Lina Badimon, Cardiovascular Research Center, C/Sant Antoni Mª Claret 167, 08025 Barcelona, Spain. Tel.: +34 93 556 58 80; fax: +34 93 556 55 59. E-mail: firstname.lastname@example.org
Summary. Background: Tissue factor (TF) and its signaling mediators play a crucial role in angiogenesis. We have previously shown that TF-induced endothelial cell (EC) CCL2 release contributes to neovessel formation.
Objective: In this study, we have investigated the signaling pathways involved in TF-induced EC tube formation.
Methods: The human microvascular endothelial cell line (HMEC-1) cultured onto basement membrane-like gel (Matrigel) was used to study TF signaling pathways during neovessels formation.
Results: Inhibition of endogenous TF expression in ECs using siRNA resulted in inhibition of a stable tube-like structure formation in three-dimensional cultures, associated with a down-regulation of Akt activation, an increased phosphorylation of Raf at Ser259 with a subsequent reduction of Raf kinase and a reduction of ERK1/2 phosphorylation ending up in Ets-1 transcription factor inhibition. Conversely, overexpression of TF resulted in an increase in tube formation and up-regulation of Akt protein. Moreover, immunoprecipitation of Akt and western blotting of the immunoprecipitates with anti-TF antibody revealed a direct interaction between TF and Akt. The effects of silencing TF were partially reversed by a PAR2 agonist that rescued tube formation, indicating that the TF-Akt pathway induces PAR2-independent effector signaling. Finally, enforced expression of Akt in TF-silenced ECs rescued tube formation in a Matrigel assay and induced Ets-1 phosphorylation.
Conclusions: In EC, TF forms a complex with Akt activating Raf/ERK and Ets-1 signaling induces microvessel formation.
Tissue factor (TF) is the principal initiator of the coagulation cascade, via the binding of factor (F)VII/FVIIa . In addition to this primary role in blood coagulation, accumulating evidence suggest that TF also plays major non-hemostatic roles as a multi-facetted transmembrane signaling receptor involved in the regulation of angiogenesis [2–4], cancer metastasis  and in inflammation . TF contributes to these pathologies by activation of the coagulation cascade as well as by coagulation-independent signaling events [7,8].
TF is linked to two different cellular signaling pathways: (i) to 7-transmembrane, G-protein-coupled receptor activation  or (ii) integrins activation . The primary signaling pathway for TF in angiogenes involves activation of G-protein-coupled, protease-activated receptors (PARs) . The cleavage of PAR2 activates Gαq, Gα12/13 and Gαi, which initiate the intracellular signaling and thus regulates gene transcription and protein translation, cell proliferation and survival, and cell motility and integrin activation. The association of PAR2 with TF-dependent signaling pathways emerged from studies that characterized the TF cytoplasmic domain phosphorylation. Cytoplasmic TF phosphorylation is dependent on PKCα activation downstream of PAR2 signaling . However, in a reciprocal way, it has been suggested that PAR2- dependent TF phosphorylation shuts off the negative regulatory function of the intracellular domain of TF in tumor cell migration and angiogenesis , indicating that TF tail phosphorylation could also be a regulator of PAR2 function.
TF also shows cross-talk with integrins. TF is associated with β1 integrin-mediated migration through the extra- and intracellular domain, when TF is phosphorylated in the cytoplasmic domain this positively regulates α3β1-dependent cell migration , suggesting that the PAR2 interaction is involved in TF binding to integrins and the disruption of these complexes decreases angiogenic signaling. However, it has been demonstrated that alternatively spliced TF induces angiogenesis through αvβ3 and α6β1 integrin ligation independently of downstream coagulation factors or PAR2 activation .
In addition, TF signaling induces calcium mobility in different cell types [14,15], activates the p42/p44 mitogen-activated protein (MAP) , p38 MAP kinase and c- Jun N-terminal kinase (JNK) . The MAP kinase family has been associated with cell proliferation. Therefore, TF/FVIIa could contribute to angiogenesis simply by stimulating cell division. Activation of the MAP kinase pathway is mediated downstream of the PARs, independent of the cytoplasmic domain of TF . Moreover, in fibroblasts, TF activates the central axis consisting of Src-like kinases, phosphatedyli-nositol 3-kinase (PI3K) and the anti-apoptotic protein c-Akt/PKB [19,20]. Akt, also known as Protein kinase B (PKB), regulates essential cellular functions such as migration, proliferation, differentiation, apoptosis, metabolism, and influences the expression and/or activity of various pro- and anti-angiogenic factors.
