Triazole RGD antagonist reverts TGFβ1-induced endothelial-to-mesenchymal transition in endothelial precursor cells

Fibrosis is the dramatic consequence of a dysregulated reparative process in which activated fibroblasts (myofibroblasts) and Transforming Growth Factor β1 (TGFβ1) play a central role. When exposed to TGFβ1, fibroblast and epithelial cells differentiate in myofibroblasts; in addition, endothelial cells may undergo endothelial-to-mesenchymal transition (EndoMT) and actively participate to the progression of fibrosis. Recently, the role of αv integrins, which recognize the Arg-Gly-Asp (RGD) tripeptide, in the release and signal transduction activation of TGFβ1 became evident. In this study, we present a class of triazole-derived RGD antagonists that interact with αvβ3 integrin. Above different compounds, the RGD-2 specifically interferes with integrin-dependent TGFβ1 EndoMT in Endothelial Colony-Forming Cells (ECPCs) derived from circulating Endothelial Precursor Cells (ECPCs). The RGD-2 decreases the amount of membrane-associated TGFβ1, and reduces both ALK5/TGFβ1 type I receptor expression and Smad2 phosphorylation in ECPCs. We found that RGD-2 antagonist reverts EndoMT, reducing α-smooth muscle actin (α-SMA) and vimentin expression in differentiated ECPCs. Our results outline the critical role of integrin in fibrosis progression and account for the opportunity of using integrins as target for anti-fibrotic therapeutic treatment.


Introduction
Fibrotic disease encloses a wide array of different pathologies both systemic such as systemic sclerosis (SSc) and sclerodermatous graft versus host disease (Scl GVHD), and organ-specific pathologies as idiopathic pulmonary fibrosis (IPF), liver cirrhosis, and progressive kidney disease. Although the etiology of these fibrotic disorders remains unexplained and may vary from disease to disease, the common pathogenetic signs are the deposition of extracellular matrix (ECM) synthesized by activated myofibroblasts and the persistence of inflammation [1,2]. High levels of cytokines, growth factors, and proteolytic enzymes produced by activated macrophages and lymphocytes restrain the resolution process and fibrosis expands gradually. Finally, matrix deposition rearranges the tissue architecture, causing organ failure and patient death. Thus, fibrotic diseases in their complex are lifetethering diseases recognized to represent an important health issue [3].
Transforming growth factorb1 (TGFb1) has been identified as the most important mediator in many types of tissue fibrosis [4]. TGFb1 induces fibroblasts differentiation into myofibroblasts providing collagen and ECM protein deposition [5]. Experimental studies have provided evidences that myofibroblasts not only originate from resident fibroblasts, but also derive from the transdifferentiation of epithelial cells, endothelial cells, and bone marrowderived cells [6,7]. Endothelial cells and bone marrowderived endothelial cells, when exposed to TGFb1, undergo endothelial-mesenchymal transition (EndoMT). As a consequence, endothelial cells lose their biological characteristic of the endothelial phenotype, acquire typical mesenchymal features, and concurrently express typical markers of myofibroblastic differentiation: a-smooth muscle actin (a-SMA), vimentin, and collagen production [8].
TGFb1 is produced as inactive pro-peptide by different types of inflammatory cells. The mature cytokine is exposed in the extracellular space non-covalently associated to a latency-associated peptide (LAP), and the LAP-TGFb1 pro-peptide forms a homodimer complex. This complex, which prevents the interaction of the mature cytokine with its receptors, is stored in the extracellular space, closely associated to the cell membrane and bounded to specific latent TGFb1 binding proteins (LTBP). LTBP fixes the system in association to fibrillin-1 and fibronectin. The release of the non-covalently bounded mature TGFb1 from the LAP complex occurs by the action of proteases (matrix metalloproteases MMP, thrombin, and plasmin) or by the effect of environmental condition changes (pH, pO2, or temperature) [9,10]. Moreover, in the recent years, the role of av integrins in TGFb1 release from the LAP complex during fibrotic disease progression became evident [11,12]. It has been demonstrated that av integrins recognize the tripeptide Arg-Gly-Asp (RGD) sequence of the LAP, and the integrin binding induces a stretch of the complex and a mechanical release of the active TGFb1 [13]. In addition, it is known that av integrins might function as docking site for MMPs contributing to the protease-dependent release of TGFb1. Thus, the maintenance of high levels of TGFb1 exacerbates tissue damages and outlines the important role of integrins in the instauration and progression of fibrosis [14,15].
In this study, we present a novel class of avb3 RGD antagonists that interfere with integrin-dependent release of TGFb1 from LTBP complex. Interestingly, we found that the RGD-2-integrin antagonist counteracts the TGFb1-induced myofibroblast differentiation of endothelial precursor cells (ECPCs) by interfering, in these cells, with the autocrine loor of TGFb1 activation.
These findings may suggest an innovation in anti-fibrotic treatment. The common anti-fibrotic treatments are based, so far, on the use of corticosteroids and immunosuppressant drugs. These drugs may retard, but do not arrest, the disease progression. Thus, the cooperation between anti-fibrotic available drugs and novel RGD antagonists might contribute to the inhibition of the vicious cycle, led by TGFb1, underlying the disease progression.

