HCPTPA, a Protein Tyrosine Phosphatase That Regulates Vascular Endothelial Growth Factor Receptor-mediated Signal Transduction and Biological Activity*

Angiogenesis is a tightly controlled process in which signaling by the receptors for vascular endothelial growth factor (VEGF) plays a key role. In order to define signaling pathways downstream of VEGF receptors (VEGFR), the kinase domain of VEGFR2 (Flk-1) was used as a bait to screen a human fetal heart library in the yeast two-hybrid system. One of the signaling molecules identified in this effort was HCPTPA, a low molecular weight, cytoplasmic protein tyrosine phosphatase. Although HCPTPA possesses no identifiable phosphotyrosine binding domains (i.e. SH2 or phosphotyrosine binding domains), it bound specifically to active, autophosphorylated VEGFR2 but not to a mutated, kinase-inactive VEGFR2. Recombinant VEGFR2 and endogenous VEGFR2 were substrates for recombinant HCPTPA, and HCPTPA was co-expressed with VEGFR2 in endothelial cell lines, suggesting that HCPTPA may be a negative regulator of VEGFR2 signal transduction. To pursue this possibility, an adenovirus directing the expression of HCPTPA was constructed. When used to infect cultured endothelial cells, this adenovirus directed high level expression of HCPTPA that resulted in impairment of VEGF-mediated VEGFR2 autophosphorylation and mitogen-activated protein kinase activation. Adenovirus-mediated overexpression of HCPTPA also inhibited VEGF-induced cellular responses (endothelial cell migration and proliferation) and inhibited angiogenesis in the rat aortic ring assay. Taken together, these findings indicate that HCPTPA may be an important regulator of VEGF-mediated signaling and biological activity. Potential interactions with other signaling pathways and possible therapeutic implications are discussed.

Angiogenesis in adult tissues is a tightly controlled process.
Typically, endothelial cells in the adult vasculature turn over extremely slowly, and the growth of new blood vessels occurs only in rare settings, such as during wound healing and during cyclic changes in the ovary and uterus. Recently, however, it has become increasingly clear that the tight controls on vascular growth in adult tissues can be breached in pathologic states such as cancer, arthritis, and diabetic retinopathy (1). Because uncontrolled vascular growth is now known to play a role in the development of these so-called "angiogenic" diseases, an intense effort has developed to determine the molecular and cellular mechanisms that control vascular growth. A recent flurry of reports has underscored the importance of signaling by endothelial receptor tyrosine kinases in vascular development. In particular, two families of endothelium-specific receptor tyrosine kinases, the VEGF receptor family and the novel Tie family of receptors, have been shown to play crucial, nonredundant roles in vascular assembly during embryonic development (2)(3)(4). Moreover, signaling by both the VEGF 1 receptor family and the Tie family of receptors has been shown to play important roles in pathologic angiogenesis in adult tissues (5)(6)(7)(8)(9)(10).
Clearly, to understand fully the precise nature of the roles of receptor tyrosine kinases in vascular development, it will be necessary to understand the signaling pathways that drive the cellular responses downstream of these receptors as well as the mechanisms by which receptor activation is modulated. In this report, the kinase domain of one of the VEGF receptors, VEGFR2 (Flk-1/KDR), was used as a bait in the yeast twohybrid system to screen for candidate signaling molecules that lie in the VEGF pathway (11)(12)(13). One of the molecules identified in this screen was a low molecular weight protein tyrosine phosphatase, HCPTPA (14). The VEGFR2 kinase domain was a substrate for HCPTPA both in vitro and in vivo, and when overexpressed, HCPTPA inhibited VEGF-mediated responses in cultured endothelial cells. Furthermore, overexpression of HCPTPA inhibited angiogenesis in the rat aortic ring assay, a model of angiogenesis that has been shown to be dependent on VEGF receptor signaling (15). Taken together, these findings suggest that HCPTPA may be an important * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 MN). Anti-Flk-1 antibodies (C-20 and 1158), anti-phosphotyrosine antibody (4G10 or Py20), and protein A-agarose were purchased from Upstate Biotechnology or Santa Cruz Biotechnology. Anti-phospho-p44/42 MAP kinase and anti-p44/42 MAP kinase antibodies were purchased from New England Biolabs. The anti-HCPTPA antibody was developed as described previously (14).
