The YSNSG cyclopeptide derived from tumstatin inhibits tumor angiogenesis by down-regulating endothelial cell migration

Abstract

We previously demonstrated that the CNYYSNS peptide derived from tumstatin inhibited in vivo tumor progression. The YSNS motif formed a β-turn crucial for biological activity. More recently, a YSNSG cyclopeptide with a constrained β-turn on the YSNS residues was designed. Intraperitoneal administration of the YSNSG cyclopeptide inhibited in vivo melanoma progression more efficiently than the native linear peptide. In the present article, we showed that the YSNSG cyclopeptide also triggered an inhibition of in vivo tumor neovascularization and we further analyzed its in vitroantiangiogenic effect. The YSNSG cyclopeptide did not alter endothelial cell proliferation but inhibited cell migration by 83% in an in vitro wound healing model. The inhibition was mediated by a decrease in active MT1-MMP at the migration front as well as a decrease in u-PA and u-PAR expression. The cyclopeptide also altered β1-integrin distribution in endothelial cell lamellipodia, induced a strong decrease in the phosphorylated focal adhesion kinase (p125^FAK), disorganized F-actin stress fibers and decreased the number of lamellipodia, resulting in a non migratory phenotype. Our results confirm the YSNSG cyclopeptide as a potent antitumor agent, through both the inhibition of invasive properties of tumor cells and the antiangiogenic activity.

Tumor angiogenesis plays a crucial role in the growth of solid tumors.1, 2 Among the various processes that regulate angiogenesis, the generation of proteolytic activity is thought to be pivotal in the regulation of endothelial cell migration and capillary tube formation. Key factors involved in proteolytic degradation are matrix metalloproteinases, especially MT1-MMP,3 as well as cell-bound urokinase-type plasminogen activator (u-PA) and plasmin.4 The u-PA receptor, u-PAR, is also largely involved in endothelial cell migration.5 u-PAR was shown to bind β1-integrin through its D2 domain, to induce outside-in signalling, FAK phosphorylation and to promote Human umbilical vein endothelial cell (HUVEC) migration.6, 7

NC1 domains from Type IV collagen, but also from other basement membrane-associated collagens, display several biological activities, mainly regulating tumor invasion and angiogenesis in various cancer types.8 NC1 α3(IV) chain, also called tumstatin, was shown to display antiangiogenic properties.9–12

In previous studies, we demonstrated that the NC1[α3(IV) 185–203] peptide exhibited both in vitro and in vivo antitumor properties.13, 14 The shorter CNYYSNS peptide, corresponding to residues 185–191 of the NC1[α3(IV)] domain, reproduced the antitumor properties.15 On the YSNS residues, the peptide adopted a β-turn conformation crucial for biological activity. Based on these results, a cyclopeptide was designed containing both the four YSNS residues forming the β-turn and a glycine residue allowing cyclization, which constrained the β-turn conformation, increased the peptide stability and therefore biological activity. In an in vivo mouse melanoma model, the YSNSG cyclopeptide was more efficient than the linear CNYYSNS peptide to inhibit tumor progression by decreasing invasive properties of tumor cells through an inhibition of MMPs and plasminogen activation system.16

In this in vivo model, the administration of the cyclopeptide also reduced the neovascularization of the tumors. Consequently, we studied the mechanisms of inhibition of the YSNSG cyclopeptide on angiogenesis. We found that, in vitro, it inhibited capillary-like pseudotube formation by endothelial cells, without affecting their proliferation or inducing apoptosis. Endothelial cell migration was largely decreased in an in vitro wound healing model. The cyclopeptide inhibited active MT1-MMP expression at the migration front, as well as u-PA and u-PAR expression. It modified β1-integrin distribution, decreased phosphorylated p125^FAK and induced a strong disorganization of cytoskeleton architecture resulting in a non migratory phenotype of endothelial cells.

