Endorepellin Affects Angiogenesis by Antagonizing Diverse Vascular Endothelial Growth Factor Receptor 2 (VEGFR2)-evoked Signaling Pathways

Background: Endorepellin specifically targets endothelial cells via dual-receptor antagonism of α2β1 and VEGFR2 to inhibit angiogenesis. Results: Endorepellin attenuates two major signaling branches of VEGFR2 to transcriptionally repress HIF-1α concurrent with stabilized and cytoplasmically localized NFAT1. Conclusion: Endorepellin suppresses signaling of VEGFR2 independent of oxygen tension to inhibit angiogenesis. Significance: Endorepellin via dual-receptor antagonism provides novel mechanisms applicable to similar angiostatic fragments. Endorepellin, the angiostatic C-terminal domain of the heparan sulfate proteoglycan perlecan, inhibits angiogenesis by simultaneously binding to the α2β1 integrin and the vascular endothelial growth factor (VEGF) receptor 2 (VEGFR2) on endothelial cells. This interaction triggers the down-regulation of both receptors and the concurrent activation of the tyrosine phosphatase SHP-1, which leads to a signaling cascade resulting in angiostasis. Here, we provide evidence that endorepellin is capable of attenuating both the PI3K/PDK1/Akt/mTOR and the PKC/JNK/AP1 pathways. We show that hypoxia-inducible factor 1α (HIF-1α) transcriptional activity induced by VEGFA was inhibited by endorepellin independent of oxygen concentration and that only a combination of both PI3K and calcineurin inhibitors completely blocked the suppressive activity evoked by endorepellin on HIF1A and VEGFA promoter activity. Moreover, endorepellin inhibited the PKC/JNK/AP1 axis induced by the recruitment of phospholipase γ and attenuated the VEGFA-induced activation of NFAT1, a process dependent on calcineurin activity. Finally, endorepellin inhibited VEGFA-evoked nuclear translocation of NFAT1 and promoted NFAT1 stability. Thus, we provide evidence for a novel downstream signaling axis for an angiostatic fragment and for the key components involved in the dual antagonistic activity of endorepellin, highlighting its potential use as a therapeutic agent.

The complex interplay between tumor cells and their vascularized stroma has profound effects on cancer cell proliferation, migration, metastasis, and angiogenesis. Moreover, the newly formed angiogenic vessels exert an instructive role as a vascular niche by providing a paracrine and "angiocrine" mode of regulation (1). This mechanism involves the secretion and processing of various growth factors and extracellular matrix constituents that influence tumor and endothelial cells in a bidirectional manner, with integrins acting as functional hubs for pathological angiogenesis (2). Heparan sulfate proteoglycans act as depots for pro-and anti-angiogenic factors (3)(4)(5)(6)(7), and in concert with members of the fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF) family and their receptors, modulate various steps of angiogenesis (8 -10). These heparan sulfate-bound factors are dynamically processed by various proteases and heparanase to target their cognate receptor and augment their biological activity during development, tissue remodeling, and cancer growth (9,(11)(12)(13)(14).
Perlecan shows a clear functional dichotomy. The parent perlecan proteoglycan is pro-angiogenic as shown in gene-targeted studies (45)(46)(47), by primarily acting as a co-receptor for FGF2 and VEGFA (48 -50). Characterization of the zebrafish perlecan knockdown provides strong genetic evidence linking perlecan to developmental angiogenesis (51). We found that angiogenic blood vessel development of the intersegmental vessels was largely inhibited in the absence of perlecan (51). Notably, knockdown of the ␣2␤1 integrin showed a vascular phenotype similar to that evoked by perlecan knockdown (52). Thus, perlecan functions at multiple levels during the angiogenic cascade influencing endothelial cell migration, proliferation, and lumen formation (53,54).
In contrast to its parent molecule, the C-terminal domain V of perlecan, named endorepellin to designate its intrinsic antiendothelial activity (55), is anti-angiogenic in in vitro and in vivo studies (56 -59). Endorepellin can be liberated by cathepsin L (60) whereas its C-terminal module LG3 can be cleaved by bone morphogenetic protein 1 (BMP1)/Tolloid-like proteases (61) releasing a smaller biologically active fragment (41,56). Specifically, endorepellin triggers a signaling cascade that leads to disruption of the endothelial actin cytoskeleton (56,(62)(63)(64). Endorepellin interacts with the ␣2␤1 integrin receptor (56,63,65), while simultaneously interacting with the ␣2␤1 integrin and VEGFR2 5 in endothelial cells (66). Importantly, systemic delivery of endorepellin to tumor xenograft-bearing mice causes a marked suppression of tumor growth and metabolic rate mediated by sustained down-regulation of the tumor angiogenic network (57). Genetic analysis using a siRNAmediated block of endogenous ␣2␤1 integrin or animals lacking the ␣2␤1 integrin receptor have definitively shown that this is a key receptor for endorepellin and thus for the perlecan protein core (58). Therefore, endorepellin represents a member of the family of cryptic domains residing within larger parent molecules of the extracellular microenvironment that act in a dominant negative manner.
