Endostatin Blocks Vascular Endothelial Growth Factor-mediated Signaling via Direct Interaction with KDR/Flk-1*

Endostatin, a fragment of collagen XVIII, is a potent anti-angiogenic protein, but the molecular mechanism of its action is not yet clear. We examined the effects of endostatin on the biological and biochemical activities of vascular endothelial growth factor (VEGF). Endostatin blocked VEGF-induced tyrosine phosphorylation of KDR/Flk-1 and activation of ERK, p38 MAPK, and p125FAK in human umbilical vein endothelial cells. Endostatin also inhibited the binding of VEGF165 to both endothelial cells and purified extracellular domain of KDR/Flk-1. Moreover, the binding of VEGF121 to KDR/Flk-1 and VEGF121-stimulated ERK activation were blocked by endostatin. The direct interaction between endostatin and KDR/Flk-1 was confirmed by affinity chromatography. However, endostatin did not bind to VEGF. Our findings suggest that a direct interaction of endostatin with KDR/Flk-1 may be involved in the inhibitory function of endostatin toward VEGF actions and responsible for its potent anti-angiogenic and anti-tumor activities in vivo.

Angiogenesis, the formation of new capillaries from the preexisting blood vessels, is critical for tumor growth and metastasis (1)(2)(3). Endostatin, a 20-kDa proteolytic fragment of collagen XVIII, was discovered as a potent inhibitor of angiogenesis originally from murine hemangioendothelioma cell medium (4). Subsequently, the recombinant endostatin was shown to inhibit tumor growth and metastasis in various animal models (4). Furthermore, repeated cycles of endostatin therapy resulted in prolonged tumor dormancy without resistance to endostatin (5). On the cellular level, endostatin has been shown specifically to block proliferation and migration of endothelial cells and to induce endothelial cell apoptosis (4,6,7). However, the molecular mechanisms of endostatin-mediated anti-angiogenesis and tumor regression are not yet clear.
Vascular endothelial growth factor (VEGF), 1 a potent mitogen for endothelial cells, is an important mediator of angiogenesis and is involved in the differentiation of endothelial cells and the development of the vascular system (8 -10). In particular, it is thought that VEGF is the most important angiogenic factor closely associated with induction and maintenance of the neovasculature in human tumors. The increased expression of VEGF mRNA has been detected in a variety of tumors, and recently tumor VEGF level was recognized as an important prognostic marker of tumor angiogenesis (10,11). VEGF exerts its effects through binding to its two receptor tyrosine kinases, KDR/Flk-1 and Flt-1, expressed on endothelial cells. KDR/ Flk-1 is related mainly to the mitogenic and chemotactic responses, whereas Flt-1 is required for endothelial cell morphogenesis (12,13). Recent studies (14 -16) have identified the VEGF-induced signaling events in endothelial cells including extracellular signal-regulated kinases (ERKs), p38 mitogenactivated protein kinase (p38 MAPK), and p125 focal adhesion kinase (p125 FAK ). In addition, endothelial nitric-oxide synthase is considered an essential mediator of VEGF-induced angiogenesis (17,18).
Previous studies have reported that endostatin inhibits endothelial cell proliferation, migration, and angiogenesis in response to VEGF (6,19), but its mechanism of action is not clearly delineated. Here we report that endostatin blocks VEGF-induced tyrosine phosphorylation of KDR/Flk-1 and activation of ERK, p38 MAPK, and p125 FAK , which are downstream events of KDR/Flk-1 signaling and are involved in the mitogenic and motogenic activities of VEGF in endothelial cells. Furthermore, we demonstrated that endostatin inhibits the binding of VEGF to endothelial cells and to its cell surface receptor, KDR/Flk-1. The binding partner for endostatin is KDR/Flk-1 but not VEGF. Our results suggest that direct interaction of endostatin with KDR/Flk-1 blocks the binding of VEGF to endothelial cells and the VEGF-induced signaling of KDR/Flk-1 itself and its downstream signaling events, resulting in the inhibition of VEGF-induced endothelial cell proliferation and migration.

