PTEN Modulates Vascular Endothelial Growth Factor-Mediated Signaling and Angiogenic Effects*

Phosphatidylinositol 3-kinase is activated by vascular endothelial growth factor (VEGF), and many of the angiogenic cellular responses of VEGF are regulated by the lipid products of phosphatidylinositol 3-kinase. The tumor suppressor PTEN has been shown to down-regulate phosphatidylinositol 3-kinase signaling, yet the effects of PTEN on VEGF-mediated signaling and angiogenesis are unknown. Inhibition of endogenous PTEN in cultured endothelial cells by adenovirus-mediated overexpression of a dominant negative PTEN mutant (PTEN-C/S) enhanced VEGF-mediated Akt phosphorylation, and this effect correlated with decreases in caspase-3 cleavage, caspase-3 activity, and DNA degradation after induction of apoptosis with tumor necrosis factor-α. Overexpression of PTEN-C/S also enhanced VEGF-mediated endothelial cell proliferation and migration. In contrast, overexpression of wild-type PTEN inhibited the anti-apoptotic, proliferative, and chemotactic effects of VEGF. Moreover, PTEN-C/S increased the length of vascular sprouts in the rat aortic ring assay and modulated VEGF-mediated tube formation in an in vitroangiogenesis assay, whereas PTEN-wild type inhibited these effects. Taken together, these findings demonstrate that PTEN potently modulates VEGF-mediated signaling and function and that PTEN is a viable target in therapeutic approaches to promote or inhibit angiogenesis.

Vascular endothelial growth factor (VEGF) 1 plays a key role in endothelial cell differentiation (vasculogenesis) and the sprouting of new blood vessels from preexisting ones (angiogenesis). Angiogenesis is critical for normal embryonic vascular development as well as a number of physiological and patho-logical conditions, including ischemic diseases, inflammation, and cancer. Binding of VEGF to VEGF receptor-2 (VEGFR-2) leads to receptor phosphorylation and subsequent activation of phosphatidylinositol 3-kinase (PI3K), phospholipase C-␥1, Src family tyrosine kinases, and other signaling proteins (1)(2)(3)(4).
The key role of PI3K in VEGF-mediated signal transduction and angiogenic responses is well established (2,5,6). Experimental evidence has shown that activation of PI3K is critical for VEGF-mediated endothelial cell proliferation, survival, and migration. Moreover, downstream activation of Akt by PI3K is responsible for phosphorylation and activation of endothelial nitric oxide synthase (eNOS) by VEGF (6,7).
PI3K catalyzes the phosphorylation of inositol phospholipids at the D3 position to generate phosphatidylinositol 3,4,5trisphosphate and phosphatidylinositol 3,4-bisphosphate. These 3-phosphoinositides act as potent signaling molecules to regulate many cellular responses that are important for angiogenesis, including cell adhesion, proliferation, vesicular trafficking, protein synthesis, and cellular survival (8 -11). Recent studies demonstrated that the 3-phosphoinositides are important substrates for PTEN (phosphatase and tensin homology deleted from chromosome 10) both in vitro (12) and in vivo (13,14). PTEN was originally identified as a candidate tumor suppressor gene based on its high frequency of mutation in a variety of tumors (15). PTEN was subsequently found to exhibit both protein tyrosine phosphatase and inositol 3Ј-phosphatase activity. However, its lipid phosphatase activity appears to be primarily responsible for its tumor suppressor effects, as mutation or loss of PTEN results in increased 3-phosphoinositides and downstream activation of Akt (13, 16 -18). Thus, PTEN functions in opposition to PI3K, and numerous studies have shown that overexpression of wild-type PTEN suppresses cell growth and proliferation through G 1 cell cycle arrest and enhances apoptosis by down-regulating PI3K/Akt signaling (16, 18 -23).