The observations summarized above suggest that perlecan/ endorepellin might be directly involved in modulating the VEGFA/VEGFR2 signaling axis. Indeed, we discovered that perlecan binds via endorepellin to both ␣2␤1 integrin and VEGFR2 (66). Endothelial cells that express ␣2␤1 integrin but lack VEGFR2 do not respond to endorepellin treatment (66). Because binding of endorepellin was distal to the VEGFA binding site on the VEGFR2 ectodomain, we favor a model where endorepellin would act as an allosteric inhibitor of VEGFR2, independent of VEGFA concentrations. This binding most likely occurs via the two proximal LG1-LG2 domains, whereas LG3 would bind to the ␣2␤1 integrin. Functionally, endorepellin activates the Tyr phosphatase SHP-1 which is bound to the cytoplasmic domain of the ␣2␤1 integrin (59). SHP-1 then dephosphorylates VEGFR2, thereby blocking endothelial cell migration, survival, and proliferation (59). This dual-receptor binding leads to rapid internalization and degradation of both receptors which, together with deactivation of VEGFR2, evokes attenuated VEGFA production and secretion (66). These findings provide a new paradigm for anti-angiogenic fragments derived from large precursors, that is a "dual-receptor" antagonism.
In this study, we demonstrate that endorepellin attenuates VEGFA-induced activation of two major signaling pathways: the PI3K/PDK1/Akt/mTOR and the PKC/JNK/AP1 pathways. VEGFA synthesis and secretion via the transcriptional activity of HIF1A promoter gene were inhibited in response to endorepellin in an oxygen-independent manner. Furthermore, VEGFA-induced transcriptional activity of nuclear factor of activated T cell 1 (NFAT1) and AP1 promoter was suppressed by endorepellin. Subsequently, endorepellin was shown to inhibit VEGFA-evoked nuclear translocation of NFAT1 while concurrently promoting NFAT1 stability. Thus, our study enhances our overall understanding of the potential functional role of endorepellin as a potent inhibitor of angiogenesis and supports the concept of dual-receptor antagonism that endorepellin requires both the ␣2␤1 integrin and VEGFR2 for its downstream signaling. These findings provide a mechanistic explanation for endothelial cell specificity, the only cells that simultaneously express these two receptors, and predict that similar biological processes could be operational for other endogenous angiostatic fragments derived from larger protein precursors.
Immunoblotting and Immunoprecipitation-Following each treatment, endothelial cells were lysed in modified radioimmune precipitation assay buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM EDTA/EGTA/sodium vanadate, 10 mM ␤-glycerophosphate, and protease inhibitors: 1 mM phenylmethanesulfonyl fluoride and 10 g/ml leupeptin/tosylphenylalanyl chloromethyl ketone/aprotinin each) for 20 min on ice (73). Insoluble material was removed by centrifugation at 14,000 ϫ g, and protein levels were determined using the DC assay (Bio-Rad). For immunoprecipitation, protein A-Sepharose magnetic beads (GE Healthcare) were absorbed with antibodies for 4 h at 4°C, and precleared cell lysates were added to the beads for 18 h at 4°C. After extensive washing, the beads were boiled in reducing buffer, and supernatants were separated by SDS-polyacrylamide gel electrophoresis. Proteins were then transferred to nitrocellulose membranes (Bio-Rad), probed with the indicated antibodies, developed with enhanced chemiluminescence technique (Thermo), and detected using Image Quant LAS-4000 (GE Healthcare) as described previously (74).
