Angiopoietin-1 inhibits endothelial cell apoptosis via the Akt/survivin pathway.

A productive angiogenic response must couple to the survival machinery of endothelial cells to preserve the integrity of newly formed vessels. Angiopoietin-1 (Ang-1) is an endothelium-specific ligand essential for embryonic vascular stabilization, branching morphogenesis, and post-natal angiogenesis, but its contribution to endothelial cell survival has not been completely elucidated. Here we show that Ang-1 acting via the Tie 2 receptor induces phosphorylation of the survival serine-threonine kinase, Akt (or protein kinase B). This is associated with up-regulation of the apoptosis inhibitor, survivin, in endothelial cells and protection of endothelium from death-inducing stimuli. Moreover, dominant negative survivin negates the ability of Ang-1 to protect cells from undergoing apoptosis. The activation of anti-apoptotic pathways mediated by Akt and survivin in endothelial cells may contribute to Ang-1 stabilization of vascular structures during angiogenesis, in vivo.

A productive angiogenic response must couple to the survival machinery of endothelial cells to preserve the integrity of newly formed vessels. Angiopoietin-1 (Ang-1) is an endothelium-specific ligand essential for embryonic vascular stabilization, branching morphogenesis, and post-natal angiogenesis, but its contribution to endothelial cell survival has not been completely elucidated. Here we show that Ang-1 acting via the Tie 2 receptor induces phosphorylation of the survival serinethreonine kinase, Akt (or protein kinase B). This is associated with up-regulation of the apoptosis inhibitor, survivin, in endothelial cells and protection of endothelium from death-inducing stimuli. Moreover, dominant negative survivin negates the ability of Ang-1 to protect cells from undergoing apoptosis. The activation of antiapoptotic pathways mediated by Akt and survivin in endothelial cells may contribute to Ang-1 stabilization of vascular structures during angiogenesis, in vivo.
During angiogenesis, endothelial cells receive cues from growth factors to initiate mitosis, migration, and organization of endothelial cells into primitive angiotubes and patent vascular networks (1,2). These processes critically depend on preservation of endothelial cell viability. Disruption of endothelial cell-matrix contacts or interference with extracellular survival signals is sufficient to initiate caspase-dependent apoptosis in endothelium, culminating with rapid involution of vascular structures (3,4). Unlike most angiogenic regulators, including fibroblast growth factor or vascular endothelial growth factor (VEGF), 1 angiopoietin-1 (Ang-1) does not stimu-late endothelial cell growth but rather promotes stabilization of vascular networks and branching morphogenesis in vivo and in vitro (5)(6)(7)(8). Little is known about the signaling requirements of these responses, and the mechanism(s) of Ang-1-induced cytoprotection are unknown (7,9).
The major goal of this paper was to elucidate a potential link between endothelial cell viability and maintenance of angiogenesis by examining the ability of Ang-1 to activate the antiapoptotic serine-threonine kinase, Akt (or protein kinase B). Moreover, we examined the relationship between Ang-1, Akt activation, and the expression of the anti-apoptotic genes, bcl-2 and survivin, in cultured microvascular endothelial cells (MVEC).

MATERIALS AND METHODS
Cell Culture and Reagents-Bovine MVEC (Vec Technologies, Rensselaer, NY) were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, L-glutamine, and antibiotics (penicillin and streptomycin). Cells (up to passage 12) were used for the experiments. In experiments examining endogenous survivin expression, human umbilical vein endothelial cells (HUVEC) were used, because the survivin antibody recognized human survivin better than bovine survivin. HUVEC were cultured on gelatin-coated tissue culture flasks in M199 medium containing 20% fetal bovine serum, 50 g/ml endothelial cell growth supplement (a commercial preparation that contains mainly acidic fibroblast growth factor), 100 g/ml porcine heparin, 10 units/ml penicillin, and 100 g/ml streptomycin. Two to three individual donors were pooled at passage one and used up to passage three. Both MVEC and HUVEC cultures had typical cobblestone morphology and stained uniformly for von Willebrand factor, as assessed by indirect immunofluorescence. Angiopoietin-1 and -2 and soluble recombinant Tie 1 and 2 receptors were provided by Regeneron. A recombinant form of Ang-1 was used in all of the experiments. This form of Ang-1 differs from the native Tie 2 ligand in that it possesses a modified NH 2 -terminal sequence and a mutation in Cys 245 that make it easier to produce and purify.
