Angiogenesis is a normal process involved in development, reproduction, and wound healing, as new blood vessels are formed from the pre-existing vasculature. Despite being a beneficial event under certain circumstances, angiogenesis is also a major contributing factor to several diseases including, rheumatoid arthritis, cancer, and ocular diseases such as diabetic retinopathy (reviewed in [1–3]). Angiogenesis is a multi-step event that requires growth factor stimulation of endothelial cells, resulting in cellular proliferation, migration, tube formation, and finally stabilization of the new vessels. As angiogenesis only initiates following angiogenic growth factor stimulation, many strategies that target primary angiogenic factors, such as vascular endothelial cell growth factor (VEGF) or its angiogenic receptor (VEGFR) have been developed and are at various stages of clinical testing. Although these types of anti-angiogenic therapies have shown some ability to control disease in certain settings, recent studies have highlighted an increased risk of severe side effects with the widely used anti-angiogenic, Bevacizumab (reviewed in [4–6]). For reasons such as this, the discovery of novel mechanisms controlling angiogenesis is necessary so that new therapeutic targets can be identified.

RhoB is a member of the Ras superfamily of GTPases, which includes proteins such as Rac1, Cdc42, RhoA, and RhoC. Rho family proteins (with the exception of RhoE) are GTPases that function by cycling through a GTP-bound activated state and a GDP-bound inactive state. Regulation of these states is achieved through GTPase-activating proteins (GAPs), guanine nucleotide exchange factors (GEFs), and guanine nucleotide dissociation inhibitors (GDIs) [7]. RhoB shares ~80% homology with its closely related family members RhoA and RhoC, however its subcellular localization was found to be very different, with almost exclusive localization to the cytosolic face of early endosomes and pre-lysosomal compartments [8]. This suggested a role in receptor trafficking, and indeed RhoB has been shown to control trafficking of a number of growth factor receptors including platelet derived growth factor receptor (PDGFR) [9], and epidermal growth factor receptor (EGFR) [10]. RhoB can contribute to growth factor receptor signaling, as it has been shown to be required for the PDGFR-driven migration of vascular smooth muscle cells via its ability to activate and traffic endosome bound Cdc42 to the cell periphery [9]. RhoB has also been shown to regulate whether EGF-bound EGFR remains in early endosomes or is transported to late endosomes for degradation, and in this manner can control duration of receptor signaling [10]. Thus, as a regulator of growth factor receptor activity, RhoB may play a significant role in mediating growth factor-induced angiogenesis.

Given its expanding role in regulating signals from angiogenic growth factor receptors, we were interested in examining the effect of RhoB on various angiogenic processes in general, and on its ability to modulate angiogenic processes induced by the primary disease-associated angiogenic factor VEGF. We hypothesized that RhoB would be required for VEGF-induced capillary morphogenesis and that the absence of RhoB would result in impaired angiogenic activities in endothelial cells. We found that VEGF stimulation upregulated expression of RhoB in endothelial cells. We also observed that although the absence of RhoB did not affect endothelial cell viability, RhoB was critically important for VEGF-induced endothelial cell migration and sprout formation. We further show that lack of RhoB in endothelial cells results in upregulation of RhoA activity, and that suppression of this activity or the activity of Rho-associated kinase (ROCK) restored VEGF-induced endothelial cell capillary morphogenesis in the absence of RhoB. We thus conclude that RhoB is required to control RhoA activity in response to VEGF stimulation to allow organization of endothelial cells during endothelial cell sprouting and capillary morphogenesis.

RhoB has been shown to play a role in growth factor receptor trafficking and through this mechanism can regulate growth factor receptor signaling under certain circumstances [9, 10]. With this in mind, we became interested in determining whether RhoB regulated VEGF-induced angiogenic processes in endothelial cells, in order to identify possible novel targets which might ultimately be useful for enhancing the efficacy of current anti-VEGF/VEGFR blocking strategies. We thus used small interfering (si)RNA silencing strategies in human umbilical vein endothelial cells (HUVEC) to determine the effects of reduced RhoB expression on the ability of VEGF to induce endothelial cell proliferation or morphogenesis, and by what potential mechanisms RhoB may regulate these angiogenic processes.

VEGF upregulates expression of RhoB

Our initial studies focused on characterizing the expression of RhoB in HUVEC as a model cell system. We found that HUVEC expressed readily detectable levels of endogenous RhoB (Figure 1A). Moreover, VEGF treatment of HUVEC resulted in increased RhoB protein levels within 8 h post-stimulation, with protein reaching maximal expression around 12 h post-VEGF stimulation (Figure 1A). To determine if increased RhoB protein levels were associated with concomitant increases in RhoB activity in these cells, we utilized the G-LISA™ Activation Assay designed for the detection of activated RhoA. This assay utilizes the Rho effector protein Rhoteckin for binding activated RhoA, and as Rhoteckin also binds RhoB, albeit much less efficiently [11], we utilized the same G-LISA™ assay but modified it for detection of the pulled-down active RhoB by using a RhoB specific monoclonal antibody for detection. Using this modified assay, we found that HUVEC do not show increased levels of active RhoB following VEGF stimulation (Figure 1B), even at time points where increased levels of RhoB protein are observed (Figure 1A).

