated multi-modal molecular analyses ('Multi-omics") has yet to be defined. Secondly, it is unknown whether comparatively small tissue samples from brain biopsies can yield sufficient information with such methods. Here we adapt stereotactic needle core biopsy tissue in two separate patients. In the first patient with recurrent GBM we performed highly resolved multi-omics analysis methods including single cell RNA sequencing, spatial-transcriptomics, metabolomics, proteomics, phosphoproteomics, T-cell clonotype analysis, and MHC Class I immunopeptidomics from biopsy tissue that was obtained from a single procedure. In a second patient we analyzed multi-regional core biopsies to decipher spatial and genomic variance. We also investigated the utility of stereotactic biopsies as a method for generating patient derived xenograft models in a separate patient cohort. Dataset integration across modalities showed good correspondence between spatial modalities, highlighted immune cell associated metabolic pathways and revealed poor correlation between RNA expression and the tumor MHC Class I immunopeptidome. In conclusion, stereotactic needle biopsy cores are of sufficient quality to generate multi-omics data, provide data rich insight into a patient's disease process and tumor immune microenvironment and can be of value in evaluating treatment responses.

One sentence summary

Integrative multi-omics analysis of stereotactic needle core biopsies in glioblastoma.

Abstract 5788: Single-cell and spatial transcriptomic mapping of human renal cell carcinoma brain metastases uncovers actionable immune-resistance targets

Introduction: The discovery of immune checkpoint inhibitors has revolutionized metastatic renal cell carcinoma (RCC) treatment. However, in patients with RCC brain metastases, response rates are low and survival outcomes poor. To understand the tumor microenvironmental differences between primary kidney tumors, extracranial metastases, and brain metastases, we developed a detailed single-cell atlas of RCC brain metastases along with their matched extracranial and primary tumors.

Methods: We performed single-nucleus RNA-seq on 27 samples (nearly 200,000 cells) from RCC patients; samples included 14 brain metastases, 8 matched primary kidney tumors, and 5 matched extracranial metastases. We performed multiplex IHC to validate selected transcriptomic findings. We used Nanostring CosMx 960-plex RNA spatial molecular imaging technique on selected samples to validate cellular interactions in a spatial context.

Results: We established a multi-tissue single-cell atlas of RCC brain metastases by identifying 9 major and 37 minor malignant, immune, and stromal cell clusters. Brain metastases had higher neuronal and glial cells interacting with immune and tumor cells. Brain metastasis tumor cells were also transcriptomically reprogrammed to adapt to the brain microenvironment through enrichment of MYC targets, MTORC1 signaling, epithelial-mesenchymal transition, fatty-acid metabolism, oxidative phosphorylation, and reactive oxygen species pathways. Moreover, cell-to-cell communication and downstream target gene expression analyses showed that brain metastasis tumor cells expressed ligands and receptors that induce tumor cell proliferation in both autocrine and paracrine fashions. Among T-cell populations, we found fewer proliferating cytotoxic T lymphocytes in the brain than in other sites. Moreover, T cells in brain metastases expressed higher levels of several targetable inhibitory checkpoints than did extracranial metastases. In addition, we found that naïve/memory T cells in brain metastases were a favorable prognostic marker for overall survival after craniotomy. Our characterization of myeloid cell populations across the 3 disease sites found fewer dendritic cells and monocytes in the brain compared to other sites. Macrophages in brain metastases more highly expressed an M2 immunosuppressive gene signature than did those in primary RCC tumors.

Conclusion: Our findings from the largest single-cell atlas of RCC brain metastases with matched primary and extracranial metastases suggest several unique targetable, immunosuppressive biological mechanisms in the brain microenvironment. These results provide a foundation for a deeper understanding of RCC brain metastasis biology and can serve as a resource for the scientific community to further explore therapeutically targetable tumor and immune-related mechanisms.

Citation Format: Elshad Hasanov, Truong Nguyen Anh Lam, Jerome Lin, Patrick K. Reville, Merve Hasanov, Anna K. Casasent, David Shih, Sahin Hanalioglu, Mehmet Asim Bilen, Omar Alhalabi, Berrin Babaoglu, Baylar Baylarov, Adeboye O. Osunkoya, Lisa M. Norberg, Joy Gumin, Tuan M. Tran, Jianzhuo Li, Anh G. Hoang, Haidee D. Chancoco, Brittany C. Parker Kerrigan, Erika J. Thompson, Betty YS Kim, Dima Suki, Melike Mut, Figen Soylemezoglu, Giannicola Genovese, Kadir C. Akdemir, Hussain A. Tawbi, Nizar M. Tannir, Florencia McAllister, Michael A. Davies, Padmanee Sharma, Jason Huse, Frederick Lang, Nicholas Navin, Eric Jonasch. Single-cell and spatial transcriptomic mapping of human renal cell carcinoma brain metastases uncovers actionable immune-resistance targets [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5788.

Gibbs process distinguishes survival and reveals contact-inhibition genes in Glioblastoma multiforme

Tumor growth is a spatiotemporal birth-and-death process with loss of heterotypic contact-inhibition of locomotion (CIL) of tumor cells promoting invasion and metastasis. Therefore, representing tumor cells as two-dimensional points, we can expect the tumor tissues in histology slides to reflect realizations of spatial birth-and-death process which can be mathematically modeled to reveal molecular mechanisms of CIL, provided the mathematics models the inhibitory interactions. Gibbs process as an inhibitory point process is a natural choice since it is an equilibrium process of the spatial birth-and-death process. That is if the tumor cells maintain homotypic contact inhibition, the spatial distributions of tumor cells will result in Gibbs hard core process over long time scales. In order to verify if this is the case, we applied the Gibbs process to 411 TCGA Glioblastoma multiforme patient images. Our imaging dataset included all cases for which diagnostic slide images were available. The model revealed two groups of patients, one of which - the "Gibbs group," showed the convergence of the Gibbs process with significant survival difference. Further smoothing the discretized (and noisy) inhibition metric, for both increasing and randomized survival time, we found a significant association of the patients in the Gibbs group with increasing survival time. The mean inhibition metric also revealed the point at which the homotypic CIL establishes in tumor cells. Besides, RNAseq analysis between patients with loss of heterotypic CIL and intact homotypic CIL in the Gibbs group unveiled cell movement gene signatures and differences in Actin cytoskeleton and RhoA signaling pathways as key molecular alterations. These genes and pathways have established roles in CIL. Taken together, our integrated analysis of patient images and RNAseq data provides for the first time a mathematical basis for CIL in tumors, explains survival as well as uncovers the underlying molecular landscape for this key tumor invasion and metastatic phenomenon.

EPCO-23. DYSFUNCTION OF LARGE H3K9ME3 DOMAINS IN ATRX DEFICIENT GLIOMAS INDUCES GENETIC REARRANGEMENTS AND LATENT DEVELOPMENTAL SIGNALING NETWORKS THROUGH SUPER-ENHANCER LANDSCAPES

Somatic alterations in ATRX (Alpha Thalassemia/Mental Retardation Syndrome X-linked), a member of SWI/SNF family chromatin regulator has been found to be frequently mutated in diffuse gliomas and defines molecular subtypes with aggressive behavior. Mechanistically, ATRX regulates incorporation of histone H3.3 into chromatin sites across the genome, maintains alternative lengthening of telomeres and establishes genomic distribution of polycomb responsive genes. To understand in depth role of ATRX in gliomas, we performed ChIP-seq and/or Cut-and-tag for histone marks that define active transcription, enhancers, repressors, gene bodies and CTCF on Atrx intact and deficient mNPCs (mouse Neural Progenitor Cells). Our integrated analysis reports depletion of H3K9me3 loci’s that coincidently enriched with Atrx binding sites and Lamina-Associated Domains with genes enriched for cell cycle, motility and chromosome organization. This chromatin perturbations at heterochromatin domains was further confirmed in our Hi-C analysis with switching of A-B and B-A compartments, reorganization of TADs with occasional CTCF marks and gain of novel interacting loops that showed gene expression leakage required for gliomagenesis using Capture Hi-C. Notably, we observed aberrant levels of endogenous retroviral elements (ERVs) family members, including upregulation of Line-1 elements in mNPCs and patient derived glioma stem cells (GSCs) and our analysis shows increased copy number variations in ATRX deficient gliomas as a consequence of Line-1 activation in these subsets of tumors. Finally, our integrated omics- analysis demonstrates enrichment of super-enhancers in Atrx deficient background with several putative druggable candidates for clinical benefits. As an example, we show presence of H3K27ac super-enhancer over HOXA locus and targeting pan-HOXA genes using small inhibitor peptides diminished proliferation and migration of mNPCs and GSCs with increased in cell apoptosis with alterations in downstream developmental signaling pathways. To summarize,our data establishes tangible links between Atrx deficiency and multiscale dysregulated cellular phenotype in gliomas with identifying novel targets for therapy.

