Nancy Boudreau - US grants

homeobox genes; developmental genetics; extracellular matrix; angiogenesis; cell migration; vascular endothelial growth factors; integrins; gene induction /repression; immunoprecipitation; chick embryo;

Brain arterial vascular malformations arise through a disruption in normal vascular homeostasis and excessive angiogenesis. Because normal angiogenesis requires that expression of a number of genes be temporally and spatially coordinated, we are investigating how the Homeobox (Hox) master transcription factor family of genes coordinates the expression of angiogenic or quiescent phenotypes in endothelial cells. We have identified Hox genes, which promote angiogenesis and those which attenuate angiogenesis and consequently favor maintenance of a quiescent differentiated state. Furthermore, the Hox genes impact the endothelial cell phenotype by coordinately modulating expression of extracellular matrix (ECM) components, their integdn receptors, as well as ECM degrading proteinases and their inhibitors. We hypothesize that an imbalance in pro-angiogenic Hox genes and angiostatic Hox genes contribute to the development of BAVMs and that restoring this balance can correct the aberrant vascular morphology and excessive angiogenesis associated with this disease. We will establish the patterns of expression of three Hox genes: Hox D3, Hox D10 and Hox A5 in normal and BAVM tissue. Secondly we will attempt to block cerebral angiogenesis in vivo by restoring expression or the angiostatic Hox D10 gene or attenuating expression of the pro-angiogenic Hox D3 gene. We will also explore whether Hox A5, known to inhibit proliferation of epithelial cells, can also inhibit endothelial cell growth and consequently angiogenesis. We will also investigate whether the downstream targets of Hox A5 are also genes involved in remodeling of the extracellular matrix.

in the preliminary data in the last application showing the impact of HoxAS on vascular stability and growth. This work was done using endothelioma cells (bEND and EOMA) in cutaneous tissue.2 Further, we have now extended these studies to examine the impact of restoring HoxAS on growth of transformed endothelial cells in the mouse brain. A manuscript has just been submitted and these data are presented in Section 3.1b (Figs. 4- 7). The purpose of the work was to provide proof-of-concept that proliferative vascular tissue in the brain could be controlled by manipulation of HoxAS. Progress was hampered in part because our first choice of endothelioma cell type (bEND) did not reproducibly form lesions in vivo. The murine EOMA endothelioma cell line formed lesions similar to those induced by the bEND cells, but was phenotypically stable. We subsequently showed that restoring HoxAS could also limit growth of the EOMA tumors in the brain, and this was accompanied by a reduction in the HoxAS target Hifla and an increase in the anti-angiogenic TSP-2 gene. In addition we are currently conducting a clinical study to screen a large number of samples from the AVM tissue bank to document the prevalence of HoxAS dysregulation in a wide spectrum of AVM cases; this is part of Aim1. We previously showed preliminary ISH data (Fig.1) and have now performed this analysis on 7 AVM patient samples. A challenge which has slowed these studies is that our OCT frozen tissue sections are difficult to section due to the complex angio-architecture, thus making quantitative or semi-quantitative analysis by ISH difficult. Moreover, there is currently only one commercially available antibody against HoxAS, and while we have attempted to use this for staining cultured cells and also attempted different antigen retrieval and processing protocols for tissue samples, we have not been successful. I have also consulted with other investigators studying HoxAS who have not been able to apply this antibody for staining. We have now successfully used this for Western blot of tissue lysates (see Fig.8) and can apply this to human BAVM tissue specimens to confirm that HoxAS protein as well as mRNA is reduced. Nonetheless, we have now included new data to show that we can detect changes in HoxAS protein by performing Western blot on tissue lysates (Sec. 3.1, Fig.8) to confirm that changes in HoxAS mRNA are reflected at the level of protein. Importantly, to more accurately quantitate differences in HoxAS expression, we have now adapted a tissue micro-dissection technique that we developed for cutaneous hemangioma tissue to perform real time PCR analysis. We microscopically removed as much of the adjacent connective tissue and extract RNA from the remaining tissue, which is enriched in aberrant vessels. Real time PCR on this tissue has confirmed our ISH results: HoxAS expression is lacking in the vascular nidus of the AVM (Fig.1). We anticipate that we can now complete these experiments within the next several months as additional samples become available from Core C. All reviewers raised concerns regarding the number of different models proposed, the lack of justification for these models, and the paucity of preliminary data to support their use or their relevance to the human disease. Dr. Young has addressed the use of our model systems in the General Introduction. Moreover, the aims in Project 3 have now been revised and simplified to use only one general model murine angiogenesis and vascular dysplasia which has been well characterized by our group. 1. To first investigate whether HoxAS can impact cerebral angiogenesis, we will use focal VEGF stimulation using AAV-VEGF direct injection into the brain of transgenic mice which selectively expresses HoxAS in PHS 398/2590 (Rev.11/07) Page 266 Continuation Format Page

