Background

Many factors are known for their angiogenic potential. Most of these bind heparin (aFGF, bFGF, VEGF, PlGF, HGF, HB-EGF, angiogenin) It has been demonstrated that heparin (or its cell surface analog, heparan sulfate) is necessary for the interactions between the factor and its receptor. In fact, heparin in low concentrations functions as an activator of the angiogenic process, presenting these molecules to their receptors. However an excess of soluble heparin can partition the factors away from the heparan sulfate of the cell surface and the matrix, and inhibit angiogenesis. The identification and characterization of angiogenic and anti-angiogenic factors is a critical step in developing antiangiogenic therapy. It is clearly important to have well characterized experimental models which permit studies on the mechanisms of angiogenesis in systems which reproduce the steps of neovascularization.

Specific Aims

The principal objectives of this project are divided as follows:

1) Developing and standardizing models in vitro and in vivo for studying tumor angiogenesis.

2) Characterization and culture in vitro and in vivo of highly vascular neoplasms

3) Use of the above models to understand the mechanisms of action of heparin-binding angiogenic factors(in particular VEGF and similar molecules) and establishment of the cellular events occurring during neoangiogenesis in matrigel implants.

4) Use of the above models to test potentially antiangiogenic substances, which in turn may be tested for therapeutic potential in experimental tumor models.

Experimental Approach

In these objectives we will employ in vitro and in vivo models to answer the following specific questions:

1) Developing and standardizing models in vitro and in vivo for studying tumor angiogenesis.

A: Stimulation of endothelial cells by tumor cell products in migration, chemoinvasion, morphogenesis and growth assays

Endothelial cell migration will be studied using the boyden chamber assay, where the factors are placed in the lower chamber and the cells in the upper chamber which is separated by a porous filter. Stimulated cells migrate to the lower compartment. Our laboratory has developed a modification of this test where the filter is covered by a thin layer of matrigel, a reconstituted basement membrane (called chemoinvasion). This "chemoangiogenesis" test allows us to measure not only the induction of migration but also their ability to degrade the basement membrane in response to an angiogenic stimulus.

Supernatants of Kaposi's sarcoma cells, a highly vascular lesion, have been shown to be able to induce endothelial cell migration and invasion in this test. This test will be used as an initial screen of supernatants from different tumor cell types for their ability to induce endothelial cell invasion.

Growth assays will be used to determine the effect of factors in the conditioned media on endothelial cell proliferation.

A morphological and differentiation assay which utilizes a three-dimensional collagen gel has been utilized in the past. Three-dimensional matrigel gels have been used to develop a new test, in which endothelial cells form capillary-like networks in vitro. Differentiation in this assay is stimulated or inhibited by various peptides from the laminin molecule. Our laboratory has demonstrated that supernatants from Kaposi's sarcoma cells promote the spreading of endothelial cells on matrigel. This assay will also be used to evaluate the effects cell products from highly angiogenic neoplasms. Collaborative studies carried out with Prof M. Presta (Univ. of Brescia) have shown that these assays can be used to characterize the phenotype of endothelial cell transfected with growth factors.

Once tumor cell types that induce a strong angiogenic response in vitro and in vivo are identified, the use of these functional tests and fractionation by FPLC of cellular supernatants will allow for the identification of new angiogenic cytokines. This technique can also be used for purification of inhibitors of angiogenesis. This has been shown in a collaboration with Dr DeScalzi (CBA, Genova) on the angiogenic and antiangiogenic properties of cultured chondrocytes during their differentiation.

B: Stimulation of neoangiogenesis in pellets of an extracellular matrix (matrigel) by purified angiogenic factors and crude tumor cell products.

Previous studies have demonstrated that injection of KS cells subcutaneously in nude mice produces a KS-like lesion formed of spindle-cells, new vessels and inflammatory elements. Our experiments in vitro have demonstrated the angiogenic capacity of cultured KS cell supernatants; inoculation of these supernatants in vivo together with matrigel and heparin in syngeneic mice induces an intense angiogenic response, with spindle cells, new vessels and inflammatory infiltrate. One objective of our research is to examine the angiogenic response of chromatographic fractions from tumor cell supernatants. bFGF, a cytokine expressed by KS cells, is also able to induce a strong angiogenic response in the presence of heparin. However, antibodies against bFGF only partially inhibit the angiogenic response to KS cell supernatants. In our studies we will focus on VEGF, a highly angiogenic molecule present in high levels in KS lesions and whose receptors, based on recent studies, are expressed by KS cells. Our interest in VEGF is not only due to its high expression levels in KS and glioblastomas, but also that VEGF is identical to VPF (vascular permeability factor), a molecule responsible for vascular permeabilty, a characteristic of Kaposi's sarcoma.