Most studies of PTEN have focused on its role in tumor cell biology; however, a recent report demonstrated that PTEN could modulate the response of cardiac myocytes to PI3K activation and thereby regulate myocyte survival and hypertrophy (24). Based on the roles of PI3K and Akt in signaling by VEGF, we hypothesized that PTEN could regulate VEGF-mediated endothelial cellular responses and angiogenesis. In this report, we demonstrate that inhibition of endogenous endothelial PTEN by adenovirus-mediated overexpression of a dominant negative PTEN mutant in cultured endothelial cells potently enhances a variety of VEGF-mediated cellular responses, including cell survival, mitogenesis, and migration. In contrast, these effects of VEGF are significantly inhibited by overexpression of wild-type PTEN. Moreover, overexpression of wild-type or dominant negative PTEN modulated endothelial tube formation in vitro and vascular sprouting in an ex vivo model of angiogenesis.
Adenovirus Construction-cDNAs encoding wild-type (WT) and catalytically inactive PTEN, in which cysteine 124 has been mutated to serine (C/S), were generously provided by Dr. Charles Sawyers (University of California, Los Angeles). To generate adenoviruses directing the expression of these proteins, cDNAs encoding PTEN were subcloned into pShuttle-CMV then recombined with pAdEasy-1 by electroporation into BJ5183 Escherichia coli (Stratagene, La Jolla, CA) (29). The recombinant adenoviral vector DNA was transfected into human embryonic kidney 293 cells with LipofectAMINE (Invitrogen), then the viruses were serially amplified in 293 cells, purified on a CsCl density gradient by ultracentrifugation, and titered as described previously (30). A control adenovirus consisting of the identical adenovirus backbone without a cDNA insert ("empty virus," AdEV) was provided by Dr. Walter J. Koch (Duke University Medical Center, Durham, NC) (31).
Adenovirus Infection-For most experiments, HUVECs were grown in EGM-MV (containing 5% FBS, Clonetics Corp.). When the cells were nearly confluent, the medium was changed to EGM containing 2% FBS, and viruses were added to the medium at a dilution of 1:1000 (multiplicity of infection ϳ100). The cells were incubated for 16 h at 37°C, then the medium was changed to serum-free endothelial basal medium (EBM, Clonetics Corp.), and the cells were treated as indicated.
Apoptosis Assays-Two assays of programmed cell death were used. To assay caspase-3 activity, HUVECs were plated in triplicate at 3 ϫ 10 5 cells/well of a 6-well plate and grown overnight in EGM-MV containing 5% FBS. The following day the medium was changed to EGM containing 2% FBS, and the cells were either left uninfected or were infected with AdPTEN-WT, AdPTEN-C/S, or AdEV overnight. The next day apoptosis was induced by serum starvation in EBM and treatment with TNF␣ (50 ng/ml). Half of the cells were treated with VEGF (50 ng/ml), and half were left untreated, and the cells were incubated at 37°C for 3 h. Cells were lysed, and an aliquot of each cell lysate was used in a fluorimetric assay of caspase-3 activity (EnzChek caspase-3 assay kit, Molecular Probes, E-13138) according to the manufacturer's instructions.
To confirm the effects of wild-type and inactive PTEN on caspase-3 activity, apoptosis was also evaluated using the cell death detection ELISA-Plus kit (Roche Molecular Biochemicals, 1774425). HUVECs were plated in triplicate wells of a 24-well plate. The cells were left uninfected or were infected with AdPTEN-WT, AdPTEN-C/S, or AdEV overnight, and apoptosis was induced as described above in VEGFtreated and -untreated cells. Cells were lysed, and cytoplasmic histoneassociated DNA fragmentation (mono-and oligonucleosomes) was detected by spectrophotometry according to the manufacturer's instructions.
Cell Migration-The migration of HUVECs was determined using a "scratch" wound assay as described previously (6,32). Briefly, HUVECs were grown in triplicate in 60-mm dishes, and when confluent, they were left uninfected or were infected with AdPTEN-WT, AdPTEN-C/S, or AdEV for 16 h. The cell monolayer was scraped with a sterile rubber policeman to create a cell-free zone, then washed once with medium and treated with or without VEGF (20 ng/ml) in EGM containing 2% FBS. HUVEC migration was quantified by measuring the width of the cellfree zone (distance between the edges of the injured monolayer) at 4 distinct positions 24 h after treatment on an Olympus IX-70 inverted microscope connected to a Diagnostic Instruments Spot RT Color Camera, and data were analyzed with NIH Image, v.1.6.2.
