Gabriele Bergers, PhD
Co-Principal Investigator and Leader of Project 1
Introduction to Project 1
The vascular stem cell niche is a microanatomical unit consisting of distinct cell types and matrices that regulate proliferation, fate specification, and protection of normal neural stem cells (NSCs). The intimate association between normal NSCs and endothelial cells has been reported to regulate self-renewal and differentiation of normal NSCs. Moreover, blood vessels are used as migratory scaffolds for neural progenitors. A vascular niche also exists in various tumors, including gliomas, in which tumor cells and tumor initiating cells (TICs) are nestled around tumor vessels. R. Gilbertson and J. Rich's groups also observed a bidirectional communication between vascular cells and TICs as CD133+ tumor cells exhibited a five-fold increase in growth, maintenance of aggregated tumor spheres, and self-renewal capacity. One might therefore reason that ablation of tumor vessels should have beneficial effects in GBM by eliminating the vascular niche. On the contrary, the survival benefits of the anti-angiogenic drug bevacizumab are short-lived and followed by aggressive tumor relapse. This could be partly because although anti-VEGF therapy is known to prevent the growth of new blood vessels, it is less efficient in tackling the established tumor vasculature.
Furthermore, while the description of the vascular niche in tumors has so far been restricted to the endothelial-TIC interaction, it is important to note that blood vessels recruit inflammatory and circulating cells to the niche. In contrast to normal vessels, GBM vessels lose the appropriate balance of pro-and anti-angiogenic signals, partly due to the overproduction of VEGF, resulting in tortuous, dilated vessels that impair blood flow and generate hypoxic areas. Hypoxia via induction of hypoxia-inducible transcription factor 1 triggers an influx of proangiogenic bone marrow-derived innate myeloid cells that both stimulate new blood vessel growth and maintain TSC due to their expression of factors known to regulate stem cell self-renewal, proliferation, differentiation and migration.
GBM recurrence following antiangiogenic therapy can also be due to the vascular niche providing signaling cues that attract tumor cells to blood vessels and induce an epithelial-to-mesenchymal (EMT) acquisition, which is enhanced during anti-VEGF therapy. This renders tumors non-responsive to the therapy and provokes an aggressive, proinvasive recurrence pattern. Our group has demonstrated that VEGF inhibits c-Met signaling through the formation of a VEGFR2:c-Met heterodimer complex and that c-Met downregulation in VEGF-deficient GBMs, which exhibit severe perivascular invasiveness, abrogated the invasive phenotype. We also show that T-cadherin is high in clustering cells in which c-Met is downregulated but low in those with high c-Met activity. Moreover, while HGF inhibits T-Cadherin expression in VEGFko-GBM cells but not in VEGFko-shMet cells, HGF upregulates the mesenchymal N-cadherin in tumor cells, supporting the proposition that HGF promotes an EMT-like acquisition leading to a T- to N-cadherin switch in astrocytoma.
In addition to HGF signaling, enhanced ECM stiffness and compression in GBM can augment and potentiate secretion and signaling efficacies of various cytokines and growth factors in the vascular niche that enable EMT acquisition in GBM cells. Compression favors highly flexible, contractile cells with reduced mechano-responsiveness because these traits permit the cells to expand and survive within a confined space andnavigate through the dense ECM, features that are shared by tumor cells that undergo EMT.
In this project we intend to test two major hypotheses regarding the regulation and function of the tumor vascular niche in GBM.
Hypothesis 1: We hypothesize that innate immune cells, recruited to the vascular niche by the tumor, support tumor stem cell maintenance and propagation. We further speculate that disruption of the vascular niche further enhances hypoxia-regulated recruitment of innate immune cells, which in turn will promote neovascularization and provide survival and growth factors for tumor stem cells that make them less dependent of the vascular niche. We also argue that anti-VEGF therapy does not ablate the majority of tumor vessels but, rather, prunes freshly growing vessels, leaving established "normalized" tumor vessels with enhanced pericyte coverage behind that are non-responsive to antiangiogenic therapy and provide sufficient niches for tumor stem cells.
Hypothesis 2: We propose that the tumor vascular niche in part due to elevated compression provides signaling cues that can promote an epithelial-mesenchymal acquisition in GBM eliciting mesenchymal and stem cell-like properties in tumor cells that also maintain tumor cells in close proximity to blood vessels. In addition we hypothesize that vascular disruption due to antiangiogenic therapy or vascular compression due to high cellularity, force and matrix stiffness can augment signaling cues in the vascular niche that endorse an EMT-like phenotype in tumor cells. This would further suggest that more committed glioma cells could dedifferentiate to a more stem-like state.
