Daniel J. Brat, MD, Mayo Clinic 1994 PhD, Mayo Clinic 1994 - US grants

DESCRIPTION (provided by applicant): The following proposal is designed to provide the primary investigator, Daniel J. Brat, M.D., Ph.D., with necessary scientific experience and mentorship to allow a transition to an independent clinician scientist. Dr. Brat received his M.D. and Ph.D. degrees from Mayo Medical and Graduate Schools, and completed Anatomic Pathology and Neuropathology training at Johns Hopkins Hospital. His academic interests center on morphologic and molecular genetic investigations of primary brain tumors, both in terms of underlying mechanisms and classification. The goal of this proposal is to demonstrate a relationship between biologic behavior of brain tumors and their patterns of genetic alterations using comparative genomic hybridization in the format of DNA micro-arrays. Comprehensive tumor genotypes will be useful for determining pathways of genetic progression in distinct types of brain tumors, and for establishing patterns of genetic alterations that discriminate subsets of CNS neoplasms based on biologic behavior, response to therapy, and outcome. Genetic alterations that define certain gliomas are currently used to direct therapy: anaplastic oligodendrogliomas with 1p and 19q losses are sensitive to specific chemotherapy regimens. Distinct alterations among astrocytoma subtypes, including glioblastoma multiforme (GBM), have also been defined, but require further investigation in order to establish molecular subsets that may define behavior. Emerging micro-array technology offers the opportunity to define primary brain tumor genotypes comprehensively and precisely. Under the guidance of Erwin Van Meir, Ph.D., the first goal will be to demonstrate genetic alterations in the format of comparative genomic DNA arrays using a limited number of probes that are well characterized in adult GBMs. Once the experimental system has been validated, micro-arrays will be expanded to include a higher density of informational markers (200-300 loci). These will include gene families of significance in CNS tumorigenesis and markers from all chromosomes so that micro-arrays are useful for investigating patterns of genetic alterations in both glial and neuronal neoplasms, including those of childhood. Specialized DNA micro-arrays will be applied to biologically distinct brain tumors in order to define unique molecular genetic subgroups, and to gliomas from patients enrolled in clinical trials to determine if any patterns discriminate between tumors with regard to behavior, response to therapy, or clinical outcome.

DESCRIPTION (provided by applicant): Glioblastoma (GBM) is the most common primary brain tumor and also the highest grade (WHO grade IV). Progression to a GBM represents an abrupt turning point, with rapid progression to death following the transition (mean, 50 weeks). Two pathologic features that distinguish GBM from lower grade tumors and are mechanistically instrumental are necrosis, typically with surrounding cellular pseudopalisades, and microvascular hyperplasia. Pseudopalisades are hypoxic and secrete pro-angiogenic factors that promote microvascular hyperplasia, an exuberant form of angiogenesis that supports the rapid tumor expansion. Mechanisms underlying pseudopalisades, hypoxia, and necrosis in GBM have not been defined. We hypothesize that vaso-occlusion and intravascular thrombosis give rise to pseudopalisades and the ensuing hypoxia-induced angiogenic cascade, accounting for the abrupt onset of rapidly progressive disease. This proposal follows our preliminary data, which indicates that pseudopalisades are neoplastic cells migrating away from central hypoxia created in part by microscopic thrombotic vascular occulsion. The initiating vascular insult that precedes intravascular thrombosis and pseudopalisading necrosis has not been determined. Ang-2, a Tie-2 receptor antagonsist that mediates endothelial apoptosis and vascular regression in the absence of VEGF, is expressed by endothelial cells of high grade gliomas and is a prime candidate for initiating vascular pathology. Also occurring during the transition to GBM are PTEN mutations and increased tumor cellularity, especially around blood vessels. We will examine whether neoplastic PTEN loss leads to the secretion of a protein capable of triggering endothelial apoptosis through Ang-2 mediated mechanisms. Candidate secreted or cell contact proteins will be identified by advanced protein separation methods and mass spectrometry. We also hypothesize that intravascular thrombosis is relevant to tumor necrosis and glioma biology and will examine whether PTEN loss and/or hypoxia, promote thrombosis through increased expression of the pro-thrombotic factors PAR-1 and tissue factor. The rationale and experiments in this proposal are novel and innovative, since the etiologies of hypoxia, pseudopalisades, and necrosis are unknown, vascular mechanisms have not been proposed, and intravascular thrombosis, while a frequent finding, has not been recognized as a potential driving force in tumor progression. Our emerging model represents a paradigm shift in the understanding of GBM and should lead directly to more effective therapies.