EXPERIMENTAL PROCEDURES
Materials-VEGF 165 , VEGF 121 , basic fibroblast growth factor (bFGF), protein A-agarose, and antibody to p125 FAK were from Upstate Biotechnology (Lake Placid, NY). Tissue culture dishes and plasticware were obtained from Falcon Plastics. M199 and heparin were purchased from Invitrogen. KDR/Flk-1-Fc, Flt-1-Fc, and antibody for VEGF were from R&D Systems (Minneapolis, MN). Anti-KDR/Flk-1 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phosphotyrosine antibody was from Transduction Laboratories (Lexington, KY). Antibodies for phospho-specific ERK (Thr-202/Tyr-204) and ERK2 were obtained from New England Biolabs (Beverly, MA). Antibodies for human IgG-HRP and mouse IgG-HRP were from Pierce. [ 3 H]thymidine, 125 I-VEGF 165 , NHS-activated Sepharose, and the chemiluminescent substrate for horseradish peroxidase were from Amersham Biosciences. All other reagents were purchased from Sigma unless indicated otherwise.
Cell Culture-Human umbilical vein endothelial cells (HUVECs) were isolated from human umbilical cord veins by collagenase treatment as described previously (20), and only passages 2-6 were used. The cells were grown in M199 supplemented with 20% fetal bovine serum (FBS), 100 units/ml penicillin, 100 g/ml streptomycin, 3 ng/ml bFGF, and 5 units/ml heparin at 37°C under a humidified 95 and 5% (v/v) mixture of air and CO 2 , respectively.
Expression and Purification of Recombinant Endostatin-The recombinant mouse endostatin was expressed and purified from HEK293 cells stably transfected with the pFLAG-CMV-1-endostatin as described previously (21). After dialysis against PBS, the purity of recombinant mouse endostatin as checked by SDS-PAGE was as high as 97%.
Immunoprecipitation-Confluent HUVECs were incubated for 6 h in M199 containing 1% FBS, and cells were stimulated by the addition of VEGF (10 ng/ml). After stimulation, cells were lysed in 1 ml of lysis buffer (20 mM Tris/HCl, pH 8.0, 2 mM EDTA, 137 mM NaCl, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 10% glycerol, and 1% Triton X-100). Lysates were clarified by centrifugation at 15,000 ϫ g for 10 min, and the resulting supernatants were immunoprecipitated with either 1 g/ml anti-KDR/Flk-1 antibody or anti-FAK antibody at 4°C for 3 h followed by the addition of protein A-agarose beads at 4°C for 1 h. Immunoprecipitates were washed three times with lysis buffer, solubilized in SDS-PAGE sample buffer containing ␤-mercaptoethanol, and further analyzed by Western blotting.
Western Blotting-Cell lysates or immunoprecipitates from HUVECs were run in SDS-PAGE and transferred to polyvinylidene difluoride membrane. The blocked membranes were then incubated with the indicated antibodies, and the immunoreactive bands were visualized using a chemiluminescent substrate.
Endothelial Cell Proliferation Assay-The counting assay was performed as described previously (19). Briefly, HUVECs were seeded at a density of 2.0 ϫ 10 5 cells/well onto gelatin-coated 6-well plates. After 24 h, the medium was replaced with M199 containing 1% FBS with or without endostatin as indicated (0, 0.1, 1, 3, 5, and 10 g/ml, respectively). After 30 min, cells were treated with 10 ng/ml VEGF 165 . After 48 h, the cells were trypsinized, and the total number of cells was counted. To further confirm the inhibitory effect of endostatin on VEGFinduced HUVEC proliferation, [ 3 H]thymidine incorporation assay was carried out as described previously (22). Briefly, HUVECs were seeded at a density of 2 ϫ 10 4 cells/well in gelatin-coated 24-well plates. After 24 h, cells were washed twice with M199 and incubated for 6 h in M199 containing 1% FBS. Cells were stimulated by the addition of 10 ng/ml VEGF for 30 h, followed by the addition of 1 Ci/ml of [ 3 H]thymidine for 6 h. High molecular weight DNAs were precipitated using 5% trichloroacetic acid at 4°C for 30 min. After two washes with ice-cold H 2 O, 3 H radioactivity was solubilized in 0.2 N NaOH, 0.1% SDS and determined by liquid scintillation counter. Each sample was assayed in duplicate, and the assays were repeated twice.
