Background information. Endothelial cells play a major role in angiogenesis, the process by which new blood vessels arise from a pre-existing vascular bed. VEGF-A (vascular endothelial growth factor-A) is a key regulator of angiogenesis during both development and in adults. HGF (hepatocyte growth factor) is a pleiotropic cytokine that may promote VEGF-A-driven angiogenesis, although the signalling mechanisms underlying this co-operation are not completely understood.
Results. We analysed the effects of the combination of VEGF-A and HGF on the activation of VEGFR-2 (VEGF receptor-2) and c-met receptors, and on the stimulation of downstream signalling pathways in endothelial cells. We found that VEGFR-2 and c-met do not physically associate and do not transphosphorylate each other, suggesting that co-operation involves signalling events more distal from receptor activation. We demonstrate that the VEGF isoform VEGF-A165 and HGF stimulate a similar set of MAPKs (mitogen-activated protein kinases), although the kinetics and strengths of the activation differ depending on the growth factor and pathway. An enhanced activation of the signalling was observed when endothelial cells were stimulated by the combination of VEGF-A165 and HGF. Moreover, the combination of VEGF-A and HGF results in a statistically significant synergistic activation of ERK1/2 (extracellular-signal-regulated kinase 1/2) and p38 kinases. We demonstrated that VEGF-A165 and HGF activate FAK (focal adhesion kinase) with different kinetics and stimulate the recruitment of phosphorylated FAK to different subsets of focal adhesions. VEGF-A165 and HGF regulate distinct morphogenic aspects of the cytoskeletal remodelling that are associated with the preferential activation of Rho or Rac respectively, and induce structurally distinct vascular-like patterns in vitro in a Rho- or Rac-dependent manner.
Conclusions. Under angiogenic conditions, combining VEGF-A with HGF can promote neovascularization by enhancing intracellular signalling and allowing more finely regulated control of the signalling molecules involved in the regulation of the cytoskeleton and cellular migration and morphogenesis.
Vascular endothelial cells provide a structural barrier between circulation and surrounding tissue, and play a pivotal role in regulating vascular permeability and blood flow (Cines et al., 1998). Endothelial cells also play major roles in angiogenesis, the process by which new blood vessels arise from a pre-existing vascular bed. Angiogenesis is essential for the formation of vasculature during embryonic and post-natal development (Conway et al., 2001). Angiogenesis also contributes to many pathological conditions, including tumour growth, ischaemic disease, diabetes, rheumatoid arthritis, atherosclerosis, and inflammatory processes. Therefore, the therapeutic correction of angiogenesis, either by promoting deficient angiogenesis or by preventing excessive angiogenesis, is a promising approach for the treatment of several human diseases.
Angiogenesis is regulated by a number of hormones, cytokines, growth factors and low-molecular-mass mediators (Carmeliet and Jain, 2000). One of these molecules, VEGF (vascular endothelial growth factor) plays a particularly important role. VEGF-A, -B, -C, -D, -E, -F and PlGF (placental growth factor) constitute a family of cytokines with a central role in the regulation of morphogenesis, specification and homoeostasis of vessels (Tammela et al., 2005). VEGF-A is an essential mediator of vasculogenesis and angiogenesis during development and in a variety of pathological situations in adults. VEGF-A exists as multiple isoforms (VEGF-A121, VEGF-A145, VEGF-A165, VEGF-A183, VEGF-A189 and VEGF-A206 in humans), of which VEGF-A165 is the most abundant and best characterized. The VEGF-A isoforms differ in their binding properties for proteoglycans and the extracellular matrix. As a consequence, they have different abilities to activate endothelial cells and induce structurally different vascularization patterns In vivo (Whitaker et al., 2001; Lee et al., 2005).
