RhoJ/TCL Regulates Endothelial Motility and Tube Formation and Modulates Actomyosin Contractility and Focal Adhesion Numbers
Objective—RhoJ/TCL was identified by our group as an endothelial-expressed Rho GTPase. The aim of this study was to determine its tissue distribution, subcellular localization, and function in endothelial migration and tube formation.
Methods and Results—Using in situ hybridization, RhoJ was localized to endothelial cells in a set of normal and cancerous tissues and in the vasculature of mouse embryos; endogenous RhoJ was localized to focal adhesions by immunofluorescence. The proangiogenic factor vascular endothelial growth factor activated RhoJ in endothelial cells. Using either small interfering (si)RNA-mediated knockdown of RhoJ expression or overexpression of constitutively active RhoJ (daRhoJ), RhoJ was found to positively regulate endothelial motility and tubule formation. Downregulating RhoJ expression increased focal adhesions and stress fibers in migrating cells, whereas daRhoJ overexpression resulted in the converse. RhoJ downregulation resulted in increased contraction of a collagen gel and increased phospho–myosin light chain, indicative of increased actomyosin contractility. Pharmacological inhibition of Rho-kinase (which phosphorylates myosin light chain) or nonmuscle myosin II reversed the defective tube formation and migration of RhoJ knockdown cells.
Conclusion—RhoJ is endothelial-expressed in vivo, activated by vascular endothelial growth factor, localizes to focal adhesions, regulates endothelial cell migration and tube formation, and modulates actomyosin contractility and focal adhesion numbers.
Endothelial cells line all blood vessels and play a critical role in the formation of new vessels, or angiogenesis. Angiogenesis is involved in physiological processes, such as the female menstrual cycle, but also plays a key role in pathologies, such as diabetic retinopathy and the growth and metastasis of solid tumors.1 During angiogenesis, endothelial cells perform a variety of functions, including degradation of the extracellular matrix, migration, proliferation, lumen formation, and vessel stabilization,2 and roles for the Rho GTPases RhoA, Cdc42, and Rac1 have been identified in many of these processes.3
Rho GTPases are molecular switches that cycle between active GTP-bound and inactive GDP-bound forms. They were initially identified as regulators of the actin cytoskeleton but have subsequently been shown to be involved in many cellular processes, such as vesicle trafficking, cytokinesis, and gene expression.4 Actin stress fibers are contractile components of the cytoskeleton that comprise bundles of actin and myosin, known as actomyosin. The small Rho GTPases, RhoA, -B, and -C, when activated, promote stress fiber formation via their activation of Rho-kinase (ROCKI/II, ROKβ/α); ROCK phosphorylates a number of targets, resulting in increased myosin light chain (MLC) phosphorylation and increased actomyosin contractility.5 Recently, it was demonstrated that established tubes have greater levels of phospho-MLC and that this is reduced on induction of sprouting.6 Indeed, inhibition of ROCK or MLC phosphorylation was accompanied by an induction of Rac1-dependent sprouting, suggesting that sprouting angiogenesis is both accompanied by and requires a downregulation of actomyosin contractility.7 Focal adhesions are large, dynamic protein complexes that mediate connections between the extracellular matrix and the intracellular cytoskeleton. They play a critical role in mechanotransduction, and their coordinated assembly and turnover are required for cell motility.8
Using a combination of bioinformatic and expression analyses using reverse transcription and quantitative polymerase chain reaction, our laboratory previously identified the small Rho GTPase RhoJ/TCL (TC10 like) as being highly and specifically expressed by endothelial cells.9 RhoJ was initially identified by 2 groups as being a novel Rho GTPase related to TC10 (RhoQ),10,11 belonging to the Cdc42 subfamily of Rho GTPases. Cdc42 has a critical role in filopodia formation,4 whereas TC10 mediates insulin-regulated metabolic events12; in contrast, overexpressing constitutively active or inactive forms of RhoJ modulated filamentous (f)-actin formation.11,13,14 A role for RhoJ in the in vitro differentiation of 3T3 L1 adipocytes has been demonstrated,15 and, like its close homolog TC10, increased levels of RhoJ-GTP were stimulated by insulin treatment.10 Subsequent work has shown that epitope-tagged, overexpressed RhoJ localized to early and recycling endosomes, and a role for RhoJ in regulating early endocytosis was identified.16
The aim of this study was to further explore the tissue distribution, investigate the activation, and determine the role of RhoJ in endothelial cells. We have identified a vascular expression pattern of RhoJ in both human tissue and mouse embryos and demonstrated that endogenous RhoJ localizes to focal adhesions. We have identified a critical role for this Rho GTPase in mediating endothelial cell motility and tube formation. We also demonstrate that RhoJ plays a role in modulating actomyosin contractility and focal adhesion numbers.
