Tumor Neovascularization by a Non-angiogenic Mechanism: Plasticity-based Contributions by Brain Stem Cells.
New Faculty I
Every year, 2,040 people in California are afflicted by brain cancer. Glioblastoma multiforme, the most prevalent and aggressive form of brain cancer in adult, is believed to stimulate the formation of new blood vessels which are required for this cancer to progress to malignancy and metastasize to other tissues. New blood vessels are believed to be generated by angiogenesis, the process whereby pre-existing blood vessels form new vascular branches that supply oxygen- and nutrient-deprived tissue. Angiogenesis has classically been proposed to be the predominant mechanism of vascular remodeling in adults, putatively linking angiogenesis to the development of solid tumors, ischemia, autoimmune disorders and Alzheimer disease. Our research focuses on neural stem cells that were originally proposed to differentiate to only neuronal cell types. Unexpectedly, we have found that these stem cells can be diverted away from neural lineages and instead be induced to become blood vessels. Thus, neural stem cells are more flexible or "plastic" than previously believed, expanding the importance of the adult neural stem cell in maintaining the cellular composition and function of brain. Our goal is to test the potential of brain stem cells to promote tumor growth of by forming blood vessels. We will also develop methods to arrest the generation of blood vessels by these stem cells so as to block the tumor growth and expansion. The discovery and continued study of neural stem cell-mediated blood vessel formation could define the neural stem cell as a relevant target of vascular therapy.
Statement of Benefit to California:
Gliomas, the most prevalent brain cancer in adult 4, are dependent upon the circulatory system to acquire growth-promoting nutrients and to metastasize to breast, lung, colon and skin. Despite the latest therapies directed against the vasculature of grade IV gliomas, an 85% recurrence rate persists 1, as does the dismal median survival of <1.5 years 4 (2). The latest therapies failed, eliciting significant side effects because they target factors common to normal and cancer blood vessel formation. We have obtained evidence that a distinct, brain stem cell-mediated pathway supplies gliomas, but not healthy brain tissue, with blood vessels. The ultimate goal of our research effort is to develop novel therapies which acutely target this stem cell to specifically block the formation and maintenance of brain tumor capillaries.
SYNOPSIS: The ultimate goal of this research effort is to develop novel therapies targeted against formation of aberrant capillaries that sustain brain tumors to understand the cellular and molecular mechanisms by which vascular networks are maintained and remodeled in normal versus cancerous brain. As a postdoctoral fellow with Fred (Rusty) Gage at the Salk Institute, the applicant discovered that neural stem cells (NSC), which were originally proposed to differentiate only into neurons and glia, can differentiate into pericytes and endothelial cells (EC), a process requiring the mixed coculture of NSCs with ECs to mimic the presence of vascular niche endothelium. The ability of stem cells from one tissue type to differentiate into cells of a completely different tissue type (called "stem cell plasticity") has been contended and a broad body of literature attributes these events to cell fusion artifacts. However, the applicant and his postdoctoral mentor took great pains in their study to preclude cell fusion events. Their article (published in Nature) makes a persuasive case for cell fusion-independent differentiation of neural stem cells to the endothelial lineage. Preliminary evidence from the applicant also indicates that glioma progenitor cells (immortalized rat C6 glioma cells) can convert to EC-marker expressing cells. However, unlike the NSCs, a hypoxic microenvironment typical of the tumor is required to drive this process. The research proposed here will test the cancer stem cell hypothesis for high grade malignant gliomas in humans, which are highly vascularized tumors. The applicant wishes to explore the notion that the endothelial cells within malignant gliomas are actually derived from glioma stem cells. He also wishes to explore the idea that the pathway from neural stem cell to endothelial cell can be traveled in both directions. In the first aim, the applicant will characterize the stem cell niche that triggers the non-angiogenic ‘conversion’ of neural progenitor cells to vascular (EC and pericyte) cells. The main goal of the experiments is to show that the conversion of C6 cells to endothelial cells is not a rat/human cell fusion event. He will isolate C6-derived EC like cells and test for the presence of human chromosomes and ribonuclear proteins. The applicant would also like to define signaling pathways and “molecular switches” that modulate the conversion event from the C6 line. For the "molecular switch" line of work the applicant will take both a directed and an unbiased approach. The directed approach will focus on the role of hypoxia inducible factors (HIF1 alpha and HIF2 alpha). An unbiased approach will use expression profiling of cells cultured with normal oxygen and low oxygen to detect potential candidate