Confocal Microscopy-Routinely, ϳ5 ϫ 10 4 HUVECs or PAE-VEGFR2 cells, grown on gelatin-coated 4-chamber slides (Nunc), were serum-starved for 2 h prior to treatment with PBS, VEGFA (2.6 nM), and endorepellin (100 nM). For experiments combining VEGFA and endorepellin, cells were first preincubated with endorepellin for 2 h. Cells were washed with PBS and fixed for 30 min in 4% paraformaldehyde at 4°C. Cells were blocked in PBS/5% BSA, incubated with various antibodies for 1 h, washed in PBS, and then incubated for 1 h with the appropriate secondary antibodies (goat anti-mouse IgG Alexa Fluor 488 or goat anti-rabbit IgG Alexa Fluor 594). Nuclei were visualized with DAPI (Vector Laboratories). Fluorescence images were acquired with a 63ϫ, 1.3 oil-immersion objective, using a Zeiss LSM-780 confocal laser scanning microscope. Merged images represent single optical sections (Ͻ0.8 m), collected with the pinhole set to 1 Airy Unit for the red channel, and adjusted to give the same optical slice thickness in the green and blue channels. Images were acquired in single confocal planes to determine co-localization precisely using the ZEN 2010 software, with filters set at 488/594 nm for dual-channel imaging, and Z-stacks acquired at 0.36-m intervals. All images were analyzed using ImageJ and Adobe Photoshop CS5.1 (Adobe Systems). To quantify co-localization of VEGFR2 and caveolin or clathrin further we utilized line scanning, a technique that measures the pixels along a singly defined axis along the specimen to determine localization of the differentially labeled fluorophores (75,76). The extent of overlap (defined as two different fluorescent labels displaying independent emission wavelengths that occupy the same pixel) between the two potentially interacting molecules serves as a qualitative assessment of a proximity-dependent localization (77).
Luciferase and Protein Kinase C (PKC) Assays-After washing cells with PBS, 150 l of lysis buffer (50 mM potassium phosphate buffer, 2% Triton X-100, 20% glycerol, 4 mM DTT) was added to each well. Cells were lysed at 25°C for 10 min, centrifuged at 2000 ϫ g for 2 min, and ϳ100 l of cleared cell lysate was dispensed into a 96-well ELISA plate, together with 100 l of luciferase assay buffer (100 mM potassium phosphate, 2 mM DTT, 8 mM MgSO 4 , 175 M coenzyme A, 750 M ATP) and 0.5 mM D-luciferin. A plate luminometer (PerkinElmer Life Sciences) was used for light measurement.
PKC activity was measured with a PepTag nonradioactive PKC assay (Promega) according to the manufacturers' instructions. Cells were treated with VEGFA (2.6 nM) and endorepellin (100 nM), washed with PBS, and lysed in PKC extraction buffer (25 mM Tris-HCl, pH 7.4, 0.5 mM EDTA, 0.5 mM EGTA, 0.05% Triton X-100, 10 nM ␤-mercaptoethanol, and leupeptin/aprotinin 1 mg/ml each). The precleared supernatants were assessed for PKC activity using a bright colored fluorescent peptide substrate highly specific for PKC. The enzyme alters the net charge of the peptide substrate from ϩ1 to Ϫ1, thereby allowing the phosphorylated and nonphosphorylated forms of the substrate to be separated on an agarose gel. The bands containing phosphorylated peptides were extracted from the gel, heated at 95°C, solubilized, acidified with glacial acetic acid, and finally evaluated by measuring the absorbance at 570 nm.
Statistical Analysis-All data were expressed as means Ϯ S.E. and statistically analyzed with Student's t test or paired t test using the Sigma-Stat software 11.0 (SPSS Inc). p Ͻ 0.05 was considered statistically significant.

Endorepellin Blocks Coupling of VEGFR2 to the PI3K Axis and Attenuates VEGFA-evoked PI3K/Akt Signaling in Endothe-
lial Cells-VEGFA-evoked activation of VEGFR2 induces autophosphorylation of VEGFR2 intracellular domain at Tyr-1175 which mediates downstream signaling via two major branches, PI3K and PLC␥ (78). Endorepellin initiates a signaling cascade by concurrently binding to VEGFR2 and the ␣2 I domain of the ␣2␤1 integrin (66). This leads to activation of SHP-1 and subsequent dephosphorylation of VEGFR2 at Tyr-1175 (59). Thus, we investigated whether any adaptor/signaling molecules known to bind to Tyr-1175 were disrupted by endorepellin. One of the key adaptor proteins linked to Tyr-1175 is Shb (79). Upon stimulation of PAE-VEGFR2 cells with VEGFA, Shb is phosphorylated and binds directly to Tyr-1175, which then mediates PI3K activity, stress fiber formation, and cell migration (79). We discovered that VEGFA-mediated Shb recruitment was attenuated by endorepellin treatment (Fig.  1, A and G, p Ͻ 0.01). Concurrently, endorepellin inhibited the recruitment of PI3K evoked by VEGFA (Fig. 1, A and F, p Ͻ 0.01).
Key downstream signaling components of the PI3K pathway include PDK1, a master kinase, which is crucial for the activation of Akt/PKB (80), and subsequent activation of eNOS and mTOR (78,81). Endorepellin alone evoked no significant changes in PDK1 phosphorylation at Ser-241, but inhibited VEGFA-evoked PDK1 phosphorylation (Fig. 1, B and H, p Ͻ 0.01). Moreover, endorepellin significantly suppressed Collectively, these results indicate that endorepellin, by directly interacting with the VEGFR2 ectodomain, antagonizes a major branch of this RTK signaling axis that controls survival, permeability, and migration (81).