DESCRIPTION: Breast cancer is the most common cancer and second leading cause of death in women. Neoadjuvant chemotherapy is increasingly used to "shrink" tumors prior to surgery and enable breast conservative approaches;however, long-term survival remains poor, in part due to factors that limit delivery of cytotoxic drugs to tumor tissue. Abnormal tumor blood vessels and altered transendothelial permeability and increased interstitial fluid pressures (IFP), conspire to limit delivery of macromolecular cytotoxic drugs. Thus, approaches that aim to alter tumor vessel hemodynamics and vascular permeability would effectively increase tissue accumulation of chemotherapeutic agents by overcoming high IFP and promoting convection driven uptake of large macromolecular agents into tissues. To this end, we have discovered a novel, endogenous pathway regulating vascular permeability that remains functional in tumor vessels. Whereas transforming growth factor beta 1 (TGF?1) restricts normal vascular permeability, inhibition of the type I TGF??receptor Alk5 expressed in vascular cells, enhances vascular permeability in normal as well as tumor vasculature (Sounni et al. manuscript submitted). Thus, we propose to administer an Alk5 inhibitor, in combination with macromolecular chemotherapeutic agents to improve breast tumor perfusion and accumulation of conventional chemotherapeutic agents. Paradoxically, while increases in tumor perfusion and improved penetration of cytotoxic agents have also been achieved by 'normalization'of tumor vasculature by blocking vascular endothelial growth factor (VEGF) to reduce vascular permeability, these effects are transient and largely limited to immature vessels more frequently associated with early stage tumors. As tumors progress, tumor vessels mature and become refractory to VEGF blockade. Importantly, we have also observed that TGFb1-mediated vascular stabilization remains functional in more mature tumor vessels, and further predict that Alk5 blockade will improve cytotoxic drug penetration in both early as well as late stage tumors. We will compare accumulation of Doxil in mammary tumor-bearing mice treated with Alk5 inhibitor as opposed to those treated with anti-VEGF antibody (DC101), and assess vascular permeability and accumulation of Doxil in both early and late stage tumors in MMTV-PymT mice. Moreover, while previous studies have used MR imaging data to asses breast cancer responses after cytotoxic drug administration, we will demonstrate that MR imaging of macromolecular contrast media (MMCM) in the presence or absence of Alk5 blockade correlates with Doxil accumulation and thus can be used to predict macromolecular drug distribution and to identify tumors most likely to benefit from cytotoxics combined with agents that improve vascular permeability. In addition, we will assess the anti-tumor impact of improved Doxil accumulation following Alk5 blockade by monitoring histopathologic characteristics of mammary adenocarcinomas, as well as tumor burden, tumor latency (to endpoint), and frequency of pulmonary metastasis, and further demonstrate that MR imaging based predictions of drug accumulation also correlates with therapeutic responses. We will also examine responses of orthotopically implanted human breast tumor cells in mice treated with Alk5 inhibitor plus Doxil versus Doxil as monotherapy, to demonstrate enhanced efficacy of Doxil in human tumor cell killing. Together these studies will establish the effectiveness of 1) improved cytotoxic drug accumulations in the presence and absence of ALK5 inhibitors at different tumor stages, 2) MRI predictions of tumor microvascular permeability and cytotoxic drug accumulation, and 3) enhanced tumor responses and reduced cytotoxicity from improved chemotherapeutic drug accumulation in tumors. Thus, by utilizing new predictive MRI correlations and exploiting a novel endogenous pathway regulating vascular permeability that remains functional in breast cancer, the delivery of chemotherapeutic agents and subsequent response of both early and late stage breast tumors will be radically improved. NARRATIVE: The major goal of our project is to examine whether specific short-term inhibition of ALK5 in vivo alters hemodynamics and tissue perfusion of mammary adenocarcinomas such that delivery of Molecular Resonance (MR) imaging compounds is improved and/or delivery of chemotherapeutic agents is enhanced, thus providing a survival advantage. Realization of this goal will improve diagnostic imaging and drug delivery into tissues containing vasculature that limits efficient tissue perfusion. The ability to transiently alter vessel stability and "open" vascular beds to facilitate intravenous or potentially intra-organ delivery of diagnostic or therapeutic agents would represent a significant advance in disease therapy and/or diagnostic imaging. SPECIAL REVIEW NOTE: In order to conform to the scientific objectives outlined in the program announcement RFA-GM-09-008, EUREKA applications submitted to the NCI were initially evaluated by a group of reviewers representing diverse scientific interests. The priority score reflects the average of all the scores given by the full committee after a thorough discussion.