Material and Methods

Reagents

The YSNSG cyclopeptide was synthesized by Ansynth Service B.V. (Roosendaal, Netherlands) and dissolved in the culture medium.In all in vitro experiments, it was used at 10 μM, the most effective concentration determined in our previous studies.16 The WST-1 reagent was from Roche Diagnostic (Meylan, France). Matrigel was from Sigma-Aldrich (Saint Quentin Fallavier, France). Plasminogen was from Calbiochem (distributed by VWR, Strasbourg, France). S-2251 was from Chromogenix (Instrumentation Laboratory S.A., Saint-Mandé, France). Anti-u-PAR (rabbit polyclonal antibody, FL-290), anti-MT1-MMP Hinge region (rabbit polyclonal antibody) and anti-p125^FAK (rabbit polyclonalantibody, Tyr 861) were from Santa Cruz Biotechnology (Tebu-bio, Le Perray-en-Yvelines, France). Anti-CD31 (rat anti-mouse monoclonal antibody, clone MEC 13.3) was from Pharmingen (BD Biosciences, Le Pont de Claix, France). Anti-β1 (mouse monoclonal antibody, clone 6S6) and anti-MT1-MMP blocking antibody (mouse monoclonal, clone LEM-2/63.1) were from Chemicon (Millipore, Saint Quentin en Yvelines, France). Anti-Caveolin-1 (rabbit polyclonal) was from Cell Signalling (distributed by Ozyme, Saint Quentin Yvelines, France). Anti-PAI-1 (goat polyclonal antibody) was from American Diagnostica inc. (Neuville sur Oise, France). Anti-MT1-MMP, propeptide region (rabbit polyclonal) was from USBiological (distributed by Euromedex, Souffelweyersheim, France). Alexa-568-conjugated phalloidin and Alexa-488-conjugated secondary antibody were from Molecular Probes (Fisher Scientific, Illkirch, France).

In vivo studies

Tumor growth measurement

Suspensions of B16F1 cells (2.5 × 10⁵ cells in 0.1 mL RPMI 1640 medium) were subcutaneously injected into the left side of different series of syngeneic C57Bl6 mice. Intraperitoneal administrations of cyclopeptide (10 mg/kg) were performed at days 3, 5 and 7. Tumor sizes were measured every 2 days. Tumor volume was determined according to v = ½ A × B², where A denotes the largest dimension of the tumor and B represents the smallest dimension.17 Mice were sacrificed at day 14 and tumors were surgically extracted for morphological and biochemical studies.

Tumor angiogenesis analysis by immunohistochemistry

Frozen tumor specimens were cut in 5-μm sections using a cryostat apparatus, put on a coated glass slide, and fixed in acetone. Sections were preincubated for 20 min with serum and incubated overnight at 4°C with a rat anti-mouse CD31 antibody, followed by biotinylated rabbit anti-rat, peroxidase-labeled streptavidin and the peroxidase substrate 3-amino-9-ethyl-carbazole (Vector Laboratories,Burlingame, CA), and H₂O₂, which gives a dark red reaction product. Tumor sections were counterstained with hematoxylin. Semiquantitative evaluation of capillary vessels (CD31 positive) was done using the Image J software.

Cell cultures

HUVEC were purchased from PromoCell (Heidelberg, Germany). They were cultured in endothelial cell growth medium (ECGM) supplemented with 0.4% (w/v) endothelial cell growth supplement/Heparin (ECGS/H), 5% (v/v) fetal calf serum (FCS), 10 ng/mL epidermal growth factor (EGF), 1 μg/mL hydrocortisone, 50 ng/mL amphotericin B.

In all experiments, cell viability was greater than 98%, as assessed by trypan blue exclusion.

Cell proliferation measurement

Cell proliferation was determined using the WST-1 reagent according to the manufacturer's instructions (Roche Diagnostic, Meylan, France). Briefly, 1 × 10⁴ endothelial cells were seeded on 96-well plates and incubated for 24, 48, 72, 96 hr with or without 10 μM cyclopeptide, in the presence of 1% FCS. 10 μL of WST-1 were added to the medium and, at the end of a 20 min incubation period, absorbance was measured at 450 nm.

Cell migration assay

Endothelial cells (1 × 10⁵) were seeded on 24-well plates. After reaching subconfluence, the cells were starved overnight and the cell layer was wounded with a sterile 100- to 1000-μL pipet tip. The scratch resulted in a cell free gap of ∼1.0 mm between two adjoining areas of endothelial cells. After washing away suspended cells, the cells were incubated with or without 10 μM cyclopeptide, in the presence of 1% FCS. The photographs (4 photographs/wound and 3 replicates for each condition) were taken on an inverted microscope immediately after the scratch and at 24 and 48 h. The size of the wound (area of the scratch) was determined using Image J software.