Thymidine Incorporation Assay-To evaluate the effect of PTEN on VEGF-mediated DNA synthesis, [ 3 H]thymidine incorporation was assayed as described previously (33). Briefly, uninfected, AdPTEN-WT-, AdPTEN-C/S-, or AdEV-infected HUVECs were plated at 25,000 cells/ well of a 24-well plate, quiesced by incubation in EBM for 24 h, then treated with or without VEGF (20 ng/ml) for 18 h. The cells were pulse-labeled with [ 3 H]thymidine (2 Ci/ml, Amersham Biosciences, Inc.) for 3 h, the DNA was precipitated, and the amount of [ 3 H]thymidine incorporation was determined by liquid scintillation counting.
Cell Counts-To confirm the effects of PTEN on VEGF-mediated cellular proliferation, cell counts were performed. Uninfected, AdPTEN-WT-, AdPTEN-C/S-, and empty virus-infected HUVECs were plated in triplicate wells of 24-well plates in a volume of 500 l of EGM-MV and incubated for 24 h at 37°C. The following day the medium was changed to EBM with or without VEGF (20 ng/ml). After 24 h the medium was replaced with fresh EBM with or without VEGF (20 ng/ml) and incubated for an additional 24 h. After 48 h, cells were trypsinized and resuspended in EGM-MV. Cell numbers in each group were counted on a hemacytometer (Fisher) using an Olympus CK2 inverted microscope. Data were expressed as the means Ϯ S.E.
Tube Formation Assay-The endothelial tube formation assay, which has been described previously for the evaluation of angiogenesis in vitro (34 -36), was performed using the in vitro angiogenesis kit (Chemicon, Temecula, CA) according to the manufacturer's instructions. HUVECs were left uninfected or were infected with AdPTEN-WT, AdPTEN-C/S, or AdEV for 16 h, then were trypsinized and resuspended in EGM containing 2% FBS. Wells of a 96-well plate were coated with ECMatrix solution, and 5 ϫ 10 3 cells were plated in triplicate wells in a volume of 50 l of EGM containing 2% FBS either with or without VEGF (25 ng/ml). The cells were incubated for 24 h at 37°C, and tube formation was evaluated by phase-contrast microscopy using an Olympus IX-70 microscope (100ϫ magnification) connected to a Diagnostic Instruments Spot RT Color Camera. To determine the viability of cells in these assays, cells were stained with Hoechst 33342 (5 g/ml, Sigma) and propidium iodide (2.5 g/ml, Sigma) for 5 min at 37°C and analyzed by fluorescence microscopy.
Rat Aortic Ring Assay-This ex vivo angiogenesis assay was performed essentially as described previously (33,37) with some modifications. Thoracic aortas were excised from 2-3-week-old Sprague-Dawley male rats and immediately placed into cold Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS. Clotted blood inside the aorta was flushed with media, and the periadventitial fibroadipose tissue was removed. Aortas were then cut into cross-sectional rings ϳ1-1.5 mm in length. Aortic rings were mock-infected or infected with 2 ϫ 10 11 viral particles of AdPTEN-WT, AdPTEN-C/S, or AdEV in 1.0 ml of DMEM, 2% FBS, at room temperature for 15 min. Rings were placed into wells of a 24-well plate containing 0.4 ml of cold growth factor-reduced Matrigel (BD Biosciences) then incubated at 37°C until the Matrigel polymerized. The wells were then overlaid with 0.5 ml of EBM without phenol red, and the rings were maintained at 37°C for up to 10 days with medium changes every 2 days. Vascular sprouting from each ring was examined daily on an Olympus IX-70 microscope (100ϫ magnification), and digital images were obtained. Quantitative analysis of endothelial sprouting was performed using images from day 5, and sprout length was quantified in NIH Image (v.1.6.2) using a calibrated micrometer. The greatest distance from the aortic ring body to the end of the vascular sprouts was measured at three distinct points per ring and in three different rings per treatment group.