Recently, we have demonstrated that TF has a key role in coordinating the formation of neovessels with a stable phenotype via CCL2 expression , and that this expression is regulated through Ets1 transcription factors . Lavenburg et al.  showed that Akt signaling regulates Ets-1 and controls different aspects of cell motility that are integrated in the precise regulation of vascular tube formation. It can thus be hypothesized that TF signals through Akt to regulate Ets-1 transcription and to stabilize new microvessel formation.
Materials and methods
Human microvascular endothelial cells (HMEC-1) were used to study the signaling process of TF-induced endothelial cell tube formation. Key experiments were repeated using primary human dermal microvascular endothelial cells (HDMEC). Three-dimensional culture systems, as well as in vivo implantation of cell-loaded Matrigel plugs in nude mice were performed to investigate vessel formation. A nucleofector device was used for RNA silencing, and enforced expression of TF or Akt-1 was performed by stable or transient transfections. RT-PCR was done to visualize mRNA levels and western blot was performed to study protein expression in endothelial cells. Finally, the interaction between TF and Akt-1 was analyzed by immunoprecipitation and confocal microscopy. Results are given as mean ± standard error of the mean (SEM) unless otherwise indicated. An expanded methods section is available in the Supporting Information.
TF induces tube-like formation by phosphorylation of Akt
We have recently demonstrated in vitro and in vivo that TF is necessary to coordinate the formation of microvessels with a stable vascular phenotype and that it is the TF expressed in endothelial cells that is the driving stimulus for complex neovessel formation . To investigate the mechanism by which TF induce neovessel formation, we used an endothelial cell line (HMEC-1) transduced with TF siRNA or TF lentivirus. When TF was silenced in HMEC-1 (reduction of 88 ± 2.5% in mRNA TF) capillary formation was inhibited (Fig. 1A). Quantification revealed a five-fold reduction in area covered by tubes in HMEC-1 TF siRNA compared with HMEC-1 scrambled siRNA (Fig. 1B). The reduction in mRNA TF was accompanied with an inhibition of TF expression (Fig. 1C). In contrast, when TF was overexpressed by lentiviral transduction, capillary tube-like structures were significantly increased (Fig. 1A). As the PI3K/Akt signaling pathway has been shown to regulate cell survival by TF/FVIIa, we analyzed the role of Akt, a downstream effector of the TF-induced signaling via activation of PI3K. The expression of phospho-Akt was abrogated in TF silenced cells (Fig. 1C,D). In contrast, Akt phosphorylation was highly increased in TF overexpressing cells. Similar results were obtained using primary endothelial cells (HDMEC) (Fig. S1). These data suggest that TF-induced neovessel formation is triggered through Akt phosphorylation.
Tube formation induced by TF is partially independent of PAR2
Signaling activities of TF have previously been reported as mainly mediated by PAR2. Under basal conditions, endothelial cells express low levels of PAR2 , but these levels are raised during angiogenesis . To study the influence of PAR2 in TF-dependent angiogenesis, TF silenced cells were stimulated with a PAR2-agonist (HA-5042; Bachem AG, Bubendorf, Switzerland) (30 μm). PAR2 activation induced a mild recovery in the formation of tube-like structures, but it was not able to rescue the inhibition induced by TF silencing (Fig. 2A,B). The PAR2 agonist induced a slight phosphorylation of Akt in TF silenced cells but did not modify TF expression (Fig. 2C). The PAR2 agonist did not have any effect by itself on tube formation or Akt-phosphorylation in wild-type cells (Fig. S2). We further used a specific PAR2 siRNA approach to abrogate PAR2 expression in the cells. PAR2 silencing decreased tube-like structure formation in Matrigel cultures; however, the effect was significantly lower than the effect of TF silencing (reduction was 22.5% vs. 71% in TF silenced cells) (Fig. 2E). Similar results were obtained using primary endothelial cells (Fig. S1). The silencing of PAR2 was accompanied with an inhibition of PAR2 protein expression and decreased Akt and TF phosphorylation (Fig. 2G,H). The decrease in Akt phosphorylation was more evident in TF silenced cells than in PAR2 silenced cells (reduction was 93.75% vs. 50% in PAR2 silenced cells). These results indicated that TF is one of the main triggers of neovessel formation.