Cell Culture-Primary cultured human umbilical vein endothelial cells (HUVEC) were obtained from Clonetics Corp. (San Diego, CA). ECRF 24, an immortalized HUVEC cell line, was provided by Dr. Hans Pannekoek (University of Amsterdam). Endothelial cells were maintained at 37°C, 5% CO 2 in endothelial growth medium (EGM, Clonetics Corp.) and grown on 2% gelatin-coated (Sigma) plates. Endothelial cells were serum-starved in EBM (Clonetics Corp.). Another endothelial cell line, EA.hy926, was provided by Cora-Jean Edgell (University of North Carolina) and was maintained in DMEM supplemented with 10% fetal bovine serum and 1 mM hypoxanthine, aminopterin, and thymidine (Sigma).
Yeast Two-hybrid Screening-All yeast plasmid vectors and yeast strains were generously provided by Roger Brent (Harvard University). For generation of bait plasmids, a cDNA encoding the entire intracellular domain of murine VEGFR2 (Flk-1) was cloned by PCR from reverse-transcribed mouse embryo mRNA. To obtain a kinase-inactive VEGFR2, site-directed mutagenesis was performed using a commercially available kit (Transformer; CLONTECH) to mutate lysine 866 to arginine (Flk-1KR). Both the wild-type and the mutated VEGFR2 kinases were subcloned into the yeast bait vector, pJK-202, downstream of the LexA DNA binding domain to create plasmids encoding a LexA/ VEGFR2 fusion protein (pJK-Flk-1 and pJK-Flk-1KR). For library screening, pJK-Flk-1 was transformed into yeast strain EGY191/ pJK103; yeast expressing the LexA-Flk-1 bait were then transformed with a human fetal heart plasmid (pJG4-5) library. Library proteins that interacted with the intracellular domain of Flk-1 were selected and tested for specificity essentially as described previously (16).
Expression and Purification of Recombinant Proteins-Recombinant Flk-1 kinase was expressed as a carboxyl-terminal fusion with glutathione S-transferase (GST) in Sf9 insect cells using the baculovirus system as described previously (17). The SHC SH2 domain (SHC-SH2) was expressed as a carboxyl-terminal GST fusion protein in Escherichia coli BL21 (FϪ, ompT, rBϪ mBϪ) strain. GST fusion proteins were purified from clarified Sf9 cells or E. coli lysates by glutathione-Sepharose 4B (Amersham Pharmacia Biotech) chromatography. Recombinant HCPTPA was produced in E. coli as described previously (14).
Receptor Dephosphorylation Assay-Briefly, GST-Flk-1 (10 g) was immobilized on 20 l of glutathione-Sepharose beads. Subsequently, the beads were incubated for 10 min at room temperature in 50 l of kinase buffer (20 mM Tris-HCl, pH 7.5, 0.1 M NaCl, 12 mM MgCl, 1 mM dithiothreitol), 50 Ci of [␥-32 P]ATP. After thorough washing in 1ϫ kinase buffer, labeled protein was eluted with 10 mM glutathione in 20 mM Tris-HCl, pH 8.0. Radiolabeled Flk-1 was then aliquoted and incubated with 4 g of BSA or increasing amounts of HCPTPA at 30°C for 10 min. The protein mixture was then separated by 8% SDS-PAGE, analyzed by autoradiography, and quantified on a Molecular Dynamics PhosphorImager.
For assaying dephosphorylation of endogenous VEGFR2 by HCPTPA, ECRF24 cells were quiesced in EBM medium (Clonetics) for 24 h and then incubated in the presence or absence of VEGF (25 ng/ml) for another 5 min at 37°C. Following VEGF stimulation, the cells were lysed, and VEFGR2 was immunoprecipitated using a commercially available anti-Flk-1 antibody (Santa Cruz Biotechnology) and immobilized on protein A-agarose. Protein A-agarose beads were washed and incubated in the absence or presence of purified recombinant HCPTPA at 30°C for 10 min. The reaction was stopped by adding 2ϫ sample buffer and boiling for 5 min. Samples were divided into 2 aliquots, separated by 8% PAGE, transferred to nitrocellulose, and probed with either an anti-phosphotyrosine antibody or the anti-Flk-1 antibody.
