Alternagin-C, a Disintegrin-like Protein, Induces Vascular Endothelial Cell Growth Factor (VEGF) Expression and Endothelial Cell Proliferation in Vitro *

Alternagin-C (ALT-C), a disintegrin-like protein purified from the venom of the Brazilian snake Bothrops alternatus , interacts with the major collagen I receptor, the (cid:1) 2 (cid:2) 1 integrin, inhibiting collagen binding. Here we show that ALT-C also inhibits the adhesion of a mouse fibroblast cell line (NIH-3T3) to collagen I (IC 50 2.2 (cid:3) M ). In addition, when immobilized on plate wells, ALT-C supports the adhesion of this cell line as well as of human vein endothelial cell (HUVEC). ALT-C (3 (cid:3) M ) does not detach cells that were previously bound to collagen I. ALT-C (5 n M ) induces HUVEC proliferation in vitro , and it inhibits the positive effect of fluoride, 1 mg/ml leupeptin, and 1 mg/ml pepstatin. ELISA assays were performed using the kit Quantikine immunoassay for human VEGF (R&D Sys-tems, Minneaopolis, MN) according to the manufacturer’s Briefly, 200 (cid:3) l of conditioned medium, cell extracts, controls or stand-ards were added on each well, previously coated with human mono- clonal anti-VEGF antibody. Afte r 2 h ofincubation, wells were washed and incubated with an enzyme-linked polyclonal anti-VEGF antibody. Following another wash, a substrate solution was added to wells and color developed in proportion to the amount of VEGF bound in the initial step. The plate was read on a Dynex plate reader with absorbance of 450 nm. antibody (1:200). Then, protein A/G-agarose (20 (cid:3) l/mg protein) was added, and samples were incubated at 4 °C under rotation for 2 h. The contents of Akt/PKB and phosphorylated Akt/PKB were analyzed by Western blot-ting. Cellular proteins (30 (cid:3) g) were subjected to 12% SDS-PAGE, trans- ferred to polyvinylidene difluoride filters (Hybond-P, Amersham Bio-sciences) and blocked with Tween-TBS (20 m M Tris-HCl, pH 7.5, 500 m M NaCl, 0.1% Tween-20) containing 1% bovine serum albumin. Primary antibodies used in Western analysis were anti-Akt/PKB (1:1000) and anti-phosphoserine (1:1000) antibodies. The polyvinylidene difluoride filters were next washed three times with Tween-TBS, followed by a 1-h incubation with appropriate secondary antibody conjugated to biotin. Then, the filters were incubated with streptavidin-conjugated horseradish peroxidase (1:1000). Immunoreactive proteins were visual-ized by 3,3 (cid:4) -diaminobenzidine (Sigma) staining. The bands were quan- tified by densitometry, using Scion Image Software (Scion Co., Frederick, MD). Statistical significance was assessed by ANOVA fol- lowed by Bonferroni’s t test, and p (cid:3) 0.05 was taken as statistically significant. Statistical Analysis of Data— All adhesion and cell proliferation as- says were analyzed for statistical significance. Each experiment was repeated three times in triplicate and mean and standard error mean were calculated. the comparison of multiple concentrations of ALT-C against the control group, Dunnett’s statistical approach was used. For the experiments with one dose of ALT-C in various cell types, each treated group was compared with its own control using one-way analysis of variance (ANOVA) and Bonferroni’s the ERK pathway promotes survival against a variety of apoptotic stim-uli, and this pathway involves the expression of the caspase-8 inhibitor, c-Flip (60). ALT-C induces ERK-2 translocation to the nucleus (28), and this activity could result in a protective effect against apoptosis. Here we show for the first time that these events triggered by a disintegrin-like protein culminate in HUVEC proliferation.

Cell attachment to the extracellular matrix depends mostly on the integrins, a large family of glycoproteins expressed at the cell surface (1). Integrins are heterodimers formed of noncovalently associated ␣and ␤-subunits (2). In many cells in culture, integrin-mediated adhesion results in specialized adhesion sites, named focal contacts (3). In these sites, structural and signaling proteins such as integrins, cytoskeletal proteins, and kinases are concentrated and initiate signal transduction pathways (4). Aggregation of integrin receptors, ligand occupancy, and tyrosine kinase-mediated phosphorylation are the key events that results in diverse processes such as cell migration and differentiation, tissue remodeling, cell proliferation, angiogenesis, and tumor cell invasion and metastasis (1,5).