Akt signaling is necessary for endothelial tube formation
To show a direct link between TF signaling and the Akt activation, we silenced Akt with a specific siRNA. First we observed that Akt silencing blocked tube-like structure formation both in HMEC-1 and HDMEC (Fig. 3A and S1), and that this effect could not be rescued by non-lipidated recombinant TF (rTF) treatment (50 nm)  (Prospec-Tany Technogene, East Brunswick, NJ, USA) (Fig. 3A). The same results were obtained when we used HA Akt DN (a kinase dead mutant) transfected cells instead of silenced cells (data not shown). The western blot analysis shows that Akt siRNA decreased Akt protein expression and that this effect was not rescued by rTF (Fig. 3B). To further demonstrate the direct link, we transiently expressed a myristylated form of Akt (myrAkt). myrAkt includes a N-terminal Src myristylation signal that targets it to the membrane, this form of Akt is constitutively active. Restoration of Akt activity with the myrAkt in TF silenced endothelial cells reversed the phenotype of these cells (Fig. 3C). Notably, tube formation was rescued by restoring Akt activity. Together, these data indicate that TF signals through Akt and regulates tube formation, which in turn influences the angiogenesis process.
In addition, Akt was immunoprecipitated from cell lysates by applying anti-Akt antibodies. Western blotting of the immunoprecipitates with anti-TF antibodies revealed a complex formation between TF and Akt, and this interaction was abolished in TF silenced cells (Fig. 3F). The same results were obtained when we immunoprecipated TF and blotted the Akt. The functional consequence of this complex was therefore investigated.
To evidence that TF induces microvessel formation in vivo through Akt, we subcutaneously inoculated nude mice with 5 × 106 scrambled-siRNA cells, TF-siRNA cells, TF overexpressing cells or Akt-siRNA TF overexpressing cells in Matrigel plugs. The cells implanted in Matrigel plugs retained their viability and assured the sustained inhibition of TF or Akt over the course of the entire experiment (data not shown). Angiogenesis was evaluated after 7 days. In the control plugs, spouting endothelial microvessels were highly evident and showed frequent branching and abundant blood-filled channels containing red blood cells (Fig. 4A). In contrast, when TF was silenced a few single small un-branched vessels were observed (Fig. 4B). When TF expression was up-regulated in endothelial cells, a thick mesh of microvessels surrounded by a branching structure of microvessels was formed and all of them contained red blood cells (Fig. 4C). Finally, when in TF overexpressing cells Akt was silenced, only small vessels without branching were found (Fig. 4D). All of these results indicate that TF signals through Akt to induce microvessel formation.
TF silencing through Akt signaling increase Raf Kinase activity
We have recently demonstrated in endothelial cells that TF signals through ERK1/2 to stimulate the transcription factor Ets-1 which regulate CCL2 expression and induces angiogenesis , and Lavenburg et al.  demonstrated that Akt activates Ets-1 and regulates vascular tube formation. It was also demonstrated that Akt negatively regulates the Ras/Raf/MEK/ERK pathway via phosphorylation and inactivation of Raf at Ser259 [27,28]. To analyze whether the interaction between TF and phospho-Akt affects the phosphorylation-state of Raf, ERK ½, Est-1 and finally CCL2, we studied the phosphorylation pattern of these molecules in TF silenced or TF overexpressed endothelial cells cultured in Matrigel for 18 h. Aliquots of cell lysates were first probed with phospho-Ser473-Akt and Akt antibodies. There was over a 90% reduction in Akt phosphorylation in TF silenced cells; in contrast, there was a 50% increase in Akt phosphorylation in TF overexpressing endothelial cells (as shown previously in Fig. 1D). A second set of aliquots from the same cell lysates was probed for phospho-Ser259-Raf and Raf. TF silencing induced a moderate increase in Raf-Ser259 phosphorylation and TF overexpression reduced its phosphorylation, demonstrating that TF signaling through Akt may thereby reduce Raf Kinase activity (Fig. 5). At the level of ERK 1/2 TF-induced phosphorylation was decreased in TF silenced cells, again suggesting an altered Raf kinase activity (Fig. 5). Finally TF silencing also decreased Ets-1-Thr38 phosphorylation. The marked effect on Akt activity followed by inhibition of ERK 1/2 and Ets-1 phosphorylation in conjunction with an increased Raf-Ser259 phosphorylation points to a regulatory role of Akt in TF-induced angiogenesis.