An expanded Methods section is available in the online Data Supplement at http://atvb.ahajournals.org.
Primary and immortalized cell lines used were human umbilical vein endothelial cells (HUVECs) (passage 1 to 6), human pericytes from placenta (hPC-PL), human aortic smooth muscle cells, HMEC-1 (human microvascular cell line-1), and HEK293T. For small interfering (si)RNA knockdown, HUVECs were transfected with 10 to 25 nmol/L negative control duplex (siControl), RhoJ siRNA duplex 1 (5′-CCACTGTGTTTGACCACTA-3′), or RhoJ siRNA duplex 2 (5′-AGAAACCTCTCACTTACGA-3′). Transfections were performed using either RNAiMAX Lipofectamine (Invitrogen) at a final concentration of 0.3% (vol/vol) or GeneFECTOR (Venn Nova Inc). RNAiMAX was used for all experiments except for the organotypic tube-forming assays and the phospho/total MLC Western blotting. The inhibitors used were Y27632 (10 μmol/L, ROCK inhibitor) or blebbistatin (5 μmol/L, nonmuscle myosin II inhibitor). For lentiviral production the plasmids psPAX2 and pMD2G were used with either pWPXL-GFP-daRhoJ(Q79L) or pWPI (green fluorescent protein [GFP]). GFP-positive HUVECs were sorted before performing assays. Changes in protein expression were assessed by Western blotting.
Scratch wounding of monolayers was achieved using a plastic pipette tip. 2D tube formation was assessed by plating cells on natural Matrigel. In the organotypic tube-forming assay, HUVECs were cocultured with confluent human dermal fibroblasts, and tubule formation was quantified using the AngioSys software.6,7 Contractility was assessed using cell-mediated contraction of a type I collagen gel.17 MLC2 phosphorylation was assessed by Western blotting for phospho-MLC2 and total MLC2.7
In situ hybridization was performed on human tissue sections, which were costained with digoxigenin-labeled riboprobes/antidigoxigenin–rhodamine combined with Ulex europaeus agglutinin I–fluorescein. The whole-mount in situ hybridization protocol was adapted from Piette et al18 and performed on embryonic day 9.5 mice using digoxigenin-labeled probes to mouse RhoJ or von Willebrand factor. Immunofluorescence was performed using anti-vinculin, anti–pFAK-Y397, purified rabbit anti-RhoJ antisera, or phalloidin-TRITC and imaged using confocal microscopy. Focal adhesions were manually counted. Levels of actin fluorescence were measured using the LSM 510 confocal software.
Details of experimental repetition are noted in the full and supplemental figure legends in the online Data Supplement; all assays were performed with at least 3 different HUVEC isolates. Data are plotted with error bars representing SEM or are displayed as box and whisker plots: these indicate the maximum, minimum, 25th and 75th percentile, and median values. To test statistical differences in defined pairs of groups, the Mann–Whitney test was performed except for the scratch-wound and gel-contraction assays, where the Student t test was used. Statistical significance is denoted as follows: ***P<0.001, **P=0.001 to 0.01, and *P=0.01 to 0.05.