DNAR-11. CHARACTERIZING THE GENOMIC CONSEQUENCES OF G-QUADRUPLEX STABILIZATION IN ATRX-DEFICIENT HIGH-GRADE GLIOMA

α-thalassaemia/mental retardation X-linked (ATRX) mutations are a critical molecular marker for high-grade glioma (HGG). These mutations lead to accumulations of abnormal DNA secondary G-quadruplex (G4) structures, thereby inducing replication stress and DNA damage. As G4s arise at GC-rich regions (i.e., pericentromeric and telomeric regions), ATRX-deficiency alters genome-wide accessibility of chromatin and causes transcriptional dysregulation. However, the genomic consequences of this in the context of ATRX-deficiency are poorly understood. Our goal is to target ATRX deficiency through G4 stabilizers, a class of novel small molecule compounds that selectively bind to and stabilize G4s. Using a combination of functional experiments such as cell viability, western blot, flow cytometry, RNA-sequencing (RNA-seq), and immunofluorescence (IF), we evaluated the mechanisms that drive selective lethality upon G4 stabilization. Patient-derived glioma stem cells (GSCs) were treated with either vehicle (DMSO) or varying doses of CX-5461 (G4 stabilizer, Senhwa Biosciences). Excitingly, ATRX-deficient GSCs demonstrate dose-dependent enhanced sensitivity to G4 stabilization, compared to ATRX-intact and vehicle controls. Cell viability assays confirmed the specificity of CX-5461 in comparison to other commercially used G4 stabilizers. G4 stabilization activated p53-independent apoptosis and exhibited G2/M arrest in ATRX-deficient GSCs and, interestingly, upregulated expression of both ATR and ATM pathways, indicating enhanced replication stress and DNA damage, respectively. IF staining confirmed enhanced induction of replication stress and DNA damage markers 53BP1 and gH2AX. Our preliminary findings suggest that ATR and ATM activation leads to Cyclin D1 degradation and inhibition of transcription factor NF-κB, thereby driving apoptosis. In fact, RNA-seq analyses revealed positive enrichment of apoptosis, DNA repair, and NF-κB pathways and negative enrichment of the G2/M checkpoint in ATRX-deficient GSCs treated with CX-5461. Our work defines mechanisms of action and efficacy of a novel therapeutic strategy for pre-clinical ATRX-deficient HGG models, with strong implications for other ATRX-deficient cancers and potential translation into clinical practice.

IMMU-10. EXPLORING THE IMMUNOLOGIC CONSEQUENCES OF ATRX DEFICIENCY IN GLIOMA

ATRX is a key chromatin regulator, which is mutated in large subsets of both adult and pediatric gliomas. Despite being a common mutation, little is known about the biological ramifications of ATRX deficiency. Recently, it has been demonstrated that ATRX deficiency drives increased replication stress, DNA damage, and global epigenetic dysregulation. Despite these advances, little is known about the impact of ATRX deficiency on the tumor microenvironment (TME). In order to explore the impact of ATRX deficiency on the TME we utilized the RCAS/nTVa system to generate a novel murine model of ATRX deficient glioma. Mice bearing allografts of these tumors displayed significantly increased survival relative to the ATRX intact lines. This survival benefit persisted in Nu/nu mice, which lack an adaptive immune system, but not in SCID mice, which lack both adaptive and innate immunity. Bulk RNA sequencing revealed the ATRX deficient tumors displayed increased expression of inflammatory and innate immune gene sets, and western blotting revealed increased phosphorylation of tank binding kinase 1 (TBK1), a key innate immune regulator. Multiplex cytokine analysis of conditioned media also revealed increased expression of inflammatory cytokines such as CCL2, CCL5, and CXCL10 in the supernatant of ATRX deficient cell lines. Taken together, these results indicate that ATRX deficiency can drive innate immune activation, and that this activation can increase survival in vivo. Further exploration is needed to characterize the upstream drivers of TBK1 activation and elucidate alterations to immune cell infiltration in ATRXdeficient tumors.

Investigating Useful Features for Overall Survival Prediction in Patients with Low-Grade Glioma Using Histology Slides

Glioma, characterized by neoplastic growth in the brain, is a life-threatening condition that, in most cases, ultimately leads to death. Typical analysis of glioma development involves observation of brain tissue in the form of a histology slide under a microscope. Although brain histology images have much potential for predicting patient outcomes such as overall survival (OS), they are rarely used as the sole predictors due challenges presented by unique characteristics of brain tissue histology. However, utilizing histology in predicting overall survival can be useful for treatment and quality-of-life for patients with early-stage glioma. In this study, we investigate the use of deep learning models on histology slides combined with simple descriptor data (age and glioma subtype) as a predictor of (OS) in patients with low-grade glioma (LGG). Using novel clinical data, we show that models which are more attentive to discriminative features of the image will confer better predictions than generic models (82.7 and 65.3 AUC RFD-Net and Baseline VGG16 model, respectively). Additionally, we show that adding age and subtype information to a histology image-based model may provide greater robustness in the model than using the image alone (3.8 and 4.3 stds for RFD-Net and Baseline VGG16 model with 3-fold CV, respectively), while a model based on image and age but not subtype may confer the best predictive results (83.7 and 82.0 AUC for RFD-Net + age and RFD-Net + age + subtype, respectively). Clinical relevance- This study establishes important criteria for deep learning models which predict OS using histology and basic clinical data from LGG patients.

Abstract 2548: The central nervous system immune cell interactome is a function of cancer lineage, tumor microenvironment and STAT3 expression

Introduction: Deconstructive immune cell profiling of central nervous system (CNS) tumors has focused on the tumor, excluding interrogation of the tumor microenvironment (TME). Integrated spatial analysis can ascertain the cell interactome and may be a key biomarker for effective anti-tumor immune responses.

Methods: En bloc resections of glioma (n=10) and lung metastasis (n=10) to preserve the tissue architecture, underwent tissue segmentation and high dimension opal 7-color multiplex imaging. Bioinformatic analysis of scRNA was used to infer immune cell functionality.

Results: CD3+ T cell frequency was equivalent between CNS cancer lineages. Within gliomas T cells were confined to the perivascular space and the infiltrating edge. In lung metastasis, T cells are confined to the tumor stroma. CD163+ macrophages predominate in brain metastasis throughout the TME (p<0.05), while CD68+ monocytes (CD68+, CD11c+CD68+, and CD11+CD68+CD163+) are more common in gliomas (p<0.05). T cell dyad and cluster immune interactions were more common in the absence of nuclear STAT3 expression. T cells usually interact with CD163+ macrophages as dyads in metastasis at the brain interface (p=0.031) and within tumor (p=0.0009); in clusters throughout the TME (interface: p=0.024; tumor: p=0.01; necrosis: p=0.045), and as STAT3+ dyads and cluster interactions in the tumor (p<0.05). Immune suppressed CD11c+CD163+ dendritic cells (tumor: p=0.036; and necrosis p=0.020) predominate in metastasis. In contrast, gliomas typically lacked dyad and cluster interactions except for T cell and CD68+ dyads in the tumor (p=0.023). Bioinformatic analysis of CD45+ scRNA seq data revealed that the majority of innate immune populations express both pro-inflammatory and immune suppressive genes and that subsets of CD68+ and CD11c+CD68+ cells expressed markers such as TMEM119, P2YR13 and CX3CR1 that identify microglia.

Conclusion: Current therapies are targeted to cell populations and singular pathways. Immunosuppressive macrophages dominate within the TME and targeting this population may create an environment that favors T cell activation and effective immune responses. Furthermore, the immune interactome, an important event for anti-tumor immune response, is a function of cancer lineage, TME, and STAT3 expression, which will gain relevance for future therapeutics directed to modulating these interactions.

Citation Format: Hinda Najem, Martina Ott, Cynthia Kassab, Arvind Rao, Ganesh Rao, Anantha Marisetty, Adam M. Sonabend, Craig Horbinski, Roel Verhaak, Anand Shakar, Santhoshi Krishnan, Frederick S. Varn, Victor A. Arietta, Pravesh Gupta, Sherise D. Ferguson, Jason Huse, Gregory N. Fuller, James Long, Dan Winskowski, Ben Freiberg, C. David James, Leonidas C. Platanias, Maciej S. Lesniak, Jared K. Burks, Amy B. Heimberger. The central nervous system immune cell interactome is a function of cancer lineage, tumor microenvironment and STAT3 expression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2548.