Antagonists of integrins have been developed in order to provide powerful therapeutic approaches for the treatment of several types of cancer, such as antibodies to the ␣ v integrin (6). Synthetic peptides with the sequence Arg-Gly-Asp (RGD) can competitively block the binding of several integrins to their ligands and efficiently reduce platelet aggregation and the number of experimental metastasis (7). RGD peptides induce the disassembly of focal contacts in melanoma cells and disrupt the actin cytoskeleton (8). Disintegrins are small peptides derived from viperidae snake venoms with an internal RGD or KGD motif (9). Disintegrins can bind to integrins and interfere with integrin function. In platelets, disintegrins inhibit the adhesion of fibrinogen to its receptor, the integrin ␣ IIb ␤ 3 , resulting in inhibition of platelet aggregation (10,11). Some RGD-disintegrins have been shown to inhibit tumor cell-extracellular matrix adhesion (12) and decrease the number of experimental metastasis (13,14). RGD-disintegrins bind mostly to ␣ 5 ␤ 1 or ␣ v ␤ 3 integrins in distinct cell types therefore inhibiting also cell adhesion to fibronectin (13). Accutin and triflavin, two RGD-disintegrins from Agkistrodon acutus and Trimeresurus flavoviridis venoms, respectively, inhibit angiogenesis and induce apoptosis in endothelial cells (15,16). The VAP1 protein (vascular apoptosis-inducing protein) isolated from Crotalus atrox venom is a metalloprotease/disintegrin that induces apoptosis (17). Since it was shown that the RGD-dependent ␣ v ␤3 integrin provides a survival signal to proliferative endothelial cells during new blood vessel growth (18,19), it is thought that the anti-adhesive activity of RGD-disintegrins on endothelial cells may contribute to their anti-angiogenic activity.
A different class of disintegrin is also found in some snake venoms that do not have the RGD motif. These proteins are larger than the RGD-disintegrins (about 30 kDa) and they have an extra C-terminal, cysteine-rich domain (20 -24). These disintegrin-like proteins do not bind to the integrin ␣ IIb ␤ 3 , ␣ 5 ␤ 1 , or ␣ v ␤ 3 , but they interact with the collagen receptor, the ␣ 2 ␤ 1 integrin therefore inhibiting cell adhesion to collagen I (23).
The majority of the RGD and non-RGD disintegrins is synthesized in the venom gland as precursor forms having a proand metalloprotease domains, and proteolytic processing of these proteins releases the disintegrin-like/cysteine-rich domain (20 -21,23). It is possible to isolate the full-length protein or the processed domains from the venoms. Related proteins (the ADAMs, 1 for a disintegrin and metalloproteinase) are found in mammals as well as in several other organisms, in which they are involved in several physiological processes such as fertilization, cell differentiation, and shedding of receptors (25). The ADAMs have a similar domain organization with extra domains including transmembrane and intracellular domains (26). Both ADAMs and snake venom metalloproteinases (SVMPs) belong to the Reprolysin protein family of metalloproteases (27).
We have previously described the isolation and characterization of ALT-C, a disintegrin-like protein from Bothrops alternatus snake venom (23). ALT-C is synthesized as a precursor form with a metalloprotease domain, from which is released after proteolytic processing, yielding a form with disintegrin and cysteine-rich domains (23). ALT-C binds to the ␣ 2 ␤ 1 integrin-transfected-K562 cells therefore inhibiting collagen I adhesion (IC 50 , 100 nM). When immobilized on plate wells, ALT-C promotes the adhesion of ␣ 2 ␤ 1 integrin transfected-K562 cells but not the adhesion of control cells. ALT-C does not bind to the integrins ␣ IIb ␤ 3 , ␣ 5 ␤ 1 , ␣ v ␤ 3 , ␣ 1 ␤ 1 , ␣ 9 ␤ 1 , and ␣ 4 ␤ 1 (23). In human neutrophils, ALT-C also induces migration via integrin signaling, with activation of focal adhesion kinase (FAK) and its association with phosphatidylinositol 3-kinase (PI3-kinase), leading to the increase in the content of F-actin and Erk-2 nuclear translocation (28). It has been demonstrated that jararhagin, a metalloprotease/disintegrin homologue from Bothrops jararaca venom mimics the collagen interaction with fibroblasts resulting in an up-regulation of the ␣ 2 ␤ 1 integrin and the matrix metalloproteases MMP-1 and MT1-MMP genes (29). Therefore, disintegrin-like proteins can activate integrins leading to significant modifications of cellular events and it may be in a different way from the RGD-disintegrins. However, these events are not well understood yet.