AIM 1. Test the hypothesis that perivascular innate immune cells in the vascular niche are critical cell constituents that regulate angiogenesis and tumor stem cell maintenance in gliomas.
We will conduct an extensive analysis of the myeloid cell populations in fresh human GBM and oligodendroglioma tissues, human GBM xenografts of the mesenchymal and proneural subtypes, and high-grade oligodendroglioma xenografts. We will obtain tumor tissues from Dr. Mitchel Berger, Chair of Neurosurgery, as overseen by Dr. Joanna Phillips, Co-Director of the Tissue Core. We will take half of the tumor for FACS analysis and cell sorting to identify and isolate the cells and reserve the remainder for tissue analysis. These studies will elicit important information about the prevalence, location and composition of myeloid cells in GBM and oligodendrogliomas and whether the distinct GBM subtypes differ in their myeloid influx and composition.
In collaboration with Drs. Persson and Weiss (Project 2), we will also test the hypothesis that infiltrating myeloid cells, like endothelial cells, secrete factors in a paracrine fashion that have the capability to mediate GBM stem cell maintenance and that they might be most efficient when primed in VDA- or anti-VEGF treated tumors with a compromised vascular niche. We will use a neurosphere assay to test different GBM stem cell populations for self-renewal capacity assay. To test the ability of myeloid cells to promote GBM stem cell proliferation and migration, we will perform co-culture experiments with cells plated under stem cell conditions in the absence or presence of myeloid cells from untreated or treated tumors. We will then determine potential candidates that drive neovascularization and stem cell renewal by using cytokine/chemokine/angiogenesis PCR arrays.
We will subsequently treat glioma-bearing mice with antiangiogenic and VDA agents alone and in combination with inhibitors that block myeloid cell recruitment or function. Besides standard histopathology and survival analysis, we will investigate the remaining GBM stem cell fraction in tumors by FACS and immunohistochemistry, as well astheir self renewal capacity. In addition, congruent with recent that GBM stem cells can differentiate into endothelial cells, we will test whether vascular disruption provokes signaling cues that induce GBM stem cell differentiation into endothelial cells
AIM 2. Test the hypothesis that the vascular niche is a mediator of epithelial-mesenchymal transition (EMT) in GBM evoking invasiveness and stem-cell-like properties.
In Aim 2 we intend to decipher vascular-niche produced factors besides HGF that can promote an EMT-like acquisition. Transforming growth factor-b (TGFb) has become a highly relevant candidate, as it is elevated in mesenchymal GBM, produced by monocytes and brain microvascular endothelial cells, and functions as a potent inducer of EMT. We will assess whether TGFb induces EMT features in TGFbR I and II-positive GBM cells by testing expression of various EMT factors. In addition, we will isolate fractionated endothelial cells and myeloid cells from untreated and anti-VEGF and VDA-treated GBM and test whether co-cultures with GBM cells induces or increases EMT signatures.
Secondly, we will test the proposition that enhanced ECM stiffness and compression in GBM can augment and potentiate secretion and signaling efficacies of various cytokines and growth factors in the vascular niche that enable EMT acquisition in GBM cells. Dr. Weaver's laboratory will assist in assaying the effect of ECM stiffness and compression force on EMT acquisition. We will then monitor levels of EMT factors under compression with EMT PCR arrays and validate the results by qPCR and protein expression.
Thirdly, we will assess the hypothesis that microenvironment-regulated EMT generates tumor cells with stem cell properties by functionally assessing whether EMT transcription factors can promote such properties in GBM. We will do this by conditionally inducing the EMT transcription factors Snail and Twist in human GBM xenografts using tamoxifen-inducible constructs. We will then assess changes in expression of EMT factors and the self-renewal capacity of the cells in tumor sphere assays. We will assess the behavior of these GBM cells upon activation of Snail and Twist in vivo by implanting the cells intracranially into mice. We will determine whether Twist or Snail is sufficient to promote an EMT-like phenotype and assess EMT factors, focusing particularly on the invasive behavior of the cells. Our preliminary data also revealed that anti-VEGF therapy increases EMT features in GBM, suggesting that anti-angiogenic therapy might negatively affect the tumor stem cell compartment but provoke more EMT and stem-cell-like features in the non-stem cell tumor population. We will therefore critically assess the stem cell traits in anti-VEGF and VDA-treated GBM and compare them to non-treated controls using a stem cell PCR array.