Capillary tube formation on Matrigel®

Matrigel® (10 mg/mL) was added to a 24-well culture plate (200 μL per well). After 30 min of incubation at 37°C, 1 × 10⁵ HUVEC were suspended in serum-free ECGM and seeded onto the gel. They were incubated with or without the cyclopeptide (10 μM). Capillary tube formation was observed after a 24 h incubation period under a phase-contrast inverted microscope and photographed. The semiquantitative evaluation of the pseudotube network was performed using the pixelisation software Image J.18

Zymography analyzes

Cell incubation with the peptides

At subconfluence, HUVEC were washed twice with Phosphate-buffered saline (PBS) to remove residual FCS and incubated for 48 h in ECGM, with or without the cyclopeptide (10 μM). Conditioned media were harvested and centrifuged at 500g for 10 min at 4°C to remove cellular debris. Protein contents of conditioned media were determined by the Bradford method, using bovine serum albumin (BSA) as a standard.19

Gelatin/plasminogen zymography

For the determination of plasminogen activators, HUVEC-conditioned media were analyzed on SDS-polyacrylamide gels containing 1 mg/mL gelatin and 10 μg/mL plasminogen.20

Plasmin generating activity measurement

Plasmin generating activity was measured in conditioned media using the H-D-Val-Leu-Lys-pNA (S-2251) peptide as a substrate.21 Briefly, 185 μL of 0.1 M Tris HCl buffer, pH 7.8 containing 0.4 mM S-2251 and 4 μg/mL plasminogen were added to 96-well low binding titration plate (Nunc) and incubated with 15 μL (20 μg of proteins) of conditioned media for 24 hr at 37°C. Absorbance was measured at 405 nm.

Cell surface biotinylation

After incubation with the different effectors, cells (1 × 10⁶) were washed twice with serum-free medium and surface proteins were labeled with 2.5 mL of 500 μg/mL EZ-LinkTM Sulfo-NHS-LC-Biotin (Pierce, Rockford) in PBS under gentle shaking at 4°C for 30 min. Cells were washed three times with 5 mL PBS and incubated with 5 mL of 100 mM glycine/PBS (pH 7.2) for 30 min under gentle shaking at 4°C. Cells were washed three times with 5 mL PBS and lysed with NP40 (Nonidet P40) buffer [50 mM Na2PO4 (pH 8.0), 300 mM NaCl, 1% (v/v) NP40, 10 mM EDTA, 1 mM PMSF, 10 μg/mL leupeptin, 10 μg/mL aprotinin]. Insoluble material was removed by centrifugation (14,000g at 4°C for 20 min). Protein concentration was evaluated and diluted to a concentration of 250 μg/mL. One mL of diluted lysate was added to 40 μL ImmunoPure® immobilized monomeric avidin gel (Pierce, Rockford) and incubated overnight at 4°C on a rotator. Avidin gel was washed five times with NP40 buffer (10,000g at 4°C for 30 s) and 1× reducing Laemmli sample buffer boiled for 5 min. After centrifugation (10,000g at 4°C for 30 s), the samples were subjected to SDS-PAGE and Western blot analysis using an anti-MT1-MMP (Hinge region, Santa-Cruz Biotechnology).

Western blot analysis

Samples were electrophoresed in a 0.1% SDS, 10% polyacrylamide gel. They were then transferred onto Immobilon-P membranes (Millipore, St. Quentin en Yvelines, France). The membranes were blocked with 5% nonfat dry milk, 0.1% tween 20 in a 50 mM Tris-HCl buffer, 150 mM NaCl, pH 7.5 (TBS) for 2 h at room temperature, incubated overnight at 4°C with anti-MT1-MMP (0.2 μg/mL), anti-β1 integrin (1 μg/mL), anti-u-PAR antibody (1 μg/mL), anti-PAI-1 anti-FAK (1 μg/mL) or anti p125^FAK Tyr 861 (1 μg/mL), and then for 1 h at room temperature with a second peroxidase-conjugated anti-IgG antibody. Immune complexes were visualized with the ECL chemoluminescence detection kit (GE Healthcare, Orsay, France).