Endorepellin Suppresses HIF-1␣ Transcriptional Activity Independent of Oxygen Concentration-Next, we investigated the role of downstream targets of mTOR. One of the established targets of mTOR is HIF-1␣, known to positively regulate VEGFA transcription (82). Because we did not detect any HIF-1␣ protein in HUVECs under normoxia (supplemental Fig. 1A) we utilized PAE cells overexpressing the human VEGFR2 (PAE-VEGFR2) (66,68). In these cells we found abundant HIF-1␣ under normoxia in contrast to either parental cells or cells expressing the human VEGFR1 (supplemental Fig. 1A). Thus, VEGFR2 overexpression can drive HIF-1␣ expression and/or enhance HIF-1␣ stability in these cells. HIF-1␣ levels were markedly down-regulated by endorepellin in a time-dependent fashion, in contrast to those of the constitutively expressed HIF-1␤ or von Hippel-Lindau protein (pVHL) ( Fig.  2A). Under normoxia, HIF-1␣ has a very short half-life (ϳ7 min), and its rapid turnover is mediated by hydroxylation of two proline residues (Pro-402 and Pro-564) by oxygen-dependent prolyl-hydroxylase (82). This leads to ubiquitination by pVHL E3 ubiquitin ligase and degradation via the 26 S proteasome (83). The lack of changes in pVHL levels, thus, suggests that endorepellin-mediated down-regulation of HIF-1␣ might be transcriptionally mediated. To address this issue directly, we treated PAE-VEGFR2 with lactacystin, an established proteasome inhibitor (84), with or without endorepellin. We found that HIF-1␣ levels were increased by lactacystin, an effect counteracted by endorepellin (Fig. 2B). Collectively, these data demonstrate not only a potent transcriptional inhibition of the HIF1A locus evoked by endorepellin, but also implicate posttranscriptional mechanisms to subdue HIF-1␣ to achieve angiostasis.
Next, we tested the VEGFR2 inhibitor and endorepellin in stably transfected PAE-VEGFR2 cells harboring a 2.6-kb genomic fragment encompassing the human VEGFA promoter cloned upstream of luciferase (66). In this case, endorepellin and SU5416 equally inhibited transcriptional activity of VEGFA promoter (p Ͻ 0.05, and p Ͻ 0.01, respectively; Fig. 2G). However, there was no significant additive effect when cells were incubated with both endorepellin and SU5416 (p ϭ 0.7; Fig.  2G). Concurrently, endorepellin inhibited secreted VEGFA levels under hypoxic conditions (p Ͻ 0.001, supplemental Fig. S2, C and D) comparable with levels achieved under normoxia (66).
To determine whether the decrease of HIF-1␣ protein levels detected above and those of VEGFA previously reported (66) were due to a decrease in transcription of HIF1A and VEGFA genes, we performed quantitative real-time PCR. In contrast to VHL, both HIF1A and VEGFA mRNA levels were significantly suppressed by endorepellin (p Ͻ 0.001, Fig. 2H). Collectively, these findings corroborate the view that endorepellin is a direct antagonist of VEGFR2 and requires an attenuation of its tyrosine kinase activity to evoke repression of HIF1A and VEGFA expression.
It is well established that the serine/threonine phosphatase calcineurin promotes HIF-1␣ expression by dephosphorylating receptor for activated C kinase 1 (RACK1) (82). This leads to inhibition of RACK1 dimerization and to a block of RACK1mediated ubiquitination and degradation of HIF-1␣ (86). Thus, we utilized INCA-6, a cell-permeable inhibitor of calcineurin activity (87). A 4-h incubation with INCA-6 using an optimal concentration (30 M) specific for calcineurin suppression (87) evoked inhibition of HIF-1␣ promoter activity even greater than endorepellin (p Ͻ 0.001, Fig. 3B). Combination of endorepellin and INCA-6 caused an even further suppression down to ϳ36% of control values (Fig. 3B). Similar effects were obtained with 100 nM cyclosporin A, an established inhibitor of calcineurin (88) (data not shown). Notably, INCA-6 also suppressed VEGFA promoter luciferase activity in a time-dependent manner (t1 ⁄ 2 ϳ3.5 h, Fig. 3C). Using a similar protocol we found that INCA-6 suppressed VEGFA levels in PAE-VEGFR2 cells and that combination of both treatments reduced VEGFA protein levels even further (Fig. 3D). Importantly, endorepellin had no repressive effects in addition to a combination of LY294002 and INCA-6 (Fig. 3E). These findings indicate that endorepellin affects both the PI3K/PDK1/Akt1/mTOR and the calcineurin pathways in down-regulating the transcription of HIF1A and VEGFA genes.