In the present work we show that ALT-C induces endothelial cell proliferation in vitro, and these effects could be mediated at least in part by an increased expression of vascular endothelial growth factor. These effects are opposite to those observed for most RGD-disintegrins, which inhibit angiogenesis and induce apoptosis (15,30). To our knowledge, this is the first report of a disintegrin acting as a survival factor.

EXPERIMENTAL PROCEDURES
Materials-The venom of B. alternatus was kindly provided by the venom commission of Instituto Butantan, Sã o Paulo, SP, Brazil. Alternagin-C (ALT-C) was purified from B. alternatus as previously described (23). Recombinant human FGF-2, VEGF 165 and the mouse anti-VEGF monoclonal antibody (clone 26503) were purchased from R&D Systems, Minneapolis, MN. Genistein, LY294002, PMA, and bisindolylmaleimide IV (BIM) were from Calbiochem, San Diego, CA. Protein A/G-agarose, anti-Akt, and anti-phosphotyrosine antibodies were from Santa Cruz Biotechnology, streptavidin-conjugated horseradish peroxidase was from Caltag Laboratories.
Cell Lines-Mouse embryo fibroblasts NIH-3T3 were from American Type Culture Collection. Human fibroblasts were obtained from Clonetics TM .
Isolation of Human Endothelial Cells-Primary human umbilical vein endothelial cells (HUVECs) were routinely obtained by treatment of umbilical veins with 0.1% collagenase IV solution (Sigma) as previously described (31) and maintained in 199 medium (M-199) with HEPES (Sigma) supplemented with antibiotics (100 units/ml penicillin, 100 g/ml streptomycin), 2 mM glutamine, and 20% fetal calf serum (FCS, Cultilab, Campinas, Brazil), at 37°C in a humidified 5% CO 2 atmosphere until they reached confluence. Cells were used in first or second passages only, and subcultures were obtained by treatment of confluent cultures with 0.025% trypsin/0.2% EDTA solution in PBS.
Inhibition of Cell Adhesion-Cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, L-glutamine, streptomycin, fungizon, and geneticin in the case of transfected cells at 37°C in a water-jacketed CO 2 incubator. Adhesion of cells labeled with 5-chloromethylfluorescein diacetate (CMFDA) was performed as described previously (23). Briefly, collagen type I (1 g/well), was immobilized on a 96-well microtiter plates (Falcon, Pittsburgh, PA) in 20 mM HEPES buffer plus 150 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , and 1 mM MnCl 2 (adhesion buffer) overnight at 4°C. Wells were blocked with 1% BSA in adhesion buffer. Cells (5 ϫ 10 6 /ml) were labeled by incubation with 12.5 M 5-chloromethylfluorescein diacetate in adhesion buffer at 37°C for 30 min. Unbound label was removed by washing with the same buffer. Labeled cells were incubated with ALT-C at several concentrations before being transferred to the plate (1 ϫ 10 5 cells/well) and incubated at 37°C for 30 min. After washing to remove unbound cells, the remaining cells were lysed by the addition of 0.5% Triton X-100. In parallel, a standard curve was prepared in the same plate using known concentrations of labeled cells. The plates were read using a Spectra-Max Gemini XS fluorescence plate reader (Molecular Devices, Sunnyvale, CA) with a 485-nm excitation and 530-nm emission filters.
Adhesion and Detachment Assays-To test if ALT-C could induce cell adhesion, plate wells were coated with ALT-C (0.5-20 g/ml), and labeled cells were allowed to adhere for 2 h at 37°C, followed by lysis and measurement of released fluorescence. For detachment assays, collagen type I (0.1-1 g/well) was immobilized on a 96-microtiter well plate in adhesion buffer overnight at 4°C. CMFDA-labeled cells (1 ϫ 10 5 cells/well) were allowed to adhere for 30 min at 37°C followed by the addition of disintegrins to the medium. After washing with adhesion buffer to remove unbound cells, the remaining cells were lysed, and the plate was read as described above. In the case of HUVECs, the percentage of adhered cells was determined by adding MTT (0.5 mg/ml, final concentration) and incubating at 37°C for 3 h at 5% CO 2 . The solution was removed and replaced by 200 l of isopropyl alcohol, and the absorbance of the solution was measured at 595 nm.