RhoJ Is Expressed in Endothelial Cells
RhoJ expression was localized by in situ hybridization both on human tissue sections and in the developing mouse embryo. U europaeus agglutinin I was used as a marker of endothelium, because it recognizes an endothelial specific glycosylation,19 and this colocalized with RhoJ in endothelial cells lining vessels in a number of adult human tissues (heart, adrenal gland, lymph node, muscle, pancreas, placenta, liver, lung, bladder cancer, bone cancer, and ovarian cancer). However, expression was not found in the vessels of testis, brain, kidney, stomach, colon, or rectal cancer (Figure 1A and supplemental Figure I in the online Data Supplement). The RhoJ probe also stained some nonendothelial cells, including muscle, liver, and some cancer cells, suggesting that RhoJ expression is not restricted to the endothelium. Because this vessel-like expression pattern may be attributable to expression of RhoJ by nonendothelial vessel cell types, its expression was tested in human placental-derived pericytes and aortic smooth muscles cells. Low levels of RhoJ message and protein were found in isolates of both cell types at a level much lower than that seen in endothelial cells (supplemental Figure II).
Whole-mount in situ hybridization was performed at embryonic day 9.5, the stage at which the vasculature is developing and angiogenesis is occurring in the developing mouse.20 RhoJ, like the endothelial expressed gene von Willebrand factor,21 is expressed in the main trunk vessels and intersomitic vessels (Figure 1B). These data are consistent with an endothelial expression pattern of RhoJ in developing vessels.
To determine the intracellular localization of endogenous RhoJ, purified rabbit polyclonal antiserum was developed and validated using RhoJ knockdown in HUVECs or RhoJ overexpression in a nonendothelial cell type (data not shown). Using this reagent, RhoJ was localized in punctuate regions of the cell, and costaining with anti-vinculin antibodies or anti-phospho(397) FAK antibodies indicated that these structures were focal adhesions (Figure 1C). Because RhoJ is closely related in structure and amino acid sequence to other members of the Rho GTPase family, we established that neither this RhoJ polyclonal antisera nor the RhoJ monoclonal antibodies used in this study for Western blotting recognized Cdc42, Rac1, or RhoA (supplemental Figure III).
RhoJ Plays a Role in Endothelial Cell Motility, Growth, and Tube Formation
To determine whether RhoJ plays any role in endothelial cell biology and angiogenesis, a series of assays were performed using HUVECs with either siRNA-mediated knockdown of RhoJ or overexpression of dominant-active RhoJ. Knockdown of RhoJ protein was performed using 2 RhoJ-specific duplexes (RhoJ D1 and RhoJ D2); as a control, cells were either mock-transfected or transfected with a negative control duplex (siControl). Angiogenesis requires both endothelial cell motility and proliferation. Chemokinetic motility was assessed using a basic scratch-wound assay. A scratch was made in a monolayer of HUVECs 2 days after transfection. Unlike the control cells (mock and siControl), cells with reduced RhoJ expression (RhoJ D1 and D2) did not migrate to close the scratch; at 24 hours, there was still a region of scratch evident (Figure 2A). The mean area of scratch remaining at different time points was quantified from 3 independent experiments (Figure 2C), and statistically significant differences observed at 24 hours by comparing either of the RhoJ siRNA duplex–treated groups with the siScramble control. Mitomycin C was added to this assay to prevent cell division, strongly suggesting that it is impaired cell motility in the RhoJ knockdown cells, rather than a proliferative defect, that prevented closure. The RhoJ knockdown efficiently knocked down RhoJ protein levels (Figure 2E) but did not alter the levels of Cdc42, Rac1, or RhoA in endothelial cells (supplemental Figure IV). The absence of any interferon response at duplex concentrations used for any of the various assays was confirmed by monitoring expression of the interferon inducible genes 2′,5′-oligoadenylate synthetase 122 and IFN-stimulated gene of 20 kDa23 (data not shown).
The effect of expressing a GFP-tagged dominant active Q79L mutant of RhoJ,11 permanently in the GTP-bound active form, was also assessed. HUVECs were transduced with lentivirus to express either GFP or GFP-daRhoJ; GFP-expressing cells were then purified by cell sorting; expression of the GFP or GFP-daRhoJ was confirmed by Western blotting (Figure 2F and supplemental Figure VA). GFP-daRhoJ primarily localized to the plasma membrane, where it was more concentrated in focal adhesions, there was also some localization to intracellular vesicles (supplemental Figure VB). By contrast to the siRNA knockdown, the overexpression of GFP-daRhoJ slightly but reproducibly accelerated wound closure in at least 3 experiments (Figure 2B and 2D). However, variations in the speed of wound closure between different HUVEC isolates meant that significant differences are not revealed by performing statistical tests on the combined data at different time points. These data strongly suggest a role for RhoJ in mediating endothelial cell movement.