Akt Phosphorylation and Activity-Cells were washed twice with PBS and lysed with cell lysis buffer (1% Nonidet P-40, 10% glycerol, 137 mM NaCl, 20 mM Tris-HCl, pH 7.4, 20 mM NaF, 2 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). 20 g of protein was separated on SDSpolyacrylamide gel electrophoresis gel and transferred onto a polyvinylidine difluoride membrane (Millipore). After blocking with PBS containing 0.2% Tween 20 containing 5% milk for 1 h, the membrane was incubated with anti-Akt antibody (Santa Cruz Biotechnology), phosphospecific Akt antibody (New England Biolabs). ECL (Amersham Pharmacia Biotech) was used for detection. For activity assays, lysates were precleared with protein G-agarose for 30 min at 4°C and immunoprecipitated for 2 h with anti-Akt antibodies in the presence of 2 g/ml bovine serum albumin with or without 16 g/ml competitor peptides (Santa Cruz Biotechnology). Immunoprecipitates were washed twice with cell lysis buffer, once with water, and once with kinase buffer (20 mM HEPES, pH 7.2, 10 mM MgCl 2 , 10 mM MnCl 2 ). Immunoprecipitated proteins were incubated in 50 l of kinase buffer containing 2 g of histone H2B (Roche Molecular Biochemicals) and [ 32 P]ATP (5 M, 10 Ci) for 30 min at room temperature. Kinase reactions were stopped by the addition of SDS sample buffer, and samples were subjected to Cerekenov counting. Parallel samples were processed to confirm equal amounts of precipitated Akt.
Fluorescence-activated Cell Sorter (FACS) Analysis of Hypodiploid Cells-MVEC were plated onto bacteriological dishes in serum-free medium in the presence of either vehicle (TBS containing CHAPS) or Ang-1 (250 ng/ml). Cells were incubated for 18 h, and both floating and adherent cells were collected. To determine the number of subdiploid cells, MVEC were fixed for 1 h in 70% ethanol and stained with a * This work was supported by Grants HL57665 and HL61372 (to W. C. S.), T32 L10183 (to D. F.), and CA78810 and HL54131 (to D. C. A.) from the National Institutes of Health and by a grant-in-aid from the American Heart Association (to W. C. S.). 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 U.S.C. Section 1734 solely to indicate this fact.
§ Contributed equally to this work. solution containing 500 g/ml RNase H and 50 g/ml propidium iodide and analyzed by using an FACS. At least 5000 events were analyzed, and the percentage of cells in the sub-G 1 population was calculated. Aggregates of cell debris at the origin of the histogram were excluded from the analysis of sub-G 1

cells as indicated in the legends to Figs. 2 and 4.
Viral Infection of MVEC-MVEC were infected with 50 -100 multiplicities of infection of herpes simplex viruses expressing ␤-galactosidase or the dominant negative ⌬p85 subunit of PI3 kinase as described (10). Alternatively, MVEC were infected with similar multiplicities of infection of adenoviruses containing the ␤-galactosidase or the hemagglutinin-tagged activation-deficient phosphorylation mutant Akt (AA-Akt). After 4 h, the virus was removed, and the cells were left to recover overnight in complete medium. In preliminary experiments with the ␤-galactosidase virus, these conditions were optimal for infecting 95% of the cultures. Infected cells were either plated in bacteriological dishes or lysed in lysis buffer for immunoblotting.
Northern Blotting-MVEC were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and serumstarved for 24 h followed by challenge with Ang-1 as described above. Total RNA was extracted from cell pellets with TRI reagent (10 6 cells/ 0.2 ml, Molecular Research Center, Inc., Cincinnati, Ohio). For Northern analysis, 10 -20 g of total RNA were separated on 1% agarose gels with formaldehyde, transferred to nylon filters (Hybond-N, Amersham Pharmacia Biotech), UV cross-linked, and hybridized with the corresponding 32 P-labeled cDNA probes (survivin, bcl-2, or ␤-actin) in Ex-pressHyb hybridization solution (CLONTECH, Palo Alto, CA). After washing, the filter was exposed for autoradiography.