Statistical Analysis-All results were expressed as the mean Ϯ S.E. Statistical analysis was performed using the one-tailed Student's t test (two sample, unequal variance), and p Ͻ 0.05 was considered statistically significant.

PTEN Is Expressed in Endothelial Cells-We first analyzed
whether PTEN is expressed in endothelial cells. Lysates from several different immortalized endothelial cell lines as well as primary HUVECs and NIH 3T3 fibroblasts were analyzed by Western blotting with an antibody against PTEN. PTEN was expressed to varying degrees in both EA.hy926 cells (25) and HUVECs, but it was not detectable in EC-RF24 cells (27) (Fig.  1). PTEN was highly expressed in Py-4-1 cells, a murine microvascular endothelial cell line (26), as well as in parental NIH 3T3 cells and 3T3 cells stably expressing chimeric fms-Tie1 or fms-Tie2 receptors (28). The expression of PTEN in endothelial cells indicates that this protein could potentially modulate VEGF-mediated signaling.
Overexpression of PTEN Modulates VEGF-mediated Akt Phosphorylation-VEGF is known to activate the PI3K/Akt pathway (5); therefore, we investigated whether PTEN overexpression could alter VEGF-mediated Akt phosphorylation. Primary HUVECs were either left uninfected or were infected with recombinant adenoviruses encoding wild-type PTEN (AdPTEN-WT), a catalytically inactive mutant of PTEN in which cysteine 124 has been mutated to serine (AdPTEN-C/S), or an empty adenovirus as a control for viral infection (AdEV). After overnight virus infection, the cells were serum-starved for 6 h and then treated with or without VEGF. Endogenous PTEN was detectable at moderate levels in both uninfected and AdEV-infected HUVECs, whereas both wild-type and inactive PTEN were overexpressed after virus infection (Fig. 2). VEGF treatment increased Akt phosphorylation in uninfected and AdEV-infected cells. Importantly, overexpression of PTEN-WT abrogated this effect, whereas PTEN-C/S enhanced phosphorylation of Akt. Similar amounts of total Akt and tubulin were observed in each lane, demonstrating that the effects on Akt phosphorylation were a result of the PTEN proteins themselves. VEGF treatment also enhanced the phosphorylation of ERK1 and -2, but this was not altered by overexpression of either wild-type or inactive PTEN (Fig. 2), suggesting that VEGF-mediated ERK activation is not dependent on the production of 3-phosphoinositides.
PTEN Modulates VEGF-mediated Anti-apoptosis-Akt is known to play a critical role in cell survival mediated by VEGF (5). To determine whether the effects of PTEN on VEGF-mediated Akt phosphorylation in HUVECs correlated with the effects on apoptosis, we evaluated caspase-3 cleavage in HUVECs induced to undergo apoptosis. Caspase-3 is a key mediator of apoptosis, and cleavage of this enzyme to its active form correlates with the onset of apoptosis. HUVECs were infected with PTEN or control viruses, and the cells were serum-starved and treated with TNF␣, which has been shown to induce endothelial cell apoptosis (38). In lysates from uninfected cells and those infected with AdPTEN-WT or AdEV, VEGF treatment blocked the TNF␣-induced cleavage of caspase-3 (Fig. 3). Overexpression of PTEN-WT increased basal caspase-3 cleavage and appeared to reduce the protective effect of VEGF in these cells. In contrast, essentially no cleaved caspase-3 was detectable in cells expressing PTEN-C/S either with or without VEGF treatment. These findings indicate that overexpression of PTEN-WT enhances endothelial cell apoptosis, whereas inhibition of PTEN is protective.