Glioblastoma (GBM) is the most common primary brain tumor and also the highest grade (WHO gradeIV). Progression to GBM represents an abrupt turning point, with death quickly following the transition. One of the most specific pathologic features that emerges during the transition and distinguishes GBM from lower grade tumors is necrosis with surrounding cellular "pseudopalisades". Pseudopalisades are composed of hypoxic tumor cells that secrete pro-angiogenic factors critical for promoting angiogenesis and tumor expansion. Mechanisms underlying the development of pseudopalisades, hypoxia, and necrosis in GBM are undefined, but we believe that understanding their origins will be critical for attempts to stabilize this disease. We hypothesize that vaso-occlusion and intravascular thrombosis give rise to pseudopalisades and the ensuing hypoxia-induced angiogenic cascade, accounting for the abrupt onset of rapid disease progression. Our preliminary data has demonstrated that thrombotic vascular occlusion within the neoplasm is associated with hypoxia-induced outward migration of glioma cells to form pseudopalisades. Mechanisms by which neoplastic cells induce endothelial damage, vaso-occlusion, and thrombosis have not been established. Ang- 2 is a Tie-2 receptor antagonist that mediates endothelial apoptosis in experimental gliomas and is a prime candidate for initiating these events. Since PTEN mutations occur during the transition to GBM, we will examine whether PTEN loss leads to the secretion of proteins that trigger endothelial apoptosis through Ang- 2. We also hypothesize that increased expression of the pro-thrombotic proteins tissue factor (TF) and protease activated receptor-1 (PAR1) promote intravascular thrombosis. We will examine whether PTEN loss or hypoxia promotes TF-mediated intravascular clotting and whether PAR1 activation leads to increased cellular migration associated with pseudopalisade formation. An animal model of astrocytoma will be used to validate the significance of PTEN loss, TF expression and intravascular thrombosis in the progression to GBM and to determine if anti-thrombotic therapies are capable of delaying the development of hypoxia and prolonging survival. Vaso-occlusion and intravascular thrombosis have not been previously recognized as driving forces in the development of hypoxia, angiogenesis and glioma progression. This proposal introduces entirely novel concepts that may explain the highly aggressive properties of GBM and suggests therapeutic approaches that could potentially stabilize its progression.

The Animal Models and Pathology Specialized Resource (Core D) of the Emory Molecular and Translational Imaging Center (EMTIC) will assist the overall project by providing service in two critical areas: animal models and histopathologic services. The Core is necessary for the success of the EMTIC and is highly integrative, being utilized by all four research projects and all five pilot projects. The animal models component will act to generate and characterize novel rodent models for molecular imaging in cancer research and will assist the individual projects by collecting and analyzing tumor data from these animals. The production and analysis of relevant animal models of cancer is central to the EMTIC because of the heavy emphasis on the development and validation of novel tracers in preclinical studies that could be utilized as markers for tumor detection, progression, and dissemination in vivo. The animal component will also serve as a repository for the preservation and distribution of any novel transgenic strains of mice that may be utilized in the Research Projects and Pilot Projects as they become necessary. The Core will perform necropsy and tissue collection for experimental animals, and will collect tumor samples from selected animals for generation of cell lines for use by EMTIC investigators. The pathology component of the Core will provide tissue and tumor procurement expertise, tissue processing, as well as histological and immunohistochemical analysis of tumor samples from human disease and animal models. Histologic and immunohistochemical evaluation of tumors from patients and animal models will be necessary in order to correlate and validate the detection of novel tracers by PET, MRI and optical methods with the targeted biomarkers in tissues. Thus the Animal Models and Pathology Core will contribute significantly to the overall goals of the EMTIC from a scientific perspective by providing these services as well as expertise. In addition, the inclusion of these services into a comprehensive core component will provide additional benefit to EMTIC as a whole, through consolidation of effort, avoidance of unnecessary experimental duplication, and by drawing upon the collective expertise of the Core personnel.