Association Assay-GST-Flk-1 fusion protein was immobilized on glutathione-Sepharose beads, autophosphorylated in vitro, and eluted from the beads using thrombin to cleave the Flk-1 kinase from the GST moiety. Following thrombin treatment, soluble Flk-1 was isolated by pelleting the agarose beads and harvesting the supernatant. To test the effect of dephosphorylation of VEGFR2 on substrate association, GST-SHC-SH2 was immobilized on glutathione-agarose beads followed by the addition of Flk-1 kinase (50 ng) that had been either untreated or treated with HCPTPA. The mixture was incubated for 1 h at 4°C, and the beads were washed thoroughly with 1ϫ PBS to remove unbound Flk-1 kinase. Protein complexes were then eluted with 10 mM glutathione, separated by SDS-PAGE, and analyzed by serial blotting with anti-Flk-1 and anti-Tyr(P).
Adenovirus Construction-To construct an adenovirus directing the expression of HCPTPA (AdHCPTP), a cDNA encoding the full-length HCPTPA fused at its amino terminus with an influenza hemagglutinin (HA) epitope tag was subcloned into an adenoviral transfer vector (18). After excision from the transfer vector, the cDNA encoding the HCPTPA/HA fusion protein and the 5Ј adenoviral sequences was directly ligated to DNA encoding an E1/E3 deleted adenoviral genome, and the ligation product was transfected into 293 cells. AdHCPTP from the transfected 293 cells was purified by banding twice on CsCl gradients; viral titers were determined by optical densitometry, and recombinant virus was stored in 10% (v/v) glycerol at Ϫ20°C. AdGFP, which directs the expression of green fluorescent protein (GFP), was also grown and purified as described above and used as a control virus. This virus was a gift from Dr. Peter Corry (William Beaumont Hospital, Royal Oak, MI).
Virus Infection-HUVECs were grown on 10-cm tissue culture plates in EGM. When confluent, the medium was changed to minimum essential medium containing 2% fetal bovine serum, and AdHCPTP or AdGFP viral stocks were added to the medium at a multiplicity of infection of 10. After 24 h, the cells were placed in quiescence medium (EBM) for 24 h before treatment with 10 ng/ml VEGF. Cells infected with AdGFP were visualized under a fluorescence microscope for expression of GFP to determine successful virus infection.
Immunoprecipitation and Western Blotting-AdHCPTP-and AdGFP-infected HUVECs in confluent monolayers were quiesced in EBM and then treated with VEGF or left untreated for 8 min. The cells were lysed in RIPA buffer (1% Nonidet P-40, 0.05% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 0.05 mM NaF, 1 mM sodium orthovanadate, 5 mM benzamidine, and 1 mM EDTA containing the protease inhibitors phenylmethylsulfonyl fluoride, aprotinin, and leupeptin). Insoluble debris was removed from the cell lysates by centrifugation at 14,000 ϫ g for 10 min at 4°C. The protein extracts were subjected to immunoprecipitation with anti-Flk-1 antibody (Santa Cruz Biotechnology, C1158). The immunoprecipitated proteins were resolved by 6% SDS-PAGE, transferred to nitrocellulose, and probed with anti-phosphotyrosine antibody to detect VEGFR2 phosphorylation. The blots were subsequently stripped and reprobed with anti-VEGFR2 antibody (Santa Cruz Biotechnology, C20) to determine equal loading. The protein extracts were also immunoprecipitated with anti-HCPTPA antibody to detect expression of the recombinant HCPTPA. To assess MAP kinase activation, 20 g of the protein lysates was resolved by 10% SDS-PAGE, transferred to nitrocellulose membrane, and incubated with anti-phospho-p44/42 MAP kinase antibody. The blots were subsequently stripped and reprobed with anti-p44/ p42 MAP kinase antibody to ensure equal loading.