Depletion of RhoB levels in HUVEC inhibits cell migration but does not affect cell viability

In order to assess the contribution of the small GTPase RhoB to processes important to angiogenesis, we employed a siRNA strategy to specifically deplete levels of RhoB in HUVEC. Two siRNAs directed to RhoB were validated for sequence specificity and used at 20 nM concentration to effectively reduce protein levels of RhoB in HUVEC. When compared to mock transfected and control siRNA transfected cells, there was an evident reduction in RhoB protein levels, as detected by western blot analysis, at 48 h post-siRNA transfection (Figure 2A).

As cell survival and migration are both important requirements for angiogenesis to occur, we next examined whether depleting RhoB affected either of these two processes. RhoB levels were thus depleted using two independent specific RhoB targeting siRNA constructs, and HUVEC growth and cell viability was examined over time following quantification of viable cell numbers using trypan blue exclusion. We observed no significant difference in HUVEC growth or viability between RhoB-depleted and control siRNA treated cells (Figure 2B). Cell migration was also assessed using a scratch wound assay. HUVEC were transfected with either RhoB-targeted siRNA or non-targeting control siRNA at concentrations of 20 nM and subsequently, the confluent HUVEC monolayer was 'scratched' and photographed to determine wound diameter at time 0. Media was then replaced with MCDB 131 minimal media with 0.05% FBS and supplemented with 50 ng/ml VEGF, thus making migration of HUVEC essentially VEGF-dependent. Cells were then allowed to migrate and fill the wound over the course of 24 h, at which time wound diameters were re-photographed and the percent wound closure in each condition was determined. When assessed in this manner, reduced levels of RhoB resulted in significant inhibition of cell migration as indicated by decreased percent wound closure after 24 h as compared to control siRNA transfected cells (Figure 2C and 2D). Taken together, our data suggest that RhoB plays an important role in modulating VEGF-induced cell migration signals, while appearing to be dispensable for VEGF-induced proliferative signals in endothelial cells.

siRNA-mediated depletion of RhoB reduces HUVEC capillary sprouting and morphogenesis

To assess the importance of RhoB to HUVEC luminal vessel-like formation we first utilized a collagen gel-based assay. In this assay cells are placed onto a collagen I matrix and induced to sprout with VEGF, resulting in polarized vessel-like structures that contain lumen. RhoB was silenced in HUVECs using the targeted siRNAs, and 24 h later transfected HUVEC were plated on collagen I gels where they were subsequently stimulated with 50 ng/ml VEGF in EGM-2 growth medium. Sprout structures were then counted over a period of 10 days. We observed a statistically significant reduction in the number of vessel structures generated by RhoB siRNA-treated HUVEC when compared to cells treated with non-targeting control siRNA or mock transfected cells in response to VEGF stimulation (Figure 3A). We also assessed the ability of RhoB-depleted cells to form capillary-like networks on basement membrane extract (BME). In this assay, which is commonly used to test the angiogenic potential of endothelial cells, cells will usually elongate and align to form a network of cord structures that are devoid of lumens. When these cord structures were quantified, RhoB appeared to be required for HUVEC capillary morphogenesis in this assay, with HUVEC depleted of RhoB showing significant reduction in the number of cord structures formed as compared to control transfected cells (Figure 3B). It should be noted however, that the cord structures that did form in RhoB-depleted cells were similar in morphology to those observed in control-treated cells (Figure 4D), and could thus have formed as a result of incomplete RhoB depletion in 100% of cells.

Pharmacological inhibition of rho-dependent kinases ROCK I/II in RhoB-depleted HUVEC partially restores capillary morphogenesis