DESCRIPTION (provided by applicant): Tumor growth and progression requires development of a new vascular supply. Many current anti-angiogenic therapies target VEGF, but targeting VEGF alone is not universally effective in combating angiogenesis induced by tumors in different organs or at various stages of the disease. Thus more global approaches to inhibit angiogenesis are needed. My laboratory has defined roles for the homeobox (Hox) master transcription factors, HoxD3, HoxA3 and HoxB3 in coordinately regulating endothelial cell (EC) invasion, proliferation, and adhesion during angiogenesis. In addition we have also demonstrated that other Hox factors, namely HoxD10 and HoxA5 act in a dominant manner to maintain cultured EC in a quiescent, differentiated state and resist angiogenic stimuli. To explore the anti-angiogenic and anti-tumorigenic potential of Hox genes in vivo, we will focus on HoxA5, which is highly expressed in quiescent differentiated EC, but expression is reduced in activated tumor endothelium. Ectopic expression of HoxA5 in cultured EC impairs migration, down-regulates several pro-angiogenic genes including VEGFR2, ephrin A1 and IL-6 and significantly increases levels of anti-angiogenic factors including Thrombospondin-2 (TSP-2). HoxA5 also stabilizes EC adherens junctions and reduces endothelial cell permeability. Our preliminary evidence suggests that EC HoxA5 may also suppress recruitment/infiltration of pro-angiogenic bone marrow derived cells (BMDC). Thus, we hypothesize that maintaining high levels of HoxA5 selectively in EC in the tumor microenvironment in vivo will stabilize the vasculature, inhibit angiogenesis and BMDC recruitment and impair tumor progression. We have developed a transgenic mouse model in which expression of HoxA5 is largely restricted to EC through TIE-2 driven expression of a tetracycline activator protein and HoxA5 can be activated in angiogenic conditions by removal of tetracycline. These mice have been crossed with 2 established de novo models of skin (K14-HPV16) and mammary MMTV-PyMT) carcinogenesis for further analysis of EC HoxA5 in reducing tumor angiogenesis and progression in complex and dynamic tumor microenvironments. We will also evaluate how EC HoxA5 influences recruitment of pro-angiogenic BMDC in these models. Finally we will develop methods to therapeutically restore HoxA5 expression in tumor tissues in vivo and explore the ability of HoxA5 to potentially coordinate the behavior of endothelial, epithelial cells and recruitment of BMDC within tissues to identify new approaches to treat tumors and their reactive stroma. PUBLIC HEALTH RELEVANCE: These studies will investigate the whether Hox genes which maintain a quiescent, mature phenotype in vascular endothelial cells, can be sustained in angiogenic environments and thus block tumor angiogenesis and progression. We will focus on the vascular stabilizing transcription factor, HoxA5, and using a novel transgenic mouse model, will sustain its expression in tumor endothelial cells in vivo to determine whether this impairs angiogenesis and ultimately limits tumor progression. Moreover, as this gene is normally expressed in quiescent resting vessels, but absent in tumor vessels, sustained expression would not interfere with normal vascular function. In addition we will develop two novel approaches to apply HoxA5 therapeutically to both skin and mammary tumors to further demonstrate its clinical applicability. Finally we will explore the notion that HoxA5 may act to coordinately stabilize the vasculature as well as the tumor epithelium and the pro-tumorigenic immune response thus represent a means to target the entire tumor microenvironment.

PROJECT SUMMARY Poorly healing wounds linked to age or diabetes have an annual health care cost of over $25 billion. Current therapies consist of application of recombinant growth factors to stimulate angiogenesis and provisional matrix synthesis, but this approach has had limited effectiveness and is very costly. Thus new approaches to correct both inadequate angiogenesis and the prolonged inflammatory phase of poorly healing wounds are needed. My laboratory has been studying the role of Homeobox (Hox) genes in pathological tissue remodeling and have identified a group of Hox3 genes (HoxA3,B3 and D3) whose expression is upregulated during normal tissue repair but not in compromised diabetic wounds. We developed a topical gene transfer methods to deliver plasmids expressing HoxD3 and HoxA3 and observed significant improvement in healing of diabetic wounds. More recently we observed that that paralogous HoxB3 is the most potent and can improve healing to rates indistinguishable from healthy mice. Our preliminary evidence indicates that along with stimulating angiogenesis, HoxB3 also corrects the prolonged inflammatory phase of poorly healing wounds and promotes influx of immune suppressive late wound macrophages and mast cells. Thus restoring of HoxB3 reprograms poorly healing wounds to allow them to transition from a prolonged inflammatory state to an immunosuppressive state that is conducive to wound resolution. We have also observed changes in a number of other immune cells in diabeetic wounds that are corrected by HoxB3 but the role of these cells has not yet been explored. Thus we propose to conduct studies to identify downstream mediators of HoxB3 and establish the functional roles of changes in immune cell trafficking induced by HoxB3 that contribute to greatly improved wound repair.