To confirm the effects of PTEN overexpression on endothelial cell apoptosis, we assayed both capase-3 activity and DNA fragmentation. HUVECs were infected with recombinant ad-enoviruses, and apoptosis was induced by serum starvation and TNF␣ treatment. Cell lysates were then used in a fluorimetric assay of caspase-3 activity. Consistent with the Western blotting results, VEGF treatment reduced caspase-3 activity in all cells, including those infected with AdPTEN-C/S (Fig. 4A). However, overexpression of PTEN-WT significantly increased, whereas PTEN-C/S significantly decreased caspase-3 activity compared with control cells, both with and without VEGF treatment. Similar results were obtained using a spectrophotometric assay of histone-associated DNA fragmentation, although the effects of PTEN-WT overexpression were not as pronounced (Fig. 4B). Taken together, these findings demonstrate that PTEN modulates the anti-apoptotic effects of VEGF.
PTEN Inhibits VEGF-mediated Proliferation-Although PTEN did not appear to alter VEGF-mediated MAP kinase activation, we evaluated the effects of wild-type and inactive PTEN on VEGF-mediated endothelial cell proliferation. HU-VECs were either left uninfected or were infected with AdPTEN-WT, AdPTEN-C/S, or AdEV, then DNA synthesis was assayed by [ 3 H]thymidine incorporation. As expected, VEGF induced an increase in DNA synthesis in both uninfected and empty virus-infected HUVECs (Fig. 5A). Overexpression of PTEN-C/S resulted in a significant increase in VEGF-stimu-  Fig. 2. The cells were then serum-starved in the presence of TNF␣ (50 ng/ml) for 3 h to induce apoptosis and treated with or without VEGF (50 ng/ml). Cell lysates were separated by SDS 8 -16% PAGE and Western-blotted with antibodies against cleaved caspase-3, PTEN, and tubulin. VEGF treatment caused a reduction in caspase-3 cleavage that was blocked by overexpression of PTEN-WT. In contrast, PTEN-C/S reduced caspase-3 cleavage in cells treated with or without VEGF. Un, uninfected cells; EV, empty virus-infected cells. lated DNA synthesis. In contrast, PTEN-WT markedly reduced both basal and VEGF-mediated thymidine incorporation. Notably, overexpression of PTEN-WT in HUVECs led to a marked increase in the number of apoptotic-appearing cells (data not shown), consistent with our earlier results demonstrating that PTEN-WT enhanced endothelial cell apoptosis.
To confirm the results of thymidine incorporation assays, we performed cell counts on uninfected HUVECs and on those infected with the PTEN or control adenoviruses. Cells were treated for 48 h with or without VEGF, then trypsinized and counted in triplicate. In all groups, VEGF induced an increase in cell number compared with untreated cells. As with changes in DNA synthesis, no differences were noted between uninfected and empty virus-infected cells (Fig. 5B). Overexpression of PTEN-WT slightly reduced the number of untreated cells, likely because of increases in apoptosis, but it significantly inhibited the VEGF-mediated increase in cell number. In contrast, expression of PTEN-C/S significantly increased the number of cells at 48 h in both the VEGF-treated and -untreated groups.
PTEN Alters VEGF-mediated Endothelial Cell Migration-The lipid products of PI3K are known to regulate cell migration and chemotaxis (10,11). To investigate whether PTEN overexpression modulates the effects of VEGF on endothelial cell migration, we performed a scratch wound assay (6, 32) on cultured HUVECs that were either uninfected or infected with the different recombinant adenoviruses used previously. Consistent with published results (1), VEGF enhanced the migration of HUVECs into the cell-free zone after wounding (Fig. 6). Overexpression of PTEN-WT significantly reduced HUVEC migration both at baseline and after VEGF treatment. In contrast, PTEN-C/S significantly enhanced VEGF-mediated cell migration, although it had no appreciable effect on basal migration. As seen in the thymidine incorporation experiment, numerous PTEN-WT-infected cells appeared apoptotic, suggesting that the adverse effects of PTEN-WT on migration could be due in part to enhanced cell death.