Endothelial Cell Migration Assay-Chemotaxis assay was performed as described previously (22). Briefly, the chemotactic motility of HU-VECs was assayed using Transwell with 6.5-mm diameter polycarbonate filters (8-m pore size). The lower surface of the filter was coated with 10 g of gelatin. The fresh M199 medium (1% FBS) containing VEGF was placed in the lower wells. HUVECs were trypsinized and suspended at a final concentration of 1 ϫ 10 6 cells/ml in M199 containing 1% FBS. Various concentrations of endostatin (0, 0.1, 1, 3, 5, and 10 g/ml, respectively) were given to the cells for 30 min at room temperature before seeding. One hundred l of the cell suspension was loaded into each of the upper wells. The chamber was incubated at 37°C for 4 h. Cells were fixed and stained with hematoxylin and eosin. Nonmigrating cells on the upper surface of the filter were removed by wiping with a cotton swab, and chemotaxis was quantified by counting with an optical microscope (ϫ200) the cells that migrated to the lower side of the filter. Ten fields were counted for each assay.
Binding of 125 I-VEGF to Its Receptors on the Surface of HUVECs-The binding of 125 I-VEGF to HUVECs was performed as described previously (23). Briefly, HUVECs were seeded at a density of 5 ϫ 10 4 cells/well onto gelatin-coated 24-well plates, incubated overnight, washed with binding buffer (25 mM HEPES/NaOH, pH 7.4, 0.1% BSA in M199), and incubated at 37°C for 2 h. The cultured cells were first treated with various amounts of endostatin for 30 min, and 125 I-VEGF 165 (0.125 nM) was added in 200 l of binding buffer. The binding was allowed to proceed at 4°C for 3 h, and the cells were washed twice with ice-cold binding buffer followed by washing once with ice-cold PBS containing 0.1% BSA. Subsequently, the cells were solubilized by the addition of 0.5 ml of 20 mM Tris-HCl, pH 7.4, containing 1% Triton X-100 at room temperature for 20 min, and receptor-bound radioactivity was determined in a ␥-counter. The amount of nonspecific binding was determined in the presence of a 100-fold molar excess of nonlabeled VEGF 165 . Each data point was assayed in triplicate, and the assays were repeated twice.
Binding of KDR/Flk-1-Fc to the Immobilized VEGF-VEGF 165 and VEGF 121 (80 ng/well) in 100 l of PBS were immobilized to 96-well plates. The wells were washed and blocked with 3% BSA in M199 for 2 h. After 10 min of preincubation of KDR/Flk-1-Fc (25 ng/ml) in M199 containing 25 mM HEPES/NaOH, pH 7.4, and 0.1% BSA with or without various amount of endostatin, the mixture (100 l) was added to each well. After 2 h, the wells were washed three times with PBST (PBS ϩ 0.05% Tween 20). The bound KDR/Flk-1-Fc was determined by incubation with anti-human IgG-HRP followed by a chemiluminescent substrate. All experiments were carried out at room temperature. Each data point was assayed in triplicate, and the assays were repeated at least twice.
Endostatin Affinity Chromatography-Endostatin affinity and negative control beads were prepared using NHS-activated Sepharose according to the manufacturer's instructions and suspended in 0.1% BSA/ PBS. KDR/Flk-1-Fc (2.5 g) in 450 l of PBS was added to 50 l of control or endostatin affinity beads (1:1 slurry), and the mixtures were incubated overnight at 4°C. Beads were washed three times with 400 l of PBS. Bound materials were eluted by boiling with 50 l of reducing SDS-loading dye. Samples (25 l) were separated by SDS-PAGE and transferred to polyvinylidene difluoride membrane, and KDR/Flk-1-Fc was detected using anti-human IgG-HRP.
Binding of 125 I-VEGF to Immobilized Endostatin-Various amounts of endostatin in 100 l of PBS were coated on ELISA plates overnight at 4°C. The wells were washed and blocked with M199 containing 3% BSA at room temperature for 2 h. 125 I-VEGF 165 (0.125 nM) was added to each well, and the binding was allowed to proceed at room temperature for 2 h. The wells were washed three times with PBST. Bound 125 I-VEGF 165 was removed by the addition of 150 l of 0.1 N NaOH at room temperature for 30 min, and the radioactivity was determined in a ␥-counter. Each sample was assayed in triplicate, and the assays were repeated twice.