Cell Proliferation Assay-HUVECs (10 4 cells/well) were incubated in a 96-well plate in 100 l of 199 medium plus 5% FBS for 2 h at 37°C and 5% CO 2 . Then, 100 l of fresh medium having FGF-2 or VEGF, ALT-C or both at different concentrations were added to each well, followed by 72 h incubation at 37°C and 5% CO 2 . Cell concentration was measured by adding MTT (0.5 mg/ml, final concentration) and incubating at 37°C for 3 h at 5% CO 2 as described above. For assays with mouse fibroblasts, FGF-2 and VEGF were omitted. Cell proliferation was also measured in response to immobilized ALT-C. 96-well culture plates (Nunc, Roskilde, Denmark) were incubated with ALT-C solutions in PBS (1-20 g/ml), overnight at 4°C. After blockade with 0.1% BSA in PBS, 1 ϫ 10 4 cells/well were seeded in 199 medium supplemented with 5% fetal calf serum and grown for 72 h. In order to investigate the possibility that the cell counts obtained after 72 h merely reflected a proportional difference in initial cell adhesion rates for each condition, an adhesion assay was run in parallel, and revealed after allowing cells to adhere for 2 h. The number of cells was then quantified by the MTT procedure as above.
Gene Expression-Plate wells were coated overnight with ALT-C (10 g/ml) or collagen type I (40 g/ml) and then blocked with 1% BSA. Human fibroblasts (passage 4) at 80% confluence in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum were seeded (2 ϫ 10 6 cells/ml) in 10-cm plastic dishes on ALT-C, collagen type I or directly on plastic and incubated for 2 h at 37°C in 5% CO 2 . Culture medium was removed, and cells were lysed with lysis buffer from QIAgen RNEasy kit for isolation of total RNA.
Target Preparation, GeneChip TM Hybridization, and Data Analysis-Labeled cRNA was synthesized from total RNA according to standard Affymetrix protocols. Briefly, 20 g per sample of total RNA and a poly(T) primer containing a T7 RNA polymerase promoter were used to generate double-stranded cDNA. A T7 based in vitro transcription reaction (Enzo BioArray HighYield RNA Transcript labeling kit) was used to generate biotin-labeled and amplified cRNA from doublestranded cDNA. Total RNA integrity and cRNA size distribution were analyzed using agarose gel electrophoresis in an Agilent Bioanalyzer. Ten microgram samples of labeled cRNA were hybridized to Affymetrix HU-95A probe arrays, containing probes sets representing ϳ10,000 genes, for 16 h, and scanned with the Affymetrix GeneArray Scanner. Data were analyzed with Affymetrix Microarray Analysis Suite v 5.0 and dChip© softwares. Genes up-or down-regulated at statistical significant levels (p Ͻ 0.05) and a fold change of Ն1.5 were considered to be a significant change of gene expression compared with the control cells.
Analysis of VEGF Expression by RT-PCR and ELISA-Human fibroblasts were grown on 6 cm dishes coated with ALT-C (10 g/ml), collagen type I (40 g/ml) or collagen I treated with soluble ALT-C (400 nM). Cells were lysed 4, 24, and 48 h after treatment, and total RNA was isolated using TRIzol (Invitrogen). Samples were treated with DNase before the experiments in order to discharge any contaminant DNA. RT-PCR was run using the Superscript one-step RT-PCR kit (Invitrogen), according to manufacturer's instructions. The following VEGF primers were used: forward, GAGCGGAGAAAGCATTTGTTT; reverse TGCAACGCGAGTCTGTGTTT. ␤-Actin primers were used in the same conditions as endogen controls (forward, CGTGGGCCGCCCTAG-GCACCAGGG, and reverse, CGGAGGAAGAGGATGCGGCAGTGG). PCR products were analyzed in a 2% agarose gel electrophoresis. For ELISA, cells were disrupted by using a cell lysis buffer made of 50 mM Tris-HCl, pH 7,4 plus 150 mM NaCl, 1.5 mM MgCl 2 , 1.5 mM EDTA, 1% Triton X-100, 10% Glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml leupeptin, and 1 mg/ml pepstatin. ELISA assays were performed using the kit Quantikine immunoassay for human VEGF (R&D Systems, Minneaopolis, MN) according to the manufacturer's instructions. Briefly, 200 l of conditioned medium, cell extracts, controls or standards were added on each well, previously coated with human monoclonal anti-VEGF antibody. After 2 h of incubation, wells were washed and incubated with an enzyme-linked polyclonal anti-VEGF antibody. Following another wash, a substrate solution was added to wells and color developed in proportion to the amount of VEGF bound in the initial step. The plate was read on a Dynex plate reader with absorbance of 450 nm.