Immunofluorescence

F-actin detection

Cells were plated on glass-slides and incubated for 24 or 48 h with the YSNSG cyclopeptide (10 μM). They were fixed with 3.7% formaldehyde for 10 min at room temperature, and permeabilized with acetone for 3 min at −20°C. The slides were washed with PBS and saturated in PBS with 3% BSA. Cells were then incubated for 1 h with Alexa-568- conjugated phalloidin diluted 1/40 in PBS with 1% BSA.

p125^FAK and β1 integrin detection

HUVEC were plated on glass-slides and incubated for 24 or 48 h with the YSNSG cyclopeptide (10 μM). They were fixed for 15 min with methanol. The slides were washed with a Tris buffered saline solution with 0.01% Tween (TBS-T) and saturated in TBS-T with 3% BSA. Cells were then incubated for 2 h at room temperature with an anti-β1 integrin subunit antibody or an anti-p125^FAK antibody diluted 1/100 in TBS-T with 3% BSA. Slides were washed in TBS-T and cells were incubated for 1 h with the Alexa-488-conjugated secondary antibody diluted 1/1,000 in TBS-T with 3% BSA. Cells were then washed with TBS-T. Control preparations were incubated with omission of the first antibody. Nuclei were counterstained with Hoechst 33,342. Immunofluorescence-labeled cell preparations were studied using a Leica DMIRE2 confocal laser scanning microscope with the 63× oil-immersion objective and Leica TCS SP2 operating system (Leica Microsystemes SAS, Rueil-Malmaison, France). Acquisitions were performed by exciting the Alexa-488 and the Hoechst dye with the 488 nm line of an argon ion laser and the 351 nm line of an air cooled argon-UV laser respectively. The emitted fluorescence was detected through the appropriate adjustment of the AOBS® (Acousto Optical Beam Splitter) system.

MT1-MMP and caveolin detection

HUVEC cells were seeded onto glass coverslips and incubated for 48h with or without YSNSG cyclopeptide (10 μM) at 37°C in presence of 5% CO₂ and then fixed in paraformaldehyde for 10 min. After three washes in PBS for 5 min, cells were permeabilized 10 min at room temperature in PBS with 0.1% triton ×100 and washed three times in PBS for 5 min. Then cells were incubated in blocking solution (PBS containing 3% BSA). Coverslips were then incubated overnight at 4°C with the primary antibodies (anti-caveolin, anti MT1-MMP (hinge region), anti pro-MT1-MMP). Immunolabelling were revealed after 1 hr-incubation with the secondary antibodies conjugated to AlexaFluor 488 (green) or Alexafluor 568 (red). Nuclei were counterstained with DAPI (blue). A Leica SP2 confocal system mounted on an DRM2 optical microscope (Leica Microsystems, Heidelberg, Germany) was used for acquisitions. All acquisitions were made using UPlan FI×63, 1.4 numerical aperture objective. We used the 488-nm of an air-cooled 100 mW argon laser for excitations of AlexaFluor 488 and a 568-nm line of an air-cooled 60 mW krypton argon laser for excitation of AlexaFluor 568. Emitted fluorescence was detected through the combination of the appropriate filter set and 40 images were captured with a 0.2 μm z-step. Images were treated with Image J.

Statistical analyzes

For in vivo experiments, volumes of primary tumors were statistically analyzed using the nonparametric u-test of Mann and Whitney and the parametric Student's t test paired with weight-matched mice. For in vitro experiments, statistical analyzes were performed by Student's t test and results expressed as means ± 1 SD. Each experiment was performed 3 times and each time in triplicates.

Results

The YSNSG cyclopeptide inhibits in vivo tumor growth and angiogenesis

B16F1 melanoma cells were subcutaneously injected into the left side of C57Bl6 mice. Intra-peritoneal injections of YSNSG cyclopeptide were performed at days 3, 5 and 7. Tumors were measured every 2 days. The cyclopeptide inhibited tumor growth by 49% at day 14 (Fig. 1a). The results were similar to those we previously published.16

Figure 1. The YSNSG cyclopeptide inhibits in vivo tumor growth and angiogenesis. B16F1 cells were subcutaneously injected to syngeneic C57Bl6 mice (2.5 × 10⁵ cells per mouse). Intraperitoneal administrations of the YSNSG cyclopeptide (10 mg/kg) was performed at days 3, 5 and 7. (a) Inhibition of tumor volume was shown at day 14. Statistical significance was determined using the nonparametric u-test of Mann and Whitney. (b) Mice were sacrificed at day 14 and tumors were surgically extracted. Tumor sections were immunohistochemically stained with an anti-CD31 antibody (positive labeling appears in dark red) and counterstained with hematoxylin. Negative control was prepared by omitting the primary antibody. Magnification, ×200. Labeling was quantified using the Image J software.