Endorepellin Inhibits VEGFA-evoked Recruitment of PLC␥ and Inhibits PKC/JNK/AP1 Pathway in Endothelial Cells-Because Tyr-1175 is the major binding site of PLC␥ on the VEGFR2 intracellular domain (78) and this residue is dephosphorylated by endorepellin-evoked activation of SHP-1 (59), we determined whether endorepellin would affect the VEGFAmediated recruitment of PLC␥ in HUVECs. Using co-immunoprecipitation we found that VEGFA-mediated PLC␥ recruitment was completely blocked by endorepellin treatment (Fig.  4A). We then investigated whether the activity of PKC, a known downstream effector of PLC␥, was also affected by endorepellin. We utilized a quantitative nonradioactive PKC assay, which measures total kinase ability to convert an artificial peptide substrate from ϩ1 to Ϫ1. In other words, upon phosphorylation of the synthetic PKC substrate by PKC, the overall net charge changes from ϩ1 to Ϫ1 to allow for a delineation of the phosphorylated and unphosphorylated forms. We found that endorepellin was capable of significantly blocking the VEGFAevoked PKC activity in HUVECs (p Ͻ 0.001, Fig. 4B). In agreement with these findings, endorepellin attenuated VEGFA-induced phosphorylation of c-Jun N-terminal kinase (JNK) at Thr-183/Tyr-185 (Fig. 4C).
Next, we determined whether AP1-luciferase endothelial cells would respond to phorbol 12-myristate 13-acetate (PMA), an analog of diacylglycerol which is released by PLC␥ and activates PKC. We found that AP1-dependent luciferase activity was induced by PMA, and endorepellin was capable of markedly suppressing this effect (p Ͻ 0.001, Fig. 4G). Consistent with these findings, treatment of these cells with Rottlerin, a PKC inhibitor, suppressed AP1 promoter activity, albeit there was an additive effect with endorepellin (supplemental Fig. S3B). Collectively, our findings implicate endorepellin in interfering with VEGFA activation of the PLC␥/PKC/JNK/AP1 signaling axis.
Endorepellin Inhibits VEGFA-evoked Nuclear Translocation of NFAT1 and Promotes NFAT1 Stability-One of the key transcription factors regulated by calcineurin is NFAT1 (89). VEGFA-evoked activation of calcineurin evokes dephosphorylation of NFAT1 at specific serine residues which leads to acti-vation of the protein and its subsequent translocation into the nuclei where NFAT1 induces several genes involved in growth control and angiogenesis (90). Thus, we hypothesized that endorepellin could interfere with this important pathway. Using confocal microscopy, we found that VEGFA evoked rapid and efficient NFAT1 translocation into the endothelial cell nuclei (Fig. 5B), in contrast to either vehicle or endorepellin (Fig. 5, A and C, respectively). However, a preincubation with endorepellin significantly blocked NFAT1 nuclear translocation (Fig. 5D). Quantification of NFAT1 nuclear translocation, based on the NFAT1 nuclear/cytoplasmic ratio, showed a significant (84 Ϯ 3%) inhibition by endorepellin (p Ͻ 0.01, Fig. 5E).
Next, we examined the intracellular levels of NFAT1 following treatment with the calcineurin inhibitor cyclosporin A (100 nM), endorepellin (100 nM), or the proteasome inhibitor lactacystin (10 M). We discovered that endorepellin caused a sta-bilization of the inactive form of NFAT1 (the upper band in Fig.  5F is the serine-phosphorylated inactive form of NFAT1). Both cyclosporin A and lactacystin stabilized NFAT1 in a similar fashion. In contrast, there was no effect on the levels of HDM2, the human homolog of the murine double-minute gene 2 (Fig.  5F), an E3 ubiquitin ligase that targets NFAT1 for proteasomal degradation (91), suggesting a transcriptional effect of endorepellin. To address this possibility directly, we utilized an NFAT-luciferase reporter construct that contains four tandemrepeated copies of NFAT/AP1 elements from the human IL2 gene cloned upstream of the minimal IFNB promoter (70). In these mass cultures of PAE-VEGFR2 NFAT-Luc cells, there was a marked induction of promoter luciferase activity by VEGFA and a marked suppression by preincubation with endorepellin (p Ͻ 0.05, Fig.  5G). Notably, endorepellin by itself significantly suppressed NFAT-luciferase promoter activity (p Ͻ 0.01, Fig. 5G). We conclude that endorepellin, by suppressing VEGFAevoked activation of calcineurin, interferes with NFAT1 nuclear translocation and evokes stabilization of the inactive NFAT1 species. Moreover, endorepellin-evoked suppression of calcineurin activity also affects the levels of NFAT1 transcriptional activity and downstream events such as VEGFA levels (compare Fig. 3).