Statistical Analysis of Data-All adhesion and cell proliferation assays were analyzed for statistical significance. Each experiment was repeated three times in triplicate and mean and standard error mean were calculated. For the comparison of multiple concentrations of ALT-C against the control group, Dunnett's statistical approach was used. For the experiments with one dose of ALT-C in various cell types, each treated group was compared with its own control using one-way analysis of variance (ANOVA) and Bonferroni's post-hoc analysis. Acceptable p levels were * (p Ͻ 0.05) and ** (p Ͻ 0.01).

RESULTS
ALT-C Inhibits the Adhesion of Mouse Fibroblasts to Collagen I-It was previously reported by our group that ALT-C inhibits the adhesion of ␣ 2 ␤1-transfected cells to collagen type I, with an IC 50 of 100 nM (23). Here we show that the binding of mouse fibroblasts (NIH-3T3) to collagen I is also inhibited by ALT-C (Fig. 1). ALT-C inhibited collagen binding to fibroblasts with an IC 50 of 2.2 M. Inhibition was not 100% even in higher concentrations probably due to the presence of other collagen receptors in fibroblasts. These results suggest that the integrin ␣ 2 ␤1 is involved in the adhesion of this cell type in the binding of collagen I.

ALT-C Supports the Adhesion of Fibroblasts and Endothelial
Cells-It was previously demonstrated that ALT-C supports the adhesion of ␣ 2 ␤ 1 -transfected K562 cells but not the control cells (23). Here we show that ALT-C also significantly supported the adhesion of fibroblasts and endothelial cells in a mass-dependent fashion (Fig. 2, A and B). Adhesion to collagen I was considered to be 100%. These results confirm the effect of ALT-C as an adhesion molecule.
In agreement with these data, ALT-C was unable to detach cells that were previously adhered to collagen I (Fig. 3A). We have done a collagen concentration x adhesion curve for both fibroblasts and HUVECs, in which we could see that even collagen concentrations lower than 0.01 g/well support cell adhesion (not shown). The collagen concentration that gives 50% of adhesion was 0.04 and 0.12 g/well for HUVECs and fibroblasts, respectively. The adhesion of mouse fibroblasts to collagen I (0.12 g/well) was not affected by incubation with ALT-C (1-4 M) for 2 h (Fig. 3A). Comparable results were obtained for HUVECs (Fig. 3B); however, lower ALT-C concentrations must be used since HUVECs from primary culture are more sensitive than the fibroblast cell line used in this work. Only at the dose of 100 nM ALT-C induced a low but significant detachment (10%). Thus, ALT-C strongly favors cell adhesion and does not reproduce the anti-adhesive actions already described for some RGD-type disintegrins (33)(34).
ALT-C Induces Endothelial Cell Proliferation-The ability to antagonize cell adhesion has been described as a main feature of anti-angiogenic disintegrins. Since the disintegrin-like ALT-C promoted endothelial cell adhesion, we further investigated its effects on endothelial proliferation. The stimulation of endothelial cell proliferation achieved when ALT-C was immobilized on plastic was proportional to the concentration of ALT-C coated to plastic surfaces (Fig. 4A). In the presence of 5% fetal calf serum, no significant variations were seen in the initial adhesion rates among the different conditions (adhesion measured after 2 h from seeding cells), including cells adhering on wells coated only with BSA. However, only in wells coated with ALT-C the number of cells increased after 72 h, and the increase was dependent on ALT-C concentration (Fig. 4A). HUVECs were also responsive to treatment with different concentrations (1-100 nM) of soluble ALT-C (Fig. 4B). The induc-tion of proliferation observed when treating cells with 5 nM ALT-C was comparable to that observed with cultures treated with 10 ng/ml FGF-2 or 10 ng/ml VEGF, two potent angiogenic factors (Fig. 4B). Interestingly, concentrations of ALT-C greater then 10 nM significantly reduced this effect. ALT-C (10 and 100 nM) significantly inhibited the VEGF and FGF-2 effects, respectively (Fig. 4C). ALT-C did not induce proliferation of mouse fibroblasts in any of the doses and periods tested (Fig. 4D).
Gene Expression Induced by ALT-C-In order to explain the effects of ALT-C on cell proliferation, we performed a gene expression assay using the GeneChip TM technology. When compared with human fibroblast growing on plastic, ALT-C induced a significant increase in several genes related to cell cycle control, including VEGF (Fig. 5) and other growth factors such as inducible early growth response (TGF␤), interleukin 11 (IL-11), early growth response 2 and 3 (EGR2 and 3), and insulin-induced gene (IIG1). A total of 45 genes were up-regu-

FIG. 2. ALT-C supports the adhesion of mouse fibroblasts and HUVEC cells.