Abstract 2395: Homozygous MTAP deletion in primary human glioblastoma is not associated with elevation of methylthioadenosine

Homozygous deletion of the CDK2NA locus frequently results in the co-deletion of methylthioadenosine phosphorylase (MTAP) in many fatal cancers such as glioblastoma multiform (GBM). In cell culture, cell lines with MTAP-deletions show elevations of its substrate metabolite, methylthioadenosine (MTA). High levels of MTA inhibit PRMT5, which sensitizes MTAP-deleted cell lines to PRMT5 and MAT2A inhibition. While extensively corroborated in vitro, the clinical efficacy of these strategies ultimately relies on equally significant accumulations of MTA in human tumors. In this work, using comprehensive metabolomic profiling, we show that MTA is primarily secreted, resulting in exceedingly high levels of extracellular MTA in vitro. We further show that primary human glioblastoma tumors minimally accumulate MTA in vivo, which is likely explained by the metabolism of MTA by MTAP-competent stromal cells. Together, these findings highlight the metabolic discrepancies between in vitro models and primary human tumors and should thus be carefully considered in the development of the precision therapies targeting MTAPhomozygous deleted GBM.

Citation Format: Yasaman Barekatain, Jeffrey Ackroyd, Victoria Yan, Sunada Khadka, Ko-Chien Chen, Raghu Kalluri, John de Groot, Jason Huse, Florian muller. Homozygous MTAP deletion in primary human glioblastoma is not associated with elevation of methylthioadenosine [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2395.

CBIO-04. G-QUADRUPLEX STABILIZATION ENHANCES REPLICATION STRESS AND DNA DAMAGE IN ATRX-DEFICIENT HIGH-GRADE GLIOMA

Loss of function mutations in α-thalassaemia/mental retardation X-linked (ATRX) are a critical molecular hallmark for invariably fatal high-grade glioma (HGG). Mutational inactivation of histone chaperone ATRX leads to accumulations of abnormal DNA secondary structures known as G-quadruplexes (G4s), thereby inducing replication stress and DNA damage. As G4s arise at GC-rich regions (i.e., pericentromeric and telomeric regions), ATRX-deficiency alters genome-wide accessibility of chromatin, leads to transcriptional dysregulation, and induces alternative lengthening of telomeres (ALT). Our goal is to target ATRX deficiency through G4 stabilizers, which represent a class of novel small molecule compounds that selectively bind to and stabilize G4 structures. However, the genomic consequences and efficacy of this therapy for ATRX-deficient HGG are poorly understood. We therefore sought to evaluate the molecular mechanisms that drive selective lethality in patient-derived ATRX-deficient glioma stem cells (GSCs), following G4 stabilization. We found that ATRX-deficient GSCs demonstrate dose-dependent enhanced sensitivity to G4 stabilization, compared to ATRX-intact controls. Cell viability assays confirmed the specificity of our G4 stabilizer in comparison to other commonly used G4 stabilizers. Interestingly, G4 stabilization activated p53-independent apoptosis in ATRX-deficient GSCs. Furthermore, ATRX-deficient GSCs exhibit upregulated expression of both ATR and ATM pathways upon G4 stabilization, indicating enhanced replication stress and DNA damage via double-stranded breaks, respectively. Our preliminary findings suggest that ATR and ATM activation leads to the inhibition of transcription factor NF-κB, which in turn drives apoptosis. Lastly, our data indicate that G4 stabilization perturbs the ALT phenotype in ATRX-deficient GSCs, likely contributing to telomeric dysfunction. Taken together, these findings suggest that G4 stabilizers could synergize with ionizing radiation, the standard of care, as they are both DNA-damaging therapies. Our work defines mechanisms of action and efficacy of a novel therapeutic strategy for ATRX-deficient HGG, with strong implications for other ATRX-deficient cancers.

EPCO-19. SYSTEMS BIOLOGY APPROACH ON MGMT-METHYLATED, IDH WILD-TYPE SHORT-TERM SURVIVORS REVEALS MUTATIONS IN THE BRCA1-MEDIATED DNA REPAIR SIGNALING PATHWAY

Numerous multi-omic studies have revealed the vast molecular complexity of glioblastoma; however, these studies have failed to identify actionable drivers for definitive therapy. Unlike conventional relational data analysis methods used to examine such data, systems biology approaches using graph databases have emerged as powerful tools with flexibility and scalability needed to reveal specific vulnerabilities. Here we demonstrate the utility of a patient-centric graph database in identifying novel factors associated with poor outcome (survival < 8 months) in patients with MGMT promoter methylated glioblastoma. Using a cohort of 112 patients from the TCGA database, we integrated high-impact mutations with MGMT methylation status. Network analysis revealed three subnetworks, consisting of (a) methylated patients, the M-network, (b) methylated and unmethylated patients, the MU-network, and (c) unmethylated patients, the U-network. In addition, querying the genes in the M-network revealed three key molecules in the DNA repair pathway, namely, FANC-A, FANC-E, and MYUTH, mutated in five patients (q-value < 2.3E-02). Moreover, we observed that BRCA1 mediates other critical signaling molecules in the MU-network. Interestingly, in intermediate (8-24 months) or long-term ( >24 months) survivors networks, this pathway is not implicated. In light of recent studies implicating BRCA1 modulating temozolomide resistance in GBM sphere-forming cells and BRCA1’s protein expression predicting survival in patients, this result suggests that mutations in BRCA1-mediated DNA repair pathways hinder response to chemotherapy despite methylated MGMT, reducing the survival in patients. Besides this novel result explaining low survival in MGMT methylated patients through a synthesis of epigenetic and genetic data, our framework provides a novel and compelling paradigm for data integration at various scales from molecular events in a single patient’s tumor to integrating the molecular profile (or genotype) with phenotypic characteristics at the population level. We will present additional insights derived from this analysis at the SNO annual meeting.

EPID-09. VARIATION IN GLIOMA INCIDENCE AMONG US HISPANICS BY GEOGRAPHIC REGION OF ORIGIN

BACKGROUND

Glioma incidence is 25% lower in U.S. Hispanics than in White non-Hispanics. The US Hispanic population is diverse and registry-based analyses may mask incidence differences associated with geographic/ancestral origins.

METHODS

County-level glioma incidence data in U.S. Hispanics were retrieved from the Central Brain Tumor Registry of the United States (CBTRUS), which includes data from the Centers for Disease Control’s National Program of Cancer Registries and the National Cancer Institute’s Surveillance, Epidemiology, and End Results program and covers ~100% of the U.S. population. American Community Survey (ACS) data were used to determine county-level proportion of the Hispanic population of Mexican/Central American origin, Caribbean origin (Puerto Rican, Cuban, Dominican), or other origin. Incidence rate ratios (IRRs) were generated to assess the association of glioma incidence in Hispanics with predominant origin group.

RESULTS

Compared to Hispanics living in predominantly Caribbean-origin counties, Hispanics in predominantly Mexican/Central American-origin counties were at lower age-adjusted risk of glioma (IRR=0.83; P< 0.0001), glioblastoma (IRR=0.86; P< 0.0001), diffuse and anaplastic astrocytoma (IRR=0.78; P< 0.0001), oligodendroglioma (IRR=0.82; P< 0.0001), ependymoma (IRR=0.88; P=0.0121), and pilocytic astrocytoma (IRR=0.76; P< 0.0001). Associations were consistent in children and adults, and when using more granular regions of origin. However, Central American origin was associated with modestly increased incidence of several lower-grade glioma histologies. Associations were only partially attenuated after adjusting for state-level estimated of European admixture in Hispanics using 23andMe data.

CONCLUSIONS

Glioma incidence in U.S. Hispanics differs significantly in association with the geographic origins of the Hispanic community, with those of Mexican/Central American origin at significantly reduced risk relative to those of Caribbean origin. U.S. Hispanics are culturally, socioeconomically, and genetically diverse. Although classified as a single ethnic group in most registry data, more granular analytic approaches could advance cancer epidemiology and disparities research.

EPCO-08. ATRX DEFICIENCY INDUCES DYSFUNCTIONAL HETEROCHROMATIN ARCHITECTURE IN GLIOMAS AND ESTABLISHES DISEASE-DEFINING TRANSCRIPTIONAL NETWORKS

Loss of ATRX (Alpha Thalassemia/Mental Retardation Syndrome X, a member of SWI/SNF family chromatin regulator is altered in diffuse gliomas and defines molecular subtypes with aggressive behavior. Mechanistically, ATRX regulates incorporation of histone H3.3 into chromatin sites across the genome, maintains alternative lengthening of telomeres and establishes genomic distribution of polycomb responsive genes. We have recently reported Atrx deficiency induces glioma oncogenic features via widespread alterations in chromatin accessibility using mouse Neural Progenitor Cells (mNPCs- Tp53 -/-,Atrx -/-). Surprisingly, Atrx was found to be associated with transcription start site and enhancer regions, suggesting their strong association with epigenome architecture. In this background, we have recently performed ChIP-seq for histone marks that define active transcription, enhancers, repressors and gene bodies and Cohesion molecules on Atrx intact and deficient mNPCs. Our integrated analysis reports depletion of H3K9me3 loci’s with enrichment of H3K27me3 marks that coincidently enriched with Atrx binding sites and Lamina-Associated Domains (LADs). GSEA confirmed that the genes corresponding to “newly formed LADs” in mNPC-to-astrocyte differentiation are significantly enriched for genes down-regulated in Atrx deficient mNPCs and belongs to Gene Ontology categories such as cell cycle, chromosome organization and DNA damage. Alternatively, genes corresponding to decreased LAD association are enriched for up-regulated genes in Atrx deficient mNPCs and for regulation of differentiation, adhesion and cell death. Additionally, whole-genome bisulphite sequencing further demonstrated loss of methylation marks at H3K9me3 sites in Atrx deficient mNPCs, suggesting perturbations of heterochromatin regions leading to activation of canonical signals that define glioma phenotype and disease-state. To summarize, our data establishes tangible links between Atrx deficiency and dysregulated chromatin and heterochromatin architecture in gliomas and suggests the role of Atrx in establishing global chromatin features and transcriptional networks. Further, our data unravel novel therapeutic molecules/pathways for engineering potential.