Statistical Analysis-The p values were calculated from Student's t test based on comparisons with control samples tested at the same time.

Endostatin Inhibits VEGF 165 -induced Tyrosine Phosphorylation of KDR/Flk-1 in
HUVECs-According to the previous reports (6,19), endostatin inhibits VEGF-induced endothelial cell proliferation, migration, and angiogenesis, but its mechanism is not clear. To investigate the molecular mechanisms of endostatin associated with its anti-angiogenic activities, the recombinant mouse endostatin was purified from HEK293 cells stably transfected with the pFLAG-CMV-1-endostatin as described previously (21). The anti-angiogenic activities of purified endostatin were evaluated (Fig. 1). Endostatin inhibited VEGF-induced proliferation of HUVECs in a dose-dependent manner (Fig. 1A). In addition, endostatin blocked VEGF-induced DNA synthesis (Fig. 1B). Endostatin also inhibited VEGF-induced proliferation of bovine aortic endothelial cells but had no effect on the proliferation of HT1080 cells (data not shown). VEGF-induced endothelial cell migration was also blocked by endostatin in a dose-dependent manner with halfmaximal activity at ϳ 1 g/ml (Fig. 1C). These results indicate that endostatin used in this study had the intrinsic properties described in other studies (6,24).
Because previous studies have shown that VEGF induces proliferation and migration through activation of its cell surface receptor, KDR/Flk-1 (10, 17), we investigated the effects of endostatin on VEGF-induced KDR/Flk-1 phosphorylation. When HUVECs were stimulated with 10 ng/ml VEGF 165 for 5 min, a protein migrating at 205 kDa was strongly tyrosinephosphorylated ( Fig. 2A). Preincubation of HUVECs with 10 g/ml endostatin (30 min) prior to VEGF 165 stimulation dramatically reduced VEGF-induced tyrosine phosphorylation of this protein. To determine whether the VEGF-induced tyrosine-phosphorylated protein is KDR/Flk-1, tyrosine phosphorylation of KDR/Flk-1 was analyzed by immunoprecipitation with anti-KDR/Flk-1 and then immunoblotting with anti-phosphotyrosine antibody. As shown Fig. 2B, VEGF 165 induced tyrosine phosphorylation of KDR/Flk-1; this phosphorylation was blocked by endostatin pretreatment. These results indicate that endostatin blocks VEGF-induced KDR/Flk-1 activation in endothelial cells.
Endostatin Inhibits VEGF 165 -stimulated ERK, p38 MAPK, and p125 FAK Activation-To further determine the inhibitory effect of endostatin on VEGF-induced KDR/Flk-1 activation, we investigated the effect of endostatin on the intracellular signaling pathways induced by VEGF. Because previous studies have demonstrated that VEGF induces a rapid activation of ERK and p38 MAPK, which are downstream signaling molecules of KDR/Flk-1 (14,15), the effects of endostatin on VEGF-induced ERK and p38 MAPK activation were assessed. Subconfluent HUVECs were preincubated with various concentrations of endostatin for 30 min and then stimulated with 10 ng/ml of VEGF 165 for 10 min. Fig. 3A shows that VEGF stimulated the phosphorylation of ERK and p38 MAPK, and endostatin inhibited VEGF-induced ERK and p38 MAPK activation in a dosedependent manner. Pretreatment of endothelial cells with endostatin (10 g/ml) for various times prior to the addition of VEGF 165 (10 ng/ml) showed that endostatin almost completely blocked VEGF-induced ERK and p38 MAPK phosphorylation after 30 min of preincubation (Fig. 3B). When HUVECs were treated simultaneously with endostatin and VEGF 165 , the inhibitory activity was greatly reduced.
To confirm the inhibitory effect of endostatin on KDR/Flk-1 signaling, we investigated whether endostatin could block p125 FAK tyrosine phosphorylation, which lies downstream of the KDR/Flk-1 receptor (15). Preincubation of HUVECs with endostatin (30 min) caused a marked decrease in p125 FAK tyrosine phosphorylation induced by VEGF 165 (Fig. 3C). These results suggest that endostatin can block multiple signaling events of VEGF in endothelial cells, presumably via VEGF receptor inactivation.
