Candece Gladson, M.D. - US grants

DESCRIPTION (provided by applicant): We have shown previously that cooperation between ligated integrin avp3 and the PDGFr promotes the motility of glioblastoma cells. Here, we propose to analyze the signaling mechanisms associated with this response. Our preliminary data implicate the cellular Src family member, Lyn: (i) The activity of Lyn is higher in glioblastoma (Grade IV) tumor biopsies, as compared to anaplastic astrocytoma (Grade III) tumor biopsies and normal brain;(") Lyn is specifically activated by the cooperation of ligated integrin otvp3 and the PDGFr on glioblastoma cells, although Fyn is the predominant cellular Src family member expressed in these cells;and (Hi) Lyn is necessary for the promotion of migration associated with the cooperation between ligated integrin avp3 and the PDGFr. My coinvestigator, Dr. Dan Flynn, has shown recently that the N-terminus of a cellular Src family member can dictate specificity in signaling that functionally differentiates cellular Src family members (c-Src and c-Yes). Thus, we hypothesize that the N-terminus of Lyn dictates the specificity in signaling that differentiates Lyn from Fyn and that the elevated levels of Lyn found in vivo in glioma tumor biopsy samples promote the motility/invasion characteristics of these tumors, thereby contributing to slioma progression. We also find that the cooperation of ligated integrin avp3 and the PDGFr on glioblastoma cells results in increased phosphorylation of focal adhesion kinase (FAK) and of HEF1, a CAS family member. Thus, we hypothesize that Lyn promotes the motility/invasion characteristic of the glioblastoma tumors through a FAK/HEF1 signaling mechanism. Here, we will use stable transfection with Lyn, Lyn mutants, and Lyn/Fyn chimeras, as well as siRNA technology to: (1) Determine the domains in Lyn that are necessary for its specific activation by the cooperation of ligated integrin ctvp3 and the PDGFr;(2) Determine whether Lyn is necessary for malignant glial cell migration/invasion in vivo, and whether the amino-terminus of Lyn (SH4-Unique-SH3-SH2 or SH4-Unique domains) is required for this effect;and 3) Determine whether FAK and the downstream effector, HEF1, are necessary for migration of vitronectin-adherent (ligated integrin avp3) and PDGF-stimulated glioblastoma cells, and SYF mouse embryo fibroblasts transfected with Lyn. The results will be of general interest, i.e., tumors other than brain tumors, as increased cellular Src family member activity also likely promotes the progression and metastasis of non-glioma tumors.

DESCRIPTION (provided by applicant): The goals of this study are to understand the mechanism by which rK5 induces apoptosis of microvascular endothelial cells (MvEC) and of malignant glioma cells, the mechanism by which radiation sensitizes these cells to rK5, and to determine whether radiation sensitizes brain MvEC and tumor cells in intracerebral mouse models of malignant glioma to rK5-induced apoptosis. rK5 (also known as Abbott-828) is a recombinant form of the fifth kringle domain of plasminogen. Our preliminary data indicate that rK5, especially when used in conjunction with prior irradiation, has the potential to kill proliferating human brain MvEC, and to reduce the size of malignant glioma tumors in a mouse model. It appears to act, at least in part, through the low density lipoprotein receptor-related protein (LRPI)-mediated internalization of its cell surface receptor GRP78. We have confirmed that GRP78 is expressed on MvEC and glioma cells in human malignant astrocytic tumor biopsies (5 of 5). We hypothesize that irradiation will sensitize brain MvEC and malignant glioma cells to treatment with rK5 resulting in induction of apoptosis through a mechanism that requires GRP78 and LRP1, and that this strategy will inhibit the growth of these tumors in vivo. We will: (1) Determine the optimal conditions for rK5 killing of brain MvEC and glioma cells and identify the apoptotic pathways utilized using inhibitors and siRNA technology, determine whether the rK5 acts through GRP78 using blocking antibodies and siRNA, and determine the role of p38 MAP kinase signaling in the pro- apoptotic signaling through analysis of the phosphorylation status, small molecule inhibitors, and the effect of downregulation with siRNA. (2) Determine the role of LRP1-mediated internalization of GRP78 in the sensitization using siRNA, co-immunoprecipitation and immuno-electron microscopy, and the role of phosphorylation of the cytoplasmic tail of LRP1. (3) We will then determine the effect of rK5 treatment, plus/minus irradiation, on malignant glioma growth, invasion and angiogenesis in vivo in a syngeneic, immune-competent, intracerebral mouse model, and in a human xenograft intracerebral nude mouse model. Relevance: This work should lead directly to a novel therapeutic strategy for the treatment of malignant astrocytic tumors, which would represent a major advance in the treatment of this devastating disease.