Results
RGD triazole-derived avb3 antagonists RGD triazole-derived antagonists of the avb3 integrin receptor were synthesized in our laboratory as previously described [16]. Solid-phase assay was used to test the ability of RGD ligands to compete with 125 I-echistatin for the binding to avb3 integrin receptor. We found that RGD-1 compound showed an IC50 value of 2.1 ± 1.3 lM, while compound RGD-2 showed an IC50 of 37 ± 11 nM. In contrast, the IC50 value of RGD-3 compound was found to be [10 lM, addressing for a less active antagonistic efficacy toward avb3 receptor ( Table 1).

Effect of the triazole RGD antagonists on ECPCs adhesion
ECPCs, isolated from human umbilical cord blood (UCB), were characterized using flow cytometry assay by surface expression of endothelial cell-specific antigens: Ulex europaeus I agglutinin (Ulex), Platelet Endothelial Cell Adhesion Molecule-1 (PECAM-1/CD31), Vascular Endothelial Growth Factor Receptor 2 (VEGFR-2/KDR), Phagocytic Glycoprotein-1 (CD44), integrin b1-chain (CD29), and integrin heterodimers avb3 (Fig. 1a). ECPCs were monitored throughout the experimental procedures for avb3 integrin expression (Fig. 1b), the maintenance of endothelial phenotype (Fig. 1c), and the in vitro tube formation ability (Fig. 1d). Sub-confluent cultures of ECPCs between the 3rd and 6th passage were exposed to different doses (10, 1, 0.1, and 0.01 lM) of three different triazole RGD antagonists and allowed to adhere to vitronectin (VN) for 1 h. At the end of the incubation, non-adherent cells were removed by a gentle wash with PBS. We found a 40 % of inhibition of ECPCs adhesion to VN for RGD-1 compound at 10 lM, whereas lower concentrations of this compound showed a weak inhibition of cells adhesion. The RGD-2 compound clearly inhibited ECPCs adhesion to VN in a dose-dependent manner, ranging from 80 % of inhibition at 10 lM to 20 % at 10 nM. Finally, RGD-3 compound did not significantly inhibit ECPCs adhesion to VN (Fig. 1e). To further evaluate RGD-2 inhibitory activity toward different RGD containing substrata, ECPCs cells were exposed to RGD-2 compound at different doses before adhesion to osteopontin (OPN), Fibronectin (FN), and Matrigel. We found that RGD-2 antagonist showed a dose-dependent inhibition of ECPCs adhesion to OPN from 75 % at 10 lM to 25 % at 10 nM. In contrast to this, inhibition of adhesion to FN and Matrigel was very weak. Matrigel was used as negative control since it is composed by laminin, collagen IV, nidogen/enactin, and proteoglycan, which are ligands for other non-RGD families of integrin receptors (laminin-type and collagen-type) (Fig. 1f).
Effect of RGD-2 triazole compound on avb3 expression in TGFb1-stimulated ECPCs ECPCs were exposed to RGD-2 compound at 1 lM and/or TGFb1 at 1 nM/ml for 24 h. Here we found an overexpression of b3 subunit mRNA in TGFb1-treated ECPCs, while the co-treatment with the RGD-2 antagonist did not show any b3 subunit increase (Fig. 2a). Along with this, we found that protein expression of avb3 receptor was increased, in TGFb1 treated ECPCs, while treatment with RGD-2 antagonist abrogates TGFb1, inducing avb3 expression in ECPCs (Fig. 2b).
Effect of RGD-2 antagonist on TGFb1 signal transduction pathway and TGFb1 expression on ECPCs cell membrane In order to evaluate the signal transduction activation of TGFb1 pathway in ECPCs cells, we investigated ALK-5 expression and SMAD phosphorylation after TGFb1 and/ or RGD-2 treatment. We observed that 24 h treatment with TGFb1 induces the increase of ALK-5 expression also in the presence of the RGD antagonist (Fig. 3a). Moreover, when we evaluated SMAD2 phosphorylation, we found an increase in Phospho-SMAD2 in TGFb1-treated cells, while in RGD-2/TGFb1-co-treated cells, SMAD2 phosphorylation was similar to untreated cells (Fig. 3b). We, also, evaluated the effect of RGD-2 antagonist on TGFb1 activation in ECPCs. Cells were exposed for 24 h to exogenous TGFb1 (1 ng/ml). After the removal of the medium, cells were allowed to grow for the next 24 h in a standard medium. At the end of the second incubation, ECPCs were lysed directly in tissue culture plates and processed for western blotting analysis. We found high levels of membrane-associated TGFb1, in TGFb1-treated cells, compared to untreated cells. While the treatment with RGD-2 antagonist (1 lM) alone did not modify TGFb1 levels, the co-treatment with TGFb1 and RGD-2 antagonist reduced significantly the expression of endogenous TGFb1 to the levels found in untreated cells (Fig. 3a).

