Patient-derived Orthotopic Xenograft Research Articles

DIPG-16. COMBINATION OF ARGININE DEPLETION AND POLYAMINE INHIBITION AS AN ANTICANCER STRATEGY FOR DIFFUSE INTRINSIC PONTINE GLIOMA (DIPG)

DIPG is an aggressive pediatric brainstem tumor, with a median survival below 12 months. Tumor cells are dependent upon arginine, a semi-essential amino acid, metabolised by arginase enzymes into ornithine, a pivotal precursor to the polyamine pathway. Polyamines, frequently upregulated in cancer, are intracellular polycations controlling key biological processes – the inhibition of which we have previously shown to be highly efficacious in preclinical DIPG models. Pegylated arginase (BCT-100) has recently been shown to significantly delay tumor development, prolonging survival of neuroblastoma-prone Th-MYCN mice. This study investigated the effects of arginine depletion therapy as a single agent and in combination with polyamine pathway inhibitors in DIPG. We found that ARG2, the gene encoding for arginase II, is expressed significantly more highly in DIPG tumors compared to normal brain. Arginine depletion via BCT-100 reduced DIPG cell proliferation and colony formation in patient-derived cell lines. Using orthotopic patient-derived xenograft models of DIPG, we found that frequent dosing of BCT-100 (4x/week) significantly delayed tumor development and increased the survival of the mice (p<0.0001). DFMO is an FDA-approved inhibitor of the enzyme ornithine decarboxylase, a key driver of polyamine synthesis. The combination of BCT-100 with DFMO led to significant enhancement in DIPG survival (p<0.005 compared to single agent treatments). Triple combination therapy with addition of the polyamine transport inhibitor AMXT-1501 led to a potent and profound improvement in survival. These data show that arginine depletion therapy using BCT-100 combined with dual polyamine inhibitory agents represents a potentially exciting new approach for the treatment of DIPG.

MODL-31. RADIATION-DERIVED TREATMENT-RESISTANT PDX AND CELL CULTURE MODELS RECAPITULATE THE CHARACTERISTICS OF MATCHED PRIMARY/RECURRENT PEDIATRIC HIGH-GRADE GLIOMA

BACKGROUNDPediatric high-grade glioma (pHGG) is the most common cause of childhood cancer death. Recurrence after therapy is a major challenge, since recurrent pHGG proliferates aggressively and resists therapy. We developed and validated preclinical models of matched primary and recurrent tumors, providing a method to study recurrence and potential therapies.METHODSWe irradiated H3K27M thalamic pHGG cells (BT245) (8 Gy/week,2Gy fractions x3 weeks) and propagated the surviving cells (BT245R). We developed a murine recurrence model by orthotopically implanting BT245 cells, irradiating the resultant tumors (4 Gy/day x2d) and propagating irradiated (BT245RM) or control (BT245CM) tumor cells at endpoint. We performed phenotypic analyses, RNA-Seq, and drug testing.RESULTSBT245R cells were more stemlike than BT245, with an 8-fold greater rate of neurosphere formation (p<0.03). Geneset enrichment analysis showed similar molecular changes in BT245RM cells and primary/recurrent H3K27M pHGG patient sample pair, including relaxation of the G2/M cell cycle checkpoint (Hallmark_G2M_Checkpoint: BT245RM NES=-5.95, FDR=0.0; patient NES=-5.86, FDR=0.0), downregulation of MYC targets (Hallmark_MYC_Targets_V1: BT245RM NES=-7.43, FDR=0.0; patient NES=-5.86, FDR=0.0), and decreased differentiation (Go_Regulation_of_Stem_Cell_Differentiation: BT245RM NES=-3.35, FDR=0.0; patient NES=-3.15, FDR=0.0). Enrichment of the protein_kinase_C_signaling in BT245RM (NES=2.18,FDR=0.03) suggested response to MAPK pathway inhibition. BT245R cells were twice as sensitive as BT245 cells to the MEK inhibitor trametinib (p<0.05).CONCLUSIONSOur neurosphere and murine orthotopic patient-derived xenograft models recapitulate gene expression changes of matched primary/recurrent pHGG. RNA-Seq analysis validated the model against patient samples and identified trametinib as potentially effective in recurrent pHGG.