genes. In the second aim, the PI will determine if glioma-derived cancer stem cells contribute to pathologic vascularization of gliomas. The applicant's prior work as a postdoctoral fellow suggests that the blood vessels within brain tumors are derived by differentiation of normal stem cells recruited into the tumor milieu. However, hypoxia - a generic interstitial feature of malignant glioma - would be predicted to suppress formation of blood vessel cells from normal neural stem cells. This leads the applicant to speculate that glioma stem cells are the sources of tumor neovascularization. Using a clonal analysis he will define and characterize the subfraction(s) of mouse and human glioma cancer stem cells (CSC) that bring about the cell types and invasive properties of gliomas and determine if these CSCs maintain vasculogenic potential. Mouse glioma models, in which avian sarcoma retroviral vectors target Nestin+ or GFAP+ cells with Ras/Akt or middle T antigen and human glioma tissue will be isolated in collaboration with physicians of Children's Hospital Oakland Research Institute (CHORI). Single-cell FACS populations will be used. Those clones that exhibit CSC characteristics will be screened for vascular (and neural) differentiation potential under hypoxic and normoxic conditions. The conversion of hippocampus transplanted CSCs to ECs and pericytes that incorporate into the vasculature of the tumor mass will be quantified at different time points In Specific Aim 3, the applicant asks if the relationship between neural stem cells and endothelial cells is symmetrical. The PI will determine whether ECs can generate NSC under the right environment and whether transforming mutations of the EC-to-NSC pathway can lead instead to cancer stem cell formation. Preliminary data indicates that endothelial cells derived from the brain can be induced to grow as neurospheres and to express neural stem cell marker proteins such as Sox2 and nestin. The PI proposes a broad-based attack to address a number of questions that are raised by these preliminary results. Fate mapping experiments, expression profiling and other tactics will be used to determine whether endothelial cells can be differentiated into stem cells in vivo and whether they can be targeted by oncogenic transformation to differentiate into glioma stem cells. Other experiments to identify extracellular and interstellar effectors of endothelial cell dedifferentiation are also proposed. STRENGTHS AND WEAKNESSES OF THE RESEARCH PLAN: Vascularization is a determinant of the growth and metastasis of solid tumors. Induction of vascular network remodeling by angiogenesis is thought to be the major mechanism for tumor vascularization. The studies seem to be more directly relevant to cancer biology than regenerative medicine, although plasticity of stem cells is of general interest to stem cell science. The strength of this research plan is the interesting scientific premise that the vasculature of malignant gliomas may actually be derived from glioma stem cells rather than from conventional angiogenic mechansims. This notion - if validated by experimental test - would have significant impact on anti-angiogenic strategies for the treatment of malignant glioma. That said, there are some significant flaws in the experimental design that diminish enthusiasm. Oxygen signaling is coming to be recognized as important for regulation of stem cell fate, but there are still considerable misunderstandings in the biology community about the normal oxygen concentrations at the tissue level in the body. This proposal suffers from some of these misconceptions, that may make the in vitro work difficult to interpret. In Aim 1 the PI states that ‘normoxia’ or atmospheric oxygen conditions are the stimulus for generation of EC, pericytes and smooth muscle cells from marked NSC co-cultured with human umbilical artery endothelial cells. Atmospheric oxygen is 21% and most incubators (because of the 5% CO2 taken up for buffering) are 20% oxygen. Arterial blood is 12% oxygen, therefore atmospheric oxygen is physiologically hyperoxic and produces measurable oxidant stress. The oxidant stress signaling is likely the operative issue here and this is not acknowledged perhaps because of the misconception of atmospheric oxygen being ‘normoxic’. (For reference, mean venous oxygen levels are about 5.3%, mean brain tissue oxygen levels across mammalian species normally are about 1.5%. In many parts of the brain 1% oxygen at the tissue level is not hypoxic.) This misconception affects numerous areas of the study plan. For example, in the array studies in Aim 1, the truly normoxic condition for gene expression is being missed. Instead the investigators are comparing oxidatively stressed cells (oxygen toxic) to oxygen deprived cells. Similarly, in Aim 3 the PI states that 1% and 40-80% hyperoxic conditions (an attempt to induce instant cell senescence?) did not induce ECs to dedifferentiate, but nothing like a physiologically normoxic condition is mentioned. Regarding the design and feasibility of the research plan, the scope of the studies proposed is unrealistically broad, and throughout the application it is very difficult to separate what has already been done from things that are proposed for the years ahead. A considerable amount of information that should have been