PTEN Modulates Tube Formation and Vascular
Sprouting-Based on the effects of wild-type and catalytically inactive PTEN on HUVEC survival, proliferation, and migration, we next examined whether overexpression of PTEN would alter the endothelial tube-forming activity of HUVECs, often referred to as an in vitro angiogenesis assay (34 -36). To do this, HUVECs were infected with PTEN-expressing adenoviruses and then plated on Matrigel and treated with or without VEGF, which is known to induce capillary morphogenesis (1). Within 24 h, uninfected and empty virus-infected cells formed an organized network of endothelial tubes, and this effect was enhanced by treatment with VEGF (Fig. 7, A and D). In contrast, PTEN-WT overexpression markedly inhibited tube formation (Fig. 7B) both in the presence and absence of VEGF. Nuclear staining of PTEN-WT-infected cells with Hoechst 33342 demonstrated nuclear condensation of many cells, consistent with enhanced apoptosis (data not shown). HUVECs infected with AdPTEN-C/S formed tubes earlier than uninfected or AdEV-infected cells, as capillary networks were observed as early as 6 h after plating on Matrigel (data not shown). Interestingly, by the following day PTEN-C/S expression had induced a dramatic change in the morphology of the endothelial networks (Fig. 7C). Although the number of tubes was essentially unchanged, the total number of endothelial cells in each field was markedly increased. Capillary networks were surrounded by clusters of viable endothelial cells that failed to properly assemble into tubes.
We next evaluated the effects of the PTEN adenoviruses in the rat aortic ring assay, an ex vivo assay of vascular sprouting that depends on the function of a variety of angiogenic growth factors and their receptors (37). Sections of rat aorta were infected with the different adenoviruses and then cultured in growth factor-reduced Matrigel, and the length of vascular sprouts was measured on day 5. Compared with uninfected or empty virus-infected rings, PTEN-WT induced a slight but statistically insignificant reduction in the length of vascular sprouts (Fig. 8). In contrast, sprouts from PTEN-C/S-infected rings were 2-3-fold longer than those in control rings. Taken together with the tube formation assays, these results indicate that PTEN has a potent inhibitory effect on angiogenesis.

DISCUSSION
Angiogenesis is a complex process mediated by several endothelial receptor tyrosine kinases and their ligands. These receptor-ligand systems regulate diverse endothelial cellular functions, including endothelial cell proliferation, migration, capillary morphogenesis, and survival, which are necessary for proper vascular development. VEGF/VEGF receptor signaling is known to regulate each of these processes (1), and many of them have been linked to PI3K signaling (2,39,40). Thus, the phospholipid second messengers generated by PI3K provide a common mechanism for multiple steps during angiogenesis. In this report, we have demonstrated that the inositol 3Ј-phosphatase PTEN, which hydrolyzes the PI3K lipid products phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-trisphosphate, modulates VEGF-mediated signaling and cellular responses in cultured endothelial cells. Furthermore, PTEN altered capillary morphogenesis in an in vitro angiogenesis model as well as vascular sprouting in an ex vivo tissue model of angiogenesis.
PTEN was identified originally as a tumor suppressor protein (15) and, as a result, its effects have been studied primarily in tumor cells. Numerous studies have demonstrated that PTEN suppresses proliferation and survival of tumor cells by inhibiting PI3K/Akt signaling (13,14,17,18,20,21,23,41). However, PI3K signaling is critical for a variety of cellular functions in diverse cell types (10). It seems likely, therefore, that PTEN should regulate PI3K signaling in other cell types in which it is expressed. In fact, a recent report demonstrated that overexpression of wild-type or catalytically inactive PTEN in cardiac myocytes had profound effects on myocyte hypertrophy and survival (24). Here, by overexpressing a dominant negative PTEN mutant (PTEN-C/S), we were able to demonstrate that inhibition of endogenous PTEN in cultured endothelial cells enhanced VEGF signaling. This finding correlated with the enhancement of a variety of VEGF-mediated cellular responses.