RGD-2 antagonist reverts TGFb1-induced EndoMT in ECPCs
We found that TGFb1 significantly increased the expression of a-SMA and vimentin in ECPCs together with a significant switch to an elongated morphology. Upon the morphological observation of ECPC-treated cultures, we found that TGFb1-treated cells present a fibroblast-like morphology, as expected, but this change was absent when the cells were treated with the RGD antagonist and TGFb1, and the culture morphology was slightly similar to that of untreated cells (Fig. 4a). RGD-2 antagonist-treated cells express a-SMA and vimentin at levels similar to the untreated cells. Thus, the treatment with the RGD-2 antagonist in association with TGFb1 significantly reduced the expression of mesenchymal markers of EndoMT, as demonstrated in western blotting analysis and in Values represent the mean ± SD immunofluorescence ( Fig. 4b). Next, we evaluated MAPK pathway activation after exposure to TGFb1 and to the cotreatment TGFb1/RGD-2. We found that ERK1/2 phosphorylation increases in ECPCs after treatment with exogenous TGFb1, and that the co-treatment with TGFb1/ RGD-2 inhibited ERK1/2 phosphorylation (Fig. 4c). To explore the effect of RGD antagonist on ECPCs biological behavior during TGFb1-induced EndoMT, we evaluated the invasive activity, wound healing ability, and in vitro angiogenesis of cells exposed to exogenous TGFb1 and RGD-2 antagonist. We found a significant reduction in ECPCs invasion through Matrigel after treatment with exogenous TGFb1. Significant but less intense reduction in ECPCs invasion was found after treatment with RGD-2 antagonist alone, while the treatment with RGD antagonist together with TGFb1 partially restored the originally migratory phenotype of the cells (Fig. 4d). Moreover, it is widely recognized that ECPCs exert a fundamental role the healing process, restoring the physiological function of vascular network. We found that, accordingly with previous finding, ECPCs cells exposed to TGFb1 lose their wound healing potency, while ECPCs cells treated with TGFb1 and RGD-2 antagonist almost completely recover their ability to heal the wound. Finally, it is known that ECPCs are characterized by their high proliferation rate and by their ability to give rise to capillary network in the in vitro angiogenic assay; thus, ECPCs were seeded on Matrigel in the presence of TGFb1 and/or the RGD-2 antagonist. After 24 h incubation, the number of branches per field was counted. We observed a significant decrease in the number of branches of ECPCs treated with TGFb1 as compared to untreated cells. The co-treatment with the RGD antagonist partially restored the ECPCs ability to form a capillary network (Fig. 4f).