Cell Proliferation Assay-AdHCPTP-and AdGFP-infected HUVECs were plated in 24-well plates at a density of 25,000 cells/well. After 24 h, the medium was replaced by EBM, and cells were incubated for an additional 24 h. The medium was then replaced with either fresh EBM alone or EBM containing VEGF (10 ng/ml). The cells were incubated for approximately 15 h, followed by a 3-h pulse labeling with 2 Ci/ml of [ 3 H]thymidine (Amersham Pharmacia Biotech). The reaction was stopped by aspirating the medium and washing the cells with Hanks' balanced salt solution (Life Technologies, Inc.). The DNA was precipitated with cold 10% trichloroacetic acid at 4°C for 30 min, followed by an absolute ethanol wash. The precipitated material was resuspended in 0.5 ml of 0.5 M NaOH, and [ 3 H]thymidine incorporation was determined from a 450-l aliquot using a Beckman LS6000SC scintillation counter. Eight wells were set up for each experimental condition.
Cell Migration Assays-The rate of migration of HUVECs was determined by using a modified Boyden chamber assay as described by Clyman et al. (19). Briefly, polycarbonate filter wells (Costar Transwell, 8-m pore size) were coated with 2% gelatin in PBS for 30 min at room temperature and then incubated with DMEM containing 0.1% BSA (DMEM/BSA). AdHCPTP-and AdGFP-infected HUVECs were trypsinized, pelleted by centrifugation, and resuspended in fresh DMEM/BSA to a final concentration of 2 ϫ 10 6 cells/ml. Aliquots of the cells (1 ϫ 10 5 ) were applied to the upper chamber of the filter wells. The filter inserts with cells were placed in wells of a 24-well culture plate containing either 600 l of DMEM/BSA alone or DMEM/BSA plus 10 ng/ml VEGF. After a 4-h incubation at 37°C, the cells that had migrated to the lower surface of the filter inserts were fixed with 10% formalin (Fisher) and stained with Harris' hematoxylin (Fisher). Six randomly selected high power (ϫ 400) fields were counted on each filter.
Rat Aortic Ring Assay-This assay was done as described by Nicosia et al. (15) with some modification. Thoracic aortas were excised from 2to 3-month-old Fisher 344 female rats. Clotted blood inside the aorta was flushed with media, and periadventitial fibroadipose tissue was removed. Aortas were then cut into small cross-sectional pieces approximately 1 mm in length. Pieces of aorta were infected with adenoviruses (AdGFP or AdHCPTP) at 1 ϫ 10 8 plaque-forming units in 200 l of DMEM at room temperature for 10 min. After viral infection, one piece of aorta was placed in each well of a 48-well plate and overlaid with 150 l of a freshly prepared fibrinogen solution (3 mg/ml) containing of ⑀-amino-n-caproic acid (6 g/ml, Sigma). The fibrin gel was polymerized by the addition of 15 l of thrombin (1 unit/ml; Sigma), for 1 to 2 h at room temperature before adding 0.5 ml of EGM plus ⑀-amino-n-caproic acid. The culture was maintained at 37°C with 5% CO 2 , changing media every other day. Rings were examined grossly each day, and digital images (ϫ 400) were obtained for quantitative analysis of endothelial sprouts on day 8. The number of sprouts was determined by counting the number of tube-like structures originating directly from the aortic ring (exclusive of branches). Sprout length was measured from the images using a calibrated micrometer. A two-tailed Student's t test was used to analyze statistical differences between AdGFP-and AdHCPTP-treated groups. Differences were considered statistically significant at p Ͻ 0.05.

Direct Interaction of VEGFR2 and HCPTPA in the Yeast
Two-hybrid System-In an effort to identify signaling molecules that associate with VEGFR2, the entire intracellular region of murine VEGFR2 was used as a bait to screen a human fetal heart library in the yeast two-hybrid system. One of the clones identified in this screen encoded a small molecular weight protein tyrosine phosphatase, HCPTPA. Demonstrating the specificity of the interaction, HCPTPA interacted with Flk-1, but it did not interact with a series of unrelated baits (data not shown). Although HCPTPA contains no identifiable phosphotyrosine binding motifs, previous studies have shown that HCPTPA can associate with phosphotyrosine residues on the PDGF receptor via its catalytic domain (20). Consistent with this notion, the HCPTPA clone encompassed the entire catalytic domain and interacted with wild-type Flk-1 but failed to interact with a bait encoding a mutated, inactive VEGFR2 kinase domain (Fig. 1). Importantly, the yeast two-hybrid system used in this study has previously been shown to be capable of detecting low affinity interactions such as enzyme-substrate interactions (21).