VEGF-A binds to two structurally related RTKs (receptor tyrosine kinases): VEGFR-1 [VEGF receptor-1; Flt-1 (fms-like tyrosine kinase 1)] and VEGFR-2 [KDR (kinase insert domain-containing receptor); Flk-1 (fetal liver kinase 1)] (Ferrara et al., 2003). Signalling through VEGFR-2 appears to mediate almost all the observed angiogenic effects of VEGF-A, including the stimulation of endothelial cell differentiation, proliferation, migration and morphogenesis. VEGFR-1 is thought to regulate the activity of VEGFR-2, either as an activator, via the transphosphorylation of VEGFR-2 (Carmeliet et al., 2001; Autiero et al., 2003), or as an inhibitor, by acting as a ‘decoy” receptor competing for VEGF-A binding (Park et al., 1994).
HGF [hepatocyte growth factor; also known as SF (scatter factor)] is a pleiotropic cytokine involved in many complex biological processes, from embryogenesis and tissue regeneration to tumour growth, metastasis and angiogenesis (Trusolino and Comoglio, 2002). HGF binds to and activates one known RTK, c-met (Birchmeier et al., 2003). c-met is expressed in several cell types, including epithelial, endothelial, neuronal and haematopoietic cells. In endothelial cells, c-met is up-regulated under angiogenic conditions (Ding et al., 2003). The activation of c-met following HGF binding triggers a number of signalling pathways, most of which are common to VEGF-A signalling (Rosario and Birchmeier, 2003). Consequently, HGF-induced angiogenic responses are similar to those induced by VEGF-A. VEGF-A and HGF, however, activate only partially overlapping subsets of target genes in endothelial cells (Gerritsen et al., 2003). Moreover, VEGF-A is a stimulator of endothelial permeability (Weis and Cheresh, 2005), whereas HGF enhances the endothelial barrier function (Liu et al., 2002; Singleton et al., 2007; Birukova et al., 2009).
HGF has been shown to promote VEGF-A-driven angiogenesis (Silvagno et al., 1995; Van Belle et al., 1998; Xin et al., 2001). The molecular mechanisms underlying this co-operation are not completely understood. HGF has been found to stimulate VEGF-A production in non-endothelial cells (Silvagno et al., 1995; Van Belle et al., 1998) and to increase the expression of VEGFRs and c-met in endothelial cells (Wojta et al., 1999; Gerritsen et al., 2003). A gene expression profiling study suggested that VEGF-A165 and HGF co-operate by inducing a number of pathways, leading to a more robust proliferative response (Gerritsen et al., 2003). To provide further insights into the molecular mechanisms underlying the VEGF-A and HGF co-operation, we analysed the effects of their combination on the activation of VEGFR-2 and c-met receptors, and on the stimulation of the downstream signalling pathways in endothelial cells.
VEGF-A and HGF synergistically induce the proliferation and chemotaxis of HUVECs (human umbilical vein endothelial cells)
In order to determine which endothelial cell responses are potentiated by combining VEGF-A and HGF, we first investigated the effect of VEGF-A165 and HGF alone, and in combination, on the proliferation of endothelial cells using a [3H]thymidine uptake assay. In our experimental conditions, both cytokines individually stimulated DNA synthesis in HUVECs (Figure 1A). The combination of VEGF-A165 and HGF, at a fixed ratio of 1:1 or 1:10, induced more DNA synthesis significantly in HUVECs (Figure 1A, shown for the ratio 1:1). To investigate whether VEGF-A165 and HGF act in an additive or synergistic manner, we used the PharmTools Pro software package (PharmSoft). This software provides a computational procedure for the analysis of drug combination data that is consistent with the isobolographic method (Tallarida, 2000). We compared experimentally obtained proliferation data to the effects that were expected for the combination of VEGF-A165 and HGF on the basis of the individual relative potencies of VEGF-A165 and HGF. Our results demonstrate that the experimental regression line for the combination was significantly different from the additive regression line, which was calculated from the individual dose—response curves (Figure 1A, inset). The experimental line was located above the calculated additive line, suggesting a synergistic interaction between VEGF-A165 and HGF in inducing DNA synthesis.