IMG-17. RADIOMICS CHARACTERIZATION OF FOUR PEDIATRIC BRAIN TUMOR SUBTYPES IN PDX MOUSE MODELS

BACKGROUNDPreviously, we have reported on the development of advanced magnetic resonance imaging (MRI) protocols for mouse brain tumors. The goal of this follow-up pre-clinical study was to develop a machine-learning MRI classifier (radiomics) for four subtypes of childhood brain tumor in patient-derived xenograft (PDX) mice.METHODSMRI scans on orthotopic medulloblastoma, ependymoma, ATRT and DIPG PDX (each n=12 animals) were performed on the animal 9.4 Tesla scanner with an in-plane resolution of 47 microns. Image segmentation, as well as shape and texture based radiomics descriptors were modeled using a modified COLIAGE software for tumor classification and to characterize tumor habitat of each tumor subtype.RESULTSThe mean tumor volumes were 11.2 mm3. Each MRI scan was segmented into three regions: (i) well defined tumor (including distant metastases); (ii) peritumoral edema; (iii) tumor necrosis. 360 radiomics features (capturing co-occurrence, grey-level dependence and directional gradients) were obtained for each region. The model classified four subtypes with high accuracy while achieving sufficient segmentation accuracy despite the small lesion size. A subset of fourteen tumoral, six peritumoral and five distant MRI radiomics features were found to be predictive of the tumor sub-type (p=0.0017) independently of tumor anatomical location.CONCLUSIONSMRI protocols followed by radiomics feature analysis discriminated among specific radiological features for four distinct orthotopic PDX models: medulloblastomas exhibit low ADC values, high angiogenesis and cortical metastases as compared to ependymomas (high levels of edema and olfactory bulb metastases), ATRT (the highest level of necrosis) and DIPG (highest T2 signal intensities and spinal metastases).

Patient-derived orthotopic xenograft models of medulloblastoma lack a functional blood-brain barrier.

Novel targeted therapies for children diagnosed with medulloblastoma (MB), the most common malignant pediatric brain tumor, are urgently required. A major hurdle in the development of effective therapies is the impaired delivery of systemic therapies to tumor cells due to a specialized endothelial blood-brain barrier (BBB). Accordingly, the integrity of the BBB is an essential consideration in any preclinical model used for assessing novel therapeutics. This study sought to assess the functional integrity of the BBB in several preclinical mouse models of MB. Dynamic contrast enhancement magnetic resonance imaging (MRI) was used to evaluate blood-brain-tumor barrier (BBTB) permeability in a murine genetically engineered mouse model (GEMM) of Sonic Hedgehog (SHH) MB, patient-derived orthotopic xenograft models of MB (SHH and Gp3), and orthotopic transplantation of GEMM tumor cells, enabling a comparison of the direct effects of transplantation on the integrity of the BBTB. Immunofluorescence analysis was performed to compare the structural and subcellular features of tumor-associated vasculature in all models. Contrast enhancement was observed in all transplantation models of MB. No contrast enhancement was observed in the GEMM despite significant tumor burden. Cellular analysis of BBTB integrity revealed aberrancies in all transplantation models, correlating to the varying levels of BBTB permeability observed by MRI in these models. These results highlight functional differences in the integrity of the BBTB and tumor vessel phenotype between commonly utilized preclinical models of MB, with important implications for the preclinical evaluation of novel therapeutic agents for MB.

Functional Precision Medicine Identifies New Therapeutic Candidates for Medulloblastoma.

Medulloblastoma is among the most common malignant brain tumors in children. Recent studies have identified at least four subgroups of the disease that differ in terms of molecular characteristics and patient outcomes. Despite this heterogeneity, most patients with medulloblastoma receive similar therapies, including surgery, radiation, and intensive chemotherapy. Although these treatments prolong survival, many patients still die from the disease and survivors suffer severe long-term side effects from therapy. We hypothesize that each patient with medulloblastoma is sensitive to different therapies and that tailoring therapy based on the molecular and cellular characteristics of patients' tumors will improve outcomes. To test this, we assembled a panel of orthotopic patient-derived xenografts (PDX) and subjected them to DNA sequencing, gene expression profiling, and high-throughput drug screening. Analysis of DNA sequencing revealed that most medulloblastomas do not have actionable mutations that point to effective therapies. In contrast, gene expression and drug response data provided valuable information about potential therapies for every tumor. For example, drug screening demonstrated that actinomycin D, which is used for treatment of sarcoma but rarely for medulloblastoma, was active against PDXs representing Group 3 medulloblastoma, the most aggressive form of the disease. Functional analysis of tumor cells was successfully used in a clinical setting to identify more treatment options than sequencing alone. These studies suggest that it should be possible to move away from a one-size-fits-all approach and begin to treat each patient with therapies that are effective against their specific tumor. SIGNIFICANCE: These findings show that high-throughput drug screening identifies therapies for medulloblastoma that cannot be predicted by genomic or transcriptomic analysis.