The Autophosphorylated VEGFR2 Kinase Domain Is a Substrate for HCPTPA-Knowing that HCPTPA associated with VEGFR2 in a phosphotyrosine-dependent manner, the next question addressed was whether or not VEGFR2 could serve as a substrate for HCPTPA. When a recombinant GST-VEGFR2 kinase domain was autophosphorylated in vitro and incubated with recombinant HCPTPA, the VEGFR2 kinase domain was dephosphorylated by HCPTPA in a concentration-dependent manner ( Fig. 2A). Preincubation of VEGFR2 kinase domain with HCPTPA not only markedly reduced autophosphorylation (Fig. 2B, lanes 1 and 2) but also markedly reduced the association of the VEGFR2 kinase domain with the SH2 domain of SHC, a signaling molecule known to associate with VEGFR2 in vivo (Fig. 2B, lanes 3 and 4).
Although recombinant, autophosphorylated VEGFR2 kinase domain was a substrate for recombinant HCPTPA in vitro, it remained possible that the endogenous VEGFR2 would not be a substrate for HCPTPA in vivo. To illustrate the potential for an HCPTPA-VEGFR2 interaction in vivo, endogenous HCPTPA expression could be easily detected in two endothelial cell lines that also express VEGFR2 (Fig. 3A and data not shown). In addition, as with the in vitro phosphorylated VEGFR2 kinase, in vivo phosphorylated VEGFR2 from endothelial cells was dramatically dephosphorylated after incuba-  2. Recombinant VEGFR2 is a substrate for HCPTPA in vitro. A, recombinant VEGFR2 is dephosphorylated by recombinant HCPTPA in a concentration-dependent manner. Recombinant GST-VEGFR2 kinase was autophosphorylated in the presence of [ 32 P]ATP as described under "Experimental Procedures" and then incubated with increasing amounts of recombinant HCPTPA. Proteins were resolved by PAGE, and VEGFR2 phosphorylation was determined by autoradiography. B, dephosphorylation by HCPTPA reduces the association of VEGFR2 with the SH2 domain of SHC. Recombinant GST-VEGFR2 kinase was immobilized on glutathione-agarose, treated with HCPTPA, or left untreated and then eluted with thrombin to remove the GST tag. Aliquots of HCPTPA-treated or untreated kinase were probed by immunoblot with anti-FLK-1 and anti-Tyr(P) antibodies (lanes 1 and 2, respectively). Other aliquots of HCPTPA-treated or untreated VEGFR2 kinase were incubated with the GST-tagged SH2 domain of SHC immobilized on glutathione-agarose. After thorough washing, VEGFR2/SHC-SH2 domain complexes were eluted in sample buffer and analyzed by immunoblot with anti-Flk-1 and anti-Tyr(P) antibodies (lanes 3 and 4). tion with recombinant HCPTPA (Fig. 3B). Because HCPTPA does display phosphotyrosine substrate specificity, this was an important finding since the tyrosine residues of VEGFR2 that are phosphorylated in vitro may not be the same as those phosphorylated in vivo (22,23).
HCPTPA Inhibits VEGFR2 Autophosphorylation, Signaling, and Biological Responses in Cultured Endothelial Cells-By having shown that VEGFR2, phosphorylated either in vitro or in vivo, could serve as a substrate for HCPTPA, the next question was whether or not HCPTPA was a downstream effector of VEGFR2 signaling in intact endothelial cells. To address this question, an adenoviral vector was generated to direct the overexpression of an HA-tagged HCPTPA (AdH-CPTP). When used to infect cultured endothelial cells (ECRF24 cells), AdHCPTP directed easily detectable expression of HAtagged HCPTPA (Fig. 4A). Compared with a control adenovirus (AdGFP), AdHCPTPA induced a marked decrease in VEGFmediated autophosphorylation of VEGFR2 (Fig. 4A). In addition to blocking VEGFR2 autophosphorylation in ECRF24 cells, overexpression of HCPTPA blocked VEGF-mediated activation of MAP kinase in primary endothelial cells (Fig. 4B). Moreover, overexpression of HCPTPA dramatically blunted VEGF-mediated survival/proliferation and migration in primary endothelial cells (Fig. 5, A and B). These findings demonstrate that HCPTPA may be a negative regulator of VEGFR2 activation, signaling, and biological responses.
