UPARANT: A Urokinase Receptor–Derived Peptide Inhibitor of VEGF-Driven Angiogenesis with Enhanced Stability and In Vitro and In Vivo Potency Free

Angiogenesis is a complex multistep process leading to the formation of new blood vessels from a preexisting vascular network. During new vessel formation, endothelial cells degrade their basement membrane, migrate into the interstitial matrix, and proliferate. An imbalance in this process contributes to numerous malignant, inflammatory, ischemic, infectious, and immune disorders (1).

The receptor for urokinase-type plasminogen activator (uPAR) plays an important role in controlling cell migration (2, 3). uPAR is a glycosylated glycosyl-phosphatidyl-inositol–anchored protein (4) formed by 3 domains (DI, DII, and DIII) connected by short linker regions (5). Besides being responsible for focalizing uPA-mediated plasminogen activation on cell surface (6), uPAR also promotes intracellular signaling, thus regulating physiologic processes such as wound healing, immune responses, and stem cell mobilization, as well as pathologic conditions such as inflammation and tumor progression (7–10). The role of uPAR in angiogenesis is well documented. uPAR is able to focus urokinase proteolytic activity on cell surface and to modulate cell migration (11, 12). We found that soluble forms of uPAR, containing the ⁸⁸Ser-Arg-Ser-Arg-Tyr⁹² sequence (uPAR_88-92), as well as the synthetic peptide Ser-Arg-Ser-Arg-Tyr (SRSRY), stimulates in vitro and in vivo angiogenesis in a protease-independent manner (13). The uPAR_88-92 sequence interacts with the formyl peptide receptors (FPR) type 1 and 2, henceforth inducing cell migration in an integrin-dependent manner (9, 14, 15). Upon binding to FPR, the synthetic peptide SRSRY causes FPR internalization and triggers vitronectin receptor activation with an inside–outside type of mechanism (16).

Recently, we found that the residue Ser⁹⁰ of uPAR is critical for uPAR signaling, and that S90P and S90E single amino acid substitutions exert opposite effects on uPAR activities. Cells expressing membrane-associated uPAR carrying Ser⁹⁰ substituted with Glu exhibit a reduced binding to and a decreased adhesion onto vitronectin, an impaired agonist-induced FPR internalization, and a dramatic reduction of in vitro and in vivo cell migration and invasion (17). To inhibit uPAR functions, we developed, a family of penta-peptides carrying the S90E substitution in the uPAR_88-92 sequence. These peptides block uPAR-dependent cell signaling by interfering with the complex cross-talk involving uPAR, FPR, and integrins. The peptide containing the N-terminal pyro-glutamic acid (pGlu)-Arg-Glu-Arg-Tyr-NH₂ (pERERY-NH₂) shares the same binding site with SRSRY and competes with fMLF for binding to FPR, thus preventing uPAR/FPR interaction. pERERY-NH₂ inhibits migration of various tumor cell lines in culture (18). Subsequently, new tetra-peptides having the general formula Ac-Arg-Glu-Arg-X-NH₂ (X = Phe, Tyr, Trp) were synthesized, and the peptide Ac-Arg-Glu-Arg-Phe-NH₂ (denoted RERF) was selected for its ability to potently prevent in vitro and in vivo cell migration and invasion (19). Recently, we found that RERF also behaves as an antiangiogenic agent. It inhibits in vitro and in vivo responses promoted either by uPAR_88-92 or by VEGF. RERF also prevents cytoskeletal organization and the recruitment of αvβ3 integrin at the focal adhesions in endothelial cells exposed to VEGF, by forcing αvβ3 in an inactive state either directly or indirectly, through FPR (20).

Starting from these pieces of biologic information, we intended to generate peptide analogues with optimized properties for therapeutic applications. We synthesized new peptides containing α-methyl-α-amino acids, and they were characterized from a biologic point of view. All these new analogues are N-acetylated and C-amidated, as being a common way of the N- and C-terminal modifications to enhance the stability to exopeptidase-mediated proteolysis (21). Among the new series of peptides, Ac-L-Arg-Aib-L-Arg-D-Cα(Me)Phe-NH₂, showed the best desired activity in preliminary experiments of VEGF-directed endothelial cell migration.

It adopts in solution a turned conformation, it is quite stable to trypsin and chymotrypsin digestion, in mice and rats blood. It is also stable to sterilization in 3 mg/mL sealed vials in autoclave for 20 minutes at 120°C, and remained stable at 25°C for 2 years.

Ac-L-Arg-Aib-L-Arg-D-Cα(Me)Phe-NH₂, named UPARANT, competes with fMLF for binding to FPR, and inhibits FPR internalization in endothelial cells. It prevents F-actin polymerization and adhesion onto vitronectin in endothelial cells exposed to VEGF. In vitro, UPARANT inhibits VEGF-dependent tube formation of endothelial cells at a ×100 lower concentration than RERF. In vivo, it prevents VEGF-dependent capillary sprouts originating from the host vessels that invaded Matrigel sponges implanted in mice and neovascularization induced by subcorneal implantation of pellets containing VEGF in rabbits.