The effect of RhoJ knockdown on endothelial cell proliferation and chemotaxis was also assessed; there was slightly reduced cell growth in RhoJ D1– or RhoJ D2–transfected endothelial cells compared with mock- or siControl-transfected cells (supplemental Figure VIA). This diminished proliferation was not associated with any changes in the cell cycle distribution, but small increases in the level of apoptosis were observed (supplemental Figure VIB). Knockdown of RhoJ also inhibited chemotaxis through the porous filters in response to chemoattractant, again suggesting a role for RhoJ in endothelial cell motility (supplemental Figure VIC).
Finally, endothelial tube formation was assessed both using a Matrigel tube-forming assay and an organotypic angiogenesis assay. HUVECs transfected with RhoJ-specific siRNA duplexes produced a more poorly connected network of cells when plated on Matrigel, a solubilized basement membrane extract, compared with the control-transfected cells. These networks were less stable, with more cell retraction observed at 24 hours (supplemental Figure VII).
In the organotypic assay, HUVECs are cocultured with human dermal fibroblasts. The interaction of these cell types results in the formation of endothelial derived tubules with lumens that are embedded in naturally produced extracellular matrix.24 This process is entirely dependent on the 2 major angiogenic factors: basic fibroblast growth factor, present in the culture media; and vascular endothelial growth factor (VEGF), produced by the fibroblasts; and the resulting tubules are highly reminiscent of capillaries formed during angiogenesis in vivo.25,26 HUVECs were either mock-transfected or transfected with control or RhoJ-specific siRNA duplexes and seeded on to a confluent monolayer of human dermal fibroblasts, and tubules were allowed to develop for 5 days. Knockdown of RhoJ resulted in highly impaired tube formation in this assay, with there being significantly fewer, shorter, and lesser-branched tubes (Figure 3A and 3B). Knockdown of RhoJ was assessed by Western blotting at later time points, transfection with RhoJ D1 and D2 at 20 nmol/L gave a knockdown of approximately 60% and 40%, respectively, after 4 days (supplemental Figure VIII).
In contrast, HUVECs overexpressing GFP-daRhoJ produced a greater number of highly branched tubules with a shorter mean length (Figure 3C and 3D). These data indicate a role for RhoJ in tubule formation in these assays. siRNA-mediated knockdown of RhoJ in the human microvascular cell line HMEC-1 (human microvascular cell line-1) also resulted in impaired scratch-wound closure (supplemental Figure IX) and tube formation on Matrigel (supplemental Figure X).
Because tube formation is VEGF-dependent in the coculture assay,26 and VEGF is one of the most potent proangiogenic factors, assays were performed to assess whether RhoJ is activated by VEGF. HUVECs were treated with VEGF, and levels of both active RhoJ and Cdc42 were determined by pull down of the active GTP-bound forms of these proteins with a glutathione S-transferase fusion of the PAK1 CRIB domain and Western blotting. Densitometry was performed and levels of active Cdc42 and RhoJ in pull-downs versus cellular lysates were determined. Activation of Cdc42 peaked at 15 minutes, and its activity declined thereafter, a result consistent with previous reports.27 RhoJ, like Cdc42, was activated by VEGF, but its level of activation was more modest, with a slower and more sustained induction (supplemental Figure XI).
RhoJ Modulates Actomyosin Contractility, Actin Stress Fibers, and Focal Adhesions
The data thus far have implicated a critical role for RhoJ in regulating endothelial cell motility, and because Rho GTPases have been shown to control the actin cytoskeleton and RhoJ localizes to focal adhesions, experiments were performed to investigate how RhoJ knockdown might affect both f-actin and focal adhesion formation.