PI3K was shown in one study to regulate VEGF-mediated endothelial cell proliferation, as the PI3K inhibitor wortmannin blocked both ERK phosphorylation and bromodeoxyuridine incorporation after VEGF treatment of HUVECs (2). In contrast to these results, Wu et al. (42) find that high dose wortmannin had no effect on VEGF-mediated ERK phosphorylation. Likewise, in our studies overexpression of PTEN-WT or PTEN-C/S had no effect on ERK phosphorylation, in contrast to the marked effects of both proteins on Akt phosphorylation. These findings indicate that PI3K and its lipid products, the 3-phosphoinositides, are not required for VEGF-mediated ERK activation. However, the effects of both wild-type and inactive PTEN on VEGF-mediated thymidine incorporation and cell counts suggest that a PI3K-dependent, ERK-independent signaling pathway is required for VEGF-mediated cell proliferation. Alternatively, the appreciable effects of PTEN-WT and -C/S on endothelial cell survival could have resulted in significant effects on cell proliferation. The effects of VEGF in both the thymidine incorporation assays and cell counts were performed after quiescence in serum-free medium, which by itself can initiate endothelial cell apoptosis. Thus, VEGF-mediated thymidine incorporation may have been blunted by PTEN-WT and enhanced by PTEN-C/S due to the respective effects of these proteins on cell survival, making it unclear whether the effects on proliferation were primary or secondary. Similarly, although PI3K is known to regulate cell migration and chemotaxis, it is possible that the effects of PTEN on VEGF-mediated cell migration were at least partly due to alterations in cell survival.
It is notable that both wild-type and dominant negative PTEN had significant effects on basal endothelial cell signaling and cellular responses in addition to their effects on VEGFmediated processes. Based on its ability to hydrolyze the lipid products of PI3K, PTEN would be predicted to play a central role in the regulation of PI3K-mediated processes downstream of a variety of receptors and in virtually any cell type in which it is expressed. For example, PTEN has recently been shown to regulate IGF-1 signaling in cardiac myocytes (24). The fact that inhibition of endothelial PTEN with the dominant negative PTEN-C/S mutant resulted in enhanced basal endothelial signaling supports a key regulatory role for this enzyme in endothelial biology.
Based on the wide range of endothelial cellular effects of PTEN, we anticipated that PTEN overexpression might have significant effects on angiogenesis. Indeed, PTEN-C/S signifi- cantly altered tube formation and vascular sprouting in two separate angiogenesis assays, whereas PTEN-WT dramatically limited tube formation, an effect that appeared to be largely because of reduced endothelial cell survival. However, the combined effects of PTEN-WT on endothelial cell proliferation, migration, and survival may have resulted in a synergistic effect greater than that of blocking any one cellular response alone. In contrast, overexpression of catalytically inactive PTEN resulted in a striking effect on tube formation that is difficult to explain based on its effects in cellular assays. Although the number of tubes and networks appeared the same as for control cells, the endothelial tubes were comprised of far more endothelial cells, as if cellular proliferation were out of proportion to migration and/or morphogenesis, and these endothelial cells appeared viable, unlike those overexpressing PTEN-WT. Notably this tube formation phenotype resembles that recently observed by Bussolati et al. (43) after the inhibition of VEGFR-1. In that report, VEGFR-1 signaling was found to inhibit VEGFR-2-mediated endothelial cell proliferation via nitric oxide (NO). Inhibition of either VEGFR-1 signaling or NO production by VEGFR-1 resulted in enhanced endothelial cell proliferation, but these cells appeared unable to contribute to normal capillary morphogenesis. It was suggested from these findings that VEGFR-2 primarily regulates endothelial cell proliferation, whereas VEGFR-1 is required for endothelial differentiation. These findings were confirmed in part by Zeng et al. (44), who demonstrated negative regulation of VEGFR-2 by VEGFR-1 that was mediated by PI3K signaling (44). A common thread between these two studies appears to be that PI3K/Akt signaling activates eNOS, resulting in increased nitric oxide release (7,45). Interestingly, these findings would appear to conflict with our results, which suggest that enhanced Akt activation by PTEN-C/S should result in increased NO production and a resultant inhibition of VEGFR-2-mediated endothelial cell proliferation. Additional studies will be required to determine the effects of PTEN on NO production and the role of NO in the PTEN-C/S-induced phenotype in the tube formation assay.
