2004 Project Summaries
Recipients of Three-year Awards
Elizabeth Maher, M.D., Ph.D. Glioblastoma (GBM) represents one of the most aggressive and deadly cancers for which there has been little improvement in treatment or basic understanding of disease over the past 25 years of clinical trials and basic science research. Although in its most common form GBM is of highest grade at the time of diagnosis, the ability to study the progression from low-grade astrocytoma to GBM provides a valuable opportunity to identify the key genetic switches governing this transition. Such knowledge of progression genetics would provide both a means of identifying those low-grade patients at increased risk for progression as well as offer novel and more directed drug development opportunities in advanced gliomas including primary and secondary GBM. The support of the Goldhirsh Foundation over the past year has led to the identification of several novel chromosomal regions of gain and loss that are associated with low grade astrocytomas and other regions associated with the transition to secondary GBM. The development of a higher resolution platform for genetic analysis over the past year coupled with the recent implementation of novel bioinformatic tools has been a major advance in our ability to perform these studies and provides even higher confidence that the genes identified will prove to be clinically relevant. The proposed expanded study is designed to ultimately identify those critical genetic elements governing disease progression, grade-specific biology, and therapeutic response.
Charles L. Sawyers, M.D. Cancer therapy is rapidly evolving to incorporate our greater understanding of the molecular basis of tumor growth into the design of molecularly targeted drugs. The greatest success to date has been in targeting a class of enzymes called kinases that are broadly implicated in cancer. Clinical success has been seen primarily in leukemias and small subsets of solid tumors, in which kinase genes are activated by mutations. We believe that similar success should be possible in glioblastoma because a significant subset of these tumors also have evidence of kinase activation. However, one huge difference is that glioblastomas comprise a molecularly heterogenous group of tumors that may contain diverse types of kinase activation. Successful application of kinase inhibitors will require precise knowledge of kinase gene status in tumors so that the appropriate inhibitor can be selected. The UCLA glioblastoma program has recently had some promising results in this area using inhibitors of a kinase called mTOR in patients with a molecular lesion in this pathway. The goal of this proposal is to expand this approach more broadly and incorporate the latest advances in molecular imaging and genomics technology to patients enrolled in these clinical trials, to maximize our understanding of how these drugs work. Specifically, we will incorporate state of the art PET scanning into these clinical trials (Aim 1), decipher signatures of gene expression that define response to kinase inhibitors (Aim 2) and survey these kinase pathways for additional mutations in glioblastoma patient that have not yet been discovered (Aim 3). This proposal represents a team approach, involving collaboration between UCLA experts in the clinical management of these patients, molecular analysis of clinical material, clinical use of molecular imaging probes and basic molecular understanding of these signal transduction pathways. Recipients of One-year Awards
Arturo Alvarez-Buylla, Ph.D. Gliomas are made up of cells that divide extensively in an uncontrolled manner. Mutations in key regulatory genes involved in proliferation and cell death are considered a principal cause of glioma formation. It is likely that these frequently mutated genes regulate the normal proliferation of endogenous progenitors cells in the adult brain. A fundamental question remains unresolved: which are the cells in the brain that give rise to tumors? Recent discoveries suggest that tumor cells share important characteristics with neural stem cells. Neural stem cells (NSCs) are immature progenitor cells present in the adult that are capable of generating the specialized cell types that make up the brain. NSCs persist throughout the lifetime of an animal, so they are very long lived. They also have the capacity to divide extensively, but in a controlled manner. Why do tumor cells act like "out of control" stem cells? Do tumors arise from stem cells, or do they appropriate the mechanisms of long-term survival and cell division normally used by stem cells? Our laboratory has identified a subpopulation of astrocytes that function as stem cells in the adult brain. These cells normally give rise to transit amplifying cells that differentiate into new neurons. We will investigate whether growth factors known to be involved in brain tumor formation, induce abnormal growth of stem cells or their immediate progeny the transit amplifying cells. Preliminary evidence indicates that this is the case. Growth factors infused into the brain induce neural stem cells to form large masses of proliferating cells and some of these cells begin invading the adjacent healthy brain just as malignant tumors do. We propose experiments to identify the cells that respond to these growth factors. NSCs divide for long periods of time and during this time could accumulate mutations that make them behave as tumor cells. Many of the genes that are mutated in gliomas are known. Therefore, by introducing some or all of these mutations in neural stem cells, we will study how they affect the proliferation and invasiveness of neural stem cells and determine if these cells can give rise to tumors. We will also mutate the same genes in other non-dividing brain cells to determine whether or not stem cells have a special predisposition for tumor formation. This information will help elucidate the role that these genes play in regulating the behavior of normal neural stem cells and will suggests how tumors start. Finding better strategies to treat brain tumors is the principal motivation behind the present proposal. Understanding the origin of gliomas will be important for developing new strategies for early diagnosis. It will also tell us how glioma cells differ from the normal cells from which they arise and this could help in developing therapies that destroy tumor cells while sparing normal brain cells.