The Human Tissue and Pathology Core (Shared Resource Category 4.06) of the Winship Cancer Institute (WCI) will be critical to the overall clinical and scientific success by enhancing tissue-based investigation through its procurement and distribution of human cancer specimens and its comprehensive histopathologic services. The Core is highly integrated within the WCI, serving individual and team science projects in all of the major programs and interfacing directly with the other Cores. The tissue banking component of the Core consists of an IRB-approved Human Tissue Procurement Service (HTPS), located within the Emory University Hospital and the WCI, with affiliated banking sites at Grady Memorial and Crawford Long Hospitals. Tissue procurement personnel are highly experienced in anatomic pathology and tumor banking protocols, with quality assurance of specimens provided by members of the Department of Pathology. Tumor and normal control tissues, along with other biospecimens associated with specialized protocols, are procured as frozen, fresh, fixed, embedded or other specialized preparation, and inventoried in caTissue Core, a caBIG tool for biospecimen repositories. Tissues are distributed to WCI investigators with IRBapproved protocols to further the clinical, translational, basic and epidemiologic research. Pathology services of the Core are provided by the Research Pathology Laboratory, a full service lab located on the 5th floor of the WCI building with experienced personnel and equipment for tissue processing, histological and immunohistochemical staining and analysis, laser capture microscopy, tissue microarrays and high throughput, automated image analysis for immunofluorescent and bright field microscopy. This lab is capable of processing high volumes of tissue samples derived from human disease and animal models and works in close collaboration with the other Cores and scientific programs of the WCI as demonstrated by research productivity and support of group science grants. The success of this Core is also highlighted by its contract with the NCI for collecting and processing glioblastoma specimens for The Cancer Genome Atlas Project. Together, the HTPS and the Research Pathology Lab provide centralized, efficent and high quality services to WCI investigators for acceleration of scientific discovery.

DESCRIPTION (provided by applicant): Glioblastoma (GBM) is the most common and most malignant primary brain tumor. Recent studies suggest that a small subset of neoplastic cells, referred to as Glioma Stem Cells (GSCs), may govern the biologic behavior of GBM. GSCs have properties of self-renewal, pluripotency and high tumorigenicity. They have thus far been identified by the expression of specific markers, such as CD133, Sox2 and nestin, yet molecular mechanisms responsible for stem-like behavior have not been clearly defined, nor have regulators of the stem/non-stem equilibrium. This proposal aims to define molecular mechanisms that confer stem-like qualities to GSCs and enriches their presence in GBMs. Pathways that direct asymmetric cellular division and stem-like behavior in the Drosophila melanogaster nervous system have been well described and may provide clues to stem cell properties and the stem/non-stem balance in malignant gliomas. The Drosophila brain tumor (brat) gene product regulates asymmetric cell division through its segregation into the daughter cell destined for differentiation, where it functions to translationally repress Myc. In brat mutants, asymmetric division and neural differentiation do not occur, leading to a massively enlarged larval brain containing highly proliferative undifferentiated neuroblastic cells with neoplastic properties. The human homolog of Drosophila brat, Trim3, shows allelic loss in over 25% of GBMs and reduced expression in nearly all. In the current proposal, we hypothesize that reduced expression of Trim3, or its interacting proteins, is critical in defining stem-like properties in human GSCs by favoring a loss of asymmetric cell division and enriching the stem cell compartment. We propose to investigate Trim3 in human gliomas in order to determine if its expression is downregulated in human GBMs; if Trim3 regulates c-Myc protein expression and activity in vitro and in vivo; if Trim3 regulates stem like properties of human in GBM neurosphere cultures and GBM resection specimens; if genetic and hypoxic mechanisms regulate Trim3 during tumor progression; and if Trim3 regulates the in vivo growth properties of GBM in animal models.