Peptides were synthesized by the solid-phase approach using standard Fmoc methodology in a manual reaction vessel, and purified by RP-HPLC-C18 column to a 99% purity. Fluoresceinated UPARANT (FITC-UPARANT) was synthesized by the on-resin procedure, using ϵ-aminocaproic acid as spacer (19). It retains the 93% of the inhibitory activity on cell migration (data not shown). Molecular weights were confirmed by mass spectrometry.

A total of 90 sealed vials containing 0.5 mL of UPARANT in PBS at a concentration of 3 mg/mL were autoclaved for 20 minutes at 120°C. One sample was immediately analyzed by liquid chromatography/mass spectrometry (LC/MS) to ascertain the lack of any modified compound. The remaining samples were kept in a thermostat at 25°C. Three vials were opened every month and analyzed by LC/MS.

Blood stability studies were performed at IRBM Science Park. Technical details are reported in the Supplementary Materials and Methods.

The peptides were subjected to digestion by trypsin at a peptide:trypsin ratio of 1,000:1 (w:w) for RERF and Icm25, whereas a 10:1 (w:w) peptide:trypsin ratio was used for Icm25B and UPARANT. The trypsin concentration is 0.6 mg/L. The peptides were also subjected to digestion by chymotrypsin at a peptide:chymotrypsin ratio of 10:1 (w:w). Icm25-4 was not studied for enzyme digestion. Digestions were followed for 1 hour at 25°C in 0.1 M Tris-HCl buffer pH 8.5. The peptide consumption was followed by RP-HPLC measuring the area peaks at different times between 0 and 60 minutes. Quantification of the intact peptide peak over time indicated that its disappearance seemed to follow first-order kinetics (data not shown), and the t_1/2 was estimated.

Nuclear magnetic resonance (NMR) experiments were performed on Bruker Avance 600 MHz spectrometer, equipped with triple resonance cryo-probe. NMR characterization was performed in H2O/D2O 90:10 (v/v) and in H2O/CF3CD2OD 70:30 (v/v) at 298 K. Details are reported in the Supplementary Materials and Methods. Restrained molecular dynamic (RMD) simulations were carried out using a hand-built starting model with structural data from NMR measurements [e.g., interproton distances from Nuclear Overhauser Effect (NOE) value]. Details on structure calculations are reported in the Supplementary Materials and Methods.

Human umbilical vein endothelial cells (HUVEC) were purchased by Lonza (C2519A, Lot No. 0000115425), which provided a certificate of analysis for each cell lot. This guarantees the expression of CD31/105 and von Willebrand Factor through 15 population doublings. HUVEC, used between the third to the seventh passage according to Arnaoutova and colleagues, (22), were grown in Eagle basal medium (EBM) supplemented with 4% FBS, 0.1% gentamicin, 1 μg/mL hydrocortisone, 10 μg/mL EGF, and 12 μg/mL bovine brain extract (Cambrex).

Cell migration of HUVEC was performed in Boyden chambers, using 8-μm-pore size polyvinylpyrrolidone-free-polycarbonate filters at 37°C, 5% CO₂ as previously described (13). Briefly, 7 × 10⁴ viable cells suspended in serum-free EBM were allowed to migrate for 4 hours toward EBM or 40 ng/mL VEGF₁₆₅ (Pepro Tech), mixed with diluents or the indicated peptides at 37°C in a 5% CO₂. At the end of assays, cells on the lower filter surface were fixed, stained with hematoxylin, and 10 random fields/filter were counted at ×200 magnification.

HUVECs (5 × 10⁵ cells/sample) were preincubated with diluents or the indicated effectors for 60 minutes at 4°C, extensively rinsed with PBS and then exposed to 10 nmol/L N-formyl-Nle-Leu-Phe-Nle-Tyr-Lys-fluorescein (FITC-fMLF) purchased from Molecular Probes for 60 minutes at 4°C. Quantification of cell-associated fluorescence was assessed by reading endothelial cell lysates with a fluorescence plate reader Victor 3 (Perkin Elmer) using 485 nm excitation and 535 nm emission filters. An aliquot of each sample was separated on 10% SDS-PAGE and subjected to Western blot analysis for the α-tubulin content.

HUVEC grown on glass slides to semiconfluence were incubated with the indicated effectors for 30 minutes at 37°C and then exposed to 10 nmol/L FITC-fMLF or 10 nmol/L FITC-UPARANT diluted in serum-free EBM for 30 minutes at 37°C as described (19). To analyze cytoskeletal organization and αvβ3 localization, cells grown on glass coverslips to semiconfluence, were starved for 60 minutes in EBM. Then, cells were exposed to EBM alone, 40 ng/mL VEGF₁₆₅, or 40 ng/mL VEGF₁₆₅ mixed to 10 nmol/L indicated peptides for 30 minutes at 37°C in humidified air with 5% CO₂. Slides were washed with PBS, fixed and permeabilized with 2.5% formaldehyde–0.1% Triton X-100 in PBS for 10 minutes at 4°C. After several washes in PBS, slides were incubated with 2 μg/mL anti-vinculin mAb (clone VIIF9 purchased from Chemicon) and then with 1:800 goat Alexa Fluor 488 antimouse immunoglobulin G (IgG; Molecular Probes) at 23°C for 60 and 45 minutes, respectively. Thereafter, 2 μg/mL LM609 anti-αvβ3 monoclonal antibodies (mAb; Chemicon) and then 1:800 diluted rabbit Alexa Fluor 594 antimouse IgG (Molecular Probes) were applied to slides at 23°C for 60 and 45 minutes, respectively. To analyze cytoskeletal organization, coverslips were incubated with 0.1 μg/mL rhodamine-conjugated phalloidin (Invitrogen) at 23°C for 45 minutes. In all cases, slides were mounted using 20% (w/v) mowiol, cells visualized with a 510 META-LSM confocal microscopy (Carl Zeiss) and z-series with 0.38-μm intervals were collected.