Endorepellin Induces Physical Down-regulation of VEGFR2 by Evoking Internalization via Caveosomes-We have previously shown that endorepellin causes a physical down-regulation of both ␣2␤1 integrin and VEGFR2 in HUVECs and PAE-VEGFR2 cells (66). Thus, we hypothesized that a possible mechanism for VEGFR2 down-regulation by endorepellin would be by inducing receptor internalization via caveosomes, a process that leads to degradation of several RTKs in contrast to clathrin-mediated internalization which allows for RTK downstream signaling and recycling to the cell surface (92). Using confocal microscopy, we discovered a progressive co-localization of VEGFR2 and caveolin-1 in HUVECs exposed to endorepellin (Fig. 6, C and E), but not in vehicle-exposed cells (Fig. 6A). By 30 min, large perinuclear vesicles were detectable in the endorepellin-treated cells (Fig. 6E). Co-localization of caveolin and VEGFR2 was further proven by using Z-stacks of multiple exposures (Fig. 6, C and E). To further prove co-localization, we generated line scanning profiles of fluorescence intensities (Fig. 6, D and F). These fluorescence intensity profiles display the intensity distribution of the confocal images along a straight line drawn in a specific region within the cells (i.e. between the white arrows in Fig. 6, A, C, and E, respectively). After endorepellin treatment, we observed a strong overlap between the VEGFR2 (red) and caveolin-1 (green) fluorescence peaks (Fig. 6, D and F) compared with controls (Fig.  6B). This clearly demonstrates spatial overlap of individual puncta positive for caveolin-1 and VEGFR2. A physical interaction between caveolin-1 and VEGFR2 following endorepellin treatment for 30 min was further corroborated by co-immunoprecipitation of the two proteins (Fig. 6G). In contrast, we found neither co-immunoprecipitation of VEGFR2 with clathrin (Fig. 6G) nor co-localization of VEGFR2 and clathrin following endorepellin treatment (supplemental Fig. S4) using both Z-stacks and line scanning profiles. Comparable results were obtained using PAE-VEGFR2 cells (supplemental Fig. S5). Thus, endorepellin leads to VEGFR2 internalization and degradation predominantly via a caveosome-mediated pathway.
We have previously established (66) corroborating evidence that clearly demarcates a time-dependent (0 -40 min) decrease in total VEGFR2 protein with endorepellin (100 nM). These data reinforce the confocal co-localization microscopy presented herein.

DISCUSSION
Our central hypothesis is that perlecan and its bioactive parts affect endothelial cell behavior by simultaneously engaging adhesion and angiogenic receptors, such as the ␣2␤1 integrin and VEGFR2, respectively. According to our current working model of dual-receptor antagonism (Fig. 7), endorepellin would act as an allosteric inhibitor of VEGFR2 by binding via LG1-LG2 domains to a region distal to the canonical binding site for the natural ligand VEGFA (i.e. Ig 2-3 ). The terminal LG3 domain would bind to the ␣2␤1 integrin thus bringing together the two receptors, perhaps in a multicomplex signaling apparatus at the cell surface of endothelial cells. This dual-receptor binding leads to rapid internalization and degradation of both recep-tors, a process that can be inhibited by integrin ␣2␤1-blocking antibodies (59) and mediated by caveolin-positive and clathrinnegative endocytic compartments. Biochemically, the engagement of the two receptors by endorepellin evokes the activity of SHP-1, a powerful tyrosine phosphatase that is recruited to the tail of the ␣2 subunit of ␣2␤1 integrin and rapidly dephosphorylates several RTKs including VEGFR2 at Tyr-1175 (59). The longstanding implications of the current work expand on dualreceptor antagonism by examining the subsequent consequences on various signaling pathways directly orchestrated by VEGFR2 while discovering novel signaling pathways attenuated by endorepellin. We further discovered that endorepellin attenuates activation of the PI3K/PDK1/Akt/mTOR, calcineurin pathway, and by preventing PLC␥ recruitment, precludes PKC/JNK/AP1 activation. These effects are attributable to endorepellin blocking VEGFA-evoked signaling through VEGFR2. Further corroborating these observations is the ability of endorepellin to inhibit VEGFA expression as a direct function of VEGFR2 antagonism leading to receptor internalization via caveosomes coincident with reducing HIF-1␣ under  DECEMBER 21, 2012 • VOLUME 287 • NUMBER 52 normoxia via transcriptional and post-transcriptional mechanisms. Moreover, consistent with a blockade of the calcineurin pathway, endorepellin prevents VEGFA-dependent nuclear translocation of NFAT with further stabilization of the inactive species. Therefore, collectively these results detail suppression of two major signaling branches culminating in an attenuation of VEGFA and HIF1A expression while stabilizing inactive NFAT1 (Fig. 7). Thus, endorepellin, by acting at the receptor level, is capable of emanating diverse signaling events that converge onto repression of VEGFA transcriptional activity. We should point out that our current working model has a dynamic multiscale nature, i.e. it is based on experiments performed for short time (10 -30 min) to address proximal receptor activity and for several hours for addressing distal transcriptional effects. Interestingly, other anti-angiogenic inhibitors, such as thrombospondin-1 (93) and TIMP-2 (94), can also modulate VEGFR2 activity and other integrin receptors, suggesting that dual-receptor antagonism might be a general biological process shared by other angiostatic proteins.