Mice were sacrificed at this time and tumors were surgically extracted. Labeling of tumor section with an anti-CD31 antibody showed a decrease in tumor neovascularization in mice treated with the YSNSG cyclopeptide versus control mice (Fig. 1b). Control experiments, performed either with an irrelevant primary antibody (data not shown) or without primary antibody, showed the absence of staining and demonstrated labeling specificity.

The YSNSG cyclopeptide decreases in vitro pseudotube formation by endothelial cells

To test the effects of the YSNSG cyclopeptide on angiogenesis in vitro, HUVEC were seeded onto Matrigel and incubated for 24 h with or without the YSNSG cyclopeptide (Fig. 2a). The presence of the YSNSG cyclopeptide decreased pseudotube formation by 52% compared with control (Fig. 2b). To determine if MMPs or plasminogen activators were involved in HUVEC pseudotube formation, cells were incubated for 24 h in the presence of protease inhibitors, Galardin (10⁻⁹ M)22 (dissolved in DMSO, final concentration 0.0001%) and Aprotinin (25 μg/mL or 3.84 × 10⁻⁹ M).23 Galardin inhibited HUVEC pseudotube formation by 46% and Aprotinin by 41%. The combined effects of Galardin and Aprotinin inhibited pseudotube formation by 50%. The use of a blocking antibody directed against the catalytic domain of MT1-MMP also inhibited pseudotube formation by 50%. These results gave evidence that MMP and plasminogen activation cascades are both involved in HUVEC pseudotube formation. Combined addition of Galardin and Aprotinin and C5 inhibited pseudotube formation by 57%, suggesting that almost all the cyclopeptide effects resulted from MMP and plasminogen activator inhibition.

Figure 2. The YSNSG cyclopeptide decreases in vitro pseudotube formation by endothelial cells. (a) Capillary tube formation on Matrigel was observed after a 24 h incubation period under a phase-contrast inverted microscope and photographed. (b) The semiquantitative evaluation of the pseudotube network was performed using the pixelisation software Image J. **: significantly different from control at p < 0.01.

The YSNSG cyclopeptide inhibits in vitro endothelial cell migration

The YSNSG cyclopeptide had no effect on cell proliferation, as measured by WST-1 colorimetric assay, and did not induce HUVEC apoptosis, as evaluated by Hoechst staining (data not shown). Its effects on HUVEC migration was analyzed in an in vitro wound assay, as described in the “Material and Methods” section. In the control dishes, cells completely filled the wound after 48 hr (Fig. 3). Treatment with the YSNSG cyclopeptide slowed down the migration by 85% after 48 hr. Treatment with Galardin (10⁻⁹ M) and Aprotinin (25 μg/mL) decreased wound repair by 70% and 40% respectively. These results suggested that MMP and plasminogen activation cascades were also both involved in HUVEC migration.

Figure 3. The YSNSG cyclopeptide inhibits in vitro endothelial cell migration. HUVEC were seeded on 24-well plates. Artificial wound was performed with a 100. to 1000-μL pipet tip (arrows) and photographed at T 0 and T 48 h under a phase-contrast inverted microscope, scale bar: 50 μm and migration was quantified using the Image J software. *: significantly different from control at p < 0.05; **: significantly different from control at p < 0.01.

Effect of the YSNSG cyclopeptide on matrix metalloproteinases and their inhibitors

Matrix metalloproteinases, especially MMP-2 and MT1-MMP (also called MMP-14), are largely involved in extracellular matrix degradation and endothelial cell migration.3, 24

No significant changes in MMP-2 secretion were detected by gelatin-zymography analysis of cell media after incubation of HUVEC with the YSNSG cyclopeptide (Fig. 4a). As well, the analysis of the secretion of Tissue Inhibitors of Matrix Metalloproteinases (TIMPs) into the conditioned cultured media, by reverse zymography, did not show any significant change (data not shown).

Figure 4. Effect of the YSNSG cyclopeptide on MT1-MMP expression, activation and cellular distribution. HUVEC were incubated for 48 h with control medium or with the YSNSG cyclopeptide (10 μM). (a) ProMMP-2 secretion was studied by gelatin zymography. (b) MT1-MMP expression and activation in whole cell extracts were analyzed by western blot. (c) Cells were fixed with formaldehyde and labeled with anti-caveolin-1 (green), anti-MT1-MMP (hinge region) (red). In merged images, yellow shows colocalization of pro and/or active MT1-MMP (hinge region) and caveolin. Nuclei were conterstained with DAPI (blue). Insert: higher magnification of cell protrusion. (d) Cells were fixed with formaldehyde and labeled with anti-caveolin-1 (green), anti-proMT1-MMP (prodomain) (red). In merged images, yellow shows co-localization of proMT1-MMP and caveolin. Nuclei were conterstained with DAPI (blue). (e) Cell surface proteins were biotinylated, isolated and submitted to western blot analysis with an anti-MT1-MMP (hinge region).