Neoplastic-Stromal Cell Cross-talk Regulates Matrisome Expression in Pancreatic Cancer.

Pancreatic ductal adenocarcinoma (PDAC) is characterized by a highly desmoplastic reaction, warranting intense cancer-stroma communication. In this study, we interrogated the contribution of the BET family of chromatin adaptors to the cross-talk between PDAC cells and the tumor stroma. Short-term treatment of orthotopic xenograft tumors with CPI203, a small-molecule inhibitor of BET proteins, resulted in broad changes in the expression of genes encoding components of the extracellular matrix (matrisome) in both cancer and stromal cells. Remarkably, more than half of matrisome genes were expressed by cancer cells. In vitro cocultures of PDAC cells and cancer-associated fibroblasts (CAF) demonstrated that matrisome expression was regulated by BET-dependent cancer-CAF cross-talk. Disrupting this cross-talk in vivo resulted in diminished growth of orthotopic patient-derived xenograft tumors, reduced proliferation of cancer cells, and changes in collagen structure consistent with that of patients who experienced better survival. Examination of matrisome gene expression in publicly available data sets of 573 PDAC tumors identified a 65-gene signature that was able to distinguish long- and short-term PDAC survivors. Importantly, the expression of genes predictive of short-term survival was diminished in the cancer cells of orthotopic xenograft tumors of mice treated with CPI203. Taken together, these results demonstrate that inhibiting the activity BET proteins results in transcriptional and structural differences in the matrisome are associated with better patient survival. IMPLICATIONS: These studies highlight the biological relevance of the matrisome program in PDAC and suggest targeting of epigenetically driven tumor-stroma cross-talk as a potential therapeutic avenue.

CBMS-04 Novel xenograft model to clarify tumor progressive mechanism and therapeutic target in primary central nervous system lymphoma

Primary central nervous system lymphoma (PCNSL) is a rare lymphoma of the central nervous system and has a dismal prognosis despite intensive chemotherapy. Recent genomic analyses have identified recurrent genetic alterations in Primary central nervous system lymphoma (PCNSL). However, lack of clinically representative PCNSL models has diminished our understanding of the pathogenic mechanisms of those genetic events. Here, we established 14 patient-derived orthotopic xenografts (PDOXs). Comprehensive analysis showed that PDOXs faithfully retained the phenotypic, metabolic, and genetic features with 100 % concordance of MYD88 and CD79B mutations present in immuno-competent PCNSL patients. Notably, orthotopic xenograft formation was consistently dependent on deregulated signaling through the RelA/p65-hexokinase 2 (HK-2) axis. MYD88/CD79B mutations and Pin1 activation, or LMP1 and Pin1 activation, converge on the RelA/p65-HK-2 signaling in immunocompetent and EBV-positive PCNSL, respectively. Genetic and pharmacological inhibition of this key signaling axis potently suppressed PCNSL tumor growth in vitro and in vivo. Additionally, our models further offer a platform for predicting clinical chemotherapeutics efficacy. Therefore, our models provide critical insights into pathogenic mechanisms and therapeutic discovery in PCNSL.

The Emerging Role of CD24 in Cancer Theranostics-A Novel Target for Fluorescence Image-Guided Surgery in Ovarian Cancer and Beyond.

Complete cytoreductive surgery is the cornerstone of the treatment of epithelial ovarian cancer (EOC). The application of fluorescence image-guided surgery (FIGS) allows for the increased intraoperative visualization and delineation of malignant lesions by using fluorescently labeled targeting biomarkers, thereby improving intraoperative guidance. CD24, a small glycophosphatidylinositol-anchored cell surface receptor, is overexpressed in approximately 70% of solid cancers, and has been proposed as a prognostic and therapeutic tumor-specific biomarker for EOC. Recently, preclinical studies have demonstrated the benefit of CD24-targeted contrast agents for non-invasive fluorescence imaging, as well as improved tumor resection by employing CD24-targeted FIGS in orthotopic patient-derived xenograft models of EOC. The successful detection of miniscule metastases denotes CD24 as a promising biomarker for the application of fluorescence-guided surgery in EOC patients. The aim of this review is to present the clinical and preclinically evaluated biomarkers for ovarian cancer FIGS, highlight the strengths of CD24, and propose a future bimodal approach combining CD24-targeted fluorescence imaging with radionuclide detection and targeted therapy.