Endorepellin Antagonizes VEGFR2 Activity
The discovery that endorepellin evokes a general dephosphorylation of VEGFR2 (59) and specifically at Tyr-1175 (66) is intriguing insofar as Tyr-1175 may be the most important residue for proper VEGFR2 signaling in endothelial cells (78,81).
Notably, mutation of mouse Tyr-1173 to Phe (corresponding to human Tyr-1175 in VEGFR2) results in a loss-of-function phenotype and embryonic lethality, indicating that signaling via Tyr-1175 is essential for VEGFR2 functions during vasculogenesis (95). The adaptor molecule Shb binds to Tyr-1175 and is required for VEGFA-induced activation of PI3K signaling in endothelial cells (78). VEGFA evokes autophosphorylation of VEGFR2 at Tyr-1175 by recruiting Shb to the intracellular domain of the receptor (78). However, endorepellin abrogates this recruitment and attenuates downstream signaling via PI3K. It is well established that both the p110␣ catalytic and p85 regulatory subunits of PI3K are directly involved in mediating VEGFA/VEGFR2 signaling (78). Hence, we sought to determine whether these two subunits would directly associate with VEGFR2 upon VEGFA treatment. We observed that both the p110␣ and p85 subunits could be co-precipitated in complex with VEGFR2 in response to VEGFA and this "association" was significantly attenuated by endorepellin, with concurrent suppression of Akt phosphorylation.
Gene-targeting studies using eNos Ϫ/Ϫ mice have shown eNOS plays a key role in VEGFA-induced angiogenesis and vascular permeability in a PI3K/Akt-dependent manner (96). Indeed, VEGFA induces NO production via eNOS phosphoryl-FIGURE 6. Endorepellin evokes VEGFR2 internalization and co-localization with caveolin 1. A, C, and E, representative confocal images of HUVECs before or after treatment with endorepellin (100 nM) for 10 and 30 min as indicated. The bottom panels show Z-stack projections (63ϫ oil objective) with XZ orthogonal views. The cells were permeabilized with 0.1% Triton X-100 for 3 min and dually labeled for VEGFR2 (red) and caveolin-1 (green). Nuclei appear blue after DAPI staining. Note the progressive co-localization of VEGFR2 within caveolin-positive intracellular vesicles. All images were captured with the same exposure, gain, and intensity. Scale bars, ϳ10 m. B, D, and F, analysis of co-localization of VEGFR2 (red) and caveolin-1 (green) profiles with or without endorepellin treatment. The line scanned profiles show the distribution of fluorescence for each channel between the white arrows of A, C, and D, respectively. Notice that in the absence of endorepellin, fluorescence intensity peaks of VEGFR2 and caveolin do not superimpose, whereas in the presence of endorepellin there is strong superimposition of the two lines scans, indicating co-localization. G, representative co-immunoprecipitation (IP) and Western blotting (WB) of HUVECs Ϯ endorepellin (100 nM) for 30 min. The cells were lysed, immunoprecipitated with an anti-VEGFR2 antibody, and the proteins separated on an SDS-polyacrylamide gel. The immunoblot was then probed with anti-VEGFR2, anti-clathrin, and anti-caveolin antibodies, as indicated.
ation at Ser-1177 by Akt, a process that is blocked by PI3K inhibitors (96). We found that endorepellin inhibited the VEGFA-induced phosphorylation of eNOS at Ser-1177 and phosphorylation of mTOR at Ser-2448. This activity is notable because mTOR regulates the expression of HIF-1␣, a DNAbinding protein that promotes adaptation and survival under hypoxia by inducing the transcription of numerous genes, including VEGFA (82). Here, we discovered that VEGFR2 overexpression was capable of driving HIF-1␣ expression and/or enhancing HIF-1␣ stability in PAE-VEGFR2 cells and that endorepellin markedly down-regulated HIF-1␣ levels under normoxic and hypoxic conditions. In contrast, the levels of the constitutively expressed HIF-1␤ or those of pVHL were not affected by endorepellin, suggesting a transcriptional control mechanism. Indeed, endorepellin caused a dose-and time-dependent suppression of HIF1A promoter activity. In agreement with these findings, endorepellin and a VEGFR2 kinase inhibitor equally inhibited VEGFA promoter-driven transcriptional activity without any additive effects. Thus, endorepellin is a direct antagonist of VEGFR2 and requires this RTK to evoke transcriptional repression of HIF1A and VEGFA genes.