To further support our previous findings with the C3 transferase Rho inhibitor, we examined whether targeting pathways activated downstream of RhoA could also restore the vessel formation defects observed in RhoB-depleted cells. As the Rho-dependent kinases ROCK I and ROCK II coordinate signaling events downstream of RhoA, we targeted this pathway as a potential signaling mechanism contributing to the impaired capillary morphogenesis in RhoB-depleted cells. We thus plated control or RhoB siRNA-treated HUVEC on BME in the presence of vehicle control or two different inhibitors of ROCK I/II activity, namely Y-27632 and H-1152. Similar to what had been observed following use of C3 transferase, addition of either ROCK I/II inhibitor to control siRNA transfected cells resulted in slightly enhanced cord forming ability of HUVEC as compared to vehicle control (Figure 5A and 5C for Y-27632; Figure 5B and 5D for H-1152). Again, similar to what we had previously observed following treatment with C3 transferase, treatment of RhoB-depleted HUVEC with either ROCK inhibitor, restored cord formation in cells treated with RhoB siRNA 1 essentially to that of control siRNA levels (Figure 5C and 5D). Cord formation was also restored in RhoB siRNA 2 treated cells by the addition of ROCK inhibitors where the impairment in cord formation was reduced to only approximately 32% compared to control cells in contrast to the approximately 43% impairment in cord formation observed in the absence of ROCK inhibitors (Figure 5C and 5D). These results lend support to the notion that an inappropriately activated RhoA/ROCK pathway contributes to the observed defective capillary morphogenesis phenotype in RhoB-depleted endothelial cells.

The contribution of RhoA protein signaling to processes that are important for angiogenesis, such as proliferation, migration, capillary morphogenesis and sprouting, has been previously identified (reviewed in [17]), however the contributions of the Rho family member RhoB remain less clear. Indications from knockout murine models that RhoB may modulate vessel sprouting in the retina of these mice [18], along with its proposed role in the trafficking and signaling of growth factor receptors [9, 10, 19, 20], suggested to us that RhoB may play an important role in pathological angiogenesis directed by VEGF/VEGFR signaling. As such, RhoB could potentially prove to be an important beneficial therapeutic target for controlling pathological angiogenesis. Although some evidence suggested that RhoB could regulate endothelial cell sprouting [18], its role in growth factor induced angiogenesis was not thoroughly examined. With this in mind, the present study aimed to determine if RhoB was necessary for processes involved in VEGF-induced capillary morphogenesis, and by what mechanisms RhoB controls these events.

Our first objective was to determine whether VEGF had any direct effects on the regulation of RhoB, as certain other growth factors such as transforming growth factor β (TGF-β), and EGF have been identified as regulators of RhoB expression in a number of human cell lines [21–23]. Interestingly, we observed increased expression of RhoB protein in HUVEC following stimulation with VEGF. Results indicated a rise in protein levels between 4 and 8 h post stimulation with levels peaking at 12 h post stimulation and slowly decreasing thereafter. The mechanism leading to increased RhoB expression after VEGF treatment in HUVEC is presently unknown, and the exact implications of raising RhoB levels in VEGF stimulated HUVEC are not understood. However, it is likely that increasing RhoB expression is necessary for proper response of endothelial cells to VEGF. Indeed, transcriptional induction of the RhoB gene is achieved by TGF-β in human keratinocytes, and depletion of RhoB through siRNA has been shown to impede TGF-β-induced migration [21], indicating the potential relevance of induced expression in that system. Our finding that RhoB expression is induced by VEGF in endothelial cells highlights RhoB as a potentially important regulator of VEGF signaling, thus warranting future mechanistic studies.

In order to assess the importance of RhoB in angiogenic processes, we employed a siRNA strategy to specifically deplete HUVEC cells of RhoB, and subsequently determined whether RhoB was necessary for endothelial cell survival, migration, sprouting or capillary morphogenesis. RhoB was found to be dispensable for endothelial cell survival, as depleting RhoB levels had no effect on cell growth or viability over time. With respect to endothelial cell migration, sprouting and capillary morphogenesis, we found that RhoB was required for VEGF induction of these processes. These findings are supported by work of others in transgenic mouse or in vitro models of angiogenesis [18, 24]. In contrast to the study by Sabatel et al. [24] where angiogenic activities were induced by a combination of basic fibroblast growth factor (bFGF) and VEGF together, our study focused specifically on VEGF-induced angiogenic processes. As such, our work supports a significant role for RhoB in modulating HUVEC migration and capillary morphogenesis in response to VEGF, a main mediator of angiogenesis in pathological settings.

Our results suggest that RhoB contributes to VEGF-induced endothelial cell capillary morphogenesis in part via its ability to negatively regulate RhoA. Historically, RhoA has been shown to be activated by VEGF in endothelial cells [15] and to contribute, along with other Rho family members, to the regulation of angiogenesis (reviewed in [17]). Our results now show that VEGF upregulation of RhoB plays a role in the negative regulation of RhoA activity, as when RhoB was absent, even low concentrations of VEGF induced substantial increases in RhoA activity, a phenomenon that did not occur when RhoB was present. These results thus suggest that RhoB may be important for limiting endothelial cell response to insignificant levels of VEGF that could otherwise lead to an inappropriately timed angiogenic response. Cross-regulation between Rho family members has been previously suggested. The family member Rac has been shown to regulate the activity of RhoA in fibroblasts, resulting in control of cell morphology and migration [13]. More recently, RhoA phosphorylation has been noted to release Rac from binding to RhoGDIalpha, resulting in translocation of Rac to the cell periphery followed by its activation [12]. Additionally, RhoB has been noted to traffic Cdc42 to the cell membrane in response to platelet-derived growth factor stimulation, and thus contribute to signaling necessary for cell movement [9]. Our data indicate that RhoB also negatively regulates the level of RhoA activation in response to VEGF. We further demonstrate that this inhibition of RhoA activity by RhoB is necessary for proper endothelial cell capillary morphogenesis. Moreover, activation of the RhoA/ROCK pathway has been shown to inhibit angiogenic processes [25, 26], thus lending support to our observations that the absence of RhoB results in impaired angiogenic activities in part via uncontrolled RhoA/ROCK activation.