In initial experiments, HUVECs were either mock-transfected or transfected with siControl, RhoJ D1, or D2, and after 2 days, cells were adhered to gelatin-coated coverslips at a low density and were stained with phalloidin to detect f-actin and with anti-vinculin antibodies to visualize focal adhesions. The control-transfected HUVECs generally showed a modest number of stress fibers and focal adhesions, and the cells had a number of cellular extensions (supplemental Figure XIIA). In contrast, cells with reduced RhoJ expression had a more circular appearance with fewer cell protrusions and increased numbers of stress fibers, which, in many cells, were observed around the cell edge. Knockdown of RhoJ also resulted in significantly more numerous focal adhesions (supplemental Figure XIIB).
Endothelial cells in vessels are normally found as a monolayer surrounded by other endothelial cells; the exceptions to this are the endothelial tip cells found at the leading end of an angiogenic sprout.28 To assess whether RhoJ knockdown differentially affected f-actin and focal adhesion numbers in endothelial cells depending on their cellular context, cells at the edge of a scratch wound at the migration front were compared with those found within a monolayer. Two days after transfection, a scratch was made, and the cells were allowed to migrate before they were fixed and stained with either fluorescently conjugated phalloidin or anti-vinculin antibodies. siRNA-mediated knockdown of RhoJ resulted in significantly increased focal adhesion numbers in cells at the edge of a scratch compared with control-transfected cells (Figure 4A and 4B); in contrast, no differences were observed in cells within the monolayer (data not shown). RhoJ knockdown did not affect the localization of VE-cadherin at the cell junctions within the monolayer, suggesting that cell–cell adhesion is not affected. Lentiviral transduction of HUVECs with GFP-daRhoJ resulted in slightly faster wound closure, and enumeration of focal adhesion numbers showed that overexpression of GFP-daRhoJ resulted in cells at the wound edge with significantly fewer focal adhesions than the GFP controls (Figure 4A and 4B). Again no differences were observed in the monolayer. These data indicate a role for RhoJ in regulating focal adhesion disassembly or turnover.
Similarly, using phalloidin staining of both RhoJ siRNA–treated HUVECs and HUVECs overexpressing GFP-daRhoJ, f-actin in cells at the wound edge and in the monolayer was examined. Whereas no differences were observed in cells in the monolayer (data not shown), knocking down RhoJ expression was found to increase the numbers of stress fibers in cells at the wound edge (Figure 5A). This was quantified by measuring fluorescence levels in each cell, and significant increases were found in cells treated with either of the RhoJ-specific duplexes compared with the controls (Figure 5B). Expression of GFP-daRhoJ had the opposite effect, significantly reducing levels of stress fiber in cells at the migration front (Figure 5C and 5D). Thus, RhoJ may play a role in negatively regulating stress fibers formation in migrating cells.
The increased actin stress fiber formation may reflect increased actomyosin contractility. Recently, it has been demonstrated that quiescent tubes have an increased contractility compared with sprouting tubes, and, indeed, a reduction in contractility was required to enable endothelial cell sprouting.7 The effect of RhoJ in regulating contractility was measured in 2 ways: by measuring endothelial cell–mediated contraction of a type I collagen gel and by determining the levels of phospho-MLC in endothelial cells forming tubes on Matrigel.
An assay previously used to measure endothelial contractility and traction was performed17; siRNA-transfected HUVECs or HUVECs expressing GFP/GFP-daRhoJ were imbedded in a collagen gel. After 2 days, the gels were released from the sides of the well, and the plates were incubated for a further 5 days, as the contractility developed within the cells, so the gel retracted, and the amount of contractility was assessed by measuring the diameter of the gel. As a control, 5 μmol/L blebbistatin, an inhibitor of nonmuscle myosin II,29 was added. Knockdown of RhoJ expression with either of the RhoJ-specific duplexes resulted in a highly significant increase in contraction, which was mostly reversed with blebbistatin. GFP-daRhoJ expression resulted in a small but significant reduction in the contractility compared with cells expressing GFP alone (Figure 5E). Knockdown of RhoJ was also associated with increased phospho-MLC in HUVECs plated on Matrigel, indicating that these cells have increased actomyosin contractility (Figure 5F). No differences were seen in the levels of phospho-MLC as a result of expressing GFP-daRhoJ.