We then investigated the effect of the combination of VEGF-A165 and HGF on HUVEC migration. Using a Transwell™ assay, we showed that VEGF-A165 and HGF individually stimulated HUVEC chemotaxis (Figure 1B). Combining VEGF-A165 with HGF further increased this effect. The analysis of the dose—response data using the PharmTools Pro software package indicated that there was a statistically significant synergy between VEGF-A165 and HGF in inducing HUVEC chemotaxis (Figure 1B, inset).
Similarly, HUVEC migration in an in vitro woundhealing assay was stimulated more strongly by the combination of VEGF-A165 and HGF than by either factor individually (Figure 1C).
Finally, using an annexin V assay, we showed that the combination of VEGF-A165 and HGF promoted endothelial cell survival in, at least, an additive manner (Figure 1D).
VEGFR-2 and c-met do not transphosphorylate each other and their levels are regulated by both VEGF-A165 and HGF
We observed no cross tyrosine phosphorylation between VEGFR-2 and c-met (Figure 2A), or any physical association between c-met and VEGFR-2 in coimmunoprecipitation experiments (data not shown).
We then analysed levels of VEGFR-2 and c-met after the incubation of HUVECs with VEGF-A165 or HGF, or both. Immunoprecipitation and Western-blot experiments demonstrated that in HUVECs incubated with VEGF-A165, VEGFR-2 was downregulated following ligand-dependent internalization and degradation, which was prevented by the pre-treatment of HUVECs with chloroquine, an inhibitor of lysosomal proteases (Figure 2B). At 1 h of incubation, HGF did not prevent the VEGF-A165-mediated VEGFR-2 down-regulation (Figure 2B). After 8 h, HGF alone up-regulated VEGFR-2 protein (Figure 2C). Accordingly, the VEGFR-2 protein was not decreased in HUVECs treated with a combination of VEGF-A165 and HGF, suggesting the de novo synthesis of VEGFR-2 protein. Consistent with this, the treatment of HUVECs with a combination of VEGF-A165 and HGF increased the levels of VEGFR-2 mRNA, as well as the levels of c-met mRNA (Figure 2D).
Similar to VEGFR-2, c-met was also down-regulated in the presence of its ligand, HGF (Figure 2C). VEGF-A165 alone increased the level of c-met, but this effect was not strong enough to compensate for the HGF-dependent c-met internalization.
Finally, using an ELISA-based approach we demonstrated that HUVECs did not produce VEGF-A or HGF under quiescent conditions or after stimulation with VEGF-A165 or HGF, or both. Consistent with this, we detected no VEGF-A and HGF mRNA in HUVECs cultured under any of the conditions tested (data not shown).
VEGF-A and HGF co-operate to induce intracellular signalling
To investigate which signalling pathways were responsible for the co-operative action of VEGF-A and HGF, we analysed the phosphorylation of MAPKs (mitogen-activated protein kinases) and other serine/threonine kinases in HUVECs stimulated with VEGF-A or HGF, or both. Using the phospho-MAPK array we found that VEGF-A165 and HGF activated a similar set of signalling kinases, including ERK1/2 (extracellular-signal-regulated kinase 1/2), p38 kinase and Akt. In contrast, we did not observe the activation of JNK (c-Jun N-terminal kinase), GSK-3α/β (glycogen synthase kinase-3α/β), RSK (ribosomal S6 kinase), MSK (mitogen- and stress-activated kinase) or p70 S6 kinase by the individual factors or the factors in combination (data not shown).