EPCO-09. LONGITUDINAL ANALYSIS OF DIFFUSE GLIOMA REVEALS CELL STATE DYNAMICS AT RECURRENCE ASSOCIATED WITH CHANGES IN GENETICS AND THE MICROENVIRONMENT

Diffuse glioma is characterized by a poor prognosis and a universal resistance to therapy, though the evolutionary processes behind this resistance remain unclear. The Glioma Longitudinal Analysis (GLASS) Consortium has previously demonstrated that therapy-induced selective pressures shape the genetic evolution of glioma in a stochastic manner. However, single-cell studies have revealed that malignant glioma cells are highly plastic and transition their cell state in response to diverse challenges, including changes in the microenvironment and the administration of standard-of-care therapy. To interrogate the factors driving therapy resistance in diffuse glioma, we collected and analyzed RNA- and/or DNA-sequencing data from temporally separated tumor pairs of over 300 adult patients with IDH-wild-type or IDH-mutant glioma. In a subset of these tumor pairs, we complemented these data with single-nucleus RNAseq and multiplex imaging mass cytometry at each time point. Recurrent tumors exhibited diverse changes that were attributable to changes in histological features, somatic alterations, and microenvironment interactions. IDH-wild-type tumors overall were more invasive at recurrence and exhibited increased expression of neuronal signaling programs that reflected a possible role for neuronal interactions in promoting glioma progression. In contrast, recurrent IDH-mutant tumors exhibited a significant increase in proliferative expression programs that correlated with discrete genetic changes. Hypermutation and acquired CDKN2A homozygous deletions associated with an increase in proliferating stem-like malignant cells at recurrence in both glioma subtypes, reflecting active tumor expansion. A transition to the mesenchymal phenotype was associated with the presence of a specific myeloid cell state defined by unique ligand-receptor interactions with malignant cells, providing opportunities to target this transition through therapy. Collectively, our results uncover recurrence-associated changes in genetics and the microenvironment that can be targeted to shape disease progression following initial diagnosis.

BSCI-07. Multiomics characterization of brain metastases in multiple histologies identifies enrichment of oxidative phosphorylation as a promising therapeutic target

Purpose

Brain metastasis (BM) is a lethal complication from systematic malignant tumors, and the incidence is approximately 10–30% of patients with advanced cancer. Extensive genomic analyses with large sample sets and the following functional studies revealed clinically relevant characteristics for BMs. However, these studies have not identified specific abnormalities driving BM in multiple tumor histologies yet. To identify molecular pathogenesis and promising therapeutic targets shared across multiple histologies of BMs, we performed multiomics molecular profiling, along with functional studies using in vitro and in vivo BM models.

Methods

Frozen tissues of patient-matched BMs and primary tumors (or extracranial metastases) from breast cancer (N= 14), lung cancer (N = 14) and renal cell carcinomas (N = 7) patients were carried out whole-exome sequencing, mRNA-Seq and reverse-phase protein array. Paired parental and brain metastatic derivatives of MDA-MB-231 and BT474 were examined to assess findings from the multiomics datasets. SCID/beige mice were inoculated with MDA-IBC3 cells via tail vein injection and administered an oxidative phosphorylation (OXPHOS) inhibitor by oral gavage daily for 96 days.

Results

The multiomics molecular profiling identified enrichment of OXPHOS shared across the histologies of BMs. Brain metastatic derivative cell lines also demonstrated enhanced oxidative metabolism, along with the sensitivity to an OXPHOS inhibitor. Moreover, in vivo studies revealed that OXPHOS inhibition significantly impaired the formation of BM, and fresh brain metastatic derivatives from the murine BM model exhibited the higher oxidative metabolism and sensitivity to the OXPHOS inhibitor as with the prior in vitro studies.

Conclusions

Our multiomics characterization of BMs demonstrates heightened oxidative metabolism shared across the multiple histologies, and the OXPHOS inhibition affects more effectively for brain metastatic derivatives rather than the parentals. Further investigation focusing on metabolic abnormalities in BM will likely develop promising therapeutic strategies against BMs.

RARE-23. DIFFUSE LEPTOMENINGEAL GLIONEURONAL TUMOR: A CASE SERIES

Introduction

Diffuse leptomeningeal glioneuronal tumor (DLGNT) is a rare diagnosis first incorporated into the WHO Classification of Tumors of the Central Nervous System in 2016. Though historically considered indolent, emerging evidence suggests that the biological behavior of these tumors may be further classified by molecular features of prognostic significance.

Methods

A retrospective review was conducted in accordance with IRB approval of patients with the histologic diagnosis of DLGNT. Demographic, clinical, and molecular data where abstracted from the medical record when available.

Results

10 patients were identified (M = 8, F = 2). Median age at diagnosis was 6 years (range 0.3–21 years), and the most common symptoms at diagnosis were related to obstructive hydrocephalus, for which 3 patients required CSF diversion. MRI findings included diffuse leptomeningeal thickening, nodularity, or coating of the subarachnoid or ependymal surfaces. Histologically, these tumors expressed variable features of neuronal and/or glial differentiation. Four patients (40%) were treated with radiation therapy (all craniospinal), which was upfront for 2 patients. Chemotherapy regimens used included temozolomide, carboplatin and vincristine and vinblastine. NTRK or BRAF-targeted therapy were used upon progression. At follow-up, 6/10 had stable disease (4/6 of whom were on second line therapy), 1 had partial response, 1 passed away from sepsis and 2 were lost to follow-up. The median progression-free survival for the four patients who developed disease progression was 26 months (range 12–34 months). Next generation sequencing of the tumor tissue performed using a high-multiplex PCR-based NGS panel detected BRAF-KIAA1549 (4 patients) and NTRK (1 patient) fusions.

Conclusions

DLGNT are rare tumors with scarce data about imaging characteristic and standard of care treatment. Our case series reinforces current literature that although these tumors appear low-grade, they can be clinically aggressive. Further study is needed regarding molecular diagnosis and profiling treatment strategies.

Window-of-opportunity clinical trial of pembrolizumab in patients with recurrent glioblastoma reveals predominance of immune-suppressive macrophages

BACKGROUND

We sought to ascertain the immune effector function of pembrolizumab within the glioblastoma (GBM) microenvironment during the therapeutic window.

METHODS

In an open-label, single-center, single-arm phase II "window-of-opportunity" trial in 15 patients with recurrent (operable) GBM receiving up to 2 pembrolizumab doses before surgery and every 3 weeks afterward until disease progression or unacceptable toxicities occurred, immune responses were evaluated within the tumor.

RESULTS

No treatment-related deaths occurred. Overall median follow-up time was 50 months. Of 14 patients monitored, 10 had progressive disease, 3 had a partial response, and 1 had stable disease. Median progression-free survival (PFS) was 4.5 months (95% CI: 2.27, 6.83), and the 6-month PFS rate was 40%. Median overall survival (OS) was 20 months, with an estimated 1-year OS rate of 63%. GBM patients' recurrent tumors contained few T cells that demonstrated a paucity of immune activation markers, but the tumor microenvironment was markedly enriched for CD68+ macrophages.

CONCLUSIONS

Immune analyses indicated that pembrolizumab anti-programmed cell death 1 (PD-1) monotherapy alone can't induce effector immunologic response in most GBM patients, probably owing to a scarcity of T cells within the tumor microenvironment and a CD68+ macrophage preponderance.

EPEN-10. SPINAL MYXOPAPILLARY EPENDYMOMA AND METHYLATION-PROFILING: THE MD ANDERSON CANCER CENTER (MDACC) EXPERIENCE

INTRODUCTION

Spinal myxopapillary ependymoma (MPE) is a rare histological variant of ependymoma, classified as WHO grade I tumor. Further interrogation of the molecular and clinical profile is warranted, to better understand the biology and clinical phenotype. We summarize our institutional experience with spinal MPE including methylation-profiling.

METHODS

A retrospective analysis of charts during the period of 2001 to 2019 of histologically proven MPE was done. We performed methylation profiling for 12 patients by Infinium MethylationEPIC Kit.