Discussion
In the present study, we investigated the role of different triazole-derived RGD antagonists in the TGFb1 loop of ECPCs activation during endothelial-mesenchymal transition (EndoMT). Above the different triazole-derived compounds, we found that the stereochemistry of the tyrosine/phenyl moiety was crucial for the antagonistic efficiency of the compounds. The guanidine substitution, in the arginine mimetic portion of the molecule with the aminopyridine moiety, increased the binding affinity toward avb3 integrin. Integrin avb3 specifically recognizes the Arg-Gly-Asp (RGD) tripeptide present in the sequence of different extracellular matrix proteins. Although ECPCs express the integrin receptors avb3 and a5b1 which recognize RGD extracellular matrix proteins as vitronectin (VN), fibronectin (FN), and osteopontin (OPN), in our conditions, ECPCs adhesion to FN was not inhibited by the RGD antagonist; this result might be explained outlining two observations. The first is that avb3 and a5b1, even (1 ng/ml) and/or 1 lM RGD-2 antagonist: a mRNA for av, b3 subunits, and GAPDH, and b avb3 protein expression and densitometric analysis Fig. 3 Effect of RGD-2 antagonist on TGFb1 signal transduction pathway and TGFb1 expression on ECPCs cell membrane. a TGFb1 and ALK-5 were evaluated in ECPCs exposed for 24 h to exogenous TGFb1 and/or RGD-2 triazole and for additional 24 h to fresh standard medium. b Protein expression and densitometric analysis of PhosphoSMAD2 (pSMAD2) evaluated in ECPCs exposed to exogenous TGFb1 and/or RGD-2 for 1 h. All experiments were conducted at least three times. Values represent the mean ± SD. *P \ 0.05 though recognize the RGD tripeptide sequence, display specific binding affinity for a given ligand depending also on few essential amino acid residues surrounding the binding site pocket of the integrin receptor itself; thus, VN and OPN exhibit higher binding affinity for avb3 integrin, while FN exhibits high binding affinity to a5b1 integrin [17]. The second observation is that different RGD antagonist might display a different affinity to the same integrin receptor depending on those interactions with receptor amino acid residues close to the binding pocket [18]. Many clinical and experimental observations confirm the key role of TGFb during the instauration and the persistence of the fibrotic disease, mostly through the autocrine loop of TGFb activation [19]. The TGFb belongs to a ligand superfamily comprising the three forms of TGFbs (TGFb1, TGFb2, and TGFb3), Activins, BMPs (Bone Fig. 4 RGD-2 antagonist reverts TGFb1-induced EndoMT in ECPCs. a Contrast microscopy representative images of ECPCs after 24 h treatment with TGFb1 and/or RGD-2 compound, morphological changes. b Expression of EndoMT markers of mesenchymal differentiation; ECPCs were exposed to exogenous TGFb1 (1 ng/ml) and/ or 1 lM RGD-2 antagonist for 24 h, and a-SMA and vimentin expression were evaluated. Upper panel: western blotting for a-SMA and vimentin; lower panel: representative immunofluorescence images for a-SMA. c TGFb1-induced phosphoERK1/2 activation, ECPCs were exposed to different treatments for 24 h and lysed. d Invasiveness through Matrigel of ECPCs after 24 h treatment with exogenous TGFb1 (1 ng/ml) and/or 1 lM RGD-2 antagonist; for quantification, migrated cells were counted in six randomly chosen fields for each filter. e Wound healing assay of ECPCs after 24 h treatment with exogenous TGFb1 (1 ng/ml) and/or 1 lM RGD-2 antagonist; the degree of healing was quantified by measuring the distance between opposing edges of the wound. Four wound/ treatment and three measurements/wound were taken. f In vitro tube formation of ECPCs after 12 h treatment with exogenous TGFb1 (1 ng/ml) and/or 1 lM RGD-2 antagonist; for quantification, the number of branches per field was evaluated at 40 magnifications, in four different fields. Data were obtained from three independent experiments. Percentage of inhibition was expressed compared to untreated cells. Values represent the mean ± SD. *P \ 0.05, **P \ 0.01 Morphogenetic Proteins), and GDFs (Growth and Differentiation Factors). In general, TGFb superfamily ligands bind to a complex of TGFb type I and TGFb type II serine/ threonine kinase receptors. In the absence of stimuli, the homodimers of type I, also known as Activin receptor-Like Kinase (ALK), and of type II receptors are expressed on the cell surface in a separate form [20].
Upon recognition of the TGFb1 by the type II receptor, whose function is to present the ligand, the activation of the downstream signal requires the association of the type I/type II receptors in a heterotetrameric complex and the transphosphorylation of the type I receptor/ALK by the serine/threonine kinase activity of type II receptor. Then, the type I receptor transfers the signal into the nucleus by the phosphorylation of the SMAD proteins. The combination of different type II/type I tetrameric receptors determines the activation of different downstream signaling pathways, in response to the same ligand [21].
In endothelial cells, soluble TGFb1 interacts with type II receptor (TGFb1RII) and, among the different types of ALKs, ALK-5 phosphorylation induces the activation of SMAD2/3 pathway involved in endothelial cells inhibition of proliferation/migration and in the promotion of the extracellular matrix proteins synthesis, while ALK-1 phosphorylation induces the activation of SMAD1 and SMAD5 pathways that leads to endothelial cells migration and proliferation [22].