HCPTPA Inhibits Angiogenesis in the Rat Aortic Ring Assay-Since overexpression of HCPTPA inhibited VEGF-mediated signaling and cellular responses in cultured endothelial cells, the next step was to determine whether overexpression of HCPTPA could inhibit angiogenesis. The rat aortic ring model recapitulates the entire process of angiogenesis and is known to be driven, at least in part, via the VEGF pathway (15, 24 -27).
To test the effect of overexpression of HCPTPA on angiogenesis in the rat aortic ring model, segments of rat aorta were incubated with either AdGFP or AdHCPTP and embedded in fibrin gels. After 7-10 days, it was clear that the AdGFP-infected rings demonstrated a luxuriant development of vascular sprouts (Fig. 6A). In contrast, AdHCPTP-infected rings produced few, if any, well formed vascular sprouts (Fig. 6B). Consistent with these gross observations, when sprout formation was quantified, infection of rat aortic rings with AdHCPTP resulted in a dramatic reduction in both the number of sprouts and the length of sprouts compared with the control AdGFP (Fig. 6, C and D). DISCUSSION VEGF receptor signaling plays an important role in angiogenesis, a process known to be required for embryonic development and for the development and progression of many clinically important diseases. The present study demonstrates that VEGFR2 associates with, and is a substrate for, a low molecular weight protein tyrosine phosphatase, HCPTPA. Overexpression of HCPTPA in cultured endothelial cells impaired VEGFR2 activation, inhibited VEGF-mediated signal transduction, and blunted VEGF-mediated proliferation and chemotaxis. Moreover, in a VEGF-dependent model of angiogenesis, the rat aortic ring assay, overexpression of HCPTPA blocked FIG. 3. VEGFR2 may be a substrate for endogenous HCPTPA. A, endothelial cell lines express HCPTPA. Clarified lysates from two endothelial cell lines (EA.hy926 and ECRF24) were analyzed by immunoblot with anti-HCPTPA polyclonal antibodies. B, endogenous VEGFR2 is a substrate for recombinant HCPTPA. VEGFR2 was immunoprecipitated from ECRF24 cells either unstimulated or stimulated with VEGF. Subsequently, VEGFR2 immunoprecipitates from VEGFstimulated cells were either left untreated or were treated with recombinant HCPTPA. All immunoprecipitates were then analyzed serially by immunoblotting with anti-Tyr(P) and anti-Flk-1. the formation of vascular sprouts. Taken together, these findings strongly suggest that HCPTPA is an important negative regulator of VEGF action.
Because of this potential importance of HCPTPA in VEGF receptor function and angiogenesis, gaining insights into how its activity is regulated will be essential. Our data demonstrate that HCPTPA is constitutively expressed in endothelial cells, suggesting that its activity may be regulated by post-translational modifications. For example, tyrosine phosphorylation of HCPTPA by the Src kinase enhances its catalytic activity (28). Taken in the context of the present study, this finding suggests that activation of Src, or perhaps related kinases in endothelial cells, might blunt signaling downstream of VEGF receptors. Conversely, low molecular weight protein tyrosine phosphatases such as HCPTPA can be reversibly inactivated by nitric oxide via S-nitrosylation of the active site cysteine (29,30). Generation of nitric oxide downstream of VEGF receptor activation has been shown to be essential for the biological activities of VEGF, including mitogenesis, permeability, and angiogenesis (31)(32)(33)(34)(35)(36)(37)(38)(39). Based on these findings and our present data, it is tempting to speculate that one of the essential functions of nitric oxide generation is to transiently inhibit the activity of phosphatases such as HCPTPA to accentuate VEGFR signaling.