DESCRIPTION (provided by applicant): Glioblastoma (GBM) is the most common primary brain tumor and the most malignant form of astrocytoma (WHO grade IV). We have demonstrated that both GBM tumor cells and its vascular endothelium express the cell surface receptor tissue factor (TF), a critical initiator of thrombosis, while the vasculature of the normal brain does not. Here we explore mechanisms related to the upregulation of TF in the vasculature of GBM and hypothesize that a cytotoxic agent conjugated to a carrier directed at TF will specifically target blood vessels of GBM, but not non-neoplastic brain. Thus, the objective of this proposal is to develop a novel therapeutic approach for GBM, in which the TF-expressing vasculature is specifically targeted. The cytotoxic agent we have developed is EF24, a synthetic curcumin analog, which will be linked to enzymatically inactive coagulation factor VIIa (fVIIa), the high affinity ligand for TF that has exquisite specificity. We hypothesize that this drug conjugate (EF24-FFRck-fVIIa) will bind to TF on vascular endothelial cells (VECs) within the GBM, enter target cells by ligand-receptor mediated endocytosis, and elicit a cytotoxic response. The disruption of the blood brain barrier (BBB) due to targeting the VECs and to the tissue-destructive nature of GBM should also permit binding of the drug-conjugate directly to TF expressing neoplastic cells. Thus, this therapeutic approach has a high likelihood of having both an anti-angiogenic and direct anti-tumor effect. Specific Aim I will determine the mechanisms by which malignant gliomas induce the expression of TF by vascular endothelial cells. We will establish whether PTEN loss and hypoxia lead to release of factors by gliomas that induce endothelial TF expression in vitro and will determine if these mechanisms have correlates in human brain tumor specimens. Aim II will establish the distribution of the drug-conjugate and the drug in glioma xenografts, their associated vasculature and the adjacent normal brain. Near infrared optical imaging will be used to determine the distribution of the drug-conjugate in vitro and in vivo using Cy5.5-labeled EF24-FFRck- fVIIa. The distribution of the drug using biotinylated EF24 combined with streptavidin histochemistry will be used to more precisely localize the drug in glioma xenografts tissue sections. Aim III will determine the efficacy of the drug-conjugate against malignant gliomas in a mouse xenograft model. We will examine the effects of the EF24-FFRck-fVIIa on survival in this model and examine the biologic correlates of drug-conjugate treatment, including tumor growth, angiogenesis and disruption of the BBB. PUBLIC HEALTH RELEVANCE: Glioblastoma (GBM) is the most common primary brain tumor, with 8700 new cases per year in the United States. These tumors are universally fatal and the average length of patient survival is only 60 weeks with current therapies. This proposal uses a highly innovative approach to specifically target the abnormal blood vessels of GBM in order to slow their growth. Although this proposal specifically addresses the abnormal vessels in GBM, the approach may find application in other diseases with abnormal blood vessel growth, such as diabetic retinopathy, macular degeneration, endometriosis, Crohn's disease, psoriasis, and other cancers.

The Cancer Tissue and Pathology Shared Resource (Category 4.06) is critical to the clinical and scientific success of the Winship Cancer Institute because of its central role in facilitating tissue-based investigation though the procurement of human cancer specimens and comprehensive histopathologic services. The Resource is highly integrated, serving individual and team projects in all of the major programs and interfacing directly with the other Shared Resources. The tissue banking component consists of an IRB approved Human Tissue Procurement Service (HTPS), located within the Emory University Hospital and Winship, with affiliated banking sites at Grady Memorial, Emory Midtown and Children's Healthcare of Atlanta at Egleston. Procurement personnel are experienced in anatomic pathology and tumor banking protocols with quality assurance provided by members of the Department of Pathology. Tumor, normal control tissues and blood, along with other biospecimens on specialized protocols, are procured according to standardized operating procedures as frozen, fresh, fixed, embedded or other specialized preparation, and inventoried in caTissue Core, a caBIG tool for biospecimen repositories. Over 55,000 specimens are available. Tissues are distributed to Winship investigators with IRB-approved protocols to advance clinical, translational, basic and epidemiologic research. Pathology services are provided by the Research Pathology Laboratory, a full service lab located on the 5th floor of the Winship building with experienced personnel and equipment for tissue processing, histological and immunohistochemical staining, laser capture microscopy, tissue microarrays and a whole slide scanning for digital pathology. The lab processes large volumes of samples derived from human disease and animal models and works in close collaboration with the other Shared Resources and scientific programs of Winship, as demonstrated by research productivity and support of group science grants. The growing success of this Shared Resource is highlighted by its leadership in multiple NCI-driven initiatives, including caBIG, TCGA and caHUB. This Shared Resource provides centralized, efficient and high quality services that are not duplicated by other facilities on campus.

