Endothelial adhesion molecules facilitate the entry of leukocytes into inflamed tissues. This in turn promotes neovascularization, a process central to the progression of rheumatoid arthritis, tumor growth and wound repair. Here we test the hypothesis that soluble endothelial adhesion molecules promote angiogenesis. Human recombinant soluble E-selectin and soluble vascular cell adhesion molecule-1 induced chemotaxis of human endothelial cells in vitro and were angiogenic in rat cornea. Soluble E-selectin acted on endothelial cells in part through a sialyl Lewis-X-dependent mechanism, while soluble vascular cell adhesion molecule-1 acted on endothelial cells in part through a very late antigen (VLA)-4 dependent mechanism. The chemotactic activity of rheumatoid synovial fluid for endothelial cells, and also its angiogenic activity, were blocked by antibodies to either soluble E-selectin or soluble vascular cell adhesion molecule-1. These results suggest a novel function for soluble endothelial adhesion molecules as mediators of angiogenesis.

Kringle 5 (K5) of human plasminogen has been shown to inhibit angiogenesis by inducing the apoptosis of proliferating endothelial cells. Peptide regions around the lysine-binding pocket of K5 largely mediate these effects, particularly the peptide PRKLYDY, which we show to compete with K5 for the binding to endothelial cells. The cell surface binding site for K5 that mediates these effects has not been defined previously. Here, we report that glucose-regulated protein 78, exposed on cell surfaces of proliferating endothelial cells as well as on stressed tumor cells, plays a key role in the antiangiogenic and antitumor activity of K5. We also report that recombinant K5-induced apoptosis of stressed HT1080 fibrosarcoma cells involves enhanced activity of caspase-7, consistent with the disruption of glucose-regulated protein 78-procaspase-7 complexes. These results establish recombinant K5 as an inhibitor of a stress response pathway, which leads to both endothelial and tumor cell apoptosis. (Cancer Res 2005; 65(11): 4663-72)

Objective. Angiogenesis is an integral component of the vasculoproliferative phase of rheumatoid arthritis (RA). Recently, a heparin‐binding cytokine termed hepatocyte growth factor (HGF), or scatter factor (due to its ability to disperse cohesive epithelial colonies), was described. We conducted this study to investigate the hypothesis that this cytokine was present in the milieu of the inflamed joint, and that it contributed to the chemotaxis of endothelial cells in the synovial tissue. Methods. We examined synovial fluid, synovial tissue, and peripheral blood from 91 patients with RA and other arthritides. We used 83 total samples in an enzyme‐linked immunosorbent assay to quantitate the HGF in synovial fluids and peripheral blood. To determine whether the HGF was biologically active, an epithelial scatter factor assay was performed. Immunohistochemical analysis was used to determine localization in synovial tissues. To define a function for synovial HGF, we preincubated rheumatoid synovial fluids with neutralizing anti‐HGF and measured the ability of these synovial fluids to induce endothelial chemotaxis. Results. Synovial fluid from patients with RA contained a mean ± SEM HGF concentration of 2.0 ± 0.3 ng/ml, while synovial fluid from patients with other arthritides (including inflammatory arthritis) contained 2.4 ± 0.7 ng/ml HGF. Osteoarthritis (OA) patient samples contained the smallest quantities of synovial fluid HGF at 0.9 ± 0.1 ng/ml. RA synovial fluid contained significantly more HGF than did RA peripheral blood (1.1 ± 0.2 ng/ml) (P < 0.05). Rheumatoid synovial fluids induced more scattering of cells than did OA synovial fluids, suggesting a role for this cytokine in rheumatoid joint destruction. Interleukin‐1β induced expression of rheumatoid synovial tissue fibroblast antigenic HGF and scatter factor activity. Immunohistochemically, HGF, as well as the HGF receptor (the met gene product), localized to significantly more rheumatoid synovial tissue lining cells than normal lining cells (P < 0.05). Both HGF and its receptor immunolocalized to subsynovial macrophages as well. Levels of synovial tissue immunoreactive HGF correlated positively with the number of synovial tissue blood vessels. Anti‐HGF neutralized a mean of 24% of the chemotactic activity for endothelial cells found in 10 rheumatoid synovial fluid samples. Conclusion. These results indicate that synovial HGF may contribute to the vasculoproliferative phase of inflammatory arthritides such as RA, by inducing HGF‐mediated synovial neovascularization. These findings point to a newly described role for HGF in the fibroproliferative phase of RA‐associated synovitis.

The heptapeptide 1, NAc-Gly-Val-DIle-Thr-Arg-Ile-ArgNHEt, a structurally modified fragment derived from the second type-1 repeat of thrombospondin-1 (TSP-1), is known to possess antiangiogenic activity. However, therapeutic utility could not be demonstrated because this peptide has a very short half-life in rodents. To optimize the PD/PK profile of 1, we initiated a systematic SAR study. The initial structural modifications were performed at positions 5 and 7 of peptide 1 and at the N- and C-termini. Out of several hundred peptides synthesized, the nonapeptide 5 (ABT-526) emerged as a promising lead. ABT-526 inhibited VEGF-induced HMVEC cell migration and tube formation in the nanomolar range and increased apoptosis of HUAEC cells. ABT-526 showed acceptable PK in rodents, dog, and monkey. ABT-526, when incorporated in an angiogenic pellet implanted in the rat cornea at 10 microM, reduced neovascularization by 92%. Substitution of DalloIle in place of DIle in ABT-526 provided nonapeptide 6 (ABT-510), which was 30-fold less active than ABT-526 in the EC migration but 20-fold more active in the tube formation assay. In comparison to ABT-526, ABT-510 has increased water solubility and slower clearance in dog and monkey. Radiolabeled ABT-510 demonstrated saturable binding to HMVEC cells at 0.02-20 nM concentrations and was displaceable by TSP-1. ABT-510 and ABT-526 were shown to significantly increase apoptosis of HUAEC cells. ABT-510 was effective in blocking neovascularization in the mouse Matrigel plug model and inhibited tumor growth in the mouse Lewis lung carcinoma model. Previous studies had shown that ABT-510 was effective in inhibiting the outgrowth of murine melanoma metastases in syngeneic mice and in blocking the growth of human bladder carcinoma implanted in nude mice. It had been also shown that ABT-510 could regress tumor lesions in pet dogs or cause unexpected stabilization of the disease in advanced canine cancer. ABT-526 and ABT-510 are the first compounds in the class of potent inhibitors of angiogenesis that mimic the antiangiogenic function of TSP-1. ABT-510 is currently in phase II clinical studies.