NADPH Oxidases.

In endothelial cells, Rac1 has been strongly linked to activation of the NADPH oxidase complex in the cellular membrane. The NADPH-oxidase (NOX) complex is well recognized for its role in killing pathogens in the phagosomes of neutrophils by generating superoxide. Interestingly, the complex is also found in vascular cells, e.g., endothelial cells, smooth muscle cells, and adventitial fibroblasts (Konior et al., 2014). Initially, pioneers in the field of NOX function in endothelial and smooth muscle cells were eager to prove that these cell types have functionally active NADPH oxidases. Being similar to leukocyte enzymes, the oxidases were shown to play an important role in maintaining functionality of the cell types and thereby vascular homeostasis (Gorlach et al., 2000). Since then, the in-depth functions of the NOX proteins have been presented in excellent reviews (Lambeth, 2004; Streeter et al., 2013; Drummond and Sobey, 2014). The NOX complex comprises several subunits, including members of the family of NOX proteins, responsible for the generation of superoxide or other reactive oxygen species (ROS). However, without critical subunits like p47, p67, or p40 and 22phox, NOX is not able to generate ROS (Hordijk, 2006). For proper assembly of the full and functional NOX complex, Rac1 activity is required (Bedard and Krause, 2007; Sumimoto, 2008). In vascular cells, ROS appear to be functionally relevant for the regulation of vascular tone and gene transcription. ROS interact with proteins, lipids, and nucleic acids and is able to alter their function. However, chronic or excessive Rac1-mediated ROS production results in oxidative stress, by which ROS destroy other molecules, and is considered detrimental (Elnakish et al., 2013). There are several isoforms of the enzymatic NOX family, of which NOX2 is considered the classic NOX. NOX1, NOX2, and NOX4 are expressed by endothelial cells (van Buul et al., 2005; Dammanahalli and Sun, 2008). In contrast to NOX1 and NOX2, NOX4 is constitutively active and its activation does not require Rac1 (Martyn et al., 2006; Meng et al., 2008).

In endothelial cells, a number of small vasoactive peptide hormones, such as angiotensin-II, endothelin-1, and atrial natriuretic peptide (ANP), are related to blood pressure regulation by influencing Rac1-dependent NOX activation. Angiotensin-II receptor type 1 (AT1R)–mediated signaling is best known to induce NOX activation and generate oxidative stress. Under physiologic conditions, this pathway increases blood pressure, yet when chronically activated, it induces vascular pathogenesis (Gregg et al., 2003). In mice, chronic infusion of angiotensin-II enhances atherosclerosis and aneurysm formation, which are both proinflammatory vascular pathologies with a strong regulatory role for endothelial cells. Decreased endothelial shear stress enhances atherogenesis and aneurysm formation (Higuchi et al., 2012). In addition, whereas AT1R deletion in smooth muscle cells has no effect on aneurysm formation in the ascending aorta, endothelial-specific depletion of AT1R attenuated the pathology (Rateri et al., 2011), indicating that endothelial-specific generation of ROS by AT1R may control aneurysm formation. In that light, we recently showed that immunosuppressive drug azathioprine is a potent Rac1-inhibitory drug in aortic endothelial cells, and efficiently reduced aneurysm formation in mice after angiotensin-II infusion (Marinkovic et al., 2013, 2014b). Endothelial cells treated with azathioprine metabolites showed reduced c-Jun N-terminal kinase (JNK) phosphorylation in vitro and in vivo (Marinkovic et al., 2013, 2014b).

The vasoactive steroid aldosterone also induces ROS generation, via mineralocorticoid receptor–regulated activation of NOX and Rac1, and thereby induces angiotensin-converting enzyme (ACE) expression. ACE will generate angiotensin-II and subsequently activate AT1R (Iwashima et al., 2008). Medication suppressing the renin-aldosterone-angiotensin-II system (RAAS), such as ACE inhibitors or AT1R blockers, are known to protect against vascular diseases in heart, blood vessels, and kidney, probably in part by indirect inhibition of NOX/Rac1 activation.

Next to ligand-receptor–induced activation, the NOX complex in endothelial cells can also be activated by fluid shear stress, where a mechanosensitive complex is present and responsible for Rac1 activation upon alterations in shear stress. This complex is composed of VE-cadherin, Par3, p67phox (part of NOX complex), Tiam1 (Rac-specific GEF), and Rac1, of which VE-cadherin is responsible for the mechano-sensing, ultimately resulting in mitogen-activated protein kinase activation (Yeh et al., 1999; Liu et al., 2013). Many vascular diseases are strongly enhanced by shear stress influencing the endothelium and ligand-receptor–mediated signals, especially in atherosclerosis, where the site specificity of the lesions is determined by reduced shear stress.