By contrast, the YSNSG cyclopeptide strongly inhibited active MT1-MMP production (Fig. 4b) as demonstrated by western blot analysis on whole cell extracts. ProMT1-MMP activation by furin takes place intracellularly after exit from the Golgi,25 and then the active enzyme is addressed to cell surface. Immunolocalization using a rabbit anti-MT1-MMP polyclonal antibody that recognized both pro and active MT1-MMP showed that, in the nontreated endothelial cells, immunofluorescence was localized into the cytoplasm and at the migration front. By contrast, in the presence of the YSNSG cyclopeptide, no immunofluorescence was detected at the migration front whereas it accumulated in the cytoplasm (Fig. 4c). Immunolocalization with an anti-MT1-MMP that recognized proMT1-MMP exclusively (Fig. 4d) showed that, in the nontreated endothelial cells, immunofluorescence was localized in the cytoplasm and absent at the migration front. We concluded that MT1-MMP was present at the migration front as an active form and that YSNSG treatment lead to the disappearance of active MT1-MMP at the migration front. This was confirmed by cell surface biotinylation studies (Fig. 4e). Only active MT1-MMP was present at the cell surface and active MT1-MMP was decreased after YSNSG treatment. More over, in Figure 4d, merge image showed that proMT1-MMP colocalized with caveolin-1, suggesting an accumulation of the zymogen form into caveolae.

Since MT1-MMP is able to degrade extracellular matrix (Type I and III collagens, fibronectin, laminins, fibrin, gelatin, nidogen, and cartilage proteoglycan core proteins) by itself,26–30 this should explain, at least partially, the strong reduction in the migration capacities of HUVEC.

Effect of the YSNSG cyclopeptide on the plasminogen activation system

Another proteolytic cascade involved in tumor angiogenesis is the activation of plasminogen into plasmin by plasminogen activators, u-PA and t-PA.4 Plasmin generated activity was analyzed by measuring the hydrolysis of the H-D-Val-Leu-Lys-pNA peptide (Fig. 5a), secretion of u-PA and t-PA by gelatin-plasminogen zymography (Fig. 5b), secretion of PAI-1 and expression of u-PAR by western blot (Figs. 5c and 5d respectively). Treatment of HUVEC with the YSNSG cyclopeptide triggered a 51% decrease in plasmin generated activity (Fig. 5a). It also induced a 27% decrease in u-PA secretion, whereas t-PA secretion was not altered (Fig. 5b). No changes in PAI-1 secretion were detected after cyclopeptide treatment (Fig. 5c). By contrast, treatment of HUVEC with the cyclopeptide induced a large decrease in u-PAR expression (−60%) (Fig. 5d). Collectively, the decrease in active MT1-MMP at the migration front and the decrease in plasmin generation may explain most of the inhibition of in vitro HUVEC migration.

Figure 5. Effect of the YSNSG cyclopeptide on the plasminogen activation system. HUVEC were incubated for 48 h with control medium or with YSNSG cyclopeptide (10 μM). (a) Plasmin generated activity was measured in conditioned media using the H-D-Val-Leu-Lys-pNA (S-2251) peptide as a substrate and absorbance was recorded at 405 nm. (b) u-PA and t-PA secretion into the conditioned media were analyzed by gelatin-plasminogen zymography. Quantifications were performed by densitometry using the Bio-1D software. Results were expressed as arbitrary units. (c) PAI-1 expression was analyzed by western blot. Quantifications were performed by densitometry using the Bio-1D software. Results were expressed as arbitrary units (AU). (d) u-PAR expression was analyzed by western blot. Quantifications were performed by densitometry using the Bio-1D software. Results were expressed as arbitrary units (AU). NS: not significantly different from the control; *: significantly different from control at p < 0.05.

YSNSG cyclopeptide alters integrin distribution and endothelial cell cytoskeleton organization

HUVEC treatment with cyclopeptide induced a clear change in cell shape (Figs. 4 and 6), associated with the inhibition of cell migration.