HUVEC grown to semiconfluence were starved for 18 hours in 0.2% FBS EBM. Then, cells were exposed to EBM alone, 40 ng/mL VEGF₁₆₅ with/without 10 nmol/L UPARANT at 37°C in humidified air with 5% CO₂. At the indicated times, cells were lysed in radioimmunoprecipitation assay buffer (10 mmol/L Tris pH 7.5, 140 mmol/L NaCl, 0.1% SDS, 1% Triton X-100, 0.5% NP40) containing 5 mmol/L Na₃VO₄ and a protease inhibitor mixture (Sigma-Aldrich). Cell lysate protein content was measured by a colorimetric assay (Bio-Rad). Western blot analysis was performed as previously described (17, 20). Briefly, 50 μg proteins were separated on 10% SDS-PAGE and transferred onto a polyvinylidene fluoride membrane. The membranes were blocked with 5% nonfat dry milk and probed with anti-phospho-Akt (Ser473; p-AKT) Ab or anti-p-ERK1/2 (Thr202/204) Ab (Cell Signaling) and then with anti-AKT mAb (R&D System) or anti-ERK1/2 mAb (Millipore). In all cases, washed filters were incubated with horseradish peroxidase–conjugated antimouse or anti-rabbit antibody and detected by Enhanced Chemiluminescence Kit (GE-Healthcare).

Cell adhesion assays were performed using 16-well plates coated with 5 μg/mL vitronectin, diluted in PBS and the xCELLigence technology (Roche Diagnostics) as described (23). HUVEC (5 × 10³ cells/well) were plated in each coated well, in serum-free EBM in the presence or the absence of 40 ng/mL VEGF₁₆₅ with/without 10 nmol/L C-terminal amidated Ac-Glu-Arg-Phe-Arg-NH₂ control peptide (ERFR), 10 nmol/L RERF or 10 nmol/L UPARANT and allowed to adhere for 4 hours at 37°C, 5% CO₂. Cell adhesion was automatically monitored by the Real Time Cell Analysis xCELLigence system (Acea Bioscience), which measures the impedance value of each well and expressed as a cell index value.

For the MTS assay, the CellTiter 96 AQueous Cell Proliferation Assay Kit (Promega) was used following the manufacturer's instruction. Briefly, HUVEC (2 × 10³/well) were seeded in 96-well tissue culture plates and left to adhere in complete media for 3 hours, then rinsed twice with PBS followed by the incubation with EBM mixed to 40 ng/mL VEGF₁₆₅ with or without peptides at the indicated concentration. Medium was replaced every 24 hours. At the indicated times, suspended cells were removed and 10 μL of the MTS reagent was added into each well and cells were incubated at 37°C for 3 hours. The absorbance was detected at 490 nm with a Microplate Reader (Bio-Rad). With a similar experimental design, endothelial cell proliferation was assessed using 16-well plates and the Real Time Cell Analysis xCELLigence technology. The impedance value of each well was automatically monitored and expressed as a cell index.

The formation of vascular-like structures was assessed on Matrigel as described (13). HUVEC were suspended in 300 μL prewarmed EBM. Diluents or effectors were added to the cell suspension before seeding cells (4 × 10⁴ cells/well) on plates coated with 300 μL/well growth factor reduced Matrigel (Becton Dickinson). When indicated, cells were preincubated for 30 minutes at 37°C with 1 μg/mL anti-human VEGF Ab, 2 μg/mL LM609 anti-αvβ3 mAbs, or nonimmune serum. Complete capillary tube networks were assessed within a low-magnification field observed under light microscopy after 6-hour incubation at 37°C in humidified air with 5% CO₂. To quantify tube formation, 5 random areas/well at ×100 magnification were imaged and the number of tubes formed by cord-like structures exceeding 100 μm in length (24) were visualized using Axiovision 4.8 software (Carl Zeiss) and counted.