C: Use of the matrix sponge model to establish the temporal events of neoangiogenesis

We intend to establish the temporal events associated with angiogenesis in vivo in matrigel sponges. Our present data indicate that an inflammatory cell response is a key step in the formation of Kaposi's lesions, and that neutrophils lead endothelial cells in invasion into the matrigel. This is in agreement with the data of Kleinman that show a reduction of the angiogenic response in neutropenic mice. We will stop the angiogenesis in vivo at various time points and analyze the results by light and electron microscopy.

2) Characterization and culture in vitro and in vivo of highly vascular neoplasms

A: Kaposi's sarcoma (particularly the mediterranean form)

Kaposi's sarcoma is a highly vascularized hyperplastic lesion found in the elderly and in immuodepressed individuals. A particularly aggressive form of this lesion is found in homosexual AIDS patients. Histological examination shows the presence of inflammatory cells, spindle cells and an abnormal, poorly organized vascularization. The spindle cells are immunohistochemically positive for markers of smooth muscle (alpha-actin) and fibroblasts (TE-7), while they are negative for endothelial markers such as EN4, PAL-E and factor VIII. No differences have been noted between the various forms of KS, these cells have invasive characteristics similar to malignant sarcomas. There are data that indicate that the origins of the endemic and epidemic KS may be different, but that the lesions are maintained by the production of autocrine factors. In this project we will focus on mediterranean KS, a lesion which has a higher incidence in certain regions of Italy, for example Sardegna. We have cells from one patient which developed a non-epidemic KS along with an Arterio-venous fistula. We will conduct studies in vitro in parallel with the clinical course and histopathology.

B: Multiform Glioblastoma

Glioblastoma is a malignant neoplasm responsible for more than 60% of cerebral tumors. Generally they are derived from transformed astrocytes and characterized by high proliferation rates, cellular polymorphism, necrosis and substantial vascularization. During tumor progression alterations in the phenotype of the tumor vascularization are noted. The high level of endothelial cell proliferation suggest an increase in VEGF simultaneously with up-regulation of tyrosine-kinase receptors for this molecule. Control of the angiogenic process appears to be the best prospective therapy for glioblastoma. We will examine histologically cultures from glioblastoma lesions, and precede with angiogenesis studies similar to our studies on KS; analyzing their angiogenic potential in vitro, in vivo, and finally identifying potential targets for antiangiogenic therapy. We will also maintain glioblastoma cells in nude mice for in vivo studies.

3) Use of the above models to understand the mechanisms of action of heparin-binding angiogenic factors(in particular VEGF and similar molecules) and establishment of the cellular events occurring during angiogenesis in matrigel implants.

A: Identification and/or isolation of angiogenic molecules.

In particular we will study VEGF, a heparin binding angiogenic protein and its mechanisms of action. This molecule appears to be expressed in KS lesions as observed by in situ hybridization studies done in collaboration with Dr Stürzl (Max-Plank, Münich). In parallel we will also study the action of the tat protein encoded by the human immunodeficiency virus which has heparin-dependent angiogenic properties as we have demonstrated in vivo. Sequence similarity between VEGF/VPF and tat has led us to propose that they have a similar mechanism of action. We will also isolate a high affinity (eluting at 2M NaCl) heparin binding factor from KS cell supernatants that is very active in angiogenesis in vitro and in vivo.

B: Signal transduction in angiogenesis

In collaboration with Dr Bussolino (Dept of Genetics and Clinical Medicine, Univ. of Torino), we are investigating the high affinity VEGF receptor flk-1. This receptor is known to be important in tumor angiogenesis and its inhibition in vivo, by antibodies or transdominant mutants, inhibits neovascularization. It is not yet known which second messengers transmit the signal from this "recent" tyrosine-kinase receptor. We will analyze which tyrosine becomes phosphorylated after VEGF (and tat) binding and if this serves as a "docking" site for molecules with SH2 domains.