Figure 6. YSNSG cyclopeptide alters integrin distribution, FAK phosphorylation and endothelial cell cytoskeleton organization. HUVEC were incubated for 24 h with control medium or with YSNSG cyclopeptide (10 μM). (a) HUVEC were fixed with methanol, labeled with an anti-β1-integrin subunit antibody and with the Alexa-488-conjugated secondary antibody. The YSNSG cyclopeptide induced an inhibition of β1-integrin subunit clustering (arrows) in endothelial cell lamellipodia. Insert: higher magnification showing β1 integrin distribution at the migration front. (b) Cell were lysed in a 10 mM Tris HCl, 150 mM NaCl pH 7.5 buffer containing 1% Triton and protein extracts were submitted to western blot analysis with an anti-β1 integrin subunit antibody. (c) HUVEC were fixed with methanol, labeled with an antiphosphorylated p125^FAK antibody (Tyr 861) and with the Alexa-488-conjugated secondary antibody. (d) Cells were lysed in a RIPA buffer and extracts were submitted to western blot analysis with an anti-p125FAK (Tyr861) or an anti-FAK antibody. (e) HUVEC were fixed with formaldehyde, permeabilized with acetone and labeled with Alexa-568-conjugated phalloidin. The YSNSG cyclopeptide caused stress fibers disappearance at the leading edges. Cell nuclei were counterstained with Hoechst 33,342. Magnification, ×378.

u-PAR was shown to bind β1-integrin subunit, to induce outside-in signalling, FAK phosphorylation and to promote HUVEC migration.7 As u-PAR was largely decreased after cyclopeptide treatment, we analyzed β1-integrin subunit, FAK phosphorylation and F-actin stress fiber distribution in HUVEC by immunostaining. The YSNSG cyclopeptide altered β1-integrin subunit distribution in endothelial cell lamellipodia (Fig. 6a). As well, the phosphorylation of p125^FAK on tyrosine 861, which was shown to be specifically phosphorylated during cell migration,31 was largely decreased in treated HUVEC vs. control cells (Fig. 6b). Identical results were obtained with an antibody directed against p125^FAK phosphorylated on tyrosine 397 residue (data not shown). Total expression of β1 integrin subunit was unaffected by YSNSG treatment (Fig. 6b). The results suggested a decrease in β1-integrin clustering. Phosphorylation disappeared in lamellipodia and was diffuse in cytoplasm. This was confirmed by western blot experiments (Fig. 6c). Total FAK expression was unaltered. YSNSG treatment lead to structural changes in endothelial cell morphology as shown by F-actin stress fiber distribution (Fig. 6d). Strong actin staining and membrane ruffling activity was seen at the leading edges of control cells whereas incubation of HUVEC with the YSNSG cyclopeptide caused stress fiber disappearance at the leading edges, characteristic of a nonmigratory phenotype.

Discussion

Tumor progression may be inhibited by various matrikines derived from the non collagenic domains of basement membrane-associated collagens, with a particular interest in those derived from the NC1 domains of the different α(IV) collagen chains.32 Angiogenesis plays a crucial role in cancer progression and several matrikines, such as endostatin, arresten, canstatin or the 54–132 fragment of tumstatin were described to limit tumor growth through antiangiogenic properties.33, 34 In previous studies, we demonstrated that the C-terminal NC1[α3(IV) 185–203] peptide exhibits both in vitro and in vivoantitumor properties13, 14 as well as antiangiogenic properties.10, 11 The shorter CNYYSNS peptide, corresponding to the N-terminal residues 185–191, shared the same properties and adopted a β-turn conformation crucial for biological activity.15 It was hydrophobic and dissolved in DMSO. After 2 hr of incubation, disulfide bridges occurred between cystein residues, and the peptide became less efficient. A cyclopeptide was designed to optimize solubility and biological activity. It contains both the four YSNS residues forming the β-turn and a glycine residue allowing cyclization to constrain the β-turn conformation.16