Moreover, CCL2 expression was significantly reduced in TF silenced cells, and rescued by transfection with myrAkt (Fig. 5C). Because CCL2 is a secreted protein, and we previously demonstrated that TF inhibition decreases CCL2 secretion in endothelial cells , the concentration of CCL2 in the supernatants of TF silenced endothelial cells overexpressing Akt cells was examined. myrAkt transfection increases CCL2 expression protein and CCL2 secretion (Fig. 5D).
Our data suggest that TF-induced CCL2 secretion and tube formation is dependent of TF–Akt complex formation that signals through Raf/ERK/Ets-1.
Akt signaling is independent of the distal region of TF cytoplasmic domain
TF signaling has been linked to TF cytoplasmic domain phosphorylation , specially phosphorylation of the three serine residues . To directly analyze the interaction between TF and Akt, we studied Akt-phosphorylation in endothelial cells where the TF cytoplasmic domain had been partially deleted (TFΔCT). Cells with TFΔCT do not express the distal cytoplasmic region from amino acids 252–263 that contains the three serine residues (Ser253, Ser258, and Ser263). First, as shown in Fig 6A, partial deletion of TF cytoplasmic domain did not induce alterations in endothelial tube-like formation compared with TF over-expressing cells (Fig. 6A). Moreover, Fig. 6B shows that Akt presents the same phosphorylation levels in TFΔCT endothelial cells than in TF overexpression cells. In order to verify whether Akt interact with TF, Akt was immunoprecipitated from cell lysates applying anti-Akt antibodies. Western blotting of the immunoprecipitates with anti-TF antibodies revealed an interaction between TF and Akt even when the distal region of TF cytoplasmic domain was deleted (Fig. 6C). The same results were obtained when TF was immunoprecipitated. To confirm these results TFΔCT cells cultured for 24 h were immunostained for phospho-Akt. Confocal microscopy showed that TF (red) and phospho-Akt (green) presented the same expression profile in cells with and without a cytoplasmic domain deletion (Fig. 6D). These data suggest that Akt phosphorylation is independent of the TF cytoplasmatic domain.
TF has important functions both dependent and independent of blood clotting. Functions unrelated to coagulation include the regulation of angiogenesis [2–4], cancer metastasis  and inflammation . The molecular mechanisms of TF signaling to induce angiogenesis have been based on TF-PAR2 signaling [30,31]. Here, we provide evidence that TF induce tube formation by Akt signaling, and that this signaling can be independent of PAR2.
In this study, we have used HMEC-1 cells, an immortalized human microvascular endothelial cell line. Previously we have demonstrated that HMEC-1 retain the morphologic, phenotypic and functional characteristics of primary human microvascular endothelial cells (HDMEC) . Cells were cultured in 3DBM to study tube-like structure formation, because as reported by Blair et al. , using scanning electron microscopy and reflective confocal microscopy, HDMECs form tube-like structures. We used TF siRNA because it is more sensitive, specific and stable than antisense RNA and it is able to silence expression of all forms of TF (full length and alternatively spliced).