Whether tyrosine phosphorylation and/or S-nitrosylation of HCPTPA importantly regulate VEGF signaling is currently open to speculation. However, if both translational modifications were to occur simultaneously downstream of VEGF receptor activation, one can imagine an elegant biological rheostat. Initially, inactivation of HCPTPA by S-nitrosylation would prevail. However, with the decline in nitric oxide production, reactivation of HCPTPA by endogenous reducing agents would ensue. Enhanced phosphatase activity would then catalyze the rapid dephosphorylation and inactivation of VEGF receptors as well as of the phosphatase itself (via autocatalysis (28)). Subsequent to this series of events, phosphatases would return to their lower level of constitutive activity, perhaps to perform a maintenance function such as policing errantly activated receptors.
In addition to negative regulatory properties, some protein tyrosine phosphatases, such as the SH2-containing phosphatases, have been shown to be essential components of positive signaling downstream of receptor tyrosine kinases (40 -44). In fact, in a recent study, association of HCPTPA was demonstrated following the formation of tetrameric, but not dimeric, complexes of endothelial Eph receptors (45). In that report, different cellular responses were induced by tetrameric versus dimeric receptor complexes, leading to the hypothesis that ephrin receptor signaling is modulated by differential recruitment of HCPTPA. Consistent with that report, we were not able to demonstrate direct association of HCPTPA with activated VEGFR2, which should form predominantly dimeric com- FIG. 5. HCPTPA inhibits VEGF-mediated endothelial cell responses. A, overexpression of HCPTPA inhibits VEGF-induced endothelial cell mitogenesis. HUVECs were infected with the indicated adenoviral vector and subsequently were either left unstimulated or were stimulated with VEGF prior to measuring thymidine incorporation as described under "Experimental Procedures." B, overexpression of HCPTPA inhibits VEGFmediated endothelial cell migration. HUVECs were infected with either AdH-CPTP or AdGFP as indicated, and chemotaxis was measured as described under "Experimental Procedures." plexes, yet overexpression of HCPTPA clearly had a negative effect on VEGF-mediated cellular responses. Considered together, these findings suggest that HCPTPA, when recruited to a dimeric complex, is catalytically active and dephosphorylates the activated receptors, thus limiting, or perhaps modulating, signal transduction. When recruited to a tetrameric complex, however, HCPTPA may be maintained in an inactive conformation, allowing prolonged rather than transient receptor activation, thus modifying the signaling output and the subsequent biological response.
In addition to regulating VEGFR2, low molecular weight protein tyrosine phosphatases may regulate signal transduction downstream of other receptor tyrosine kinases. For example, HCPTP binds to PDGF receptors in a phosphotyrosine-dependent manner, and PDGF receptors are substrates for recombinant HCPTP in vitro (20,46). In addition, overexpression of HCPTP in fibroblasts reduces PDGF receptor autophosphorylation and inhibits PDGF-stimulated mitogenic signaling (46). Importantly, PDGF receptors are expressed on certain populations of endothelial cells, suggesting that endothelial HCPTPA could play an important regulatory role in multiple angiogenic signaling pathways (47,48). Consistent with this notion, HCPTPA interacted with Tie2, another endothelial receptor tyrosine kinase with crucial roles during angiogenesis, in the yeast two-hybrid system. 2 Based on these data, it may be that the negative regulation of multiple angiogenic pathways in addition to the VEGF pathway contributed to the profound inhibition of angiogenesis seen in the aortic ring assay. In contrast, HCPTPA did not interact with either VEGFR1 (Flt-1) or Tie1, suggesting some specificity in the action of HCPTPA on endothelial signaling pathways.
Considering the potential importance of angiogenesis in the development and progression of clinically important diseases, along with the crucial role that VEGF receptors and other receptor tyrosine kinases play in the angiogenic process, understanding the role of downstream effectors such as HCPTPA will be increasingly important. The results of this study suggest that in diseases characterized by vascular insufficiency, inhibition of HCPTPA activity might enhance revascularization. Conversely, in the myriad of diseases that are driven by the growth of new blood vessels, such as cancer, arthritis, and diabetic retinopathy, it is possible that enhancing the activity of HCPTPA will inhibit angiogenesis.