ECPCs might contribute substantially to the overexpression of TGFb that forces the maintenance of a dysregulated reparative process [23]. Thus, it has been demonstrated, in lung fibroblasts, that TGFb1 induces overexpression of avb3 integrin which potentiates TGFb1 responsiveness of the cells [24]. Although our findings are not completely exhaustive, indicate that ALK-5 expression is upregulated by exogenous TGFb1 treatment, but that the downstream activation of signal transduction, through the ALK-5 receptor, may depend on avb3-mediated release of endogenous TGFb1. Moreover, it is known that the type I receptor, in addition to Smad 2/3 protein phosphorylation, activates other non-Smad signaling pathways involving ERK1/2, TGF-b-activated kinase-1 (TAK-1), JNK, p38, Rho GTPases, and the PI3 K-AKT pathways [25]. We observed that while the interaction between the RGD antagonist and the integrin receptor did not induce any downstream signal transduction activation, the co-treatment with triazole RGD antagonist reduced significantly the ERK1/2 activation induced by TGFb1.
TGFb1, being the most important mediator in tissue fibrosis, is recognized to induce in vitro the endothelialmesenchymal transition in endothelial cells and circulating endothelial precursor cells [26][27][28][29]. Despite other authors demonstrated that a cyclic RGD antagonist enhances in vitro and in vivo vascularization in a model of EPCs [30], we found a mild but significant reduction in the capillary formation after the treatment with the triazole RGD antagonist. This result was in agreement with our previous observation on mature human umbilical vein endothelial cells (HUVEC) [31]. We suggest that our findings depend by the contribution of two main different mechanisms of interaction between the RGD-2 antagonist and the avb3 receptor. The first is the direct role that integrin avb3 exerts in endothelial cells invasion and in vitro angiogenesis, resulting in a reduction in invasion and angiogenesis operated by the linear triazole peptidomimetic antagonists alone. The second is the interfering effect of RGD-2 antagonist in the TGFb1/avb3 crosstalk that reverts EndoMT in ECPCs and exerts a fundamental anti-fibrotic and pro-resolution effect. It has to be noted that the interest in the synthesis and identification of bioactive RGD antagonists originates from the need in improving anti-angiogenic treatment of cancer, since the outcomes of drugs against vascular endothelial growth factor and relative receptors have shown a partial lack of efficacy both in vitro and in vivo [32][33][34]. The av RGD mimetic cyclic pentapeptide, Cilengitide (EMD 121974), has been identified as the first synthesized anti-angiogenic small molecule, and its effect on endothelial cells has been demonstrated [35]. Cilengitide showed encouraging activity in patients with glioblastoma and melanoma, and is currently used in clinical trials [36,37]. In contrast to this, other authors found that Cilengitide enhances angiogenesis, and promotes tumor growth and cell invasion [38]. In our model, despite the strong effect of the RGD triazole on the inhibition of cell adhesion, we found a weak effect on the inhibition of invasion and angiogenesis that might be explained by the specific and unique activity of this linear triazole RGD antagonist against the avb3 receptor. Along with this, we should take into account that invasiveness and tube formation assays are performed on Matrigel substrate, against which the RGD triazole effect is not specific, and that the spatial occupation of the integrin receptor binding site by the RGD triazole may not be sufficient to overcome integrin redundancy in invasiveness and tube formation processes. Indeed, during the ECPCs invasion and migration process, many interactions between ECM and other integrin receptors, as a5b1 and a6b1, have been found to be involved [39][40][41], as well as many interactions between ECM and cell-derived proteases. Thus, the avb3 integrin receptor activity might be only partially involved in ECPCs invasion and tube formation [42].
In a fibrotic lesion, TGFb1 is released by activated fibroblasts and macrophages, or endothelial cells themselves, and induces a local overload of TGFb1 driving to stroma activation. Circulating precursor endothelial cells, recruited by pro-inflammatory cytokines, differentiate in activated fibroblasts, and endogenous TGFb1 amplifies the fibrotic vicious circle. To date, the therapeutic interventions on fibrotic diseases are the use of anti-inflammatory mediators, immunomodulatory drugs, or agents that directly block TGFb1 activity, such as specific antibody. Unfortunately, anti-inflammatory treatment only retards but does not resolve fibrosis, while TGFb1 antibody might compromise important activity of this cytokine in other tissues [43]. Our findings are in agreement with the importance of integrin involvement in fibrosis [44,45]. Here, we introduced a new synthesized non-peptidic compound with anti-fibrotic properties in an in vitro model of tissue fibrosis. This RGD antagonist not only prevents the avb3-mediated TGFb1 release dampening the autocrine loop of ECPCs activation, but also promotes a mesenchymal-to-endothelial transition (MEndoT), reverting the TGFb1-mediated phenotype of activated endothelial cells. It should be noted that a avb6 monoclonal antibody is involved in a phase 2 study for idiopathic pulmonary fibrosis (STX-100, NCT01371035) [46]. This observation confirms the need of further investigations to delineate the role that integrin antagonists might have in the designing of anti-fibrotic therapies. On the whole, our results might support the hypothesis for an EndoMT reversion mediated by RGD antagonist. Exogenous TGFb1, released in the fibrotic microenvironment, induces avb3 receptor overexpression in ECPCs. The presence of high levels of avb3 receptors enhances the release of active TGFb1 from LAP. Endogenous ECPCs-derived TGFb1 interacts with its receptors on ECPCs themselves promoting EndoMT (Fig. 5a). The RGD antagonist, by occupying the binding site of the avb3 receptor, reduces the traction exerted by the heterodimer on the RGD sequence of the LAP, thus reducing endogenous TGFb1 release and dampening the autocrine loop of TGFb1 activation (Fig. 5b).