Endostatin Blocks the Binding of VEGF 165 to Its Receptors-To determine the mechanism by which endostatin blocks VEGF-mediated KDR/Flk-1 activation, we investigated the effect of endostatin on the binding of VEGF 165 to HUVECs. Fig.  4A shows that pretreatment of endostatin (30 min) prior to the addition of 125 I-VEGF 165 blocked the binding of VEGF to HUVECs in a dose-dependent manner (IC 50 Ϸ 1.25 g/ml). When endothelial cells were preincubated with endostatin for 30 min, the specific binding of 125 I-VEGF 165 (0.125 nM) was reduced to 3.7% by the presence of 10 g/ml of endostatin (Fig.  4B). The inhibitory activity of endostatin (10 g/ml) was markedly reduced without preincubation (52.0%). These results suggest that preincubation of endostatin prior to VEGF treatment is necessary to block the VEGF binding to endothelial cells.
To further confirm the inhibitory activity of endostatin on the interaction between VEGF and its receptors, using an ELISA binding assay we examined the effect of endostatin on the binding of VEGF with KDR/Flk-1-Fc, a soluble fusion protein containing the extracellular domain of KDR/Flk-1. As shown in Fig. 4C, preincubation of endostatin (10 min) with KDR/Flk-1-Fc blocked the interaction between VEGF and KDR/Flk-1-Fc in a dose-dependent manner (IC 50 Ϸ 2.5 g/ml). The binding of Flt-1-Fc to immobilized VEGF 165 was also blocked by endostatin. We next investigated the effect of preincubation time of endostatin with KDR/Flk-1-Fc on the binding of VEGF to KDR/Flk-1-Fc. The longer preincubation showed less binding of KDR/Flk-1-Fc to the immobilized VEGF (Fig. 4D). We also examined whether endostatin can affect the

FIG. 1. Endostatin inhibits VEGF-induced proliferation and migration of HUVECs.
A, HUVECs were pretreated for 30 min with various concentrations of endostatin before exposure to VEGF 165 (10 ng/ml). Control (CON) means no treatment with VEGF and endostatin. After incubation for 48 h, total cell numbers were counted under a microscope. B, cultured HUVECs were stimulated with 10 ng/ml VEGF 165 in the absence or presence of 10 g/ml endostatin and allowed to proliferate for 36 h. [ 3 H]Thymidine incorporation was measured during the last 6 h of incubation. C, various concentrations of endostatin (0.1, 1, 3, 5, and 10 g/ml) were pretreated for 30 min prior to treatment with 10 ng/ml VEGF. After a 4-h incubation, chemotaxis was quantified by counting the cells that migrated to the lower side of the filter with optical microscopy at 200ϫ magnification. The basal migration in the absence of VEGF was 65 Ϯ 3 cells/field. Each sample was assayed in duplicate, and the assays were repeated twice. *, p Ͻ 0.05; **, p Ͻ 0.01 versus VEGF. dissociation rate of KDR/Flk-1-Fc from the preformed VEGF and FKDR/Flk-1-Fc complex. Endostatin even at a high concentration (10 g/ml) had no significant effect on the dissociation rate of KDR/Flk-1-Fc from VEGF and its receptor complex (data not shown). These results suggest that preincubation of endostatin prior to VEGF treatment is necessary to block VEGF binding to endothelial cells and that endostatin inhibits VEGF-induced signaling via the direct blockage of VEGF and its receptor KDR/Flk-1 interaction.
Endostatin Blocks the Binding of VEGF 121 to KDR/Flk-1 and VEGF 121 -stimulated ERK Activation-VEGF 165 and VEGF 121 are two major isoforms of VEGF generated via an alternative splicing mechanism from a unique gene (10,17).  (lower panel in B). The arrowheads indicate tyrosine-phosphorylated protein with a molecular mass of 205 kDa.