A novel patient-derived orthotopic xenograft (PDOX) mouse model of highly-aggressive liver metastasis for identification of candidate effective drug-combinations

Liver metastasis is a recalcitrant disease that usually leads to death of the patient. The present study established a unique patient-derived orthotopic xenograft (PDOX) nude mouse model of a highly aggressive liver metastasis of colon cancer. The aim of the present study was to demonstrate proof-of-concept that candidate drug combinations could significantly inhibit growth and re-metastasis of this recalcitrant tumor. The patient’s liver metastasis was initially established subcutaneously in nude mice and the subcutaneous tumor tissue was then orthotopically implanted in the liver of nude mice to establish a PDOX model. Two studies were performed to test different drugs or drug combination, indicating that 5-fluorouracil (5-FU) + irinotecan (IRI) + bevacizumab (BEV) and regorafenib (REG) + selumetinib (SEL) had significantly inhibited liver metastasis growth (p = 0.013 and p = 0.035, respectively), and prevented liver satellite metastasis. This study is proof of concept that a PDOX model of highly aggressive colon-cancer metastasis can identify effective drug combinations and that the model has future clinical potential.

STEM-11. INTERROGATING INTRATUMORAL TRANSCRIPTOMIC HETEROGENEITY AND PLASTICITY AS RESISTANCE MECHANISMS IN GLIOBLASTOMA

Abstract Intratumoral heterogeneity is one of the major barriers in the treatment of many tumor types. In glioblastoma (GBM) – the most aggressive brain tumor - survival of patients following standard therapy has stagnated to an average of 14 months, largely due to the inherent heterogeneity and phenotypic plasticity. Patient-derived orthotopic xenografts (PDOX) are important tools for preclinical cancer studies. We have shown that they recapitulate histological and molecular features of patient tumors, thus producing more clinically relevant study outcomes when compared to other preclinical models. To understand the impact of treatment on intratumoral heterogeneity and interactions of tumor cells with the microenvironment, we applied single-cell RNA-seq in GBM PDOX models generated from naive and treatment exposed patient tumors. Here we show that PDOX models recapitulate all the major cell types and transcriptional programs reported in GBM patient samples. We demonstrate that GBM tumor cells occur in distinct phenotypic states dictated by microenvironmental cues and treatment-induced pressure. Non-neoplastic cells present within xenografts undergo transcriptomic adaptation similar to the GBM microenvironment in patients. Analysis of treated PDOXs and longitudinal PDOX models derived from patients prior and after treatment allows us to investigate short-term and stable long-term transcriptomic adaptations. We aim to reveal key molecular regulators enhancing tumor cell plasticity and resistance, that could be targeted in combinatory treatments. Our GBM PDOX models provide a platform for high-quality pre-clinical research and an improved clinical relevance in the development and testing of novel therapeutics.

TMOD-03. PAN-CANCER ANALYSIS OF ORTHOTOPIC PATIENT DERIVED XENOGRAFTS FROM BRAIN METASTASES

Abstract Brain metastases (BM) are a leading cause of cancer death and prognosis remains poor despite treatment advances at other sites. Models are central to therapeutic development, but few orthotopic patient-derived xenograft (PDX) models of BM exist. To represent diversity across BM types, we established a program to create orthotopic PDX at scale from all BM patients. To date BM were received from 100 patients and PDX attempted by direct brain injection (PDX, n=89) or injection of low passage patient-derived cell lines (PDCLX, n=11). We created 65 successful BM PDX from 13 cancers: 17 lung (55% take), 15 breast (68%), 6 melanoma (75%), 5 CNS lymphoma (83%), 3 gastrointestinal (75%), 2 esophageal (40%), 2 ovarian (67%), 1 sarcoma (100%), 1 laryngeal (100%), 1 prostate (100%), 1 pancreatic (100%), 1 uterine adenosarcoma (100%), and 1 yolk sac tumor (100%). Take rate was similar for models derived from patients with prior chemotherapy-only versus immune/targeted therapy-only (63 vs 58%). Fifteen patients had live tumor and matching PBMCs archived for modeling in vitro immunotherapy responses. Mean time to moribund among different cancer types ranged from 27 days (yolk sac tumor) to 177.5 days (ovarian). BM PDX had a favorable timeline for preclinical study (90% moribund at 180 days). All PDX retained high fidelity to the patient driver SNVs and copy aberrations, even at &amp;gt;P4. No significant differences noted by immunodeficient strain (SCID versus NSG) or injection site (orthotopic versus heterotopic). Explants from BM PDX were able to generate long-term cell lines (60%) or short-term cultures with qualitative concordance of model-to-patient responses to targeted therapy (Osimertinib, EGFRi) and immunotherapy (Pembrolizumab, PD1i). Genomic and clinical data were used to create the DFCI BM PDX cBioPortal for public release and models distribution will be available through the DFCI Center for Patient Derived Models.