A, 96-well plates were coated with ALT-C (10 -100 g/ml) or collagen I (1 g/well) in adhesion buffer at 4°C. After blocking with 1% BSA, CMFDA-labeled mouse fibroblasts (NIH-3T3) cells (10 5 cells/well) were seeded in the wells. The plates were incubated at 37°C for 30min, washed, lysed and read as in Fig. 1. The results were obtained from three independent experiments in triplicate. The means that are significantly different from those of cells growing on collagen using Dunnett's test were shown by ** (p Ͻ 0.01). B, 96-well plates were coated with ALT-C (1-20 g/ml) or collagen I (1 g/well) in adhesion buffer at 4°C. After blocking with 1% BSA, HUVEC cells were seeded and the experiment was performed as described earlier. The results for ALT-C were normalized by the collagen values in each experiment. The error bars show the S.E. of six samples with less deviation from the mean.

FIG. 3. ALT-C does not induce detachment of mouse fibroblasts (A) and HUVECs (B) from collagen I.
96-well plates were coated with collagen type I (0.12 g/well) in 0.1% acetic acid overnight at 4°C. After 1% BSA blockage, cells (10 5 cells/well) were seeded in the wells. The plates were kept at 37°C for 30 min and then incubated with ALT-C at 37°C for 2 h. After washing, remaining cells were counted. Adhesion to collagen was considered 100%. Results were obtained from three independent experiments in triplicate. Error bars show the S.E. of six samples with less deviation from the mean obtained by three independent experiments in triplicate. The mean significantly different from the control using Dunnett's test is shown by *, p Ͻ 0.05. lated and 30 genes were down-regulated with this experiment. The expression of VEGF may explain the positive effect of ALT-C on HUVEC proliferation.
The effect of immobilized ALT-C is probably similar to the collagen effect since the differences in gene expression of these two proteins were much smaller (Fig. 6). A total of 8 genes were up-regulated, and 4 genes were down-regulated when comparing the expression induced by ALT-C and collagen I. The significance of this difference is not understood yet. VEGF expression induced when human fibroblasts were grown on both ALT-C-or collagen-coated dishes for 24 h was also confirmed by RT-PCR (Fig. 7A). However, striking differences were seen when collagen-bound fibroblasts were treated with soluble ALT-C. VEGF expression was significantly increased after 48 h of incubation, while cells growing on immobilized ALT-C or collagen showed no difference (Fig. 7B).
ALT-C Induces Akt/PKB Phosphorylation in HUVECs-Phosphatidylinositol 3-kinase (PI3K)-Akt/PKB signaling axis is activated by many angiogenic growth factors (36). To determine whether the Akt/PKB activating pathways were essential for the proliferative effect induced by ALT-C on endothelial cells, we incubated HUVEC monolayers with ALT-C (5 nM) for 15 min, in the absence or presence of different kinase inhibitors. The interaction of ALT-C with endothelial cells strongly activated Akt/PKB phosphorylation, as compared with control cultures (Fig. 8). Akt/PKB is a downstream target of activated PI3K (37) and PKC has been pointed as a key mediator of endothelial cell proliferation and differentiation (38,39). Accordingly, ALT-C-activated Akt/PKB phosphorylation was abrogated by both LY294002, a selective inhibitor of PI3K, and BIM, a broad inhibitor of PKC family members. Tyrosine kinases may also contribute to the activation of Akt/PKB by ALT-C, since the treatment of HUVECs in the presence of genistein strongly inhibited Akt/PKB phosphorylation (Fig. 8). DISCUSSION We have previously demonstrated that ALT-C is a ligand for the major collagen I receptor, the integrin ␣ 2 ␤1 and also supports K562 cell adhesion mediated by this integrin (23). In this report we show that ALT-C also supports the adhesion of other cell types such as mouse fibroblasts and HUVECs. Moreover, the adhesion of fibroblasts to collagen I is inhibited by ALT-C, supporting the evidence that the ␣ 2 ␤1 integrin is a major collagen receptor in these cells. As would be expected for an adhesive protein, ALT-C failed to induce fibroblast and HU-VEC detachment from collagen-coated surfaces, even in low collagen concentrations. It has been suggested that jararhagin, a metalloprotease with disintegrin-like and cysteine-rich domains, acts as a collagen agonist of the ␣ 2 ␤ 1 integrin, causing the activation of this integrin and producing collagen-like cell signaling events such as the up-regulation of matrix metalloproteases (29). Since ALT-C does not have the metalloprotease domain, the results presented here provide strong evidence that the disintegrin and cysteine-rich domains are responsible for integrin activation. Recently, it has been demonstrated that two peptides derived from the Cys-rich domain of jararhagin and from a homologue, atrolysin a from Crotalus atrox snake venom, inhibit collagen I binding to platelets. This effect does not involve GPIV, another collagen receptor in platelets (40).