Twelve 6- to 8-week-old C57BL/6J male mice (Charles River Laboratories) of 23 to 25 g 6 to 8 weeks old, were maintained in accordance with institutional guidelines complying with national and international laws and policies. Briefly, VEGF₁₆₅ (100 ng/mL) and heparin (50 U/mL) diluted in PBS, with or without 75 or 150 μg/kg UPARANT, were added to unpolymerized liquid Matrigel at 4°C, to a final volume of 500 μL. The Matrigel suspension was slowly injected subcutaneously into the flanks of mice, using a cold syringe, where it quickly polymerizes to form a solid gel. Matrigel with buffer alone was used as negative control. Five days after injection, the animals were killed and the gels were removed, minced, and then diluted in water for hemoglobin content measurement with a Drabkin Reagent Kit (Sigma). The final hemoglobin concentration was normalized to 100 mg of recovered gel and calculated from a standard calibration curve after spectrophotometric analysis at 540 nm.

Corneal pocket assays were performed at PRIMM. Twelve female New Zealand White rabbits (Charles River) weighing 2.5 to 3.0 kg were anesthetized by intramuscular injection of acepromazina (1 mg/kg), ketamine hydrochloride (35 mg/kg), and xylazine hydrochloride (5 mg/kg) and 3 to 4 drops of 0,4% ossibuprocain chlorohydrate solution were topically applied to the eye before micropocket surgery. A wire speculum was positioned in the eye, and a sucralfate-hydron pellets containing PBS, 5 μg UPARANT, or 180 ng VEGF with/without 5 μg UPARANT were implanted into the cornea after making a micropocket in the stroma, using standard surgical tools (25, 26). Tobradex (0.3% tobramycin–0.1% dexamethasone) was applied to the surface of the cornea after pellet implantation to prevent infection. Observation and quantification of the angiogenic responses were performed by a slit-lamp stereomicroscope. The angiogenic activity was evaluated on the basis of the number and growth rate of newly formed capillaries. A density value of 1 corresponded to 0 to 25 vessels per cornea, 2 from 25 to 50, 3 from 50 to 75, 4 from 75 to 100, and 5 for >100 vessels.

The results are expressed as the mean ± SD of the number of the indicated determinations. Data were analyzed by one-way analysis of variance and post hoc Bonferroni modified t test for multiple comparisons. P < 0.001 was accepted as significant.

The research work with mouse model has been approved by Institutional Ethical Committee of Istituto Nazionale Tumori “Fondazione G. Pascale”-IRCCS, Naples, Italy (protocol no. 09, December 20, 2010).

Previously, we reported the NMR structure of RERF in solution (19). RERF, although displaying some conformational flexibility, preferentially adopts an α-turn (type I-α_RS) conformation (27) in water/trifluoroethanol (TFE) mixed solvent. Cα-methyl-α-amino acids with well-defined stereo chemical properties impose local restrictions on backbone conformation, thus conferring structural stability (28–29). Aib residue (α-amino isobutyric acid) is the prototype of Cα-methyl-α-amino acids, and is well known to be characterized by a restricted conformational freedom (30) in the 3₁₀-α-helical region of the Ramachandran plot. Other Cα-methyl-α-amino acids, including Cα-methyl-phenylalanine (31) are β-turn and helix inducers, much stronger than the unmethylated, phenyl-containing, protein amino acid Phe. Cα-methy-α-amino acid, when incorporated into peptide sequences containing coded residues, strongly facilitates and stabilizes type III β-turns or helicogenic α-turns as previously described (32, 33). In addition, Cα-methyl-α-amino acid substitutions could provide enhanced binding to a molecular target, resistance to proteolytic degradation and longer persistence in circulation. On the basis of this information, we have synthesized 4 peptide analogues of RERF containing commercially available Cα-methyl-α-amino acids, namely:

Icm25-4 and Icm25-4c are both tetrapeptides analogues of RERF in which the Glu residue has been replaced by Aib. Icm25-4c also contains a second substitution of the terminal Phe residue with D-Cα(Me)Phe (S-configuration). Icm25 and Icm25B are both pentapeptides corresponding to N-terminal elongation of Icm25-4 and Icm25-4c with a second Aib residue.

The peptides were subjected to digestion by trypsin and chymotrypsin at different peptide:enzyme ratios, depending on the digestion rates. The enzyme concentration is 0.6 mg/L, and is similar to trypsin concentration determined in plasma 117 to 637 μg/L (34). Icm25-4 was not studied for enzyme digestion. The peptide consumption was determined by measuring the HPLC area peaks during digestion. It seemed to follow first-order kinetics (data not shown), and the t_1/2 were estimated. We found that RERF and Icm25 at a peptide:trypsin ratio of 1,000:1 are rapidly digested, the t_1/2 being 3.05 and 6.19 minutes, respectively. No digestion of Icm25B and Icm25-4c was detected at a peptide:trypsin ratio of 1,000:1, whereas at a peptide:trypsin ratio of 10:1 the t_1/2 were 37.5 and 11.4 minutes for Icm25B and Icm25-4c, respectively. Both RERF and Icm25, containing the natural Phe residue at the C-terminus, were digested by chymotrypsin with t_1/2 being 23.6 and 20.0 minutes, respectively. Conversely, no digestion of Icm25B and Icm25-4c, both containing the unnatural C-terminal Cα(Me)Phe, was detected within 60 minutes.