In the experimental mouse melanoma model, intraperitoneal injections of the cyclopeptide at days 3, 5 and 7 strongly reduced tumor growth. Anti-CD31 antibody labeling of tumor sections showed that tumor angiogenesis was also strongly decreased in cyclopeptide-treated mice. To elucidate the biochemical bases of the antiangiogenic activity, we analyzed the YSNSG cyclopeptide effects on endothelial cells (HUVEC) in vitro. Pseudotube formation by HUVEC was strongly decreased. This could result from an inhibition of HUVEC proliferation, an induction of cell apoptosis or a decrease in endothelial cell migration. We showed that the YSNSG cyclopeptide had no effect on HUVEC proliferation and apoptosis. By contrast, their migration was prevented by the peptide, as demonstrated in the in vitro wound assay, where YSNSG strongly slowed down the wound process. In this model, Galardin and Aprotinin, inhibitors of MMPs and plasminogen activation system respectively,22, 23 also largely inhibited wound healing, suggesting the involvement of both MMPs and plasminogen activator/plasmin system in HUVEC migration. This result is in accordance with experimental results from Collen and coll.35 which showed that simultaneous inhibition of plasmin and MMPs by aprotinin or BB94 caused a large inhibition of pseudotubeformation by endothelial cells.

During angiogenesis, quiescent endothelial cells are activated and acquire a migratory phenotype.1 They degrade surrounding ECM through the action of specific proteolytic cascades such as the u-PA/plasmin system and MMPs. Among these, MT1-MMP is required for efficient endothelial cell migration through its own proteolytic activity or by initiating an intracellular signalling pathway through its intracytoplasmic domain.36, 37 The critical role of MT1-MMP was also evidenced by using specific antibodies or siRNA that block angiogenesis.24 Active MT1-MMP is localized at the leading edge of cell during migration.38 The incubation of HUVEC with the YSNSG cyclopeptide induced a strong inhibition of active MT1-MMP at the migration front, as demonstrated by using specific anti-proMT1-MMP antibody. This might explain one part of the inhibitory effects of YSNSG cyclopeptide on endothelial cell migration and angiogenesis.

In vitro39, 40 and in vivo41, 42 studies showed that the u-PA/plasmin system also plays an important role in the angiogenesis process. Wounding of an endothelial cell monolayer triggered a marked, rapid and sustained increase in expression of u-PAR on the surface of migrating cells.5 The u-PA/plasmin system is involved in endothelial cell migration through the proteolytic activity generated on the cell surface but also through proteolytic activity-independent mechanisms.43, 44, 6 u-PA is concentrated on cell surface by a cell membrane u-PA receptor (u-PAR). The YSNSG cyclopeptide inhibited u-PA synthesis and induced a large decrease in u-PAR expression on HUVEC surface, resulting in a decrease in cell surface generated plasmin activity. This result may also explain a part of the inhibitory effect of YSNSG cyclopeptide and aprotinin, by decreasing plasmin proteolytic activity.

u-PAR is linked to cell membrane by a glycosylphosphatidylinositol anchor and its occupancy by u-PA stimulates cell migration independently of proteolytic activity. Even though u-PAR does not comprise transmembrane domains, binding of u-PA onto u-PAR was shown to induce intracellular signalling leading to p125^FAK phosphorylation, MAP kinase activation and endothelial cell migration.45 In addition, u-PAR was shown to interact with β1, β2 or β3-integrin families46 and to promote β1-integrin functions in cell migration.47 Incubation of HUVEC with the YSNSG cyclopeptide induced an alteration of β1-integrin distribution in lamellipodia. In addition, the phosphorylation of p125^FAK on tyrosine 861 was decreased and its distribution largely altered, leading to a profound disorganization of F-actin stress fibers, corresponding to a non migratory phenotype of endothelial cells. Adenovirus-mediated antisense u-PAR gene transfer or u-PAR antagonists were shown to inhibit in vitro capillary-like pseudotube formation by endothelial cells in fibrin gels and in vivo angiogenesis-dependent tumor growth.48, 49

Collectively, our results demonstrate that the YSNSG cyclopeptide exerts potent antiangiogenic activity by a novel molecular mechanism involving a down-regulation of MT1-MMP and u-PA/plasmin system activation through proteolysis inhibition and the induction of a non migratory phenotype of endothelial cells. Other matrikines derived from NC1 domains of α(IV) collagen chains, canstatin or the 54–132 tumstatin fragment, were shown to exert potent antiangiogenic both in in vitro and in vivo models by inducing endothelial cell apoptosis and inhibition of protein synthesis.32, 33 The antiangiogenic activity of the YSNSG cyclopeptide reinforce its potential use as a strong inhibitor of tumor progression by exerting both antitumor and antiangiogenic properties.

Acknowledgements

The confocal microscope observations were performed at INRA, UMR FARE, Reims.