FIG. 3. Endostatin blocks VEGF 165 -induced MAP kinase and p125 FAK activation. A, endostatin inhibits VEGF 165 -induced ERK and p38
MAPK activation in a dose-dependent manner. HUVECs were pretreated with various concentrations of endostatin for 30 min and then stimulated with 10 ng/ml VEGF 165 for 10 min. B, preincubation effect of endostatin on the inhibition of VEGF-induced ERK and p38 MAPK activation. HUVECs were pretreated with 10 g/ml endostatin for the indicated time periods and then stimulated with 10 ng/ml VEGF 165 for 10 min. In A and B, activation of ERK and p38 MAPK by VEGF was determined by Western blotting using antibodies against phosphorylated forms of ERKs (P-ERK1 and P-ERK2) and p38 MAPK (P-p38). The membranes were stripped and reprobed with antibodies against ERK2 and p38 MAPK. C, endostatin inhibits VEGF-induced tyrosine phosphorylation of p125 FAK . HUVECs were pretreated with 10 g/ml endostatin for 30 min and then stimulated with 10 ng/ml VEGF for 10 min. Cell lysates were immunoprecipitated with anti-p125 FAK antibody. The immunoprecipitates (IP) were analyzed by immunoblotting with anti-phosphotyrosine antibody to assay for p125 FAK tyrosine phosphorylation (P-p125 FAK ) or anti-p125 FAK antibody for p125 FAK protein levels (p125 FAK ). tin (10 min) with KDR/Flk-1-Fc blocked the binding of KDR/ Flk-1-Fc to the immobilized VEGF 121 in a dose-dependent manner (Fig. 5A). The IC 50 value was ϳ3.5 g/ml, similar to that of VEGF 165 . The longer preincubation of endostatin with KDR/ Flk-1-Fc showed a lower level of binding of KDR/Flk-1-Fc to VEGF 121 in the presence of the same amount of endostatin (Fig. 5B). As shown in Fig. 5C, both VEGF 121 and VEGF 165 stimulated phosphorylation of ERK in endothelial cells, and these effects were blocked by the treatment of endostatin. These results suggest that endostatin can inhibit the interaction between VEGF 121 and its receptor KDR/Flk-1 and block VEGF 121 -induced ERK activation in endothelial cells.
Endostatin Binds to KDR/Flk-1 but Not to VEGF-Because endostatin blocks the interaction between VEGF and its receptor KDR/Flk-1, we investigated whether endostatin directly binds to either KDR/Flk-1 or VEGF. Endostatin affinity chromatography showed that KDR/Flk-1 bound to the endostatinimmobilized beads but not to control beads (Fig. 6A). Flt-1-Fc also bound to the endostatin-immobilized beads (data not shown). In an ELISA assay, 125 I-VEGF 165 did not bind to im-mobilized endostatin (Fig. 6B). Affinity chromatography also showed that there was no interaction between endostatin and VEGF (data not shown). These results suggest that direct interaction of endostatin with KDR/Flk-1 but not with VEGF may be involved in the inhibitory activity of endostatin on VEGF-induced signaling, proliferation, and migration of endothelial cells. DISCUSSION Several reports demonstrate that endostatin, a potent inhibitor of angiogenesis, specifically blocks the proliferation and migration of endothelial cells induced by angiogenic factors including VEGF (6,19) and inhibits tumor growth and metastasis in various animal models (4). However, the molecular mechanisms of endostatin-mediated anti-angiogenesis and tumor regression are not yet clear. To delineate the mechanisms involved in the anti-angiogenic activity of endostatin, we investigated the effect of endostatin on VEGF action by cell biological and biochemical experiments. VEGF is the most important angiogenic molecule associated with tumor-induced neovascu- FIG. 4. Endostatin blocks the binding of VEGF 165 to its receptors. A, endostatin blocks the binding of VEGF to its receptors on the surface of HUVECs in a dose-dependent manner. Cultured endothelial cells were pretreated with various concentrations of endostatin (30 min), and 125 I-VEGF 165 (0.125 nM) was added to the cells. After incubation, the amount of cell-bound radioactivity was determined. Nonspecific binding was determined in the presence of a 100-fold molar excess of unlabeled VEGF. Data were presented after subtraction of nonspecific binding. B, preincubation of endostatin with cells is necessary for the efficient blocking of VEGF binding to endothelial cells. Endothelial cells were treated with 10 g/ml endostatin (0 or 30 min preincubation), and the specific binding of 125 I-VEGF 165 to HUVECs was determined. C, endostatin inhibits the binding of KDR/Flk-1-Fc and Flt-1-Fc to immobilized VEGF 165 in a dose-dependent manner. Various concentrations of endostatin were preincubated with 25 ng/ml KDR/Flk-1-Fc or Flt-1-Fc for 10 min, and then the mixtures were added to the VEGF 165 -coated wells for 2 h. The