We next studied the kinetics of activation of signalling kinases in HUVECs treated with VEGF-A165, HGF, or both (Supplementary Figure S2 http:www.biolcell.orgboc101boc1010525add.htm). We found that HGF activated strongly ERK1/2 and Akt pathways, whereas VEGF-A165 treatment resulted in higher phosphorylation of p38 kinase and HSP27 (heat-shock protein 27). For both VEGF-A165 and HGF, the maximal stimulation was observed between 5 and 15 min after stimulation. Studies of the dose—response relationships confirmed that HGF was a more potent activator of the ERK1/2 and Akt (127- and 12-fold stimulation respecively compared with 26- and 3-fold stimulation respectively for VEGF-A165), whereas VEGF-A165 was a more potent activator of p38 kinase and HSP27 (Figure 3 and Supplementary Figure S3 http:www.biolcell.orgboc101boc1010525add.htm). Treatment with VEGF-A165 and HGF together (1:1 ratio) resulted in a greater activation of ERK1/2 and p38 kinase, but not of Akt and HSP27. Despite the fact that combining VEGF-A165 with HGF produced only a moderate increase in ERK1/2 and p38 kinase phosphorylation (compared with HGF or VEGF-A165 alone), the statistical comparison using the PharmTools Pro software demonstrated that there was a statistically significant synergistic interaction between VEGF-A165 and HGF in the activation of ERK1/2 and p38 (Figure 3, insets). We could not analyse the activation of HSP27, because the treatment with HGF alone did not generate a dose—response curve (Figure 3).
Similarly, we found a statistically significant synergistic interaction between HGF and VEGF-A121 in the activation of ERK1/2 and p38 kinase, but not for the activation of Akt (Supplementary Figure S4 http:www.biolcell.orgboc101boc1010525add.htm).
VEGF-A165 and HGF have different effects on the kinetics of FAK (focal adhesion kinase) phosphorylation and the localization of phosphorylated FAK in HUVECs
We found that treatment with either VEGF-A165 or HGF increased the basal phosphorylation of FAK, a key signalling molecule that regulates cellular adhesion (Schlaepfer et al., 2004). VEGF-A165, however, induced a longer lasting FAK phosphorylation, which persisted until at least 1 h after stimulation (Figure 4A). In contrast, HGF induced a longer lasting activation of ERK1/2.
Immunocytochemical experiments demonstrated that treatment with either VEGF-A165 or HGF increased the recruitment of activated FAK to focal adhesions (Figure 4B). However, the localization and morphology of these focal adhesions was different. In VEGF-A165-stimulated HUVECs, we observed phosphorylated FAK within elongated structures at the ends of stress fibres (Figure 4B, P-FAK panels, insets). In HGF-treated HUVECs, however, phosphorylated FAK was found in smaller structures, mostly at the periphery of the cell near the protruding lamellipodia. Co-stimulation of HUVECs with VEGF-A165 and HGF induced the recruitment of phosphorylated FAK to additional contacts on the ventral surface of the cells (Figures 4B, open arrowheads, and 4C).
VEGF-A165 and HGF control distinct morphogenic aspects of the cytoskeletal remodelling associated with selective activation of Rho or Rac
Because many cellular responses depend on the remodelling of the actin cytoskeleton, we examined the changes to the actin cytoskeleton induced by VEGF-A165 or HGF, or both. We found that VEGF-A165 induced the formation of stress fibres, running along the major cell axis. These fibres were observed as early as 5 min after treatment, were most abundant after 30 min (Figure 4B, closed arrows) and persisted until at least 1 h after stimulation (data not shown). In contrast, HGF increased the number of cortical actin bundles (Figure 4B, open arrows) and induced the formation of actin-rich peripheral lamellipodia (Figure 4B, closed arrowheads). The treatment of HUVECs with the combination of VEGF-A165 and HGF induced a mixed phenotype, with stress fibres, cortical bundles and peripheral lamellipodia all being formed (Figure 4B).
We next used GST (glutathione transferase)–rhotekin and GST—PAK-1 (p21-activated kinase-1) pull-down assays to compare the effects of VEGF-A165 and HGF on the activities of Rho and Rac1, two small GTPases that are involved in the regulation of stress fibre formation and in membrane ruffling and lamellipodium extension respectively (Nobes and Hall, 1995). Consistent with our immunohistochemical observations, we found that VEGF-A165 stimulation of HUVECs selectively increased Rho activity, whereas Rac1 activity was slightly decreased (Figure 5A). In contrast, HGF increased Rac1 activity and had no effect on Rho activity.