Gabriele Bergers, Ph.D. Glioblastomas (GBM), grade IV astrocytomas, develop from astrocytes and are the most malignant brain tumors because of their fast and diffuse growth pattern. The reason for their aggressive behavior is that GBMs create their own blood vessel supply (angiogenesis) that allows them to rapidly accumulate and invade deep into the brain, making it impossible to surgically remove the complete tumor mass. Despite radiation and chemotherapy, mean survival is only about a year because residing tumor cells will accumulate and regrow. Tumor cells have different ways to invade but one hallmark of all GBM is their ability to move along normal blood vessels in the brain parenchyma to disperse significant distances (perivascular invasion). We observed that genetically modified glioblastomas (by deleting the hypoxia-inducible factor HIF-1alpha) become extremely motile and preferentially use blood vessels as a freeway system to spread deep into the brain. Surprisingly, when the same cells are placed into a different microenvironment, the vascular-poor skin, tumors do not become invasive and grow very slowly. Our observations question the novel idea of using anti-HIF agents to block tumor growth because GBMs adapt by becoming severely invasive. However, this effect is also dictated by the brain microenvironment supporting the notion that tumor growth does not occur in a vacuum but requires the orchestrated action of the surrounding organ and its cell constituents. Perivascular invasion of GBMs requires at least two cell types: the endothelial cells that form the vascular tubes and the tumor cells. In addition, the extracelular matrix (ECM) in which the tumor is embedded, may also be of importance. It is a network of fibers with attached growth factors and othe molecules that enable tumor cells to migrate. We will investigate how GBMs are attracted to blood vessels and how they move along them by studying the communication circuits between tumor cells, endothelial cells and the matrix. We will utilize genetically engineered mouse models of glioblastomas and in vitro assays to understand the orchestrated action of the different cell types. The future goal will be then to transfer the knowledge into therapeutic approaches to successfully block such invasive capacities of GBMs and thereby prolong survival of GBM patients.
R. Jude Samulski, Ph.D. Adeno-associated virus (AAV) is a single-stranded(ss) human DNA Parvovirus virus that is the only known human DNA virus classified as non-pathogenic. While over 80% of the population has been infected by AAV, no evidence of disease has ever been reported. For these and other reasons, this virus has been developed and tested as a viral vector for human gene delivery. A number of AAV isolates exist in nature that have the ability to infect specific cell types. We have developed a novel technique that allows us to generate in a laboratory setting AAV variants that have diverse and unique ability to infect specific cell types (tropism). The objective of this proposal is to engineer an AAV viral vector that has specific tropism for astrocytoma cells. We have working protocols in the laboratory and preliminary data that suggest high degree of success. We have access to unique animal model that will allow further testing of these novel vector reagents in vivo. It is anticipated that by using these protocols and the proposed approach, we can evolve an optimum AAV vector for astrocytoma gene therapy. The long-term objective is to develop alternative modalities (i.e. genetic approaches) to help combat neurological cancer by first developing viral delivery reagents that have specific tropism. These reagents and studies will be essential for eventual testing in human clinical trials.
Harald W. Sontheimer, Ph.D. Astrocyte-derived tumors, which include astrocytomas and glioblastomas, are primary brain tumors that are often collectively called gliomas. Although these tumors share a number of features and genetic alterations with other cancers, they are unique in a number of aspects. For example, they are the only tumors that grow in a restricted space, namely the bony cavity of the skull, and they are also the only cancer that spreads by active cell migration rather than by hematogenous spread. Not unexpectedly such unique biology requires some unique adaptations. One of these adaptations is being examined by this grant with the hope to develop novel interventions that selectively interfere with the growth of these tumors. Specifically this proposal investigates the role of glutamate in the growth of astrocyte-derived tumors. Glutamate is the most abundant amino acid in the nervous system where it serves as a neurotransmitter, i.e. a signal between normal nerve cells in the brain. Glutamate signaling is important in all aspects of brain function, including learning and memory. However, it had also been demonstrated that excessive glutamate can kill nerve cells in a process termed excitotoxicity. Over the past years, we and others have found that tumor cells synthesize and release glutamate and also use it to kill nerve cells adjacent to the tumor probably in an effort to make room for tumor growth, which would otherwise be limited as brain tissue cannot be pushed back. We have identified the principle pathway by which glutamate is being released from gliomas cells as system Xc, a cystine-glutamate exchange protein. This exchanger is abundantly expressed by brain cells and serves to import the essential amino acid cystine which is needed for protein biosynthesis, cell growth and the production of glutathione, a cellular antioxidant that protects cells from reactive oxygen species. We have also been able to identify two chemicals that specifically inhibit this cystine-glutamate exchanger. Preliminary studies suggest that inhibition of cystine uptake causes tumor cells but not normal brain cells to die, most likely due to a depletion of cystine. In light of this we are proposing preclinical studies to demonstrate that gliomas can be targeted by drugs that inhibit cystine uptake via system Xc and show efficacy in animal models of the disease and elucidate the mechanism of cell death following treatment with such drugs. These studies should move us a step closer to a new clinical trial for the treatment of this devastating disease.
Terry A. Van Dyke, Ph.D. Astrocytomas include a varied group of tumors that can infiltrate normal brain. These cancers cannot be "cured" by surgery and are variably responsive to radiation and chemotherapies. There are no effective treatments for these cancers, and little is known about how they develop. Because these tumors often have mixed characteristics, it is not clear whether they develop in similar or distinct ways. In addition, since the brain represents a complex environment, we do not know what cell type(s) is(are) the target for cancer development or whether neighboring cells have a role. In order to understand these cancers well enough to develop effective therapies, we must be able to model these diseases in an experimental animal model. Currently the only model with all necessary tools is the lab mouse. Current technologies allow us to alter the genes of live animals such that we can engineer into the mouse the genetic events observed in the human cancers. These mice, called genetically engineered mice or GEM serve as a valuable resource for both understanding human diseases and for examining diagnostic and therapeutic tests. Recent significant advances in modeling astrocytomas in genetically engineered mice (GEM), including studies by Dr. Van Dyke, now provide an avenue for defining the molecular and cellular mechanisms underlying the pathogenesis of astrocytic cancers and to develop reasonable therapies for their treatment. In this proposal, we will further develop a mouse model of high-grade astrocytoma established in the lab utilizing a combination of GEM techniques and introduction of genes into the adult mouse brain by injections of viruses that can infect tumor cells or normal brain cells. Our goal is to closely model the human astrocytoma disease(s). |
 |
 |