Survivin Promoter Studies-pLuc-cyc1.2 (ϩ1 to Ϫ268) was generated by polymerase chain reaction with the human survivin promoter sequence as a template and confirmed by DNA sequencing. pLuc-42 was generated by inserting the first 42-base pair fragment of the 3Ј-end of the human survivin promoter upstream of the luciferase gene and confirmed by sequencing. Transient transfection of MVEC was performed using Lipofectin reagent (Life Technologies, Inc.) as described previously (16). Briefly, MVEC were seeded in a 12-well plate (1-2 ϫ 10 5 cells/well) in 1 ml of medium and grown to 50 -80% confluence. 50 l of Opti-MEM I (Life Technologies, Inc.) containing 1 g of various plasmid DNAs was combined with 50 l of Opti-MEM I containing 4 l of Lipofectin reagent. The combined mixture was overlaid onto cells that were preincubated with serum-free medium for 20 -30 min. The transfected cells were then incubated at 37°C for 4 -6 h. The DNAliposome complex was replaced with complete medium, and luciferase activity/␤-gal expression (internal control) were measured within 36 -48 h post-transfection.
Transfection of MVEC with GFP Survivin Constructs-MVEC were transfected with the cDNAs for GFP, GFP-survivin (survivin), or GFP-C84A survivin (C84A survivin) for 24 h. Fusion of survivin with GFP does not interfere with its biological activity or localization. The survivin-GFP (Cys 84 -Ala) construct is a mutation in the Bir1 domain that is targeted to the mitotic spindle but is devoid of anti-apoptosis function. In experiments using GFP-survivin, approximately 30% of the cells were transfected, and apoptosis, under the various conditions tested, was determined by propidium iodide staining and flow cytometry. The percentage of cells with hypodiploid DNA content quantified in the GFP-expressing population is shown in each histogram. Aggregates of cell debris at the origin of the histogram were excluded from the analysis of sub-G 1 cells. In some experiments, cells were imaged on an inverted microscope (Zeiss, Axiovert) using DIC optics.

Ang-1 Inhibits Cell Apoptosis 9103
analogous to wortmannin. Ang-1 stimulated Akt phosphorylation in a time-dependent manner with maximal activation occurring within 15-30 min and sustained phosphorylation lasting for up to 2 h (Fig. 1C). Ang-1-stimulated phosphorylation of Akt on Ser 473 was antagonized by preincubation of Ang-1 with soluble Tie 2 receptor but not by incubation with soluble Tie 1 receptor bodies (Fig. 1D). In addition, Ang-1-induced Akt phosphorylation was partially blocked by the physiological antagonist of Ang-1, angiopoeitin-2 (Ang-2; 11). Interestingly, Ang-2 alone weakly activated Akt in MVEC. Our results are consistent with data from heterologous expression systems and transformed endothelial cells documenting that Ang-1 activation of PI3 kinase is important for cell migration and survival (9,12). Therefore, Ang-1 via the Tie 2 receptor stimulates Akt activation through a PI3 kinase-dependent mechanism. Next, we asked if Ang-1 could influence endothelial cell apoptosis induced by detachment from the matrix, i.e. anoikis (13). MVEC in serum-free media was plated onto Petri dishes for 18 h and underwent extensive apoptosis as determined by appearance of a hypodiploid cell population (ϳ25% versus 2% of control, adherent cultures) by propidium iodide staining and flow cytometry (Fig. 2A). Incubation of MVEC cultured under these conditions with Ang-1 inhibited apoptosis by 75% in a reaction abrogated by WM ( Fig. 2A). To examine whether Akt was required for Ang-1 cytoprotection, we infected MVEC with adenoviral ␤-galactosidase or activation-deficient Akt (AA-Akt; 14) and determined the degree of apoptosis by FACS analysis. Transduction of MVEC with AA-Akt abrogated the cytoprotec-tive effect of Ang-1 against anoikis, whereas a control adenovirus encoding ␤-galactosidase was ineffective. Moreover, Ang-1 stimulated Akt phosphorylation while MVEC were in suspension (Fig. 2B). WM also prevented Akt phosphorylation on Ser 473 induced by Ang-1 in suspended endothelial cells. Collectively, these data indicate that Ang-1 mediates endothelial cell protection through an integrin-independent, PI3 kinase/Akt-dependent pathway.