Collectively, these data suggest that RhoJ may be involved in promoting endothelial cell motility and tube formation by negatively regulating MLC phosphorylation and reducing actomyosin contractility. In this case, pharmacological inhibition of ROCK, a kinase that catalyzes MLC phosphorylation,5 should reverse the impaired motility and tube formation induced by RhoJ knockdown. Indeed, we found that the ROCK inhibitor Y27632, when added to RhoJ siRNA–treated endothelial cells, restored motility both in the scratch-wound assay (Figure 6A and 6B) and Boyden chamber chemotaxis assay (supplemental Figure XIII). RhoJ knockdown cells migrated to close the scratch when treated with the ROCK inhibitor at a similar rate to that of the negative control duplex treated cells (Figure 6A and 6B). Additionally, ROCK inhibition restored the impaired tube formation on Matrigel of RhoJ siRNA–transfected cells, resulting in a more connected and stable tube formation, equivalent to that seen in the negative control duplex–treated cells (Figure 6C and 6D). Addition of 10 μmol/L ROCK inhibitor Y27632 resulted in reduction in phospho-MLC in all samples (supplemental Figure XIV). Similar data were obtained with a structurally unrelated ROCK inhibitor H1152, with the other RhoJ siRNA duplex D1 (data not shown) and with 5 μmol/L blebbistatin, a myosin II inhibitor (supplemental Figures XV and XVI). These data strongly suggest that RhoJ is a negative regulator of actomyosin contractility.
These data are the first to determine the intracellular localization of RhoJ, demonstrate endothelial expression in tissues, show activation by VEGF, and identify a pivotal role for this Rho GTPase RhoJ in endothelial cell motility, tube formation, and modulation of the actin cytoskeleton and focal adhesion formation.
Previous bioinformatic and primary cell line expression studies performed by our laboratory identified RhoJ as having an endothelial specific mRNA expression pattern.9 In this present study, we have demonstrated expression by endothelial cells in vessels in a number of tissues and cancer types, although it is absent from brain vessels and from normal and cancerous tissue of the digestive tract. We also demonstrate a vascular expression pattern of RhoJ in the mouse embryo, at a stage in development when angiogenesis is occurring. This expression pattern suggests that it plays a role in the normal functioning of endothelial cells in a number of tissues, as well as potentially in angiogenesis. Although expression levels of RhoJ are high in endothelial cells, we have observed low levels of expression in nonendothelial cell types, indicating that expression of RhoJ is not restricted to endothelial cells.
The function of RhoJ in endothelial cells was probed by using both siRNA-mediated knockdown of RhoJ expression and expression of a constitutively active form of RhoJ, locked into its GTP-bound form. As expected, these 2 approaches gave contrasting phenotypes. Knockdown of RhoJ resulted in impaired migration and tube formation, whereas daRhoJ promoted motility and excessive sprouting during tube formation. Knocking down RhoJ expression using siRNA duplexes impaired cell migration and reduced proliferation, processes involved at early time points in the formation of tubules in the organotypic model. In addition, the RhoJ siRNA–induced increased contractility would have also been likely to be detrimental to tube formation, because previously increased actomyosin contractility resulted in cell detachment and cell death in this model of tubule formation.6 In contrast, the chaotic appearance of the tubules observed with the daRhoJ is reminiscent of abnormal, sprouting endothelium, a process dependent on VEGF receptor 2.30 The use of the organotypic assay enabled the identification of a role for RhoJ in tubulogenesis in a physiological setting; however, we are unable to conclude downstream of which soluble or extracellular matrix factor RhoJ is acting in this system. Knocking down RhoJ did not affect the expression of the other major Rho GTPases, RhoA, Rac1, or Cdc42, and we conclude that RhoJ is playing a distinct role in endothelial cells.