RERF has been shown to inhibit endothelial cell migration driven by VEGF (20). We compared the effects of RERF and Aib containing peptides in inhibiting endothelial cell migration directed toward 40 ng/mL VEGF₁₆₅. By conventional Boyden chamber assays, we found that 40 ng/mL VEGF₁₆₅ elicited endothelial cell migration up to 235% of the basal cell migration. When used as chemoattractant at 10 nmol/L concentration, Aib containing peptides, as well as RERF, slightly reduced basal cell migration, whereas the control peptides ARARY or ERFR were ineffective (Table 1). As expected, a 57% and 54% inhibition of VEGF-directed endothelial cell migration was elicited by anti-VEGF Ab and by RERF, respectively. The combination of VEGF₁₆₅ with 10 nmol/L concentration of Icm25, Icm25B, Icm25-4, or Icm25-4c reduced endothelial cell migration by 59%, 62%, 49%, and 68%, respectively (Table 1). Because the peptide Icm25-4c showed a long t_1/2 in enzyme digestions, and exerted the strongest inhibition of endothelial cell migration, it was named UPARANT, and further characterized for its biochemical properties, conformational preferences and biologic activity.

We examined the stability of UPARANT during standard autoclave sterilization. Ninety sealed vials containing 0.5 mL of UPARANT in PBS at a concentration of 3 mg/mL were autoclaved for 20 minutes at 120°C. One sample was immediately analyzed by LC/MS to ascertain the lack of any modified compound. The remaining samples were kept in a thermostat at 25°C. Three vials were opened every month and analyzed by LC/MS. As a result, no degradation after autoclaving and standing at 25°C in sealed vials for 2 years was observed (data not shown).

The stability of UPARANT in physiologic solution, in mice and rats blood was also studied at 0.3 and 1 μg/mL concentrations. Blood stability was followed for 2 hours. UPARANT resulted undegraded as shown in Supplementary Tables S1 and S2.

We investigated the conformational preferences of UPARANT in water and in water/TFE by NMR spectroscopy (Supplementary Tables S3 and S4 and Fig. S1). ROESY spectra in water and in water/TFE essentially display the same pattern of sequential and medium-range NOE effects (see Supplementary Materials and Methods), indicating that the peptide has a quite similar conformational behavior in both solvent systems. However, detailed analysis was performed in water/TFE solution only, because of more dispersed amide signals and fewer resonance overlaps. The pattern of NOEs and their relative intensities are summarized in Fig. 1A. RMD calculations, using the NMR experimental data as conformational restraints, gave an ensemble of conformers quite similar in terms of backbone conformation, underlining the presence of one predominant and stable structure for UPARANT. Figure 1B reports the backbone atoms superposition of the ensemble of conformers along the MD trajectory. Figure 1C reports the average backbone torsion angles of UPARANT as obtained from RMD simulation in vacuo at 300 K.

The peptide conformation corresponds to an incipient 3₁₀-helix characterized by 2 consecutive type III β-turn, followed by a type I β-turn. This structure is stabilized by the typical CO_i–HN_i+3 hydrogen bonds (see Fig. 1C): a weak interaction Ac-CO···NH-Arg³, a H-bond Arg¹-CO···NH-Cα(Me)Phe⁴, and Aib²-CO···NH-amide. Along the trajectory of the RMD simulation, the Arg¹ and Arg³ guanidine groups experience intraresidue H-bonds, and the Arg¹ guanidine group is also H-bonded to Cα(Me)Phe⁴-CO group.

We have previously documented that RERF competes with fMLF for binding to FPR, being FPR the main binding site of RERF (19). To test whether similar to RERF (20), UPARANT competes with fMLF for binding to FPR, HUVEC were preincubated at 4°C (to avoid receptor internalization) with diluents, 100 nmol/L concentration of fMLF, RERF, or UPARANT for 60 minutes at 4°C, and then exposed to 10 nmol/L N-formyl-Nle-Leu-Phe-Nle-Tyr-Lys-fluorescein (FITC-fMLF) for additional 60 minutes at 4°C. Quantification of cell-associated fluorescence was assessed by reading endothelial cell lysates with a fluorescence plate reader. An aliquot of each sample was subjected to Western blot analysis with anti-α-tubulin to control that equal amount of proteins/sample has been read (inset to Fig. 2A). Measurement of cell-associated fluorescence revealed that the control peptide ERFR did not modify binding of FITC-fMLF, whereas an excess of unlabeled fMLF, RERF (used for comparison), or UPARANT dramatically reduced binding of fluorescent FPR-agonist to endothelial cells at a similar extent (Fig. 2A). To evaluate the effect of UPARANT on FPR internalization in response to agonist-stimulation, binding experiments were also performed at 37°C. Upon exposure to FITC-fMLF, FPR seemed mainly internalized with the appearance of fluorescent spots localized in the cytoplasm as shown by z-stack analysis of confocal images (Fig. 2B and C). Similar to fMLF or RERF, 100 nmol/L UPARANT strongly reduced internalization of fluorescent agonist in all cell population (Fig. 2B). To assess whether UPARANT itself causes FPR internalization, HUVECs were preincubated with diluents, or an excess of fMLF, RERF, or UPARANT for 30 minutes at 37°C, and then exposed to 100 nmol/L FITC-UPARANT for additional 30 minutes at 37°C. As shown by z-stack analysis of confocal images, endothelial cells exhibited only a slightly internalization of FITC-UPARANT, which was prevented by unlabeled fMLF, RERF, as well as by UPARANT (Fig. 2D and E). All together, these findings suggest that uPARANT competes with fMLF for binding to FPR on cell surface and inhibits FPR activation by preventing its internalization.