amount of bound KDR/Flk-1-Fc or Flt-1-Fc was determined with anti-human IgG-HRP using a chemiluminescent substrate. D, inhibitory activity of endostatin on VEGF and KDR/Flk-1 interaction depends on the preincubation time of endostatin with KDR/Flk-1. Endostatin (5 g/ml) and KDR/Flk-1-Fc (25 ng/ml) were preincubated for various time periods, and the mixtures were added to the VEGF 165 -coated wells. The amount of bound KDR/Flk-1-Fc was determined as described in C. larization. In this paper, we present several novel observations. First, endostatin blocks the VEGF-induced tyrosine phosphorylation of KDR/Flk-1 in endothelial cells. Second, endostatin suppresses the VEGF-induced activation of ERK, p38 MAPK, and p125 FAK , which are downstream events of the KDR/Flk-1 signaling and are involved in the mitogenic and motogenic activities of VEGF in endothelial cells. Third, endostatin inhibits the binding of VEGF to endothelial cells and to its cell surface receptor, KDR/Flk-1. Finally, endostatin directly binds to KDR/Flk-1 but not to VEGF. Our findings clearly indicate that direct interaction of endostatin with the VEGF receptor KDR/Flk-1 blocks the binding of VEGF to its receptor, the VEGF-induced signaling of KDR/Flk-1 itself, and its downstream signaling such as ERK, p38 MAPK, and p125 FAK , resulting in the inhibition of VEGF-induced proliferation and migration of endothelial cells.
Endostatin Suppresses VEGF-induced Endothelial Cell Signaling-Previous reports showed that endostatin inhibits VEGF-induced proliferation and migration of endothelial cells (6,19). Because recent studies suggest that the VEGF receptor  6. Endostatin binds to KDR/Flk-1 but not to VEGF. A, endostatin binds to KDR/Flk-1. KDR/Flk-1-Fc was incubated with control or endostatin-immobilized beads. After bound materials were eluted, samples were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes, and KDR/Flk-1-Fc was detected using anti-human IgG-HRP. B, endostatin does not bind to VEGF. 125 I-VEGF was added to the wells of an ELISA plate coated with various amounts of endostatin. After a 2-h incubation, the bound radioactivity was determined. The wells coated with anti-VEGF antibody were used as a positive control.
KDR/Flk-1 and not Flt-1 is involved in proliferation and migration of endothelial cells in response to VEGF (12,13), we first revealed that endostatin could interfere with VEGF-induced activation of its receptor KDR/Flk-1. Endostatin blocked VEGF 165 -induced phosphorylation of KDR/Flk-1 in human endothelial cells. This finding was further supported by our data that endostatin clearly suppressed the downstream events of VEGF 165 -induced KDR/Flk-1 signaling, such as activation of ERK and p38 MAPK and p125 FAK tyrosine phosphorylation, that are requisite for the mitogenic and motogenic activities of VEGF in endothelial cells (14 -16). Endostatin also inhibited VEGF 121 -induced activation of ERK in HUVECs. These results supposed that the anti-angiogenic action of endostatin against VEGF is associated with its ability to block VEGF receptor KDR/Flk-1 activation.
Endostatin Blocks the Interaction between VEGF and Its Receptor, KDR/Flk-1, and Binds to KDR/Flk-1 but Not to VEGF-How does endostatin block the VEGF-mediated endothelial cell signaling events that result in the blocking of VEGF-induced proliferation and migration of endothelial cells? There are at least three different possible mechanisms: (i) endostatin directly interferes in the interaction between VEGF and its receptor, KDR/Flk-1, via its direct binding either to VEGF or to KDR/Flk-1; (ii) binding of endostatin to its receptor-like molecules present on the surface of endothelial cells transfers intracellular signals to block VEGF-mediated signaling; and (iii) internalized endostatin blocks the VEGF-mediated signaling. Our results showed that endostatin blocks the binding of VEGF to its receptors present on the surface of HUVECs and also interferes in the interaction between VEGF and KDR/Flk-1-Fc in a dose-dependent manner. It is worthwhile to note that endostatin blocked VEGF-induced tyrosine phosphorylation of KDR/Flk-1, proliferation, and migration of endothelial cells and activation of ERK, p38 MAPK, and p125 FAK as well as the binding of VEGF to endothelial cells and to KDR/Flk-1 with an IC 50 in the g/ml range. This working concentration was at a level similar to that shown in other reports (7,25). However, some groups have demonstrated that endostatin blocks VEGF-induced endothelial cell migration with an IC 50 in the ng/ml range (6,26). Thus, the working concentration of endostatin is still controversial.