DESCRIPTION (provided by applicant): Anti-angiogenic approaches that eliminate the neovasculature by inducing apoptosis would represent a significant advance in the treatment of glioblastomas. Tumor necrosis factor a (TNFa) can act to induce apoptosis of cultured primary human brain microvessel endothelial cells (MvEC) through a mechanism that requires expression of the TNF-receptor 1 (TNF-R1) on the MvEC. Immunohistochemical analysis of biopsies indicates that, in most patients, the expression of TNF-R1 and TNFa is significantly higher in the glioblastoma tumor endothelial cells as compared to the normal brain endothelial cells and the levels of tumor-associated angiogenesis in tumors developed by injection and propagation of mouse malignant glioma cells in the white matter of the mouse brain is significantly higher in TNF-R1-null mice than in their wild-type counterparts. Based on these and other data, we hypothesize that the upregulated expression of TNF-R1 on brain endothelial cells associated with malignant glioma tumors is a host anti-angiogenic response to the tumor, and that TNFa therapy targeted to tumor endothelial cells will inhibit tumor angiogenesis and tumor growth. We propose to test these hypotheses by identifying the cell surface signaling events that elicit, and regulate, TNF-R1-mediated apoptosis in glioblastoma MvECs. In parallel, we will establish the feasibility of therapeutic manipulation of TNF-R1 with a TNFa fusion protein that is targeted to tumor MvECs by fusion with a peptide that binds CD13. We will use two mouse models to analyze the specificity of the effects and the magnitude of the responses in vivo: an immune competent mouse model of glioblastoma and a xenograft model based on the use of human glioblastoma stem cells. We will: (1) Establish whether TNF-R1 is preferentially expressed in the brain tumor MvEC and is colocalized with molecules that may regulate its ability to signal apoptosis, using biopsies from patients with glioblastoma and normal brain;(2) Determine whether TNF-R1 functions as an anti-angiogenic molecule in the brain in response to a malignant glioma tumor and establish whether TNF- R2 contributes to, or modulates, this effect using TNF-R1-null, TNF1-null, and TNF-R2-null mice;(3) Determine whether the activation state or expression of integrin av[unreadable]3 on the brain MvEC modulates the response of these cells to the pro-death signaling of TNF1;and (4) Test the ability of a TNFa fusion protein targeted to CD13 on tumor endothelial cells with the Cys-Asn-Gly-Arg-Cys peptide to inhibit tumor angiogenesis and tumor growth, and to promote survival, in vivo. RELEVANCE: The results should identify a novel anti-angiogenic therapy that can be used in conjunction with other therapies to more effectively eliminate malignant glioma tumors and prevent their recurrence. The studies also will provide data concerning biomarkers that may be used to predict which glioblastoma patients may benefit from this strategy and biomarkers for non-invasive monitoring of its efficacy. PUBLIC HEALTH RELEVANCE: The survival of patients with glioblastoma tumors is dismal (15-18 month median survival) despite all current therapy. We propose to test a novel therapeutic strategy to preferentially kill the endothelial cells (MvECs) in the newly formed blood vessels that feed a glioblastoma tumor by targeting tumor necrosis factor a (TNFa) to these blood vessels through a peptide (NGR) that binds CD13. CD13 is upregulated on tumor endothelial cells. Importantly, we will include innovative analyses of the molecules that may regulate the responsiveness of the MvECs. This approach would be a significant advance over the therapies currently being tested that retard the growth of the blood vessels, but do not eliminate them.