Synthesis of triazole-based non-peptide RGD peptidomimetic
The triazole-based RGD ligand was achieved using the click chemistry by combining azide and alkyne in the Cucatalyzed azide-alkyne cycloaddition, followed by acidmediated hydrolysis of the protecting groups of side chain isosteres [16,31,47]. Three bioactive compounds were selected starting from a group of avb3 integrin ligands.

Solid-phase integrin binding assay
The inhibition of 125 I-echistatin-specific binding to avb3 integrins by RGD antagonists was evaluated as previously reported [47]. Briefly, 125 I-echistatin with a specific activity of 2000 Ci/mmol was purchased from Perkin Elmer, and integrin avb3 from human placenta was purchased from Millipore. Purified avb3 integrin (Millipore) was diluted in coating buffer [20 mM Tris (pH 7.4), 150 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, 1 mM MnCl2] at concentrations of 500 or 1000 ng/mL. Aliquots of avb3 integrin (100 lL/well) were added to a 96-well plate (Perkin Elmer), and plates were incubated overnight at 4°C, followed by washings with blocking/binding buffer [20 mM Tris (pH 7.4), 150 mM NaCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 1 mM MnCl 2 , and 1 % BSA], and then incubated at room temperature for an additional 2 h. After two washing with the same buffer, aliquots of 125 I-echistatin (0.05 nM) were added to each well with different concentrations of RGD compounds (from 0.01 to 100 nM). Non-specific binding was defined as 125 I-echistatin bound in the presence of an excess (1 lM) of unlabeled echistatin. After 3 h incubation at RT, plates were washed three times with blocking/ binding buffer and counted in a Top-Count NXT microplate scintillation counter (Perkin Elmer) using 200 lL/