presented in the preliminary results section is instead embedded within the study plan. Preliminary results did not have figure legends, and thus the significance of results is difficult to judge. The study plan is presented rather superficially as very little attention has been paid to controls, potential pitfalls and alternative approaches. Another concern is the reliance on undergraduate researchers for critical procedures to be executed without a clear plan by the PI for supervision. Much of the work proposed involves the C6 rat glioma cell line. This line is not well accepted within the neuro-oncology community as a valid model of human glioma, and the experiments in Aim 1 suffer from choice of this cell line. There may be considerable heterogeneity in human glioma stem cells as there are in virtually all other cancer stem cells. Even if the rat glioma cells fuse, does that mean that human cells will? The research would be more compelling if the aims were reversed, and various human stem cell clones isolated and tested for the potential to generate vascular cells. Similarly, the rat glioma cells may be responsive to oxygen conditions in a way that is different from human cells depending on the nature of the mutations that led to transformation. Finally, the PI makes the point that plasticity must be distinguished from fusion, but does not discuss transdifferentiation. As tissue sources for Aim 2, the applicant will use commercially available mouse glioma models (the RCAS system) and human glioma tissue obtained from collaborators at CHORI. The human studies should be emphasized if the PI has access to tumor samples, and the mouse work minimized, because the heterogeneity of human glioma CSC are the interest. Although the tissue sources for the study seem well thought out, the rest of Aim 2 requires more detail. How will the applicant identify glioma stem cell subtypes that drive vascularization? The strategy for isolating the CSC clones is also unclear. The PI does not state whether cells will be put at a clonal density after the percoll gradient or after passaging later. Would the search for a glioma CSC be helped by a sort after the percoll gradient utilizing a stem cell marker? While single cell sorting experiments are described, the markers used for cell sorting are not identified. It is also unlikely that the products of FACS sorting experiments will retain viability and be tumorigenic when implanted into the cranium of animal hosts. Practical experience from a number of laboratories suggests that these studies will be very difficult. In Aim 3, the applicant argues that the origins of adult neural stem cells are poorly understood, although some would contend this. He proposes that vascular ECs may represent a cellular precursor for progenitor cells of the CNS, and he has enriched single-cell/clonal preparations of primary brain ECs that maintain the marker expression profile of ECs. The PI has made the most interesting observation that culture of ECs in serum-free, basal neural culture conditions (SFNSC) and a Permanox (PX) substratum reproducibly converted ECs from strictly monolayer growth to neurosphere-growth within 5 days. These neurospheres maintain proliferation, lose the expression of EC markers and upregulate the stem cell markers Sox2 and Nestin. This indicates the possible acquisition of stem cell properties by ECs. Moreover, forskolin-initiated differentiation of EC-derived neurospheres downregulated Sox2 and Nestin and elicited neuronal morphologies and marker expression. Assuming that this system can be robustly repeated, the PI proposes both microarray analysis of this differentiation process (a better use of microarray technology than that proposed for Aim 1) as well as a focused look at integrin signaling. Is it possible that cell surface charge differences of the different plastics are the cause of the signaling? What else is known about the chemistry of the tissue culture plastics that would inform the design of the studies? If the Permanox were used in the chamber/slide set up, the thickness and curvature of plastic may be the important signal. It is hard to justify such an expensive foray into the differentiation mechanism without more scholarship in support of the research design. Additional experiments in Aim 3 are presented as a bit of a laundry list and the feasibility and likelihood of success of each experiment is not clear. Retrovirally-GFP-labeled NSC-like cells will also be engrafted into the hippocampus of adult mice to ascertain contributions to neuronal, astrocyte, ECs and pericyte lineages. The PI will also test whether EC-targeted oncogenic transformation induces EC dedifferentiation to glioma CSCs using conditional activation of Ras/Akt, EGFRvIII or ndPTEN transgenes in combination with EC-specific CreER Lentivirus. These EC oncogenic transformation studies would benefit from a control using NSC exposed to the same activated forms of oncogenes. The PI will also attempt to identify media-borne factors critical to the dedifferentiation of ECs described above. The identification of media constituents that differ between EC and NSC medium does not seem to take into account the different secretory profiles of the two cell types. In addition, he will try to define the role of PX in converting ECs to NSC-like cells and then define mediators of EC