STEM-18. DEVELOPMENT OF A NOVEL GENE THERAPY APPROACH TARGETING GLIOBLASTOMA FOLLOWING AI-DIRECTED IDENTIFICATION OF A MASTER REGULATORY GENE NETWORK

Abstract BACKGROUND Profound heterogeneity has severely hampered therapeutic advancements in glioblastoma (GBM). Remarkably, GBM exhibits broad clinical and histopathologic overlaps suggesting the presence of a common regulatory state. The GBM state embodies restructuring of the master regulatory gene network (MRGN) forced by founding mutations and perpetuated in subclones of GBM stem-like cells (GSC). Successful targeting and altering of the MRGN promise to circumvent the heterogeneity. METHODS To decipher the common MRGN in GSC, we applied a robust AI suite, NETZEN, that integrates deep neural networks with gene network-based ranking to the reference GBM network generated using TCGA’s entire GBM RNAseq dataset and a dataset containing &amp;gt; 30 diverse GSC and paired GBM differentiated cell lines, 6 normal astrocyte and 3 normal neuronal precursor cell lines. To develop a gene therapy against the MRGN, we performed a screen of a rAAV capsid library using GBM patient-derived orthotopic xenografts (PDX) to identify capsid variants with specific tropism for GBM cells to deliver shRNA-based cassettes. RESULTS We discovered a putative MRGN in GSC, anchored by developmentally restricted master regulators (MR). To confirm its critical role, we deconstructed it using shRNA in a panel of PDX and uniformly observed attenuated tumor growth and mortality compared to controls (p &amp;lt; 0.05 in all lines). More notably, when normal astrocytes were forced to reconstruct the MRGN, transformation into GSC-like cells occurred, as measured by single-cell analysis, sphere formation in vitro, and most importantly, development of lethal infiltrating brain tumors in 15/15 mice. Finally, selected novel AAV capsids with 10-40-fold higher specificity for GBM cells were utilized in a shRNA-based AAV platform to target key MR in the validated MRGN. CONCLUSIONS The GBM state is established by a developmental subnetwork permitting a first-of-its-kind, heterogeneity-agnostic AAV-based treatment approach. This AI-directed R&amp;D program can be expanded to target other cancers.

TAMI-03. IDENTIFICATION OF NOVEL DRIVERS OF LUNG-TO-BRAIN METASTASIS THROUGH IN VIVO FUNCTIONAL GENOMICS

Abstract Brain metastasis, the most common tumor of the central nervous system, occurs in 20-36% of primary cancers. In particular, 40% of patients with non-small cell lung cancer (NSCLC) develop brain metastases, with a dismal survival of approximately 4-11 weeks without treatment, and 16 months with treatment. This highlights a large unmet need to develop novel targeted therapies for the treatment of lung-to-brain metastases (LBM). Genomic interrogation of LBM using CRISPR technology can inform preventative therapies targeting genetic vulnerabilities in both primary and metastatic tumors. Loss-of-function studies present limitations in metastasis research, as knocking out genes essential for survival in the primary tumor cells can thwart the metastatic cascade prematurely. However, gene overexpression using CRISPR activation (CRISPRa) has the potential for overcoming dependencies of gene essentiality. We theorize that an in vivo genome-wide CRISPRa screen will identify novel genes that, when overexpressed, drive LBM. We have developed a patient-derived orthotopic murine xenograft model of LBM using primary patient-derived NSCLC cell lines (termed LTX cells) from the Swanton Lab TRACERx study. We are now poised to transduce LTX cells with a human genome-wide CRISPRa single guide RNA (sgRNA) library, and to subsequently inject the cells into the lungs of immunocompromised mice. We will then track the process of LBM using bioluminescent and MRI imaging until mice reach endpoint. Sequencing of primary lung tumors and subsequent brain metastases promises to uncover enriched sgRNAs, which may represent novel drivers of primary lung tumor formation and LBM. To the best of our knowledge, this study is the first in vivo genome-wide CRISPRa screen focused on identifying novel drivers of LBM, and can inform future preventative therapies to improve survival outcomes for NSCLC patients.