Similar effects of VEGF-A165 and HGF on the cytoskeletal remodelling and Rho and Rac1 activation were observed in HDMECs (human dermal microvascular endothelial cells). As seen in HUVECs, treatment of these cells with VEGF-A165 increased Rho activity and stimulated the formation of stress fibres, whereas HGF activated Rac1 and induced the extension of lamellipodia (Supplementary Figure S5 http:www.biolcell.orgboc101boc1010525add.htm).
VEGF-A165 and HGF induce structurally distinct vascular-like patterns in vitro in a Rho- or Rac-dependent manner
We used Y-27632, a Rho kinase inhibitor, and NSC-23766, a specific inhibitor of Rac1, to analyse the involvement of Rho and Rac1 in regulating the cellular responses of HUVECs to VEGF-A165 or HGF, or both. The inhibition of Rho activity with Y-27632 at concentrations ranging from 0.2 to 10 μM had no effect on HUVEC proliferation (Supplementary Figure S6A http:www.biolcell.orgboc101boc1010525add.htm) or migration (Supplementary Figure S6B); however, these concentrations are high enough to completely prevent the Rho-dependent formation of stress fibres (Ishizaki et al., 2000). In contrast, the inhibition of Rac1 with NSC-23766 reduced, in a dose-dependent manner, both basal HUVEC proliferation, and that induced with VEGF-A165 or HGF, or both (Supplementary Figure S6A). NSC-23766 at a concentration of 100 μM, previously established to be an effective concentration in NIH 3T3 cells (Gao et al., 2004), significantly decreased the basal and growth-factor-induced HUVEC migration (data not shown). At 10 μM, NSC-23766 had no effect on the basal HUVEC migration, but significantly inhibited HUVEC migration induced by HGF (Supplementary Figure S6B).
We then used an in vitro Matrigel™ assay to evaluate the role of Rho and Rac in the VEGF-A165- and HGF-dependent formation of vascular-like structures. Three-dimensional Matrigel™ induces cultured endothelial cells to form a capillary-like network that is similar to the process occurring In vivo. This process involves the transition of the cell morphology to a spindle-shaped appearance, the alignment of cells into multicellular cord-like structures and the interconnection of cords into a polygonal network. The culture of HUVECs on Matrigel™ in the presence of VEGF-A165 or HGF results in structurally distinct vascular-like patterns (Figure 5B). Thus, VEGF-A165 induced a less-branched network with large interconnections, whereas HGF stimulated the formation of a thinner and more branched network. These effects can be quantified by measuring tubular length and width and the number of branch points. HGF increased tubular length and number of branch points to a greater extent compared with VEGF-A165, whereas VEGF-A165 slightly increased tubular width (Figure 5C). When HUVECs were cultured on Matrigel™ in the presence of both VEGF-A165 and HGF, a network with a mixed phenotype was formed. This network contained enlarged cords and an increased number of branch points (Figures 5B and 5C). The inhibition of Rho resulted in a disruption of the tubular network induced by VEGF-A165 and a formation of HGF-like tubular patterns in the presence of combination of VEGF-A165 and HGF (Figures 5B and 5C). Conversely, the inhibition of Rac1 resulted in a disruption of the HGF-induced tubular network and a formation of VEGF-A165-like tubular patterns in the presence of both VEGF-A165 and HGF.
Similarly, using a three-dimensional collagen assay we demonstrated that the combination of VEGF-A165 and HGF induced a more completed branched network compared with VEGF-A165 or HGF alone (Supplementary Figure S7 http:www.biolcell.orgboc101boc1010525add.htm). Inhibiting Rho or Rac prevented the tube formation in the presence of VEGF-A165 or HGF respectively, and led to the establishment of the HGF-like or VEGF-A165-like tubular patterns respectively when VEGF-A165 and HGF were added in combination.