RESULTS

26 patients with spinal MPE were identified, median age of diagnosis was 34.2 years with a range of 11 to 59.9 years. Ten patients were below 30 years of age, lumbar spine location was commonest and 6 had leptomeningeal spread at diagnosis. All the patients underwent surgery and 11 received radiation following surgery. Eight patients below the age of 30 received radiation due to residual disease or metastases. Methylation profiling revealed 11,752 CpGs differentially methylated between the younger and older patients (p < 0.05), however only one CpG cg22496254 associated with gene NCAPG/DCAF16 (role in promoting mitosis) was detectable with FDR < 0.25 that overly methylated in the younger age group. This is a new finding in MPE.

CONCLUSIONS

Spinal MPE is a rare spinal tumor. Our study though limited by numbers, showed younger patients had aggressive phenotype, most requiring radiation. Methylation profiling reaffirmed this finding and trend in the younger patients. Prospective studies in a larger cohort of patients with methylation profiling are needed to identify prognostic variables and new targets for treatment.

IMMU-18. INTERPLAY BETWEEN IDH1 AND ATRX MUTATIONS GOVERN INNATE IMMUNE RESPONSES IN GLIOMAS

Innate-based immunotherapies are becoming increasingly important for treating brain tumor patients. About 50% of WHO grade II and III gliomas carry mutations in IDH1 and ATRX genes. Mutant IDH1 results in the production of 2-hydroxyglutarate, an oncometabolite that promotes global metabolic and epigenetic alterations. ATRX is a SWI-SNF chromatin remodeling protein that is involved in cell cycle regulation and maintenance of genomic stability. Both IDH1 and ATRX mutations have been implicated in dysfunctional immune signaling in cancer cells. However, the interplay between these mutations in mediating innate immune responses has not been investigated in gliomas. We have derived human and mouse glioma cell lines carrying mutations in IDH1 (IDH1R132H) and/or ATRX, which we then used to generate both immune competent and nude mice models. Treating the ATRX knockout (KO) cell lines with dsRNA-based innate stimuli led to an early induction in phospho-IRF3, and late induction in phospho-STAT1 and ISG15, suggesting that ATRX deletion may enable a potent activation of type I interferon production and sensitize glioma cells to dsRNA-based innate stimuli. Our syngeneic murine glioma models confirm a survival advantage for mice carrying ATRX-KO tumors. These tumors also exhibit enhanced infiltration of T-cells and expression of activated macrophage markers. On the other hand, presence of IDH1R132H led to a suppression in baseline expression of key innate immune players, which could be rescued by the mutant IDH1 inhibitor, BAY-1436032. Cells harboring IDH1R132H and ATRX-KO retained sensitivity to dsRNA indicating that IDH1R132H does not dampen the ATRX KO-mediated sensitivity to dsRNA. Our models are under evaluation with a combination of BAY-1436032 and clinically relevant dsRNA-based innate therapies. Our data indicates the presence of an interplay between IDH1 and ATRX mutations that may regulate innate signaling in gliomas. Understanding the mechanisms governing this interplay may aid in designing therapies that exploit this interplay.

TAMI-62. ANGIOGENESIS INHIBITORS STRONGLY SYNERGIZE WITH THERAPEUTICS TARGETING TUMOR METABOLISM

Angiogenesis inhibition has become a mainstay of oncology despite having fallen short of its early promise. As originally envisioned, angiogenesis inhibition would cut off the blood supply, deprive tumor cells of key nutrients, leading to their demise. In practice, while there is evidence that tumors under angiogenesis treatment do in fact exhibit some degree of metabolic stress, this is stress is not sufficient to induce significant cancer cell death. We posit that the full potential of angiogenesis inhibition can be realized by the combination of angiogenesis inhibition with emerging tumor metabolism targeting therapies. Because tumors under angiogenesis inhibition are already in a state of nutrient stress, the effects of metabolically targeted therapies such as amino acid depletion (e.g. asparginase, methionine restriction), inhibitors of stress adaption (AMPK and GCN2 inhibitors) or energy metabolism (e.g. IACS-010759, Metformin, POMHEX) stand to dramatically increase in potency whilst remaining selective for (angiogenic) tumor versus (non-angiogenic) normal tissue. Here, we provide proof-of-principal for this thesis. First, we performed metabolomic profiling of angiogenesis-inhibited tumors, which corroborates as state of nutrient stress in angiogenesis-inhibited tumors. Second, we demonstrate dramatic anti-neoplastic synergy (effectively curing of xenografted tumor-bearing mice, irrespective of initial tumor size), without enhanced adverse toxicities, between the OxPhos inhibitor IACS-010759 and the angiogenesis tyrosine kinase inhibitor, Tivozanib. The same results were recapitulated with the anti-VEGFA antibody, Avastin, and the OxPhos inhibitor could be substituted with the Enolase inhibitor HEX, with similar effects. The synergy was observed in a broad range of tumor types, even those without clear genetic susceptibilities. Together, these results suggest that Angiogenesis inhibitors synergize broadly with cancer therapies targeting metabolism, allowing the realization of the full potential of these previously disappointing drugs. Our results warrant systematic combination clinical trials between angiogenesis inhibitors and established, as well as emerging anti-metabolic cancer therapies.

CSIG-09. ATRX DEFICIENCY IN GLIOMA IMPACTS TRANSCRIPTIONAL PROFILES AND THE IMMUNE MICROENVIRONMENT IN VIVO

Current treatment for diffuse astrocytoma fails to address its underlying molecular mechanisms leading to inevitable disease progression and eventual patient death. Genomic studies have implicated ATRX alterations as critical to low grade glioma biology. Our lab has previously shown in vitro that ATRX influences glioma motility, cellular differentiation state, and epigenetic programming, however, the influence of ATRX alterations in vivo remains unclear. Here, we leveraged an RCAS/tva mouse tumor model to probe the role of ATRX deficiency in glioma. Atrx deficient murine tumors exhibited lower histopathological grade and were associated with longer survival than Atrx-intact counterparts, and syngeneic allografts of cell lines derived from primary tumors mirrored the differential degrees of aggressiveness seen in primary tumors. Tumor-derived Atrx-deficient cell lines showed increased susceptibility to G-quadruplex stabilizing compounds, pointing to increased replication stress and recapitulating a key phenotype of ATRX-mutant gliomas in humans. Transcriptional profiling revealed enrichments in MYC target genes, E2F targets as well as G2/M checkpoint pathways in Atrx-intact tumors and cells, and enrichment in RAS signaling in Atrx-deficient tumors and cells. Finally, Atrx deficient murine gliomas displayed increased levels of NK cells, a phenotype recapitulated in ATRX-mutant human gliomas, and primary Atrx-deficient glioma lines exhibited increased levels of NK cell-attracting cytokines. These latter findings suggest that ATRX deficiency could influence interactions between glioma cells and their immune microenvironment by way of phenotypically relevant molecular mechanisms.

CBIO-18. G-QUADRUPLEX STABILIZATION TARGETS ATRX-DEFICIENT HIGH-GRADE GLIOMA VIA INDUCTION OF p53-INDEPENDENT APOPTOSIS

Gliomas are the most common primary malignant brain tumor in adults and mutational inactivation of histone chaperone ATRX is a critical molecular marker in the classification of high-grade glioma (HGG). ATRX loss occurs with concurrent mutations in TP53 and IDH1/2, altering genome-wide accessibility of chromatin and inducing replication stress and DNA damage via accumulations of abnormal G-quadruplex (G4) DNA secondary structures. While G4 stabilizers in particular hold strong therapeutic promise, the genomic consequences and efficacy of this treatment are poorly understood. We previously showed that chemical stabilization of G4s in ATRX-deficient normal human astrocytes (NHAs) results in lethality due to induction of replication stress, but it is unknown what drives this lethality in ATRX-deficient patient-derived preclinical models. We therefore sought to evaluate the mechanisms that underlie cell death in ATRX- and p53-deficient preclinical in vitro models following G4 stabilization. We found that ATRX-deficient glioma stem cells (GSCs) demonstrated dose-dependent enhanced sensitivity to G4 stabilization, compared to ATRX-intact controls. Evaluation of cell death mechanisms following G4 stabilization revealed a significant increase in cleaved caspase 3 expression and no p21 expression in ATRX-deficient GSCs, suggesting p53-independent apoptotic activation. Cell cycle flow analysis demonstrated G2/M checkpoint arrest in ATRX-deficient GSCs upon G4 stabilization, suggesting that p53 is nonfunctional at the G1/S checkpoint. Our preliminary findings now suggest that p73, a functional and structural homologue of p53, is activated and drives apoptosis in these ATRX-deficient GSCs. Furthermore, ATRX-deficient GSCs demonstrated upregulated expression of both pATR/pChk1 and pATM/pChk2, indicating enhanced replication stress and DNA damage via double-stranded breaks, respectively. These findings indicate that G4 stabilizers could potentially synergize with ionizing radiation, the current standard of care, as both therapies are DNA-damaging. Taken together, this study elucidates mechanisms of cytotoxicity and efficacy of a novel therapeutic strategy in ATRX-deficient preclinical models.