DESCRIPTION (provided by applicant): Angiogenesis is a prominent characteristic of glioblastoma tumors (GBM) and is considered a promising therapeutic target. Currently available anti-VEGF therapy does not change overall survival, however, and the issue of potential differences in angiogenesis among the molecular subtypes of GBM has not been addressed. Our preliminary data support the concept that glioma stem cells (GSCs) play a key role in GBM-associated angiogenesis and they do so, at least in part, by enhancing the migration of endothelial cells (ECs). ECs isolated from GBM exhibit greatly enhanced migration as compared to ECs isolated from normal brain and this heighted motility is promoted by GSC-conditioned media (GSC-CM). The effects of GSC-CM on the ECs appear to be associated, in part, with upregulation of the pro-angiogenic molecule ephrin-B2 on the ECs based on array analysis and blocking studies using recombinant ephrin-B2-Fc and EphB4-Fc (the ephrin-B2 receptor). This suggests that factors secreted by GSCs, including VEGF, act by upregulating ephrin-B2. In addition, the data indicate that a direct interaction between GSCs and ECs further promotes EC motility and that this interaction is mediated, in part, by an interaction between integrin ¿v¿3 on ECs and L1 cell adhesion molecule (L1CAM) on GSCs. Notably, as ephrin-B2 promotes the activation and signaling of VEGFR2 and as VEGFR2 and integrin ¿v¿3 reciprocally activate each other, these two mechanisms could intersect resulting in ephrin-B2 indirectly activating integrin ¿v¿3. Thus, we propose a hypothetical model in which angiogenesis in GBM is promoted by increased EC motility that is driven by GSC-secreted factors that upregulate pro- angiogenic molecules on EC and that this mechanism enhances the effects of a direct interaction between L1CAM on the GSCs and integrin ¿v¿3 on the ECs. This hypothesis will be tested using GSCs and ECs isolated from GBM of the mesenchymal and pro-neural subtypes using recombinant proteins, antibodies and downregulation approaches to: (1) Determine whether GSC-secreted factors (including VEGF) promote EC migration, proliferation and tubulomorphogenesis and expression of pro-angiogenic molecules (including ephrin-B2 and EphB4) on ECs. (2) Determine whether the interaction of integrin ¿v¿3 on ECs with L1CAM on GSCs transmits a signal promoting EC migration, proliferation and/or tubulomorphogenesis as determined by co-culture of ECs and GSCs; stimulation with GSC-CM; and injection of ECs and GSCs into brain slices followed by 2-Photon Laser Scanning microscopy and individual cell motility tracking. (3) The mechanism will be tested in vivo by determining if the absence/blocking of integrin ¿v¿3 or ephrin-B2 on ECs and L1CAM on GSCs alters angiogenesis, tumor volume and/or survival using PDGF-B-driven GFP-GBM cells or shL1CAM- GFP-GBM cells propagated in immunocompetent integrin ¿3-mutant or control mouse brain. RELEVANCE: These studies have the potential to provide a paradigm shift in the understanding of tumor-associated angiogenesis in general and will suggest novel regimens for improved anti-angiogenesis therapy in GBM.

SUMMARY: Glioblastoma (GBM) is an aggressive cancer with a dismal prognosis (median survival ~15 mo). Although antibody-drug conjugates (ADC) hold great promise, this approach is unexpectedly ineffective in a large number of patients. Genomic analyses have identified alterations in multiple genes that regulate endocytic pathways in GBM tumors but the possibility that reconfiguration of the endocytic pathways can compromise the efficacy of ADC therapy has not been explored. The goal of the proposed study is to test the hypothesis that tumor-associated reconfiguration of endocytic trafficking in perivascular cancer stem-like cells (CSLCs) influences the outcome of ADC therapies in GBM. To test this hypothesis, we are utilizing a highly collaborative transdisciplinary approach and focusing on ABT-414 as the prototypic ADC. ABT-414 is a fully humanized IgG1 mAb directed toward EGFRvIII and amplified-wt-EGFR that is conjugated to monomethyl auristatin F. It is effective in ~50% of selected patients with recurrent GBM. We have verified that GBM CSLCs are sensitive to ABT-414 and assembled a well-characterized panel of GBM patient-derived xenografts (PDX) that show a range of sensitivities to the therapy in vitro and in vivo. We also have verified, in vitro and in vivo, that CSLCs in the perivascular space in GBM internalize ABT-414, and that it is trafficked by endocytic pathways to recycling compartments as well as to the lysosome, where release of the warhead occurs. We propose to manipulate two genes, RBSN and EHD3, that are associated with recycling and exhibit dysregulated expression in a proportion of GBM tumors. Preliminary data indicate that alterations in expression of these genes in CSLCs can reduce diversion of ABT-414 to the recycling compartments and enhance CSLC killing. We propose two Specific Aims in which we will determine: (1) The effects of manipulation of RBSN in GBM-derived CSLCs on the trafficking fate of ABT-414 and CSLC survival in vitro after deletion/downregulation of RBSN; blockade of endosome maturation using shVps39; and expression of tagged mutant-RBSN. The effects of RBSN deletion downregulation, and the impact on ABT-414 therapeutic efficacy, will be tested in vivo using an orthotopic GBM PDX model. (2) The effects of manipulation of EHD3 in GBM-derived CSLCs on the trafficking fate of ABT-414 and CSLC survival in vitro after deletion or overexpression of EHD3, and expression of mutant EHD3. In parallel, we will determine if hypermethylation of the EHD3 gene promoter represses EHD3 expression in GBM biopsies. The effects of overexpression of EHD3 or mutant-EHD3, and their impact on ABT-414 therapeutic efficacy, will be tested in vivo using an orthotopic GBM PDX model.