Angiogenesis involves the proliferation, migration and differentiation of vascular endothelial cells. VEGF is a key regulator of angiogenesis both during development and in adults. VEGF and VEGFRs have long been considered promising therapeutic targets for inhibiting and stimulating angiogenesis. However, clinical trials using VEGF-A165 in the treatment of ischaemic disease have revealed the limitations of such a unifactorial approach (Carmeliet, 2000; Simons and Ware, 2003). It is now apparent that no single factor alone can induce the formation of structurally and functionally normal blood vessels; rather, combinations of complementary growth factors are required for functional revascularization. The combination of VEGF-A165 and PlGF have been analysed extensively (Carmeliet et al., 2001; Autiero et al., 2003). Several reports have also demonstrated an improved angiogenic response induced by the combinations of VEGF-A165 and HGF (Van Belle et al., 1998; Xin et al., 2001). However, the molecular mechanisms underlying the VEGF-A/HGF co-operation have not been entirely elucidated.
In the present study, we have demonstrated that the combination of VEGF-A165 and HGF synergistically stimulated endothelial cell proliferation and chemotaxis. This implies an interaction between the VEGF-A165 and HGF receptors that increases the biological outputs triggered by ligand binding. We found, however, that c-met and VEGFR-2 do not physically associate and do not transphosphorylate each other. In contrast, VEGF-A165 and HGF are capable of increasing the c-met and VEGFR-2 protein levels respectively in HUVECs. This could result from the stimulation of the transcription of the corresponding genes. Indeed, in accordance with the previously reported data, we demonstrated that both VEGF-A165 and HGF slightly increased the c-met and VEGFR-2 mRNA levels in HUVECs. The receptor protein levels also depend on the signalling mechanisms that regulate the ligand-dependent receptor turnover. Some of these mechanisms are common to VEGFR-2 and c-met (Duval et al., 2003; Hammond et al., 2004; Murdaca et al., 2004). We showed, however, that the combination of VEGF-A165 and HGF had no effect on the c-met protein levels, compared with the treatment with HGF alone. The VEGFR-2 protein levels increased in HUVECs treated with the combination of VEGF-A165 and HGF, but only at the later time points. During the early hours of incubation, HGF alone did not protect from the VEGF-A165-dependent VEGFR-2 down-regulation. Thus our results indicate that neither VEGF-A165 nor HGF interferes with the signalling mechanisms involved in ligand-dependent receptor internalization and degradation.
The interaction between VEGF-A/VEGFR-2 and HGF/c-met signalling systems could be mediated by neuropilins that bind both VEGF-A and HGF and function as co-receptors, potentiating the pro-angiogenic activity (Sulpice et al., 2008). VEGF-A and HGF may influence the signalling induced both by VEGFR-2 and c-met through the recruitment of neuropilins which results in a more efficient activation of downstream adaptors and effectors. Indeed, we report in the present study that VEGF-A165 and HGF stimulate a similar set of MAPKs, although the kinetics and strengths of the activation differ depending on the growth factor and pathway. An enhanced activation of the signalling pathways was observed when HUVECs were treated with a combination of VEGF-A165 and HGF. Moreover, combining VEGF-A with HGF resulted in a statistically significant synergistic activation of the ERK1/2 and p38 kinase pathways.
We demonstrated that VEGF-A165 and HGF regulated, in different ways, the signalling molecules involved in the regulation of the cytoskeleton and cellular adhesion. Thus we showed that VEGF-A165 and HGF activated FAK with different kinetics and induced the recruitment of phosphorylated FAK to different subsets of focal adhesions. Most importantly, we demonstrated that VEGF-A165 and HGF regulate distinct morphogenic aspects of the cytoskeletal remodelling: VEGF-A165 induced the formation of stress fibres, whereas HGF induced the formation of actin-rich peripheral lamellipodia in HUVECs and HDMECs. Interestingly, we also found that VEGF-A165, but not HGF, stimulated the phosphorylation of HSP27; this protein has been implicated in the stabilization of actin filaments (Garrido et al., 2003).