What is the binding partner for endostatin? Our experiments, including endostatin affinity chromatography, showed that the binding partner for endostatin is KDR/Flk-1 but not VEGF. Preincubation of endostatin with endothelial cells or KDR/Flk-1-Fc was necessary for the efficient blockage of interaction between VEGF and its cell surface receptors or KDR/ Flk-1-Fc. These results may be the reason why preincubation of HUVECs with endostatin prior to VEGF treatment is required for the efficient inhibition of VEGF-induced phosphorylation of KDR/Flk-1 and intracellular signaling pathways. Previously reported results also show that preincubation of endostatin with endothelial cells is necessary for the blockage of VEGFinduced migration (6,19). The requirement of preincubation for the efficient activity of endostatin also supports the finding that the binding partner for endostatin is KDR/Flk-1 not VEGF. Although most of the VEGF-induced responses in endothelial cells are mediated through KDR/Flk-1, a recent study suggested that Flt-1 but not KDR/Flk-1 may be involved in VEGF-induced formation of capillary networks via the stimulation of nitric oxide release in HUVECs (27). Endostatin blocks the tube formation of endothelial cells (28). Our results showed that endostatin also bound to Flt-1 and blocked the interaction between VEGF and Flt-1. It is worthwhile to note that endostatin neither competes with the binding of bFGF to human tissues nor affects FGF receptor signaling (29, 30), although endostatin was originally identified based on its inhibitory activity of bFGF-induced endothelial cell proliferation (4). Although we could not completely rule out the other mechanisms, our observations suggest that direct interaction of endostatin with KDR/Flk-1 can block the binding of VEGF to its receptor present on the surface of endothelial cells and may be involved in the inhibitory activity of endostatin on VEGFinduced angiogenesis.
Model for Anti-angiogenic Activity of Endostatin-Although several groups have found endostatin-binding proteins and suggested models to delineate the mechanism of endostatin action, the molecular mechanisms of endostatin-mediated antiangiogenesis and tumor regression are not fully understood. Endostatin induces tyrosine kinase signaling through the Shb adaptor protein and enhanced apoptosis in bFGF-treated endothelial cells (30). Endostatin inhibits pro-MMP-2 activation mediated by membrane type 1 MMP and the catalytic activities of both MMP-2 and membrane type 1 MMP resulting in the inhibition of endothelial and tumor cell invasion (21). Endostatin also binds with ␣ 5 -and ␣ v -integrins on the surface of endothelial cells and inhibits integrin-dependent endothelial cell migration (25). The interaction of endostatin with tropomyosin causes disruption of microfilament integrity leading to inhibition of endothelial cell migration, induction of endothelial cell apoptosis, and ultimately inhibition of tumor growth (31).
A recent study has shown that endostatin interacts with endothelial cell surface glypicans via its glycosaminoglycan chain and that glypicans are necessary for the inhibitory activity of endostatin on VEGF 165 -but not VEGF 121 -induced endothelial cell migration; endostatin, however, efficiently blocks both VEGF 165 -and VEGF 121 -induced endothelial cell migration (26). Another study shows that heparin binding is not involved in the function of endostatin (29). An in situ-binding assay showed that endostatin binds predominantly to its functional target (blood vessels) and co-localizes largely with bFGF in human breast carcinoma tissues (29). However, in situ endostatin binding to blood vessels is resistant to treatment with heparinase and is not affected by bFGF and heparin. These results indicate that binding of endostatin to blood vessels is not mediated by heparan sulfate proteoglycans and that heparin binding domain of endostatin may not be involved in the in vivo anti-angiogenic and anti-tumor action of endostatin. We have shown herein that endostatin blocks the binding of both VEGF 165 and VEGF 121 to KDR/Flk-1 and their intracellular signaling events in endothelial cells. These findings provide a possible mechanism by which endostatin inhibits angiogenesis induced by VEGF in vitro and in vivo. The detailed molecular characteristics of endostatin/KDR binding are under investigation.