dedifferentiation by gene profiling. Finally, he proposes to use transgenic methodology to trace dedifferentiation of endogenous ECs. The idea is a good one (using transgenic mice expressing GFP-Cre from EC-specific promoters, such as those from VE-Cad, Flk1 or Tie1 to activate a reporter), but the proposed experiments are not ideal since Cre, rather than CreER, will be used. QUALIFICATIONS AND POTENTIAL OF THE PRINICIPAL INVESTIGATOR: The PI received a Bachelor's degree in Economics from Brown University in 1991. After a three-year hiatus as a research technician at Tufts Medical School he enrolled in the graduate program at UC San Diego and received his PhD in Cell and Molecular biology in 2001, working with Dr. Scot Emr on the role of Phosphatidyl inositol 3 kinase activity in the regulation of vacuolar size and membrane dynamics in yeast. He conducted postdoctoral research with Dr. Fred Gage at the Salk Institute from 2001-2005 where he discovered that adult neural stem cells possess the capacity to differentiate into the endothelial cells and smooth muscle cells that comprise the brain vasculature. In 2006 he joined the faculty of UC Berkeley where he now serves as Assistant Professor of Molecular and Cell Biology. The PI's dissertation research culminated in four first-author papers and these papers appeared in well-regarded journals, and his postdoctoral studies culminated in a pair of first author publications in top tier journals (Science and Nature). The applicant has given some thought to his career development. A key problem that he recognizes at UC Berkeley is translating basic science into clinical applications -- an important goal of stem cell research. To address this problem, he has established a key collaboration at Children's Hospital of Oakland Research Institute (CHORI). Ultimately he envisions a collaborative effort with CHORI to take the form of developing therapeutic approaches which through the deliberate up- or down- regulation of brain stem cell vascular potential will improve patient diagnosis. He states that research support from CIRM would constitute a critical first step in forging a meaningful interaction of CHORI research clinicians and UC Berkeley, bolstering stem cell programs at both institutions. The PI might also consider contacting neurosurgeons at UCSF for the possibility of obtaining more tumor explants for CSC isolation, and as potential collaborators. Also, other Bay Area hospitals are likely to provide collaborators. He attends a bimonthly neural stem cell research discussion and plays an integral part in converting a graduate level developmental biology course into a stem cell focused journal club forum. Randy Schekman, was appointed his faculty mentor. Another important relationship that he has forged is with Dr. David Schaffer at Berkeley's Department of Chemical Bioengineering. The identification of a specific mentor would strengthen the career plan. INSTITUTIONAL COMMITMENT TO PRINCIPAL INVESTIGATOR: UC Berkeley has a strong stem cell and biology environment. Professor Steven Martin, the Chairman of Molecular and Cell Biology at UC Berkeley, has written a letter documenting strong institutional commitment to the candidate. The applicant has received a generous startup package to retrofit his research laboratory and has the appropriate environment, equipment, and interactions around him for conduct of the research. He has been given a year's grace from teaching duties and assigned a faculty mentor to monitor his career development. A group of scientists at Berkeley has informally organized a neural stem cell interest group which is a good addition to the strong intellectual environment at Berkeley. The institutional track record of UC Berkeley in fostering the career development of its young faculty is exemplary. Recruitments in the area of stem cell biology are continuing with Henk Roelink being the latest recruit, and will be housed in the new Health Sciences building. DISCUSSION: The PI tackles the difficult and important clinical problem of gliomas by attempting to characterize the niche that supports the transition to malignant cells. One reviewer feels that the proposal has some charm. This is a strong investigator from a top-tier university who has done solid work, and there is some validity to the claim that this PI's transdifferentiation work came along when the term itself was still in dispute. The hypothesis is interesting, but unfortunately it suffers from a fatal flaw. In the studies as proposed, the PI will never attain a truly normoxic state. The stroke-in-a-dish studies have never yielded therapies because investigators used the hyperoxic control, and they never examined the dynamics of oxygen concentrations in the intermediate ranges which are comparable to physiologic oxygen states. Aims 1 and 3 are dependent on oxygen environment, and cancer stem cells often reside in the hypoxic core of a tumor. If the PI wants to isolate glioma cancer stem cells, he will not get them unless he creates physiologically relevant normoxic conditions. There was some concern that the PI was obtaining gliomas from a pediatric source, which might not yield sufficient number of cases to conduct the study. The PI should look at adult tumors for more samples. Also, the heterogeneity issues will be an enormous problem in assessing samples.