VEGF-triggered endothelial cell motility is the result of marked cytoskeletal reorganization and accumulation of stress fibers associated with new actin polymerization, and rapid formation of focal adhesions, which do not occur in the presence of RERF (20). To analyze the effects of UPARANT on cytoskeletal organization induced by VEGF, HUVEC were exposed to 40 ng/mL VEGF₁₆₅, in the presence/absence of 10 nmol/L UPARANT, and then double stained for vinculin and F-actin. As expected, VEGF induced a marked reorganization of actin into stress fibers spanning the length of the cells, most of which colocalized with vinculin-positive focal adhesions, that was prevented by 10 nmol/L RERF (Fig. 3A). A total of 10 nmol/L UPARANT fully abrogated VEGF-induced effects on cytoskeleton. In the latter case, endothelial cells had a condensed, rounded morphology; similar to untreated cells, the F-actin was condensed into fewer fibers and was completely absent from the leading edges of the cells (Fig. 3A). It is known that αvβ3 integrin has a prominent role in the activity of VEGF. In response to VEGF, β3 integrin regulates integrin-dependent actin reorganization, thus leading αvβ3-VEGFR2 complexes to localize at new formed focal adhesions (35). UPARANT caused disappearance of αvβ3 at focal adhesions and the appearance of thin, αvβ3-positive linings at the cell edge, similar to untreated cells, whereas the addition of 10 nmol/L ERFR did not modify VEGF-dependent integrin redistribution. (Fig. 3B). RERF has been proven to inhibit VEGF-dependent signaling by affecting αvβ3 integrin-dependent cell adhesion onto vitronectin (20). We investigated whether αvβ3 integrin is indeed involved in the inhibitory activity of UPARANT. HUVEC suspended in EBM mixed to diluents or 40 ng/mL VEGF₁₆₅ with/without 10 nmol/L ERFR, 10 nmol/L RERF, or 10 nmol/L UPARANT, were allowed to adhere for 4 hours at 37°C on 16-well plates coated with vitronectin. We found that RERF and UPARANT reduced endothelial cell adhesion onto vitronectin at a similar extent, either in the presence or in the absence of VEGF, whereas the control peptide ERFR was almost ineffective (Fig. 3C). All together, these findings show that UPARANT prevents VEGF-driven αvβ3 relocalization and ligand-dependent αvβ3 activation.

To better characterize the inhibitory effect of UPARANT on VEGF-dependent cell migration, HUVEC were allowed to migrate in Boyden chambers toward 40 ng/mL VEGF₁₆₅ with/without increasing concentrations of RERF used as positive control, ERFR used as negative control, or UPARANT. We found that unlike ERFR, both UPARANT and RERF inhibit VEGF-directed migration of endothelial cells in a dose-dependent manner. In both cases, the inhibition starts in the low fmol/L concentration range and plateaus in the nmol/L range (Fig. 4A). RERF has been proven to prevent VEGF-induced phosphorylation of many proteins, such as AKT and ERK1/2 (20). To investigate the effects of UPARANT on the phosphorylation of these VEGF intracellular targets, HUVEC were exposed to EBM or 40 ng/mL VEGF₁₆₅ with or without 10 nmol/L UPARANT from 0 to 60 minutes, and phosphorylated AKT and ERK1/2 were detected by Western blot. As shown in Fig. 4B, UPARANT consistently decreased the VEGF₁₆₅-induced and time-dependent increase of Akt and ERK1/2 phosphorylation. Moreover, ANOVA statistical analysis revealed that growth of endothelial cells exposed to VEGF was not modified by 10 nmol/L or 10 μmol/L UPARANT up to 72 hours as shown by an MTS assay (Fig. 4C) as well as by proliferation curves automatically monitored with the xCELLigence system (Fig. 4D). This is in agreement with our previous data showing that 10 μmol/L RERF does not modify the growth of endothelial cells. Although we have not identified yet the exact molecular mechanism by which RERF and UPARANT inhibit VEGF-triggered signaling, our data suggest that RERF and UPARANT impinge on common downstream signaling mediators.

To investigate the ability of UPARANT to inhibit VEGF-triggered angiogenesis in vitro, endothelial cells were plated on Matrigel in the presence of 40 ng/mL VEGF₁₆₅ mixed to diluents or increasing concentrations of RERF, ERFR, or UPARANT. As expected, endothelial cells exposed to VEGF, formed a 3-dimensional network of tubes resembling capillary-like structures, which reached at 6 hours the 257 ± 10% of braches counted in the absence of any angionenic stimulus (Fig. 5A and B). Unlike ERFR control peptide or nonimmune serum, endothelial cell incubation with anti-VEGF Ab or anti-αvβ3 mAb reduced VEGF-dependent tube formation to the basal level (Fig. 5B). UPARANT reduced the extent of VEGF-dependent tube formation in a dose-dependent manner. Similar to RERF, the inhibitory effect of UPARANT starts in the femtomolar concentration range and levels off in the nanomolar range. Remarkably, UPARANT caused an overall 75% inhibition of VEGF-dependent tube formation above 10 nmol/L, when compared with RERF, which caused a 54% inhibition. Also, the IC₅₀ of UPARANT seems to be ∼100-fold lower as compared with RERF (Fig. 5C), indicating again that UPARANT inhibits VEGF signaling more efficiently than RERF.