DESCRIPTION (provided by applicant): Glioblastoma (GBM; WHO grade IV) is the most frequent primary brain tumor and has a dismal prognosis. A common goal in neuro-oncology is to develop therapies targeted at molecular mechanisms of tumor progression. One potential detractor to such approaches is the tremendous heterogeneity within a given patient's GBM, such that molecular targets could vary depending on the micro-environment from which they are sampled. Following neurosurgical resection of these highly infiltrative tumors, it will be criticalto direct therapies at druggable targets present at the residual invasive tumor near the resection margin, rather than at those of the bulk resected tumor, since these could differ substantially. Our preliminary data and the literature suggest that transcriptional programs and EGFR/PDGFRA amplification events vary across regions within GBM. Downstream signaling networks and druggable targets likely show similar variation. We initiated a Phase II clinical tria of 5-Aminolevulinic Acid (5-ALA), a fluorescent compound that accumulates in glioma cells, thereby enhancing visualization and neurosurgical resection and allowing definition of macro-environments that include 1) perinecrotic glioma, 2) bulk glioma, and 3) glioma margin. Using this novel neurosurgical platform, we propose to define spatial and temporal molecular variations of human GBMs including transcriptional and phospho-protein profiles, EGFR/PDGFRA amplification and downstream network activation by multiplex quantum dots. We also investigate molecular alterations that evolve within recurrent GBM in patient samples as compared to those present at the tumor margin of the primary tumor. A GBM xenograft model in which hypoxia and necrosis are induced using photo-activated compounds that cause vaso-occlusion will be used to precisely monitor spatial and temporal evolution of molecular variation. We hypothesize that the transcriptional profiles, genomic amplification, and druggable tyrosine kinases vary spatially and temporally, and will differ at the residual tumor margin as compared to the bulk tumor.

CANCER TISSUE AND PATHOLOGY SHARED RESOURCE PROJECT SUMMARY/ ABSTRACT The Cancer Tissue and Pathology Shared Resource (Tissue SR; Category 4.06) facilitates tissue-based investigation through the procurement of human cancer specimens and comprehensive histopathologic services. The resource is highly integrated, serving individual and team projects in Winship's four research programs and interfacing directly with Winship's other shared resources. The tissue procurement and banking component consists of an IRB-approved Human Tissue Procurement Service (HTPS) located within the Emory University Hospital (EUH) and Winship Building C, with affiliated banking sites at Grady Memorial Hospital (Grady), Emory University Hospital Midtown (EUHM), Emory Saint Joseph's Hospital (ESJH), and Aflac Cancer and Blood Disorders Center of Children's Healthcare of Atlanta (Aflac). Procurement personnel are highly experienced in anatomic pathology and tissue banking protocols, with quality assurance provided by Department of Pathology faculty members. Tumor, adjacent non-tumor tissues, blood, and other biospecimens on specialized protocols are procured according to standardized operating procedures as frozen, fresh, fixed, or other specified preparation and inventoried and tracked using the Nautilus LIMS? system. Over 75,000 specimens are stored and available for use. Tissues are distributed to Winship investigators with IRB-approved protocols to support clinical, translational, basic, and epidemiologic research. Pathology services are provided by the Research Pathology Laboratory (RPL) component, a full service lab located in Winship Building C with experienced personnel and equipment for tissue processing, histological and immunohistochemical staining, laser capture microscopy, tissue microarrays, whole slide scanning for digital pathology, and standard and advanced capabilities for pathology image analytics. The lab processes large volumes of samples derived from human disease and animal models and works in close collaboration with the other shared resources and with Winship's research programs, as demonstrated by research productivity and support of group science grants. The success of this shared resource is highlighted by its national leadership in several NCI-driven initiatives, including The Cancer Genome Atlas (TCGA), Cancer Human Biobank (caHUB), and Minority Biospecimen/Biobanking Geographic Management Program (BMaP). This shared resource provides centralized, efficient, and high quality services that are not duplicated by other Emory University facilities.