Isolation of endothelial colony-forming cells and culture conditions
Endothelial Colony-Forming Cells (ECPCs), a subpopulation of Endothelial Precursor Cells (EPCs), were isolated from [50 ml human umbilical cord blood (UCB) of health newborns, as described [48,49], for the banking established by the Umbilical Cord Bank of Careggi Hospital (Florence, Italy) after maternal informed consent in accordance with the Declaration of Helsinki and in compliance with Italian legislation. ECPCs were analyzed for the expression of surface antigens (CD31, CD44, CD29, ULEX, KDR, and avb3) and were grown in complete EGM TM -2 BulletKit TM (CC-3162 Lonza) with 10 % Fetal Bovine serum (FBS) (Hyclone). Confluent cell cultures were propagated every 3 days, for no more than 10 passage, and seeded on gelatin-coated tissue culture plates at a density of 5 9 10 5 cells/cm 2 in a 5 % CO2 humidified incubator at 37°C. ECPCs cultures, grown to subconfluence, were treated for 24 h with 1 ng/mL of hrTGFb1 (Peprotech), or triazole RGD antagonist (1 lM), or both in EGM-2 medium (Endothelial cells Growth Medium-2) in the absence of growth factors and in the presence of 2 % FBS (Fetal Bovine Serum). In some experiments, treated cells were exposed to fresh medium for an additional 24 h treatment before protein extraction.

Cell adhesion assay
Sub-confluent cultures of ECPCs between the 3rd and 6th passage were used for the inhibition of adhesion assay. Plates (96 wells) were coated with Matrigel TM Matrix (BD Biosciences) (10 lg/ml), vitronectin (10 lg/mL), fibronectin (1 lg/mL), or osteopontin (0.5 lg/mL), by overnight incubation at 4°C. Plates were washed with PBS and then incubated at 37°C for 1 h with PBS-1 % BSA. After being washed, ECPCs were counted, suspended in serumfree medium, and exposed to triazole RGD antagonists (final concentration was 0.01, 0.1, 1.0, or 10 lM) at 37°C for 30 min to allow for the ligand-receptor equilibrium to be reached. ECPCs were plated ((4-5)10 4 cells/well) and incubated at 37°C for 2 h. All the wells were washed with PBS to remove the non-adherent cells, and 0.5 % crystal violet solution in 20 % methanol was added. After 2 h of incubation at 4°C, plates were examined at 540 nm in an ELX800 counter (Bio TEK Instruments). Data were expressed as percentage of inhibition compared to untreated cells. Experiments were conducted in triplicate and were repeated at least three times [49].