CBIO-03. ATRX LOSS IN GLIOMA RESULTS IN EPIGENETIC DYSREGULATION OF CELL CYCLE PHASE TRANSITION

Gliomas are a leading cause of cancer mortality in children and adults, and new targeted therapies are desperately needed. ATRX is a chromatin remodeling protein that is recurrently mutated in H3F3A-mutant pediatric glioblastoma (GBM) and IDH-mutant grade 2/3 adult glioma. We previously showed that loss of ATRX in glioma results in tumor growth and additional tumor mutations. However, the mechanism driving these phenotypes has not been fully established. We found that in ChIP-Seq/ChIP-qPCR of mouse neuronal precursor cells (NPCs) and GBM cells with isogenic ATRX loss, ATRX binds regulatory elements for cell cycle phase transition gene sets, and ATRX loss subsequently results in reduced expression. Furthermore, human GBM cells with ATRX knock-out demonstrate higher rates of cells in S and G2/M phases, with clusters of cells demonstrating reduced expression of cell cycle regulatory gene sets by single-cell sequencing (scSeq) analysis. In human and mouse GBM in vitro models, ATRX-deficient cells exhibit a seven-fold increase in mitotic index at 16 hours after sub-lethal radiation and enhanced activation of the master cell cycle regulator ATM with radiation. Treatment of ATRX-deficient gliomas with ATM inhibitors results in a selective increase in dysfunctional cell cycling and increased radio-sensitization in ATRX-deficient glioma cells. Using an ATM-luciferase reporter in orthotopically-implanted human GBM cells, both AZD0156 and AZD1390 demonstrate in vivo pathway inhibition. Mice intra-cranially implanted with ATRX-deficient GBM cells demonstrate a doubling of median survival compared to radiated controls (p=0.0018) when treated with AZD0156 combined with radiation; this is not seen in ATRX-sufficient models. This study demonstrates that ATRX-deficient high-grade gliomas display epigenetic dysregulation of cell cycle phase transitions, which opens a new window for therapies targeting this unique phenotype.

EXTH-06. INTEGRATED MOLECULAR PROFILING REVEALS TARGETABLE MOLECULAR ABNORMALITIES SHARED ACROSS MULTIPLE HISTOLOGIES OF BRAIN METASTASIS

Brain metastases (BMs) occur in approximately 20–40% of patients with advanced cancer, and the estimated prevalence of new BMs in United States is between 200,000–300,000 per year. While the incidence of BM has increased over the past decades due to improvements in brain tumor detection technology, the prognosis is still very poor with the median overall survival times from weeks to few months. Therefore, identification of the precise molecular landscape and therapeutic targets for BMs is absolutely essential in tangible improvement of patient management. Here, we performed integrated genomic, transcriptional, and proteomic profiling in a cohort of lung, breast, and renal cell carcinomas consisting of both BMs and patient-matched primary or extracranial metastatic tissues to identify shared cellular and molecular factors driving BMs across distinct primary tumor histologies. Although the comprehensive analysis identified the unique genomic, transcriptional and proteomic landscapes according to the different histopathologies, elevated PI3K/AKT and RAS/MAPK signaling was observed as a generalizable feature across the entire specimen cohort, along with relative immunosuppression and metabolic upregulation of the electron transport chain (ETC). Interestingly, immunosuppression via T cell depletion was significantly associated with unfavorable prognosis of patients with BMs, and ETC inhibition as the prospective therapeutic target for BM patients was demonstrated using in vitro and in vivo disease models. Taken together, our findings suggest that abnormalities involving oncogenic signaling, metabolism, and the immune microenvironment are shared across multiple histologies of BMs, and may be amenable to therapeutic targeting.

Abstract 5704: Pan-cancer genomic characterization of patient-matched primary, extracranial, and brain metastases

Brain metastases (BM) occur in 10-30% of patients with cancer. Approximately 200,000 new cases of brain metastases are diagnosed in the United States annually, with median survival after diagnosis ranging from 3 to 27 months. Recently, studies have identified significant genetic differences between BM and their corresponding primary tumors. It has been shown that BM harbor clinically actionable mutations that are distinct from those in the primary tumor samples. Additional genomic profiling of BM will provide deeper understanding of the pathogenesis of BM and suggest new therapeutic approaches.

We performed whole-exome sequencing of BM and matched tumors from 41 patients collected from renal cell carcinoma (RCC), breast cancer, lung cancer, and melanoma, which are known to be more likely to develop BM. We profiled total 126 fresh-frozen tumor samples and performed subsequent analyses of BM in comparison to paired primary tumor and extracranial metastases (ECM). We found that lung cancer shared the largest number of mutations between BM and matched tumors (83%), followed by melanoma (74%), RCC (51%), and Breast (26%), indicating that cancer type with high tumor mutational burden share more mutations with BM. Mutational signatures displayed limited differences, suggesting a lack of mutagenic processes specific to BM. However, point-mutation heterogeneity revealed that BM evolve separately into different subclones from their paired tumors regardless of cancer type, and some cancer driver genes were found in BM-specific subclones. These models and findings suggest that these driver genes may drive prometastatic subclones that lead to BM. 32 curated cancer gene mutations were detected and 71% of them were shared between BM and primary tumors or ECM. 29% of mutations were specific to BM, implying that BM often accumulate additional cancer gene mutations that are not present in primary tumors or ECM. Co-mutation analysis revealed a high frequency of TP53 nonsense mutation in BM, mostly in the DNA binding domain, suggesting TP53 nonsense mutation as a possible prerequisite for the development of BM. Copy number alteration analysis showed statistically significant differences between BM and their paired tumor samples in each cancer type (Wilcoxon test, p &lt; 0.0385 for all). Both copy number gains and losses were consistently higher in BM for breast cancer (Wilcoxon test, p =1.307e-5) and lung cancer (Wilcoxon test, p =1.942e-5), implying greater genomic instability during the evolution of BM.

Our findings highlight that there are more unique mutations in BM, with significantly higher copy number alterations and tumor mutational burden. These genomic analyses could provide an opportunity for more reliable diagnostic decision-making, and these findings will be further tested with additional transcriptomic and epigenetic profiling for better characterization of BM-specific tumor microenvironments.

Citation Format: Olivia W. Lee, Akash Mitra, Won-Chul Lee, Kazutaka Fukumura, Hannah Beird, Miles Andrews, Grant Fischer, John N. Weinstein, Michael A. Davies, Jason Huse, P. Andrew Futreal. Pan-cancer genomic characterization of patient-matched primary, extracranial, and brain metastases [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5704.

Pan-cancer analysis of whole genomes

Cancer is driven by genetic change, and the advent of massively parallel sequencing has enabled systematic documentation of this variation at the whole-genome scale1-3. Here we report the integrative analysis of 2,658 whole-cancer genomes and their matching normal tissues across 38 tumour types from the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium of the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). We describe the generation of the PCAWG resource, facilitated by international data sharing using compute clouds. On average, cancer genomes contained 4-5 driver mutations when combining coding and non-coding genomic elements; however, in around 5% of cases no drivers were identified, suggesting that cancer driver discovery is not yet complete. Chromothripsis, in which many clustered structural variants arise in a single catastrophic event, is frequently an early event in tumour evolution; in acral melanoma, for example, these events precede most somatic point mutations and affect several cancer-associated genes simultaneously. Cancers with abnormal telomere maintenance often originate from tissues with low replicative activity and show several mechanisms of preventing telomere attrition to critical levels. Common and rare germline variants affect patterns of somatic mutation, including point mutations, structural variants and somatic retrotransposition. A collection of papers from the PCAWG Consortium describes non-coding mutations that drive cancer beyond those in the TERT promoter4; identifies new signatures of mutational processes that cause base substitutions, small insertions and deletions and structural variation5,6; analyses timings and patterns of tumour evolution7; describes the diverse transcriptional consequences of somatic mutation on splicing, expression levels, fusion genes and promoter activity8,9; and evaluates a range of more-specialized features of cancer genomes8,10-18.