Our present data also show that ALT-C provides a suitable support for the adhesion of HUVECs. Moreover, in response to growing concentrations of ALT-C, either immobilized to plastic or incubated with HUVECs in the soluble form, it strongly induced endothelial cell proliferation. ALT-C up-regulates the expression of VEGF in human fibroblasts, which could explain the increase in HUVEC proliferation. We do not know if ALT-C could induce VEGF expression in HUVEC cells but it remains a possibility to be confirmed. Interestingly, ALT-C did not induce proliferation of mouse fibroblasts, thus suggesting a cell-specific effect.
The proliferative effect of ALT-C alone was similar to those exerted by VEGF and FGF-2, and the presence of ALT-C partially inhibited the endothelial cell proliferation induced by VEGF and FGF-2. Taken together, these data suggest that these proteins may partially act by a common cross-talk of signaling cascades (41).
ALT-C also up-regulates the expression of other growth factors involved in cell proliferation such as IL-11, TGF␤, and EGR2 and 3, which are probably also involved on its positive effect on HUVEC proliferation. IL-11 induces proliferation of human T-cells (42), stimulates hematopoiesis and inhibits apoptosis in a variety of cells (43). EGR2 and 3 stimulate the activities of several transcription factors that are associated with cell proliferation such as c-Fos, SRF, and c-Myc (44). Interestingly, TGF␤ regulates several inhibitory cell-cycle proteins such as p27 and p15 (45). It seems that a delicate balance between the levels of different growth factors may exist, and the effect of factors that activate the cell cycle overwhelms the effect of factors with opposite actions.
Cell detachment usually results in anoikis, a form of apoptotic cell death that occurs upon loss of matrix attachment (46), except for transformed cells expressing activated Src and Ras oncogenes (47). As demonstrated for fibroblasts, ALT-C does not induce endothelial detachment from collagen, gelatin or fibronectin-coated surfaces (data not shown) and, apparently depending on cell type, it induces cell proliferation, therefore acting as a survival factor. It has been shown that integrins and growth factor receptors coordinately regulate the expression of pro-apoptotic proteins to prevent anoikis (48). The small differences in the gene expression pattern of cell adhesion induced by immobilized ALT-C or collagen I are in agreement with the idea that this disintegrin-like protein could act as a collagen-mimetic (29). However, the effect of soluble ALT-C was very different from collagen, since it induced an important and prolonged increase in VEGF expression. These results suggest a role for ALT-C on HUVEC signaling by providing cells with sustained survival signals.
The serine/threonine protein kinase Akt/PKB has been described as a key regulator of cell viability (49,50). Activation of Akt/PKB occurs through the direct binding, in the plasma membrane, of the phosphoinositide products of PI3K reaction to its pleckstrin homology domain (51). It has been demonstrated that matrix adhesion and Ras transformation both activate the PI3K 3 Akt/PKB survival pathway (52) and that Akt/PKB is targeted for destruction by caspases during anoikis of endothelial cells (53). VEGF activation of Akt/PKB signaling in endothelial cells is also dependent on matrix attachment, and constitutively active Akt/PKB blocks the apoptosis induced by endothelial cell detachment (54).
On the other hand, many mitogens are activators of PKC family members, which have also been implicated as important mediators of endothelial cell viability (55). Phorbol esters, which mimic diacylglycerol and hence activate PKC isoenzymes, activate endothelial proliferation and induce the formation of tube-like structures by endothelial cells in vitro (38,39).
Since ALT-C acted as a strong promoter of endothelial cell FIG. 7. ALT-C induces VEGF expression in human fibroblasts. Fibroblasts were grown on ALT-C (10 g/ml), collagen type I (Collagen, 40 g/ml) or collagen I treated with ALT-C (400 nM) (Coll ϩ ALT-C). Cells were lysed 4, 24, or 48 h after treatment for total RNA and protein isolation. A, detection of VEGF mRNA. RNA samples (24 h) were treated with DNase before the experiments in order to discharge any contaminant DNA. RT-PCRs were run using the VEGF primers as described under "Experimental Procedures." ␤-Actin primers were used as endogenous control. B, detection of VEGF by immunoassay. ELISA assays for VEGF detection were performed using the kit Quantikine immunoassay for human VEGF as described under "Experimental Procedures." Error bars show the S.E. of two independent experiments in triplicate. The mean significantly different using Bonferroni's post-hoc is shown by ***, p Ͻ 0.001.