Project Summary/Abstract Glioblastoma (GBM) is the most common and most malignant primary brain tumor and has a dismal prognosis. Abundant evidence now indicates that a small subset of neoplastic cells, referred to as Glioma Stem Cells (GSCs), governs biologic behavior and resistance to therapy. GSCs inhabit specific biologic niches and have properties of self-renewal, pluripotency and high tumorigenicity. They can be identified by expression of markers, such as CD133, CD15 and nestin, yet molecular mechanisms responsible for their specific stem-like behavior in glioblastomas have not been fully defined. This proposal aims to define mechanisms underlying fundamental biological properties of GSCs, including: 1) their marked accumulation following the development of necrosis and tendency to localize to the hypoxic niche; and 2) a disrupted program of asymmetric cell division that favors self-renewing division over differentiation. We have developed novel in vitro and in vivo models and analytic techniques to study the differential behavior of stem and non-stem glioma cells within the tumor micro-environment, especially as it relates to their accumulation in regions of hypoxia. These include an in vivo orthotopic xenograft model in which stem cells are interrogated for patterns and mechanisms of accumulation following the induction of necrosis using a photo-activated dye. We also dissect pathways that direct asymmetric cellular division in Drosophila nervous system provide clues to understand the stem/non-stem dynamics in malignant gliomas. We have previously demonstrated that the human ortholog of Drosophila Brat, Trim3, is a tumor suppressor that regulates glioma stem cell dynamics and promotes stemness when lost. Here we propose to investigate novel regulatory mechanisms that arise during transition to the stem cell phenotype that can be antagonized therapeutically. Regulatory networks and potential therapeutic targets are further explored in xenografts and genetically engineered mouse models that recapitulates human gliomas to show efficacy.

In nearly all forms of human cancer, the development of necrosis is tightly linked with malignant progression. Whether necrosis accelerates progression or is largely passive remains an open question, yet modeling these events to establish mechanisms and therapeutic vulnerabilities in animals has been challenging. In glioblastoma (GBM; WHO grade IV), the most malignant primary brain tumor, the rapid, radial growth phase that leads quickly to death is consistently preceded by the development of central necrosis. While genetic alterations of GBM are known in great detail, the biological properties that result from their acquisition and lead to this accelerated growth phase require deeper investigation. The tumor microenvironment (TME) changes dramatically following the onset of necrosis, from a sheet-like growth of infiltrating cells with relatively constant growth properties to a highly complex and evolving 3-D microsystem composed of diverse cell types and spatially segregated signaling networks. To better understand the dynamic temporal and spatial changes that promote progression, we propose to advance mouse models that closely parallel these events in human gliomas, since many mouse models of GBM lack necrosis. We developed a novel method to induce focal necrosis within high grade gliomas in vivo and will study TME restructuring and its impact on glioma growth in real time using multiphoton microscopy. As translational applications, we will demonstrate how hypoxia and necrosis promote the enrichment of glioma stem cells (GSCs) in their peri-necrotic niche and lead to the dramatic influx of tumor-associated macrophages (TAMs), which increase in number over 10-fold in the human disease. We propose both genetically characterized patient-derived GBM xenografts grown in mice with humanized immune cells, as well as an immunocompetent RCAS/tv-a model, and will determine how antagonizing these processes impact disease progression and outcomes. Our preliminary data and the literature indicate substantial differences between pre-necrotic and necrotic gliomas with regard to GSC and TAM enrichment and their impact on biological properties, but the mechanisms and evolution have not been studied in depth, in large part due to the absence of a credible animal model. Our model will capture glioma growth dynamics, GSC enrichment, and TAM influx, and facilitate the development of therapies that antagonize these mechanisms to improve outcomes.