RNA extraction and RT-PCR
Total RNA was extracted from ECPCs using RNAgents (Total RNA Isolation System, Promega, Madison, WI) and quantified spectrophotometrically. A volume of 500 ng of RNA was then retrotranscribed using ImProm-II reverse transcriptase (Promega, Madison, WI). Aliquots of 2 ll of the cDNA were used for PCR amplification. The specific primers used for the identification of human av, b3, and GAPDH were designed according to published human cDNA sequences in the Genbank database, using FastPCR software [50]: av (forward5 0 -CTA TGA GCT GAG AAA CAA TGG TCC-3 0 and reverse 5 0 GCT GCT CCC TTT CTT GTT CTT C-3 0 690-bp product); b3 (5 0 -GGG GAC TGCC TGT GTG ACT C-3 0 and reverse 5 0 -CTT TTC GGT CGT GGA TGG TG-3 0 610-bp product); GAPDH (forward: 5 0 -ACC ACA GTC CAT GCC ATC AC-3 0 and reverse: 5 0 -TCC ACC ACC CTG TTG CTG TA-3 0 , 452-bp product). PCR was carried out on a Perkin Elmer Thermal cycler. Ten microliters of each PCR products were visualized after electrophoresis in a 2 % agarose. cDNA products were evaluated on the basis of a standard PCR marker (Promega) and quantified by densitometric analysis using ImageJ software (NIH).

Immunofluorescence
ECPCs cells were cultured on 25-mm coverslips pre-coated with 10 lg/ml of VN. After 24 h of incubation, in the presence of TGFb1, RGD-2 antagonist, or both, in complete EGM-2 medium, cells were washed in PBS, were fixed in 4 % paraformaldehyde, and membranes were permeabilized in 0.1 % Triton X-100 solution. Coverslips were incubated in blocking solution (PBS supplemented with 4 % BSA and 1 % horse serum) and then incubated at 4°C overnight with anti-a-SMA primary antibodies, washed and incubated for 1 h with goat anti-mouse AlexaFluor-488 antibodies (Invitrogen). Cell nuclei were counterstained with DAPI (1 lg/ml for 10 min at 37°C). Following two washes in PBS, coverslips were mounted with propylthiogallate on glass slides, and the cells were observed with an inverted confocal Nikon Eclipse TE2000 microscope equipped with a 960S-Fluor oil immersion lens.

ECPCs migratory activity
In order to investigate the migratory activity of ECPCs, Boyden chamber assays were performed using Millicell cell culture inserts (Millipore 8-lm pore size, 12 mm diameter). A 3D barrier of 50 lg/cm 2 of Matrigel was stratified on the filters, and ECPCs were loaded into the upper compartment (5 9 10 4 cells/well) in 400 ll of complete EGM-2 containing TGFb1, RGD-2 antagonist, or both and placed into 24-well culture dishes containing 600 ll of EGM-2 complete medium. After overnight incubation at 37°C, non-invading cells were removed mechanically using cotton swabs, and micro-porous membrane containing the invaded cells was fixed in 96 % methanol and stained with Diff-Quick staining solutions. Migratory activity was evaluated by counting the cells which migrated toward the lower surface of the filters (six randomly chosen fields for each filter).

In vitro tube formation assay
The effects of TGFb1, RGD-2 antagonist, or both on the ability of ECPCs to reorganize and differentiate into capillary-like network were assessed, thereby Matrigel morphogenesis assay. Briefly, 50 ll of Matrigel (1 mg/ml) was added into wells of a 96-well plate and polymerized for 1 h at 37°C. After 24 h of incubation, in the presence of TGFb1, RGD-2 antagonist, or both, in complete EGM-2 medium, cells were washed once with PBS, harvested by trypsinization, and collected by centrifugation. Then, cells were resuspended in 200 ll of EGM-2 complete medium and placed into Matrigel-coated wells (6 9 10 5 cells/well). After 12 h incubation on Matrigel at 37°C, the plates were photographed under a phase contrast microscope. The degree of tubule formation was quantified by counting the branching points in four randomly chosen fields from each well [51].

Wound healing assay
Cell migration was evaluated by an in vitro wound healing assay. Cells were grown at 80-90 % confluence in 35-mm dishes; the cell layer was wounded with a sterile 200-ml pipette tip and incubated in 0.1 % FBS culture medium for 24 h. The wound was observed after 18 h, and pictures were taken using phase contrast microscope. The degree of healing was quantified by measuring the distance between opposing edges of the wound. Four wound/treatment and three measurements/wound were taken. Percentage of inhibition was expressed compared to untreated cells.

Statistical analysis
Results were analyzed using a 2-tailed Student's t test to assess statistical significance. Statistical differences are presented at probability levels of P \ 0.05 or P \ 0.01.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.