FIG. 8. ALT-C induces Akt/PKB phosphorylation in HUVECs.
Cells were incubated with 5 nM ALT-C or PMA 15 min at 37°C, with or without inhibitors genistein (80 M), LY294002 (3 M), or BIM (10 nM). Cell lysate was immunoprecipitated with anti-Akt/PKB antibodies and probed with anti-Akt/PKB and anti-phosphotyrosine antibodies by Western blotting. Statistical significance was assessed by ANOVA followed by BonferroniЈs t test, and p Ͻ 0.05 was taken as statistically significant.
adhesion and proliferation, we examined whether the interaction of ALT-C with HUVECs promotes the phosphorylation of Akt/PKB or involves the participation of PKC. Under the conditions of our assays, ALT-C promoted an important increase in Akt/PKB phosphorylation, which was blocked by the presence of PI3K and PKC inhibitors. Moreover, in the same assays, Akt/PKB phosphorylation was increased by PMA (a phorbol ester) to levels comparable to those resulting from ALT-C treatment of HUVECs. These data demonstrate that the binding of ALT-C to endothelial cells, or that the cell adhesion to this disintegrin-like molecule, initiates signaling cascades leading to the activation of PI3K 3 Akt/PKB survival pathway, and that PKC family members also participate in this response. In fact, it was shown that PMA activation of HUVECs also leads to the activation of PI3K (56).
The fact that genistein, a potent tyrosine-kinase inhibitor, also resulted in the blockade of ALT-C-induced Akt/PKB phosphorylation in endothelial cells is consistent with our previous observations (28). It has been demonstrated that disintegrins can induce intracellular signaling events such as an increase in the content of F-actin and cytoskeleton organization (57), protein tyrosine phosphorylation (58), and up-or down-regulation of ECM-related genes (29). In human neutrophils, ALT-C induces migration via integrin signaling, with activation of focal adhesion kinase (FAK) and its association with PI3K (28). In addition to PI3K, it is possible that ALT-C could be activating other upstream key proteins of the cross-talk signaling network common to integrins and tyrosine kinases receptors (growth and survival factors receptors) such as FAK and Ras, triggering the cell cycle (1,5,59). The activation of the ERK pathway promotes survival against a variety of apoptotic stimuli, and this pathway involves the expression of the caspase-8 inhibitor, c-Flip (60). ALT-C induces ERK-2 translocation to the nucleus (28), and this activity could result in a protective effect against apoptosis. Here we show for the first time that these events triggered by a disintegrin-like protein culminate in HUVEC proliferation.
Striking differences were observed when results were compared with BaG, a dimeric metalloprotease/disintegrin protein from B. alternatus venom (61). BaG (4 M) induced a significant detachment (64%) of fibronectin-bound K562 cells after 2h incubation (61). The proteolytic activity of BaG was completely inhibited by treatment with o-phenanthroline therefore a cleavage effect on cell surface proteins can be discharged. BaG is a ligand of ␣ 5 ␤ 1 integrin, therefore it inhibits cell binding to fibronectin. Also, saxatilin, an RGD-disintegrin from Gloydius saxatilis snake venom inhibited HUVEC proliferation (62). It was reported that ␣ v integrin antagonists with an RGD sequence induce the disassembly of focal contacts in melanoma cells, disruption of actin cytoskeleton and cell detachment (8). The snake venom RGD-disintegrin salmosin induces apoptosis by disassembly of focal adhesions in bovine endothelial cells (30). Accutin, an RGD-disintegrin also induces apoptosis in endothelial cells (15). These results show significant differences between ALT-C and the RGD-disintegrins, regarding the effects on cell adhesion and proliferation. Such opposite effects have not been demonstrated yet.
Members of the ADAMs protein family in mammals and other species have been demonstrated to bind to several integrins and also supporting cell adhesion through the disintegrin/cysteine-rich domains (35,63). However, very little is known about intracellular signaling or gene expression mediated by the interaction of ADAMs and integrins. Given the homology of snake venom disintegrin-like proteins and the disintegrin domains of the members of the ADAMs protein family, these results suggest that similar functions could be asserted to the last ones, such as signaling via integrins leading to key events as gene expression and affecting cell proliferation. Due to the positive effect on cell proliferation, we suggest that ALT-C could be called a pro-integrin. If induction of cell proliferation is a characteristic of all disintegrin-like proteins remains to be elucidated.