TAMI-56. HYPOXIA-DRIVEN NECROSIS DRIVES DYNAMIC TUMOR PROGRESSION IN GLIOBLASTOMA

Abstract The onset of necrosis correlates with aggressive malignant progression in solid tumors. Whether necrosis arises from or promotes accelerated tumor proliferation and progression remains unknown, primarily due to inadequate models systems that fail to induce or capture these dynamic changes as they develop. In glioblastoma (GBM; WHO grade IV), the most malignant primary brain tumor, vascular pathology and central necrosis precedes rapid, radial expansion resulting in patient mortality. While genetic alterations in GBM have been highly characterized, biological adaptations arising from sustained hypoxia/anoxia require further mechanistic investigation. This sustained nutrient deprivation leads to necrosis which dramatically changes the tumor microenvironment (TME). To reveal the dynamic temporal and spatial changes promoting expansive progression, we are generating mouse models that more appropriately capture events found in human gliomas, accounting for unique microenvironmental stressors often lacking in GBM animal models, specifically central necrosis. Combining hypoxia-induced focal necrosis within high grade gliomas with intravital microscopy captures TME restructuring to understand its impact on glioma progression. Our studies use both genetically characterized patient-derived orthotopic GBM xenografts, alongside an immunocompetent RCAS/tv-a model, to determine how antagonizing these processes impacts disease progression and outcomes across multiple GBM subtypes. Complementary to this, we are investigating how increasing hypoxia and necrosis-driven signaling events promotes tumor-associated macrophages (TAMs) influx and reprogramming during glioma progression. Our preliminary data indicate substantial differences pre- and post-necrosis regarding TAM enrichment, immunosuppressive phenotypic conversion and their biological impact, however these mechanisms have not yet been elucidated. Our models capture glioma growth dynamics, reactive neuroinflammation and therapeutic resistance, facilitating innovative pre-clinical interventions to improve patient outcome.

EXTH-51. ANTI-INVASIVE EFFICACY AND SURVIVAL BENEFIT OF THE YAP-TEAD INHIBITOR VERTEPORFIN IN PRECLINICAL GLIOBLASTOMA MODELS

Abstract Glioblastoma (GBM) remains an incurable disease, in large part due to its malignant infiltrative spread, and current clinical therapy fails to target the invasive nature of tumor cells in disease progression and recurrence. Here, we use the YAP-TEAD inhibitor Verteporfin to target a convergence point for regulating tumor invasion/metastasis and establish the robust anti-invasive therapeutic efficacy of this FDA-approved drug and its survival benefit across several preclinical glioma models. Using patient-derived GBM cells and orthotopic xenograft models (PDX), we show that Verteporfin treatment disrupts YAP/TAZ-TEAD activity and processes related to cell adhesion, migration and epithelial-mesenchymal transition. In-vitro, Verteporfin impairs tumor migration, invasion and motility dynamics. In-vivo, intraperitoneal administration of Verteporfin in mice with orthotopic PDX tumors shows consistent drug accumulation within the brain and decreased infiltrative tumor burden, across three independent experiments. Interestingly, PDX tumors with impaired invasion after Verteporfin treatment downregulate CDH2 and ITGB1 adhesion protein levels within the tumor microenvironment. Finally, Verteporfin treatment confers survival benefit in two independent PDX models: as monotherapy in de-novo GBM and in combination with standard-of-care chemoradiation in recurrent GBM. These findings indicate potential therapeutic value of this FDA-approved drug if repurposed for GBM patients.