GENE-34. THERAPEUTICALLY TARGETING EPIGENOMIC AND TRANSCRIPTIONAL DYSFUNCTION IN ATRX-DEFICIENT GLIOMA

Diffusely infiltrating gliomas feature loss-of-function mutations in the chromatin remodeler gene ATRX as defining molecular alterations delineating major adult and pediatric disease subtypes. We recently reported that Atrx deficiency drives glioma-relevant phenotypes, such as increased motility and astrocytic differentiation profiles, by directly modulating epigenomic landscapes and the corresponding transcriptional profiles in glioma cells of origin. In particular, Atrx deficiency was associated with disruptions in H3.3 histone content at key genetic loci. To further understand the downstream epigenomic dysfunction induced by ATRX deficiency, we compared genome-wide chromatin-state maps of Atrx+ and Atrx- primary murine neuroepithelial progenitors (mNPCs). This ChIP–seq analysis revealed major differences in the localization of heterochromatin repressive marks H3K9me3 and H3K27me3. Specifically, we identified peculiar locations in the genome displaying H3K9me3 depletion and gain of H3K27me3 upon Atrx inactivation. Interestingly, these regions were flanked by Atrx binding sites and perfectly co-localized with Lamina-Associated Domains, known to play important roles in tissue lineage specification. To better target this dysfunction, we utilized the Broad Institute Connectivity Map (CLUE analysis) to identify compounds likely to revert the unique transcriptional perturbations induced by Atrx deficiency. We found that HDAC inhibitors, as a compound class, yielded expression profiles strongly anticorrelated to those driven by Atrx deficiency in these datasets. Further integrating existing gene expression data from our mNPCs and the TCGA LGG project with our CLUE findings highlighted SIRT2, a class III HDAC, as a top potential target. SIRT2 expression was significantly upregulated in both Atrx- mNPCs and in ATRX-mutant gliomas and its specific chemical inhibition normalized cellular motility in both Atrx- mNPCs and ATRX-mutant, patient derived glioma stem cells. These findings indicate that SIRT2 inhibition represents a viable strategy to revert the epigenetic effects of ATRX deficiency on facultative heterochromatin and their transcriptional and phenotypic consequences.

GENE-28. LONGITUDINAL MOLECULAR TRAJECTORIES OF DIFFUSE GLIOMA IN ADULTS

Treatment options for adult patients with glioma has remained largely unchanged over the past three decades. Targeted inhibitors and immunotherapies have improved outcomes for many cancer types but their relevance in glioma is unclear. The inevitability of glioma disease recurrence demands an understanding of mechanisms driving therapy resistance. The Glioma Longitudinal Analysis (GLASS) Consortium was initiated to establish a definitive portrait of the recurrence process and to discover vulnerabilities that render the tumor sensitive to therapeutic intervention. GLASS is a community-driven effort that seeks to overcome the logistical challenges in constructing adequately powered longitudinal genomic glioma datasets by pooling data from patients treated at institutions worldwide. Currently, the GLASS Data Resource comprises DNA sequencing data (exome and/or whole-genome) from 288 patients of whom high-quality data in at least two time points are present from 222 patients (n = 134 IDHwt, n = 63 IDHmutant-noncodel, n = 25 IDHmutant-codel). We inferred longitudinal mutation, copy number, clonal frequency, and neoantigen profiles and demonstrated that driver genes found at initial disease persisted into recurrence. Treatment with alkylating-agents resulted in a hypermutator phenotype at different rates across glioma subtypes, most frequently among IDHmutant-noncodels, and hypermutation was not associated with differences in overall survival. Acquired aneuploidy was frequently detected in recurrent IDHmutant-noncodel gliomas and further converged with acquired cell cycle pathway alterations and poor outcomes. We showed that the clonal architecture of each tumor remains largely intact over time and that genetic drift was associated with increased survival. Finally, we found that neoantigens were exposed to stable selective pressures throughout a tumor’s progression. Our results collectively suggest that the strongest selective pressures occur early during glioma development and that current therapies shape this evolution in a largely stochastic manner. The GLASS Data Resource provides a genomic reference to study the patterns of glioma evolution.

RARE-47. CHARACTERIZATION OF ADULT MEDULLOBLASTOMA; A RETROSPECTIVE REVIEW OF 200 PATIENTS AT MD ANDERSON CANCER CENTER

INTRODUCTION

Prospective studies for adult medulloblastoma (MB) are scarce and management guidelines are largely derived from pediatric experience and retrospective data. However, adult MB is a distinct disease for which limited data exist regarding clinical characteristic, prognostic factors, outcome based on upfront treatment, or patterns of recurrence.

METHODS

200 patients (≥18 years at diagnosis) were identified from 1978 to 2017. Survival analysis was performed using the Kaplan-Meier method.

RESULTS

Median age at diagnosis 29 (18–63); 111 (55.5%) standard-risk (SR), 59 (29.5%) high-risk (HR), 30 (15%) unknown. 188 (94%) and 97 (48.5%) underwent radiotherapy and upfront chemotherapy, respectively. Median PFS: 6.28 years (4.9–11.2); median OS: 8.8 (7.8–12.4) years. Median time to first recurrence: 8.2 (5.1-not reached); 5-year recurrence-free rate: 71% in SR and 45% in HR patients. Survival after first recurrence: 2 years (1.64–2.4). Documented recurrence and risk group were associated with survival (HR 30.4, p< 0.0001 and 2.33, p< 0.001), but metastatic status and local vs. distant recurrence were not. The effect of upfront chemotherapy on survival did not reach statistical significance in SR patients. The use of chemotherapy regimens with or without vincristine or containing cisplatin vs. carboplatin, did not alter the survival outcome. In a subgroup of 57 SR patients who were seen at initial diagnosis and for whom complete staging and treatment information were available 3 and 5-year PFS were 94% (88%, 100%) and 76% (66%, 86%); OS at 5 years, 94% (87%, 100%).

CONCLUSION

No statistically significant difference was noted in outcome of patients treated without vincristine or with carboplatin instead of cisplatin, suggesting attenuated upfront regimens could be considered to reduce toxicity. Adult MB patients recur late and have a poor post-recurrence survival, therefore long-term follow-up, development of molecular risk-adjusted upfront treatment, and optimization of rescue treatments are needed.

GENE-17. ATRX LOSS IN GLIOMA RESULTS IN EPIGENETIC DYSREGULATION OF THE G2/M CHECKPOINT AND SENSITIVITY TO ATM INHIBITION

Gliomas are a leading cause of cancer mortality in children and adults and new targeted therapies are desperately needed. ATRX is a chromatin remodeling protein that is recurrently mutated in H3F3A-mutant pediatric GBM and IDH-mutant grade 2/3 adult glioma. We previously showed that loss of ATRX in glioma results in tumor growth and additional tumor mutations. However, the mechanism driving these phenotypes has not been fully established. We found that in ChIP-Seq datasets of mouse neuronal precursor cells (NPCs) and experimental models of human glioma cells, ATRX binds and regulates the chromatin state of promoters and enhancers for gene sets associated with regulation of the cell cycle G2/M checkpoint. In line with this, analysis of single-cell seq (sc-seq) data from IDH-mutant gliomas (n=16) shows that ATRX-mutant tumors (IDH-A) demonstrate a population of cycling cells with dysregulated cell cycle phase gene set expression when compared to ATRX-wildtype tumors (IDH-O). In glioma models, ATRX-deficient cells exhibit a seven-fold increase in mitotic index at 16 hours after sub-lethal radiation and enhanced activation of the master cell cycle regulator ATM with radiation. Treatment of ATRX-deficient gliomas with ATM inhibitors results in a selective increase in dysfunctional cell cycling and increased radio-sensitization in ATRX-deficient glioma cells. Using an ATM-luciferase reporter in orthotopically-implanted human GBM cells, both AZD0156 and AZD1390 demonstrate in vivo pathway inhibition. Mice intra-cranially implanted with ATRX-deficient GBM cells demonstrate a doubling of median survival compared to radiated controls (p=0.0018) when treated with AZD0156 combined with radiation. This study demonstrates that ATRX-deficient glioma display epigenetic dysregulation of the G2/M checkpoint, which opens a new window for therapies targeting this unique phenotype.

TMIC-60. COMPREHENSIVE SPATIAL CHARACTERIZATION OF IMMUNE CELLS IN THE CNS BRAIN TUMOR MICROENVIRONMENT

Previous immune profiling in brain tumors has mostly focused on the high-density tumor areas, and as such, little is known about the nature and types of immunological responses that occur across the tumor landscape, including at the tumor-central nervous system (CNS) interface. En bloc resections of glioblastomas (n=10) and CNS lung metastases (n=10) were oriented on slides as whole mount wedges spanning three anatomical areas including the invasive edge, tumor region, and necrotic core. Tumor segmentation was performed and regional differences were immunologically analyzed for 770 immune genes using the NanoString nCounter System with CIBERSORT analysis to delineate immune gene signatures. The analysis was validated using multiplex immunohistochemistry (IHC). The top upregulated immune genes in the GBM necrotic core were associated with macrophages, including the CD163 lineage marker, chemotactic factors (such as CCL18 and SAA1), and the phagocytosis stimulatory factors (such as IL-8 and MARCO). The necrotic core downregulates GBM antigens (such as IL13RA2 and MAGEB2), markers of dendritic cells (such as LILRA4), and immune stimulatory processes including MHC, IFN, IL-12, TNF, and ICOS expression. In direct contrast, the infiltrating edge of the GBM relative to the tumor is enriched with stimulators for NK cytotoxicity (i.e., CD244, the fractalkine receptor for immune cells), chemokines for thymocytes and dendritic cells, and immune stimulatory IL-12 receptors. Glioblastoma has rare focal isolated areas of CD3 T-cell reactivity within the tumor. Similar to GBM, the necrotic center of lung metastases is enriched in immune suppressive macrophages, as reflected by CD163 IHC staining and arginase expression; however, they are more frequently infiltrated with M1 macrophages. Yet the majority of lung cancers are more diffusely infiltrated with CD3 T cells, especially at the infiltrating edge. In general, we noted distinct inter- and intratumoral immune gene signatures, with macrophages dominating the brain tumors, especially the necrotic core.