Given our results suggesting that RhoB negatively regulated RhoA activity, we were also interested to determine whether RhoB had negative effects on the activity of other Rho family members. To this end, we evaluated the effects of RhoB depletion on the level of activity of RhoC in endothelial cells. We were intrigued to observe that in contrast to our results with RhoA activation, RhoC activity was somewhat reduced in the absence of RhoB. Thus together, our results suggest that RhoB regulates the activity of RhoA and RhoC in a reciprocal manner. Although studies with RhoC-null mice did not indicate any angiogenic defects associated with primary mammary tumors [27], a more recent study showed that treatment of human dermal microvascular cells with siRNA directed to RhoC, inhibited migration and tube formation, suggesting that RhoC activity may be necessary for angiogenesis under specific conditions [16]. As RhoC can also contribute to processes such tumorigenesis and metastasis [28, 29] in addition to angiogenesis [16], the RhoB regulation of RhoC activity could also be of significance in tumor growth and tumor-associated angiogenesis.

Interestingly, cross-regulation of RhoB by RhoA has been previously suggested, as it was shown that depletion of RhoA led to increased RhoGDIalpha binding to RhoB thereby resulting in RhoB protein stabilization [14]. Our study, however, demonstrates the reverse interaction, namely that RhoB negatively regulates the activation of RhoA to promote endothelial cell capillary morphogenesis. It is possible that regulation in our system is achieved via similar mechanisms as suggested by Ho et al. [14], whereby RhoA and RhoB compete for RhoGDIalpha binding, although this has not yet been demonstrated. Indeed, the importance of regulation among Rho family members mediated by binding to RhoGDIs is becoming increasingly evident, with recent studies cautioning that modulation of one Rho protein can affect others by shifting the balance of RhoGDI binding [30, 31]. Additionally, it is becoming more evident that Rho proteins are spatially and temporally regulated in regards to their activity [32–34]. Evidence of a short transcript half-life for RhoB [35] indicates a high degree of regulated expression, stressing that even the induction of minor changes in RhoB expression could lead to significant effects on cell signaling. Accordingly, it is possible that when RhoB is removed, a cell can no longer control the spatially regulated activation of RhoA, resulting in delocalization of RhoA-mediated signaling events required for directed cell migration and vessel formation.

Due to their close sequence homology, RhoA and RhoB are known to bind similar protein regulators and effectors. For example, the regulator XPLN, a GEF, has been shown to specifically interact with RhoA and RhoB, but not RhoC [36]. As regulators such as GEFs and GAPs undergo translocation in response to extracellular stimuli [37, 38], and in some cases to specific sites within cells occupied by their corresponding GTPase (reviewed in [7]), it is possible that competition between RhoB and RhoA for these regulators of activation is responsible for RhoB's control over RhoA activity levels in response to VEGF. Indeed, although not directly demonstrated, this has been hypothesized as a likely means of RhoB cross-regulation [39]. However, our data suggests that RhoB negatively regulates RhoA, but appears to have a positive regulatory function with respect to RhoC, suggesting that competition for activating factors or effector proteins is not the regulatory mechanism in place for this latter interaction. With this in mind, future studies will be directed towards better understanding the relationship between RhoA, RhoB, RhoC, and the various binding partners that may function to allow RhoB to regulate angiogenesis.

In conclusion, we have demonstrated that depletion of RhoB in HUVEC results in deleterious effects on processes important to angiogenesis, such as endothelial cell migration and capillary morphogenesis. These defects are, in part, due to inappropriately increased levels of activated RhoA following VEGF stimulation in the absence of RhoB. Lack of RhoC activity may also contribute to the observed defects. Based on these results we suggest a novel mechanism whereby RhoB exerts control over endothelial cell capillary morphogenesis through the negative regulation of RhoA and the positive regulation of RhoC activity in response to the pro-angiogenic growth factor VEGF.

¹Program for Cancer Therapeutics, Ottawa Hospital Research Institute, Ottawa, ON, Canada.

Departments of ²Medicine and ³Biochemistry, Microbiology and Immunology, University of Ottawa, Ottawa, ON, Canada.