Our current studies demonstrate that PTEN can significantly modulate angiogenic cellular responses through direct Overexpression of wild-type PTEN resulted in decreased tube formation, with numerous apoptotic-appearing cells (arrowheads), as determined by nuclear staining (not shown). In contrast, PTEN-C/S induced a marked increase in the number of endothelial cells in the tubes (arrows), which appeared viable by nuclear staining (not shown). Magnification, ϫ100.
FIG. 8. PTEN modulates angiogenesis in the rat aortic ring assay. Cross-sections of rat aorta were left uninfected (Un, A) or infected with recombinant adenoviruses encoding WT (B) or catalytically inactive PTEN (C/S, C) or with empty virus (EV, D) then embedded in growth factor-reduced Matrigel. Five days later, vascular sprouting was quantified by digital microscopy as the maximal length of sprouts from the body of the aortic ring at three different points per ring (e.g. arrows) and on three rings per group (E). Data are expressed as the means Ϯ S.E. PTEN-WT significantly reduced sprout length, whereas PTEN-C/S increased it by ϳ3-fold. *, p Ͻ 0.05 versus AdEV-infected cells. Magnification in A-D, ϫ100. effects on the endothelium. However, several recent reports have indicated that PTEN expressed in tumor cells can have indirect effects on angiogenesis as well through at least two distinct mechanisms. Expression of VEGF and other angiogenic growth factors is regulated by tissue hypoxia via the hypoxia-inducible factor-1 (HIF-1) transcription factor complex (1). Loss of PTEN in glioblastoma cells has been correlated with increased stabilization of HIF-1␣ and resultant increases in VEGF expression (46). In another study, reconstitution of wildtype PTEN expression in U87MG glioma cells had no effect on cell proliferation in vitro but reduced tumor growth and angiogenesis in vivo, and these findings were attributed to increased expression of the angiogenesis inhibitor thrombospondin in PTEN-expressing cells (47). In a study of prostate cancers, however, PTEN deletion was found to correlate with increased tumor microvessel density, but loss of PTEN did not correlate with thrombospondin-1 expression (48), suggesting that additional mechanisms exist through which PTEN regulates tumor angiogenesis.
To our knowledge, no other studies have investigated the possibility that PTEN could directly regulate angiogenesis by modulating endothelial cellular functions. Taken together with results from tumor cells, our findings suggest that overexpression of PTEN in whole tumors could disrupt tumor angiogenesis through both tumor cell-mediated effects and effects on endothelial cells. This would result in induction of apoptosis in both cell types, which would be expected to be more potent than targeting either cell type alone. Another important implication of our results is that PTEN could be an important target for therapeutic pro-angiogenesis in ischemic heart and vascular diseases. Because PI3K signaling regulates many of the endothelial cellular responses required for angiogenesis, inhibition of PTEN could significantly enhance these responses. PI3K is activated downstream of other endothelial receptor tyrosine kinases that are required for vascular development, including Tie2 (28) and Tie1, 2 which play important roles in vascular maturation. The observed effects on basal endothelial signaling and cellular processes in this study indicate that PTEN plays a central role in the regulation of a variety of angiogenic cellular responses. Inhibition of PTEN might allow potentiation of PI3K-mediated responses downstream of numerous receptors in an appropriate temporal and spatial context. This contrasts with the delivery of one specific growth factor, such as VEGF, which may be insufficient to induce appropriate vascular maturation of nascent blood vessels. Notably, a preliminary study has suggested that adenovirus-mediated delivery of constitutively active Akt to the endothelium enhances angiogenesis in a model of hindlimb ischemia (49). Targeting PI3K or PTEN may have advantages over this approach, since many of the effects of PI3K are independent of Akt. Studies are currently under way to test the effects of PTEN inhibition on angiogenesis in vivo.