DDRE-14. HEMP DERIVED EXTRACELLULAR VESICLES (EVS): A POTENTIAL ANTI-GLIOMA THERAPY

Abstract INTRODUCTION Extracellular vesicles (EVs) have been harvested from many plant sources, some of which have anti-cancer effects and some could be used as therapeutic nanodelivery vectors. Hemp plant is a natural source of cannabinoids, of which delta 9-tetrahydroxicannabinol (THC) and cannabidiol (CBD) have proven anti-cancer proprieties. HYPOTHESIS We hypothesized that hemp EVs are enriched in cannabinoids and their application will reduce glioblastoma (GBM) tumor progression. APPROACH EVs were isolated from the hemp plant using ultracentrifugation. Nanotracking analysis, electron microscopy and liquid chromatography tandem mass spectrometry (LC-MS/MS) were utilized to characterize EVs. GBM cell lines were cultured in the neuropshere assay to evaluate hemp EVs anti-glioma effects. Fluorescent-labelled EVs were used to evaluate their brain tissue distribution in orthotopic patient-derived GBM xenografts. RESULTS Hemp EVs have a median diameter of 112.6nm with a typical lipid-bilayer structure. LC-MS/MS have shown that while cannabidiolic, cannabigerolic, and tetrahydroxicannabinolic acids represent 69.1 ± 2.1%, 19.1 ± 1.6%, 6.5 ± 0.54% of the total cannabinoids in hemp EVs, CBD and THC only make 4.75 ± 0.26%, and 0.5 ± 0.3%. Hemp EVs are potent anti-glioma agents with a 7-day LD-50 of 1.04µM and 2.4µM [based on EVs total cannabinoid content] for KR-158 and L0 GBM lines, respectively. Compared to the vehicle, overnight incubation of L0 cells with 1µM hemp EVs significantly reduced GBM cell migration (630.3 ± 61.43 vs 143.7 ± 8.7). Intranasal administration of hemp EVs led to a widespread distribution in tumor bearing brain including GBM tumor core. CONCLUSION Based on these results, hemp EVs with enriched cannabinoid content exert antiglioma effect in-vitro and when delivered intranasally, are widely distributed throughout the brain and within the tumor of PDX animals. Further experiments are ongoing to address the impact of nasally-delivered hemp EVs on tumor progression and compare to the application of purified acidic cannabinoids.

EXTH-73. HETEROGENEITY IN CELL SURFACE ANTIGEN AND HLA CLASS I EXPRESSION IN PEDIATRIC BRAIN TUMORS AND ITS IMPACT ON T-CELL IMMUNOTHERAPY

Abstract Immunotherapy with tumor antigen-specific chimeric antigen receptor (CAR) and/or ab (T-cell receptor) TCR T-cells has the potential to improve clinical outcomes of patients with pediatric brain tumors. As a prerequisite for successful T-cell therapies, we must determine which cell surface antigens are expressed and targetable in these tumors, and if HLA Class I is present, which is necessary for ab TCR T-cell recognition. Therefore, in this study we systematically analyzed pediatric patient-derived orthotopic xenograft (PDOX) brain tumor samples for cell surface expression of five known CAR targets: IL13Ra2, HER2, EphA2, B7-H3, and GD2, as well as HLA Class I. We established and validated a flow cytometry-based method of profiling tumor-associated antigens. Fifty-three PDOX samples have been profiled to date, including medulloblastoma, high grade glioma (HGG), diffuse intrinsic pontine glioma (DIPG), atypical teratoid rhabdoid tumor (ATRT), and ependymoma, among others. Our results showed high variability within and between individual samples. B7-H3 was the most consistently expressed, seen in 98% of the samples tested. We validated these results by conventional immunohistochemistry staining for B7-H3 and found comparable RESULTS: HLA Class I was highly expressed on all HGG samples but was undetectable on 47.8% of other brain tumor samples. This suggests that down-regulation of HLA class I is one mechanism by which brain tumors evade conventional T-cells, and that HLA-independent CAR T-cells would be useful therapies. We also compared expression of antigens in fresh patient samples and corresponding PDOX tumors and saw that they were indeed similar. To our knowledge, this is the largest group of pediatric brain tumor PDOX samples methodically analyzed for potential CAR target antigens and HLA Class I. Taken together, our data demonstrate that elimination of tumors by CAR T-cell immunotherapies will require targeting multiple antigens, and our profiling method could inform how to circumvent antigen-negative relapse.

BIOM-03. INVASIVE HISTOPATHOLOGY DRIVES POOR OUTCOMES IN SURGICALLY RESECTED BRAIN METASTASES

Abstract BACKGROUND Surgery as a single modality for the treatment of brain metastases (BrM) results in local recurrence (LR) in 60% of patients. These failure rates are reduced by half with post-operative radiotherapy. The non-invasive nature of BrM has led to the assumption that local recurrence is caused by spillage of cancer cells into the surgical cavity at the time of surgery. We present evidence suggesting that invasion of metastatic cancer cells into the adjacent brain is present in the majority of BrM and is associated with LR, leptomeningeal metastasis (LM), and overall survival (OS). METHODS We assessed the histopathological growth pattern (HGP) of 164 surgically resected BrM. HGP was correlated with LR, LM and OS. Single-cell transcriptomics (scRNAseq) was performed on 15,615 cells from metastasis center (MC) and surrounding brain (SB) adjacent to the tumor. N=30 orthotopic patient-derived xenograft models (OPDX) were established from BrM. RESULTS 56/164 (34%) BrM specimens showed a minimally invasive (MI) HGP between the tumor and adjacent brain while 108/164 (66%) showed significant invasion of tumor lobules or single-cells into the brain (HI-HGP). HI-HGP was associated with LR, LM and shortened OS in BrM patients. scRNAseq identified abundant cancer cells in SB that overexpressed pathways and genes involved in cell survival and stress adaptation compared to matched cancer cells in MC. Validation of these targets with immunohistochemistry in patient and OPDX tissues revealed cold-inducible RNA binding protein (CIRBP) overexpression in HI-HGP patient and OPDX BrM. Modulation of CIRBP expression in OPDX and cell line models of HI-HGP BrM delayed BrM progression and extended OS. CONCLUSION HI-HGP is a poor prognostic indicator in patients with surgically resected BrM, establishing HGP as an important prognostic factor that should be considered by clinicians treating BrM patients. We identify CIRBP as a functional mediator of this process.