The effects of VEGF-A165 and HGF on the cytoskeleton correlated well with our observation of the preferential activation of the Rho GTPase by VEGF-A165 and Rac by HGF. Our results are consistent with previously reported data demonstrating the involvement of the Rho pathway in the VEGF-A165-induced effects in vitro and In vivo (van Nieuw Amerongen et al., 2003; Hoang et al., 2004). Additionally, Birukova et al. (2009) reported similar effects of VEGF-A165 and HGF on the cytoskeleton and Rho and Rac activities in pulmonary artery endothelial cells, another endothelial cell type.
Rho and Rac are members of the Rho family of GTPases and have been extensively characterized as major regulators of the polymerization and organization of actin filaments, and consequently of cell polarity, motility, division and survival (Jaffe and Hall, 2005). They often have an antagonistic effect on cellular functions and are found in discrete subcellular locations (Nobes and Hall, 1999; Pertz and Hahn, 2004). In the cardiovascular system, Rho and Rac are the key regulators that act antagonistically to control endothelial barrier function, with Rho enhancing endothelial permeability and Rac counteracting the effects of Rho (Wojciak-Stothard and Ridley, 2002). Rho GTPases also control branching morphogenesis, the complex programme that leads to the formation of a network of tubes by endothelial cells (Lubarsky and Krasnow, 2003). We demonstrate in the present study that VEGF-A165 and HGF stimulate the organization of endothelial cells into structurally distinct capillary-like networks and that this process is dependent on selective activation of Rho and Rac by VEGF-A165 and HGF respectively. The inhibition of Rho results in the switch from VEGF-A165-like patterns to HGF-like patterns. Conversely, the inhibition of Rac results in the formation of a VEGF-A165-like phenotype. The vascular-like patterns that we observed in HUVECs stimulated with VEGF-A165 are consistent with the capacity of Rho to induce the more elongated cell morphology with actin stress fibres along the main axis of the cell and stimulation of lamellar protrusions on the axial ends (Pankov et al., 2005). These elongated cells aligned locally form thick and extended cords, with few transversal interconnections. In contrast, Rac favours the formation of peripheral lamellae, which become dominant elsewhere on the cell (Pankov et al., 2005). This could explain the more branched phenotype of the capillary-like network induced by HGF in a Rac-dependent manner.
We found no evidence that the treatment of HUVECs with Y-27632, an inhibitor of Rho GTPase, affected cell proliferation or migration. These results are consistent with the previously reported data that details the involvement of Rho essentially in regulation of the organization of endothelial cells in vascular network in vitro and In vivo (Hoang et al., 2004). In contrast, the treatment of HUVECs with high concentrations of Rac1 inhibitor (NSC-23766) suppressed both the basal and growth-factor-induced proliferation and migration HUVECs. At low concentrations, NSC-23766 inhibited only the HGF-dependent HUVEC migration. Our observations indicate that basal Rac activity is essential for proliferation and migration of HUVECs, and the HGF-induced Rac activity contributes to the endothelial cell migration.
In conclusion, our results indicate that, under angiogenic conditions, combining VEGF-A with HGF will promote neovascularization by enhancing intracellular signalling and increasing the proliferation, migration and survival of endothelial cells (Figure 6). Combining VEGF-A with HGF may also allow for the more finely regulated control of the signalling molecules involved in the regulation of the cytoskeleton and cellular adhesion, particularly Rho and Rac. It has been suggested that during angiogenesis, an increase in Rho activity together with the modulation of Rac may provide a cumulative improvement in both angio-architecture and barrier function (Nagy and Senger, 2006). Our results show that the combination of VEGF-A and HGF induces a more branched and strengthened vascular-like network. Together with the recent findings by Birukova et al. (2009), this suggests that combining VEGF-A with HGF is a promising approach for therapeutic angiogenesis that could facilitate the establishment of a stable and functional vasculature with improved barrier function.