Next, we examined a potential link between Ang-1 and expression of two known anti-apoptotic genes, survivin and bcl-2 (15,16). Treatment of MVEC with Ang-1 rapidly induced a time-dependent increase in survivin mRNA levels (17), which peaked 12 h after stimulation and remained sustained for up to 24 h (Fig. 3A). In contrast, Ang-1 did not up-regulate bcl-2 mRNA expression in MVEC (Fig. 1A). Consistent with a receptor-mediated response, preincubation of Ang-1 with soluble Tie 2 receptor abolished Ang-1 induction of survivin RNA in MVEC (Fig. 3B). When MVEC were transfected with a survivin-luciferase promoter construct (18), Ang-1 stimulated a 3-7-fold up-regulation of survivin transcriptional activity, which persisted for up to 24 h after stimulation (Fig. 3C). Ang-1 induced the expression of survivin protein in HUVEC, an effect abrogated by WM, or by transduction with adenoviral AA-Akt (Fig.  3D). In contrast, Ang-1 did not increase the expression of bcl-2 protein expression in MVEC (Fig. 3D). These data demonstrate that Ang-1 stimulates survivin expression in endothelial cells via a PI3 kinase/Akt-dependent mechanism.
To determine whether survivin can mediate the anti-apoptotic function of Ang-1, we transfected MVEC with cDNAs containing GFP fused to wild-type survivin (GFP-survivin) or to a dominant negative Cys 84 3 Ala survivin mutant (GFP-C84A survivin) and determined cytoprotection in response to apoptosis-inducing stimuli (19). Treatment with Ang-1 or expression of GFP-survivin, alone or in combination with Ang-1, suppressed the appearance of MVEC with hypodiploid DNA content induced by TNF␣-/cycloheximide or by anoikis (Fig. 4,

Ang-1 Inhibits Cell Apoptosis 9104
A and B). In contrast, transfection of MVEC with GFP-C84A survivin abrogated the cytoprotective effect of Ang-1 against TNF␣-cycloheximide-or anoikis-induced cell death (Fig. 4, A  and B). Consistent with the above analysis, Ang-1 alone or in combination with transfected GFP-survivin resulted in healthier morphology of any adherant cells, in contrast to cells transfected with GFP alone or GFP-C84A survivin plus Ang-1 (Fig.  4C) These data identify survivin as a novel PI3 kinase/Akt-dependent target gene for Ang-1 and demonstrate that survivin is necessary for the anti-apoptotic effect of Ang-1 in endothelial cells.
In summary, Ang-1 prevents endothelial cell apoptosis by activating a critical survival messenger, Akt, and by up-regulating a broad spectrum apoptosis inhibitor, survivin. Although Akt activation is required for survivin expression and interference with survivin function by the C84A survivin mutant abolishes Ang-1 cytoprotection, activated Akt may also execute parallel anti-apoptosis pathways through phosphorylation of caspase-9 and/or Bad (20,21). Recent studies have suggested that VEGF increases survivin expression in endothelial cells (22,23). Complementing and extending these findings with VEGF, our study with Ang-1, a non-mitogenic survival factor, may have far-reaching implications for angiogenesis when endothelial cells need to loosen their focal contacts with the underlying matrix prior to emigration, proliferation, and reorganization into patent structures that can accommodate blood flow. In this scenario, inhibition of apoptosis by Ang-1/Akt/ survivin may protect the endothelium during this complex transition and maintain a critical anti-apoptotic environment during stabilization of vascular networks. Targeted manipulation of this mechanism may be exploited to improve endothelial cell viability and favor therapeutic angiogenesis in vivo.