C: Modulation of angiogenic factor function by heparin and structurally modified heparins

Many angiogenic factors bind heparin and heparan sulfate. Numerous studies on aFGF and bFGF, as well as studies on VEGF, have shown that heparin binding is important for their interactions with their high affinity receptors and for proliferation. Heparin also appears to be an important cofactor in angiogenesis induced by these factors. It has been shown that migration to a and bFGF (as well as other factors) occurs at concentrations one log lower than that required to produce proliferation. It is not known if the same high affinity receptors are also involved in migration and invasion, and if heparin/heparan sulfate are required cofactors. We will examine the migration response to bFGF and VEGF and the role of heparin in modulating this response. In addition, the heparin structural characteristics necessary will be determined in collaboration with Dr Presta (Univ of Brescia), utilizing defined structurally modified heparins.

D: Involvement of matrix-adhesion molecules (integrins) and proteolytic enzymes (collagenase IV) in angiogenesis

During the process of neovascularization endothelial cells activated by angiogenic cytokines acquire properties similar to malignant cells; not only in migratory capacity but also in production of proteases conferring the capacity to breach the basement membrane and exit the vessel. In order to block this process, and hence tumor angiogenesis, we have tried to characterize the enzymes released by stimulated endothelial and smooth muscle cells. We have found that the 72kDa collagenase IV (MMP-2), a metalloprotease already known to be secreted by highly metastatic cell lines, is also released by vascular cells activated by KS supernatants. To block this process we have used two specific inhibitors of MMP-2:

-Pep74, a synthetic peptide which reproduces a sequence found in the propeptide

segment of all metalloproteases. This peptide contains an unpaired cysteine which coordinates the zinc atom in the active site of the enzyme and keeps the enzyme in an inactive form. Pep74, which we have already demonstrated inhibits metastatic cell invasion, is also a good inhibitor of HUVE cell invasion in a dose-dependent manner.

-TIMP-2 is a natural inhibitor of MMP-2, and as such has been demonstrated to be effective inhibitor of metastatic cell invasion in vitro. We have also shown that TIMP-2 is an effective inhibitor of endothelial cells in vitro and in vivo. We intend to apply this knowledge to new therapeutic approaches.

Other key molecules in migration are the integrin receptors for extracellular matrix molecules. We were the first to demonstrate that the a5b1 fibronectin receptor was involved in chemotaxis. More recently we have done collaborative studies which show the b4 and b1 subunits are essential in migration. Preliminary data show that the a3 subunit is involved in migration to its numerous ligands (laminin, collagen, fibronectin). Integrins have been studied by others as possible therapeutic targets. Endothelial cells have integrins and we plan to examine if blocking adhesion or migration mediated by these molecules could be potential targets for therapy.

4) Use of the above models to test potentially antiangiogenic substances, which in turn may be tested for therapeutic potential in experimental tumor models.

A: Use of a combination of interferons and retinoic acid as an antiangiogenic therapy.

There are data in the literature that interferon alfa, aside from being an antiviral agent and immune cell stimulator, is also an inhibitor of angiogenesis. In fact, combinations of interferon alfa and retinoic acid are used as a local therapy for KS. We had previously observed that interferon beta is an effective inhibitor of malignant cell invasion. Some data from other laboratories indicate that interferon beta also inhibits endothelial cell stimulation by bFGF and production of collagenase IV. We will test if the less toxic interferon beta in both our in vitro and in vivo models, can be used alone or in combination with retinoic acid in the local treatment of KS and eventually glioblastoma.

B: Control of metalloproteinases as targets in antiangiogenic therapy

We will test pep74 for potential angiogenic/growth inhibitory activity in vivo against glioblastoma. In addition we will test if N-acetyl-cysteine alone, the key amino acid in the pep74 and a well tolerated molecule already in use, could be used as a cofactor in an antiangiogenic therapy.

C: Setting up protocols for antiangiogenic gene therapy with vectors expressing interferons or TIMP-2.

For the tumors of interest (KS and glioblastoma) we intend to develop gene therapy protocols in which interferon could be transcribed in vivo via plasmid or retroviral transfection. We also intent to test TIMP-2 in a gene therapy protocol in an attempt to inhibit endothelial cell invasion. We already have the expression vectors and have begun experiments with transfected cells. The next step will be the use of retroviral vectors for therapy in vivo.