To assess the effect of UPARANT on angiogenesis in vivo, first a quantitative Matrigel sponge assay was performed. A cocktail of VEGF₁₆₅ (100 ng/mL) and heparin (50 U/mL) promoted a hemorrhagic vascularization of the Matrigel sponge, which was clearly detectable at 5 days postimplantation (Fig. 5D). The presence in the sponge of 75 or 150 μg UPARANT produced a marked reduction of capillary sprouts originating from the host vessels that invaded Matrigel sponges as detected by visual inspection of the gels and quantification of hemoglobin recovered from sponges (Fig. 5D). A 64% and 72% reduction of hemoglobin was found in sponges loaded with 75 and 150 μg, respectively.

Furthermore, antiangiogenic activity of UPARANT was investigated by a corneal pocket assay in rabbits. Neovascular growth was evaluated in 12 rabbits upon subcorneal implantation of slow release pellets containing PBS, 5 μg UPARANT, or 180 ng VEGF with/without 5 μg UPARANT. Implantation of pellets containing 180 ng VEGF stimulated ingrowth of blood vessels in the rabbit corneas starting from day 2, and continuing to grow progressively up to 15 days before regressing. At day 15, rabbit corneas probed with pellets impregnated with VEGF exhibited higher angiogenic score then the average score of those impregnated with vehicle only (PBS; 3.3 ± 0.6 vs. 0.1 ± 0.1 with P < 0.0001; Fig. 5E and F). When UPARANT was tested at 5 μg/pellet alone, it did not elicit inflammatory response. Conversely, UPARANT elicited a 75% inhibition of VEGF-induced vascularization (0.84 ± 0.4 vs. 3.3 ± 0.6 with P < 0.0001; Fig. 5E and F). Taken together, our findings indicate that UPARANT behaves as an antiangiogenic factor, which prevents VEGF-induced angiogenesis in vivo more effectively than RERF.

This work is based on previous evidence showing that the chemotactic sequence of the human urokinase receptor (uPAR_88-92) drives in vitro and in vivo angiogenesis in a protease-independent manner (13). A few years ago we developed the peptide Ac-Arg-Glu-Arg-Phe-NH₂, namely RERF, that prevents, in vitro and in vivo, migration and invasion of tumor cells by inhibiting the uPAR_88-92-dependent signals (19). More recently, we have documented that RERF prevents both uPAR_88–92- and VEGF-induced angiogenesis in vitro and in vivo (13). However, RERF is expected to have some pharmacokinetic drawbacks as drug candidate, which are common to peptides: they are usually very susceptible to proteolytic degradation in vivo and are rapidly cleared from the circulation in minutes (36). To overcome these drawbacks, we have generated a series of peptide analogues of RERF containing Cα-methyl-α-amino acids, a structural modification used in the past to overcome these limitations (37). Among the enormous scenario of possible peptide modifications developed in the last 50 years, we have chosen Cα-methyl-α-amino acid substitutions on the basis of the previously reported NMR solution structure of RERF. In fact, when it is necessary to modify the peptide composition to improve the pharmacologic profile, it is also mandatory to keep, and possibly reinforce, the bioactive conformation, and henceforth the pharmacophore orientation as in the original bioactive peptide. The NMR solution structure of RERF revealed that this peptide, although displaying some conformational flexibility, preferentially adopts a turned structure, therefore Cα-methyl α-amino acid substitutions were selected for their well-known propensity to strongly induce turned structures. Among the series of newly synthesized analogues (Icm25, Icm25B, Icm25-4, and UPARANT), UPARANT displayed in a preliminary evaluation of the biologic activity the strongest inhibition of endothelial cell migration and showed a long t_1/2 in enzyme digestions. UPARANT was first characterized from a structural point of view by NMR spectroscopy.