TF through PAR2 signaling induces proangiogenic and immuno modulating cytokines and growth factors expression. Our results support the role of PAR2 signaling in the angiogenesis processes in accordance with previous results . However, we report that TF has a significant PAR2-independent role in neovessel formation because in TF silenced cells PAR2 activation ligand binding was not able to rescue angiogenesis. PAR2 signaling induces phosphorylation of the cytoplasmic domain Ser258 of TF . The non-phosphorylated TF cytoplasmic domain negatively regulates PAR2 signaling. Indeed, PAR2-dependent TF phosphorylation shuts off the negative regulatory function of the intracellular domain signaling of TF . Phosphorylation of the TF cytoplasmic domain results in activation of various MAPK family members and triggers gene transcription . As TF induce MAPK it was suggested that TF might contribute to the angiogenic processes simply by stimulating cell division . However, we have previously shown that TF plays a role in angiogenesis stabilizing the new tube formation . It has been described that, in fibroblast, TF signaling activates the axis Scr-like kinase, PI·3 kinase-Akt signaling and Rac  leading to cytoskeletal arrangement. Moreover, in neuroblastoma cells, TF signaling activates a downstream PI3/Akt pathway mediating the anti-apoptotic and migration processes . These studies suggest that the PI3 kinase-Akt pathway in several cell types has an important role in TF-mediated signaling events. Our results revealed that TF signals through Akt to induce formation of tube-like structures with stable phenotype in 3D cultures. Inhibition of endogenous TF expression in endothelial cells resulted in a down-regulation of Akt activation and inhibition of microvessel formation. In addition, enforced expression of Akt in TF-silenced endothelial cells rescued the tube formation in a Matrigel assay. Indeed, loss of Akt has already been associated with increased vascular permeability and reduced maturation of vessels [36,37]. Furthermore, our results in endothelial cells demonstrate that TF-dependent activation of Akt results in Raf-Ser259 dephosphorylation, decreasing Raf kinase activity and subsequently ERK1/2 phosphorylation. The Raf-1 kinase is regulated by phosphorylation, and Ser259 has been identified as an inhibitory phosphorylation site, the dephosphorylation of Ser259 is an essential part of Raf-1 kinase activation . In a human breast cancer cell line, phosphorylation of Raf by Akt inhibited activation of the Raf-MEK-ERK signaling pathway , and inhibition of TF/PAR2 blocked activation of the ERK pathway in glioma cells . Previously we have seen that, in endothelial cells, TF signals through ERK1/2 to induce Ets-1, which regulates CCL2 gene expression by binding to its promoter region and thereby the secreted CCL2 attracts smooth muscle cells to the endothelial cells forming stable microvessels . These results are consistent with the proposal that Akt signaling regulates Ets-1 expression and control vascular tube formation , and the observation that in Akt-/- mice the vascularization was defective in perycite recruitment .
Based on these studies, it seems to the TF-Akt axis plays an important role in human endothelial cell microvessel formation. Indeed, phosphorylation of Akt plays a pivotal role; however, the total levels of Akt and Ets proteins are also regulated by TF levels and may also be important in human EC function. Future experiments will need to dwell in these mechanisms not yet understood.
TF-induced angiogenesis has been linked to its TF cytoplasmic domain and to integrin ligation . The role of TF cytoplasmic domain in angiogenesis has been widely studied. TF cytoplasmic tail interacts with ABP-280 providing a molecular pathway that supports metastasis and vascular remodeling . However, some investigators have found that deletion of TF cytoplasmic tail does not result in abnormal embryonic development despite the essential role of full length TF in normal angiogenesis [41,42]. In our studies, we evidence that the lack of cytoplasmic domain does not abrogate TF signaling though Akt. In fact Akt can be activated through integrin ligation as shown in glioma cells were TF regulates angiogenesis through αvβ3 . Indeed, αvβ3 regulates migration by PI3-kinase/akt activation . Recently Srinivasan et al.  demonstrated in endothelial cells that TF via integrin mediates PI3K/Akt-NFkB signaling regulating expression of adhesion molecules and facilitating the recruitment of monocytes. The functional relation among TF-integrin-Akt in the angiogenic process requires further investigation.
In conclusion, as shown in Fig. 7, which depicts a model of the TF signaling, we suggest that in endothelial cells TF activates Akt signaling that through ERK1/2 induces Ets-1 transcription factor and the up regulation of CCL2 gene expression. Then CCL2 promotes vascular smooth muscle cells recruitment that stabilizes the sprouting endothelial tubes. We have shown that these mechanisms are partially independent of PAR2.
Because neovessel formation and stabilization are important pathophysiological features of inflammation, atherosclerosis and ischemic injury, our findings help to ascertain the molecular regulation of the process and may have significant therapeutic implications.
We thank an Olaya Garcia and Mónica Pescador for their excellent technical assistance. This work was supported in part by grants from Ministry of Science and Education of Spain (SAF2010-16549), and Instituto de Salud Carlos III (CIBERobn-CB06/03) (to LB), and (CP07/00224) (to GA). We thank the Fundación de Investigación Cardiovascular and the Fundación Jesus Serra for their support.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.