The umbilical cord specimens were obtained from the Obstetrics Department, Lariboisère Hospital (Paris, France). HUVECs were isolated and cultured as described previously (Ding et al., 2003). Approval was obtained from the Local Ethical Committee for the present study. HDMECs were purchased from Promocell (Heidelberg, Germany) and cultured according to the manufacturer's instructions.
[3H]Thymidine incorporation assay
HUVECs were plated in 24-well plates at a density of 2×104 cells/well and starved for 24 h in serum-free medium. Cells were then incubated with various concentrations of growth factors for a further 24 h, followed by an incubation with [3H]thymidine (37 κBq/well) for the last 24 h. The incorporated radioactivity was measured as described previously (Ding et al., 2003).
HUVECs stained with CellTracker™ CM-DiI (Invitrogen) were seeded on to the upper side of FluoroBlok inserts (8 μm pore size; BD Biosciences), at 5×104 cells per insert. The inserts were placed in 24-well plates containing medium supplemented with various concentrations of growth factors and incubated for 6 h at 37°C, in an atmosphere containing 5% CO2. The number of cells that had migrated to the lower surface of the filters was evaluated by measuring the fluorescence using a Victor3 spectrofluorimeter (PerkinElmer). Each experiment (n=3) was performed in triplicate.
HUVECs were grown to 100% confluence in six-well plates and then incubated for 20 h in serum-free EBM-2 (endothelial basal medium 2). A linear wound was made by scratching the monolayer with a plastic tip. The wounded monolayers were incubated in EBM-2 medium with or without growth factors. After 16 h, cells were fixed with 4% (w/v) paraformaldehyde and stained with Harris haematoxylin. Cells were observed with an Axiovert 25 microscope equipped with A-Plan 10×/0.25 Ph1 objective (Zeiss). Migration was quantified by counting the number of cells migrating into the denuded area in each randomly chosen microscope field. Five fields were counted at each experimental point, and each point was performed in duplicates.
HUVECs were grown in six-well culture plates until 60% confluence and apoptosis was then induced by serum deprivation. Complete growth medium was replaced by M199 medium containing 1% foetal bovine serum (starvation medium) and the cells were incubated for 40 h in the presence or absence of VEGF-A165 or HGF, or both. Apoptosis was then assessed using the annexin V—FITC Apoptosis Detection Kit (BD Biosciences) and a FACSCalibur analysis system (Becton Dickinson), according to the manufacturer's instructions.
in vitro Matrigel™ assay
HUVEC were incubated for 24 h in starvation medium, stained with CM-DiI and seeded (2×104 cells/well) on the top of a pre-formed Matrigel™ gel in 48-well plate. Capillary-like tube formation was induced by the addition of 100 ng/ml of VEGF-A165 or HGF, or both, in the presence or absence of 10 μM Y-27632 or 100 μM NSC-23766. After a 16 h incubation, cultures were observed using an Axio Observer Z1 inverted microscope equipped with a Plan-Neofluar 2.5×/0.075 objective (Carl Zeiss MicroImaging). Images were taken with a KY-F75U digital camera (JVC) and quantified using Histolab software (Microvision Instruments).
Data are presented as the means±S.E.M. Statistical analyses were performed in Excel using Student's t test. P<0.05 was considered statistically significant. PharmTools Pro software was used to perform the VEGF-A and HGF combination analysis (Tallarida, 2000). Dose—response data were obtained for VEGF-A, HGF and the fixed-ratio (1:1 and 1:10) combinations of VEGF-A and HGF and were examined by regression analysis. The individual dose—response data sets were used to calculate the theoretical composite regression line that was then compared with the experimentally derived regression line of the combination using PharmTools Pro software package. A significant difference between the regressions was interpreted as a departure from simple additivity.
We thank the Department of Obstetrics, Lariboisière Hospital, Paris, France, for providing umbilical cord specimens.
This work was supported by the Association pour la Recherche sur le Cancer [grant number CL 3124]; and the Agence Nationale de la Recherche [grant number COD-022]. S.D. held a fellowship from the Association Franco-Chinoise pour la Recherche Scientifique and Technique.