STEM-01. TARGETING BRAIN METASTASIS-INITIATING CELLS: A PREVENTATIVE APPROACH

Abstract BACKGROUND The incidence of brain metastases (BM) is tenfold higher than primary brain tumors. BM commonly originate from primary lung, breast, and melanoma tumors with a 90% mortality rate within one year of diagnosis. Current standard of care for BM includes surgical resection with concurrent chemoradiation, but does not extend median survival past 16 months, posing a large unmet need to identify novel therapies against BM. METHODS From a large in-house biobank of patient-derived BM cell lines, the Singh Lab has generated murine orthotopic patient-derived xenograft (PDX) models of lung, breast, and melanoma BM that recapitulate the stages of BM progression as seen in humans. Using these three PDX models, we identified a population of “pre-metastatic” brain metastasis-initiating cells (BMICs) that are newly arrived in the brain but have yet to form detectable tumors. Pre-metastatic BMICs are not detectable in human patients but are important therapeutic targets with the potential to prevent BM in at-risk patients. RESULTS RNA sequencing of pre-metastatic BMICs from all three PDX primary tumor models with subsequent Connectivity Map analysis identified novel compounds that have the potential of killing all three types of BMICs. In particular, we identified two compounds that have selective killing of BMICs in vitro from all three primary tumor cohorts while sparing non-cancerous cells. We further characterized their ability to inhibit the self-renewal and proliferative properties of BMICs. Ongoing in vivo work will investigate the compounds’ preclinical utilities in preventing BM. CONCLUSION Identification of novel small molecules that target BMICs could prevent the formation of BM completely and dramatically improve the prognosis of at-risk cancer patients.

EPCO-16. LACTIC ACID IS AN EPIGENETIC METABOLITE THAT DRIVES GLIOBLASTOMA SURVIVAL AND GROWTH

Abstract Glioblastoma (GBM) is the most common primary malignant brain tumor with an unfavorable prognosis and a reprogrammed metabolism. While tumors utilize glucose, there are other carbon sources at their disposal. Originally considered as a waste product of glucose catabolism, lactate accumulates to a significant amount in tumor tissue. We launched our studies with the central hypothesis that lactate is metabolized by GBM cells to promote their survival via modulation of the epigenome. We showed that lactate rescued patient-derived xenograft (PDX) GBM cells from nutrient deprivation mediated cell death and inhibition of growth. Transcriptome analysis, Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq), and CHIP-seq. showed that lactic acid exposure entertained a signature of cell cycle progression, oxidative phosphorylation (OXPHOS) and MYC target expression. LC/MS analysis demonstrated that U-13C-Lactate elicited substantial labeling of TCA-cycle metabolites, acetyl-CoA and histone protein acetyl-residues in PDX derived GBM cells. Given that acetyl-CoA is pivotal for histone acetylation we observed a dose-dependent elevation of histone marks (e.g. H3K27ac), which was rescued by genetic and pharmacological inhibition of lactic acid-uptake, ATP-citrate lyase, p300 histone-acetyl-transferase and OXPHOS, resulting in reversal of lactate mediated protection from cell death or facilitation of GBM growth. CHIP-seq. analysis demonstrated that lactic acid facilitated enhanced binding of H3K27ac to gene promoters and cis-regulatory elements (e.g. super-enhancers). Consistently, ATAC-seq. analysis highlighted enhanced accessibility of the chromatin by lactic acid. Finally, we assessed whether lactic acid is actively metabolized in vivo, utilizing an orthotopic PDX model of GBM. In a combined tracer experiment (U-13C-glucose and 3-C13-lactate), we made the fundamental observation that lactic acid carbons were predominantly labeling the TCA cycle metabolites over glucose, implying a critical role of lactic acid in GBMs and establishing lactic acid metabolism as a novel drug target for GBM that may be targeted with epigenetic drugs.