To understand the molecular mechanisms by which RhoJ was acting, the effect of modulating its activity on focal adhesions numbers and contractility was investigated. Reducing the activity of RhoJ resulted in increased numbers of focal adhesion and stress fibers in cells migrating at the edge of scratch; additionally, we observed an increase in the physical contractility, as measured by the contraction of a collagen gel. Expressing daRhoJ again gave the converse phenotype. Interestingly, the stress fiber and focal adhesion phenotype was only evident in migrating cells, either plated sparsely or at a scratch-wound edge, and not in cells within a monolayer. This suggests that RhoJ plays a critical role in migrating cells, whereas in cells within a monolayer, the effects of modulating RhoJ expression may be inhibited through signaling from neighboring cells. This lack of phenotype observed in the monolayer may explain why siRNA-mediated downregulation of RhoJ did not inhibit early wound closure (Figure 2A and 2C), with the accumulation of focal adhesions and f-actin occurring only after the cells had started to move and had reduced contact with neighboring cells.
We have additionally demonstrated that RhoJ knockdown results in an increase in MLC phosphorylation. MLC can be phosphorylated by a number of kinases including RhoA-dependent ROCK and Ca2+-dependent MLC kinase, which result in increased myosin ATPase activity and increased stress fiber contractility.31 Using 2 structurally unrelated inhibitors of ROCK or an inhibitor of nonmuscle myosin II, the motility defect of RhoJ siRNA–treated endothelial cells was reversed in both the scratch-wound and Matrigel tube-forming assays. There is evidence that reduced contractility is required for endothelial sprouting to occur,6,7 and our data suggest that RhoJ may play a role in this process.
At present, we do not know the molecular basis by which RhoJ affects contractility and focal adhesion numbers. There is a complex interplay and reciprocal regulation between focal adhesion maturation, turnover, and actomyosin contractility; thus, components of focal adhesions have been shown to regulate contractility, and, conversely, modulating myosin II activity can influence focal adhesion size and distribution.32 Thus, RhoJ may regulate actomyosin contractility via dynamic regulation of focal adhesions.
This hypothesis is consistent with the localization of endogenously expressed RhoJ to focal adhesions. Previously, only localization of overexpressed RhoJ to early and recycling endosomes had been reported,16 and, indeed, we have observed the same. Because overexpressing proteins can often drive aberrant localization, we developed a purified polyclonal antiserum that, unlike the commercially available anti-RhoJ monoclonal antibodies, recognized endogenous RhoJ. Because a role for RhoJ in early endocytic recycling has been reported, we investigated whether RhoJ knockdown affected surface levels of VEGF receptor 2 either in resting cells or after VEGF stimulation; in neither case were any differences noted (data not shown).
In conclusion, we have identified RhoJ as a new player in endothelial cell biology. We have demonstrated that this endothelial-expressed Rho GTPase is activated by VEGF and is able to modulate actomyosin contractility and focal adhesion assembly. It plays a critical role in endothelial cell motility and tube formation and will likely prove to be a key player in angiogenesis in vivo.
Sources of Funding
This work was supported by the British Heart Foundation, Cancer Research UK, and University of Birmingham College of Medical and Dental Sciences.
We acknowledge Dr Michael Tomlinson for critical reading of the manuscript and Sharon Timms and Dr Zsuzsanna Nagy for developing the fluorescent in situ hybridization method.
- Received February 22, 2010.
- Accepted November 23, 2010.
- © 2011 American Heart Association, Inc.
- Legate KR,
- Wickstrom SA,
- Fassler R
- Herbert JM,
- Stekel D,
- Sanderson S,
- Heath VL,
- Bicknell R
- Chiang SH,
- Hou JC,
- Hwang J,
- Pessin JE,
- Saltiel AR
- Vignal E,
- De Toledo M,
- Comunale F,
- Ladopoulou A,
- Gauthier-Rouviere C,
- Blangy A,
- Fort P
- Abe T,
- Kato M,
- Miki H,
- Takenawa T,
- Endo T
- Nishizuka M,
- Arimoto E,
- Tsuchiya T,
- Nishihara T,
- Imagawa M
- de Toledo M,
- Senic-Matuglia F,
- Salamero J,
- Uze G,
- Comunale F,
- Fort P,
- Blangy A
- Walls JR,
- Coultas L,
- Rossant J,
- Henkelman RM
- Straight AF,
- Cheung A,
- Limouze J,
- Chen I,
- Westwood NJ,
- Sellers JR,
- Mitchison TJ
- Jain RK