We noticed that blocking the PI3K pathway with LY294002 did not completely block the downstream effects of endorepellin. Thus, we investigated whether endorepellin could affect calcineurin, a serine/threonine phosphatase that enhances HIF-1␣ levels by dephosphorylating RACK1, inhibiting RACK1 dimerization, blocking RACK1-mediated ubiquitination and degradation of HIF-1␣ (82). We discovered that blocking calcineurin activity with either cyclosporine A (88) or INCA-6 (87) down-regulated HIF1A promoter activity, and a combination of endorepellin and INCA-6 caused an even further suppression.
INCA-6 also suppressed VEGFA-promoter luciferase activity in a time-dependent manner, and this correlated with a decline in VEGFA protein. Importantly, endorepellin could not inhibit beyond the inhibition observed for the combination of both LY294002 and INCA-6, indicating that endorepellin affects both the PI3K/PDK1/Akt1/mTOR and the calcineurin pathways in down-regulating the transcription of HIF1A and VEGFA genes.
Calcineurin is a serine/threonine calcium-dependent phosphatase that dephosphorylates NFAT1 and causes its nuclear translocation where NFAT1 activates many genes including HIF1A (89). In this study, we found that VEGFA induced NFAT1 translocation into endothelial cell nuclei, and endorepellin efficiently counteracted this activity. Moreover, endorepellin induced stabilization of the inactive (phosphorylated) species of NFAT1 and blocked VEGFA-evoked transcriptional activity of NFAT1-promoter luciferase.
We would like to propose a unifying hypothesis for endorepellin activity: we favor a scenario where endorepellin would concurrently evoke receptor degradation and trigger a negative signaling emanating from the VEGFR2 itself. First, we discovered that endorepellin evoked VEGFR2 intracellular internalization primarily via caveosome-mediated endocytosis, FIGURE 7. Working model of endorepellin as a dual-receptor antagonist. Endorepellin could act as an allosteric inhibitor of VEGFR2 by binding via LG1-LG2 domains to a region distal to the canonical binding site of the natural ligand VEGFA (i.e. Ig 2-3 ). The terminal LG3 domain would bind to the ␣2␤1 integrin and could bring together the two receptors. This dual receptor binding leads to rapid internalization and degradation of both receptors, activation of SHP-1, and dephosphorylation of key Tyr residues in the VEGFR2 intracellular domain, importantly Tyr-1175. This biological process causes downstream attenuation of two main signaling axes initiated by engagement of PI3K and PLC␥ to the Tyr-1175 of VEGFR2 intracellular domain. This leads to suppression of HIF1A, NFAT1, and AP1 transcriptional activities which negatively affect VEGFA transcription. DECEMBER 21, 2012 • VOLUME 287 • NUMBER 52 in contrast to VEGFA that primarily evokes a clathrin-mediated endocytosis and recycling (96). Physical down-regulation of this RTK, essential for endothelial cell survival and motility/ proliferation, provides an attractive explanation for a mechanism to control pathological angiogenesis that could be applicable to other matrix-derived angiogenic inhibitors. Our findings are supported by recent studies that showed a dual receptor antagonism of engineered VEGFA variants that could simultaneously bind VEGFR2 and ␣v␤3 integrin to elicit a robust anti-angiogenic effect (97). Second, endorepellin by binding distally (Ig 3-7 ) to the canonical VEGFA-binding (Ig 2-3 ) site on the VEGFR2 could act as an allosteric inhibitor, independent of VEGFA concentrations. This binding could prevent dimerization as shown recently for monoclonal antibodies targeting the non-ligand binding site of VEGFR2/3 and preventing homo-and heterodimerization, signal transduction, and microvascular sprouting (98). In conclusion, this would negatively affect endothelial cell biology and lead to a profound antiangiogenic effect. Therefore, endorepellin and perhaps similar compounds targeting the noncanonical ligand binding site of VEGFR2 could be used together with established therapies, such as antibodies toward the ligand or receptor, as a novel anti-angiogenic approach in preclinical studies or directly in clinical use.