INNV-40. TARGETED NEXT GENERATION SEQUENCING OF PEDIATRIC HIGH-GRADE GLIOMA AND ITS THERAPEUTIC IMPLICATIONS, MD ANDERSON EXPERIENCE

INTRODUCTION

The new understanding of molecular pathways in cancer is paving the way towards personalized cancer medicine, especially in refractory disease. High-grade gliomas (HGG) are common pediatric brain tumors that tend to recur, with no available standard therapy upon recurrence. HGG are challenging tumors with illusive biology and poor outcome. We report here the molecular testing of 27 pediatric HGG patients.

MATERIALS AND METHODS

An analysis of pediatric patients with HGG treated at UT MD Anderson Cancer Center (MDACC) who underwent molecular genetic profiling using next generation sequencing with different genomic panels (AmpliSeq™Cancer Hotspot and Oncomine Panels – by Thermo Fisher Scientific).

RESULTS

27 patients with HGG (median age 14 years, range 3–18 years old) underwent genomic profiling. Primary diagnoses were glioblastoma multiforme (n=22), anaplastic astrocytoma (n = 2), gliosarcoma (n= 1), anaplastic pleomorphic xanthoastrocytoma (n= 1) and anaplastic oligoastrocytoma (n= 1). There are 46 genes common to the panels used. The most common mutation was in TP53 (73%). Other mutations included PIK3CA (19%), IDH1 (11.5%), 7.7% for ATM, EGFR and PTEN, and 3.8% for BRAF, FGFR1 and FGFR2. 24 out of 27 patients were tested at initial diagnosis and 3 upon relapse/progression. Patients at initial diagnosis received standard of care therapy of radiation and temozolomide. Only 5 patients received targeted therapy upon progression/recurrence. Some challenges of genomically-matched therapy included lack of clinical trials accepting pediatric patients, unavailability of a liquid form of a drug, and insurance disapproval for off-label use.

CONCLUSION

The next generation of therapy for childhood cancers will be based upon in-depth molecular phenotyping that may facilitate the development of rational risk-adapted and target-based therapies. This cohort, though limited by sample size, highlights the opportunity to perform molecular testing and identification of alterations in actionable genes.

Metachronous Medulloblastoma in a Child With Successfully Treated Neuroblastoma: Case Report and Novel Findings of DNA Sequencing

Metachronous neoplasms have rarely been reported in patients with neuroblastoma. This report presents the clinical case of a 23-month-old child who was diagnosed with an anaplastic medulloblastoma 5 months after completing treatment for stage IV neuroblastoma. The patient was treated with complete surgical resection and adjuvant chemoradiation followed by maintenance chemotherapy at an outside institution and came to our institution for further management. A pathologic diagnosis and review of both the suprarenal and posterior fossa masses were performed, as well as a genetic analysis of both cerebellar tumor tissue and blood using next-generation gene sequencing. At our institution, the patient was submitted to induction chemotherapy followed by high-dose chemotherapy and autologous stem cell transplantation and remains free of disease 2 years after completion of treatment. Genetic analysis revealed multiple somatic copy number variations with most deleted genes located in 2q37, a region which harbors genes involved in epigenetic regulation and tumor suppression. A homozygous deletion was found in the TSC2 gene, which is a clinically actionable gene, and patients with activating deletions in TSC2 can potentially be eligible for basket clinical trials with mTOR inhibitors. Germline single nucleotide variants were also identified in multiple genes involved in cancer (ALK, FGFR3, FLT3/4, HNF1A, NCOR1, and NOTCH2/3), cancer predisposition (TP53, TSC1, and BRCA1/2), and genes involved in DNA repair (MSH6, PMS2, POLE, and ATM). Metachronous neoplasms are rare and challenging to treat, hence genetic analysis and referral are needed to exclude hereditary cause. DNA sequencing of the tumor and germline can help identify alterations that increase predisposition or can be used to guide treatment decisions on recurrence and when standard options fail.

Ultrasmall Core-Shell Silica Nanoparticles for Precision Drug Delivery in a High-Grade Malignant Brain Tumor Model

PURPOSE

Small-molecule inhibitors have revolutionized treatment of certain genomically defined solid cancers. Despite breakthroughs in treating systemic disease, central nervous system (CNS) metastatic progression is common, and advancements in treating CNS malignancies remain sparse. By improving drug penetration across a variably permeable blood-brain barrier and diffusion across intratumoral compartments, more uniform delivery and distribution can be achieved to enhance efficacy.

EXPERIMENTAL DESIGN

Ultrasmall fluorescent core-shell silica nanoparticles, Cornell prime dots (C' dots), were functionalized with α_v integrin-binding (cRGD), or nontargeting (cRAD) peptides, and PET labels (¹²⁴I, ⁸⁹Zr) to investigate the utility of dual-modality cRGD-C' dots for enhancing accumulation, distribution, and retention (ADR) in a genetically engineered mouse model of glioblastoma (mGBM). mGBMs were systemically treated with ¹²⁴I-cRGD- or ¹²⁴I-cRAD-C' dots and sacrificed at 3 and 96 hours, with concurrent intravital injections of FITC-dextran for mapping blood-brain barrier breakdown and the nuclear stain Hoechst. We further assessed target inhibition and ADR following attachment of dasatinib, creating nanoparticle-drug conjugates (Das-NDCs). Imaging findings were confirmed with ex vivo autoradiography, fluorescence microscopy, and p-S6RP IHC.

RESULTS

Improvements in brain tumor delivery and penetration, as well as enhancement in the ADR, were observed following administration of integrin-targeted C' dots, as compared with a nontargeted control. Furthermore, attachment of the small-molecule inhibitor, dasatinib, led to its successful drug delivery throughout mGBM, demonstrated by downstream pathway inhibition.

CONCLUSIONS

These results demonstrate that highly engineered C' dots are promising drug delivery vehicles capable of navigating the complex physiologic barriers observed in a clinically relevant brain tumor model.

BSCI-12. COMPREHENSIVE GENOMIC ANALYSIS OF BRAIN METASTASES FROM MULTIPLE CANCER TYPES

PURPOSE: Brain metastases occur in approximately 8–10% of patients with cancer, and the incidence has increased over the past decades. The most common primary tumors responsible for brain metastases are lung cancer, melanoma, renal cell carcinoma (RCC), breast cancer and colorectal cancer. The precise mechanisms by which genomic and transcriptional abnormalities drive the formation of brain metastases remain unclear. Here, we conducted comprehensive genomic and transcriptional analysis with paired primary tumor tissue (or extracranial metastasis tissue) and brain metastasis tissue using whole-exome sequencing (WES), mRNA-Seq and global methylation profiling. METHODS: Frozen, paired brain metastasis tissue and primary tumor tissue (or extracranial metastasis tissue) and white blood cells were acquired from RCC (n=12), breast cancer (n=17), lung cancer (n=15) and melanoma (n=14) patients, followed by extraction of DNA and RNA. WES and mRNA-Seq were performed on the Illumina HiSeq4000 platform. For methylation profiling, DNA was analyzed using Illumina Infinium MethylationEPIC Beadchip arrays. RESULTS: Somatic mutations or methylation of VHL gene were identified in 81.8% of RCC patients. Gene Set Enrichment Analysis revealed significant enrichment for hypoxia pathway transcripts in RCC brain metastases relative to primary tumors. The most common alterations in breast and lung cancer patients were TP53 mutations with frequencies of 50.0% and 73.3%, followed by ERBB2 alterations (43.8%) in breast cancer patients and mutually exclusive alterations of EGFR (33.3%) and KRAS (26.7%) in lung cancer patients. Mutually exclusive alterations of NRAS (42.9%) and BRAF (42.9%) were also observed in melanoma patients. Gene expression and epigenetic analysis revealed characteristics of brain metastases depending on primary cancer types. CONCLUSIONS: Comprehensive genomic analysis of brain metastases from four different cancer types revealed that brain metastasis tissues have unique genomic, transcriptional and epigenetic profiles according to histopathology groups. Therefore, the therapeutic strategies should be designed based at least in part on tumor histiogenesis.