The solution structures of UPARANT and RERF reveals common features as can be seen by comparison of Fig. 6A–C. The backbone atoms superimposition for all residues of RERF and UPARANT gave root-mean-square deviation (RMSD) values of 0.76 Å, which underlines their structural similarity. Both RERF and UPARANT adopt in solution a helicogenic-turned structures; an α-turn (type I-αRS) is found for RERF and an incipient 3₁₀-helix for UPARANT (Fig. 1C). However, UPARANT shows a more stable and compact turned structure with respect to RERF. As anticipated from the design, the presence, in the tetrapeptide UPARANT, of Cα-methyl-α-amino acids, such as Aib and Cα(Me)Phe, induces a significant stabilization of the turned structure. UPARANT and RERF share a common structural motif: the aromatic ring of the Phe residue is flanked by 2 positively charged residues. A major difference regards the orientation of Arg¹ side chain. In the structure of RERF, the guanidinium group is involved in ion pairing with Glu², whereas in UPARANT the same group is involved in H-bonding with the amide terminal CO group. In addition, it is worth mentioning that RERF and UPARANT were both studied by NMR spectroscopy in 2 solvent systems, water and water/TFE, for the well-known secondary structure inducing effect of TFE. RERF is poorly structured in water and its turned structure clearly appears in the presence of TFE only. Conversely, UPARANT is well structured in both solvent systems. We have foreseen that stabilization of the turned conformation for UPARANT, similar to that found for RERF, would have beneficial effects on biologic activity. We found that UPARANT inhibits VEGF-driven angiogenesis in vitro and in vivo more efficiently as compared with RERF. In vitro, UPARANT inhibits VEGF-dependent tube formation of endothelial cells at a 100× lower concentration than RERF. Remarkably, a 75% reduction of tubes formed by endothelial cells exposed to VEGF is achieved by UPARANT at 10 nmol/L concentration whereas only a 54% inhibition has been observed in the presence of 10 nmol/L RERF. In vivo, UPARANT reduces to the basal level VEGF-dependent capillary sprouts originating from the host vessels that invaded Matrigel sponges implanted in mice and prevents neovascularization induced by subcorneal implantation of pellets containing VEGF in rabbits.

Our previous work has documented that RERF prevents uPAR/FPR and fMLF/FPR interactions, being FPR the main binding site of RERF (19). Although we have not investigated UPARANT/FPR association and disassociation process, we found that an excess of UPARANT prevents fMLF binding to FPR, suggesting that fMLF and UPARANT share the same binding site. The pivotal role of αvβ3 in mediating VEGF-dependent proangiogenic stimulus is well established (35). The involvement of αvβ3 integrin in the UPARANT inhibitory effect is supported by the findings that cell adhesion to vitronectin is greatly reduced in the presence of UPARANT. Taken together, our results confirm the involvement of FPR in binding and mediating inhibition by UPARANT and suggest a mechanism by which αvβ3 is forced into an inactive state by UPARANT and, consequently, VEGF downstream signaling mediators are not activated.

Our findings unravel a complex picture of the mechanistic effects of UPARANT on endothelial cells and, more generally, highlight the role of uPAR, integrins, and VEGF-induced signaling events in the regulation of endothelial cell functions. At molecular level, inhibitory effect of UPARANT on VEGF signaling is shown by the decreased amount of phospho-AKT and phospho-ERK1/2 in VEGF-treated endothelial cells.

Pharmacologic control of angiogenesis is considered one of the most promising approaches for the treatment of diseases sustained by excessive angiogenesis, such as cancer and inflammation. However, several side effects have been ascribed to antiangiogenic drugs such as bevacizumab that significantly increases the risk of cardiac ischemic events in cancer patients (38). With respect to this, UPARANT may be considered a good drug by virtue of its ability to selectively inhibit FPR unlike bevacizumab, indicating that UPARANT behaves differently from other VEGF inhibitors. Mechanistically, UPARANT blocks VEGF-triggered signaling by preventing FPR and αvβ3 integrin activities without affecting cell survival. In this respect, although we have not ascertain whether a direct interaction between UPARANT and VEGF occurs, it is possible to assume that UPARANT bocks integrin αvβ3 activity, thus inhibiting only a part of the VEGF/VEGFR signaling. This suggests a potentially wider, but still target-specific activity for UPARANT.

Furthermore, unlike RERF, UPARANT is stable to sterilization, is stable in blood, and displays a long-time resistance to enzymatic digestion. Peptide-based therapeutics is one of the fastest growing class of new drugs (39). Peptides have unique advantages as therapeutic agents. They have high activity per unit mass, and low manufacturing costs. They also offer great potency, selectivity, and specificity, have low off-target toxicity and low drug–drug interaction potential. In this frame, UPARANT could be considered a promising therapeutic agent for the control of diseases fueled by excessive angiogenesis, such as cancer and inflammation.

M.V. Carriero has ownership interest (including patents) in Pharmaphelix s.r.l. V. Pavone is the President of Pharmaphelix s.r.l. and also has ownership interest (including patents) in the same. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.V. Carriero, V. Pavone

Development of methodology: K. Bifulco, L. Lista, L. Mele, G. Di Carluccio

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): O. Maglio

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.V. Carriero, K. Bifulco, M. Minopoli, O. Maglio, V. Pavone

Writing, review, and/or revision of the manuscript: M.V. Carriero, O. Maglio, V. Pavone

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Bifulco, L. Lista

Study supervision: M.V. Carriero, M. De Rosa, V. Pavone

The authors thank M. Paro at PRIMM for corneal pocket assays in rabbits and Edith Monteagudo, Head of Preclinical Research, and the staff of IRBM Science Park for blood stability studies.

This work was supported by Associazione Italiana per la Ricerca sul Cancro 2013, project 14225 (M.V. Carriero) and by Italian Ministry of Health RF-2010-2316780 (M.V. Carriero).

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.