In mice, the role of NOX2 in the pathophysiology of atherosclerosis was studied in NOX2 deficient mice in a proatherogenic apolipoprotein E-deficient background. These experiments revealed that although NOX2 deficiency showed reduced production of superoxide, lesion formation in the aortic sinus was not affected (Kirk et al., 2000; Judkins et al., 2010), but total lesion area of the aorta was reduced by 50% (Judkins et al., 2010). Endothelial cell–specific NOX2-overexpressing mice, also in an apolipoprotein E-deficient background, showed increased macrophage recruitment, yet no difference in development of atherosclerosis in the sinus or total aorta (Douglas et al., 2012). Possibly, there is a local difference in NOX activation of the various NOX isoforms in the different cell types, which contributes to atherosclerosis.

In line with this, NOX4 deficiency increases various pathologic vascular phenotypes, also in endothelial cells, since loss of NOX4 resulted in a reduction of endothelial nitric oxide synthase expression (eNOS) and nitric oxide production. This was mediated by reduced transcriptional activity of the antioxidant nuclear factor erythroid 2–related factor 2 (Nrf2) transcription factor (Schroder et al., 2012). Thus, the Rac1-independent NOX4 seems to play an important role in the homeostasis of endothelial cells and thereby is protective against vascular disease, whereas other NOX family members that do require Rac1 for their activity cause a proinflammatory phenotype.

In atherosclerotic patients, lipid-lowering statins are used as standard clinical care. Statins reduce low-density lipoprotein cholesterol levels by inhibition of the rate-limiting enzyme in cholesterol synthesis, which is 3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) reductase. Statins have pleiotropic therapeutic effects; for example, they promote the degradation of Rac1 by targeting the anchoring of Rac1 to the membrane via small GTP-binding protein dissociation stimulator upregulation (Tanaka et al., 2013). Once in the cytosol and not protected by Rho-GDIs, Rac1 is rapidly degraded. Statins have been reported to decrease NOX activity and ROS generation via inhibition of Rac1, as summarized by Antonopoulos et al. (2012).

However, it has also been shown that simvastatin potently enhances Rac1 activity as a direct consequence of downregulating cholesterol synthesis (Kou et al., 2009). A decrease in the intermediate isoprenoid metabolite geranylgeranyl pyrophosphate was responsible for Rac1 activation by limiting the substrate for the enzyme geranylgeranyl transferase-1 (GGTase-1) (Kou et al., 2009). GGTase-1 transfers geranylgeranyl to specific motifs in many proteins, including RhoA, Cdc42, and Rac1, which alters protein function. Interestingly, GGTase-1 knockout macrophages have highly increased RhoA, Cdc42, and Rac1 activity, and showed a proinflammatory phenotype (Khan et al., 2011). In that light, it is not surprising that macrophage-specific GGTase-1 knockdown in mice resulted in increased inflammation in a rheumatoid arthritis model (Khan et al., 2011). In addition, in proatherogenic low-density lipoprotein receptor–deficient mice, GGTase-1 deficiency in macrophages increased T lymphocyte influx (inflammation) into the vessel wall. Surprisingly, it reduced atherosclerosis by 60%, which was caused by a RhoA-dependent increase in cholesterol efflux from lesion macrophages (Khan et al., 2013). Inhibition of GGTase-1 in rat arteries induced eNOS and reduced NOX and ROS (Zuckerbraun et al., 2005), suggesting that statins would improve endothelial cell function and help reduce cardiovascular disease, yet its influence on Rac1 activity is not resolved completely.

Because Rac1 is the “molecular engine” of most NOX activity, it seems important to control Rac1 activity to prevent excessive oxidative stress–induced pathologies (Elnakish et al., 2013; Konior et al., 2014). Many (Moldovan et al., 2006), but not all, downstream effects of Rac1 activation involve NOX activation. In the next paragraphs, cell type–specific functions of “vascular” Rac1 are discussed.

Leukocyte Transmigration.

In endothelial cells, Rac1 activation requires a proinflammatory stimulus, and thereafter may induce downstream transcription-dependent or -independent actions to facilitate leukocyte attraction and transmigration locally. The most well known downstream target of Rac1 is serine/threonine p21-activating kinase 1 (PAK1), a critical effector that induces cytoskeleton remodeling and nuclear signaling via activation of the NOX, JNK, or nuclear factor (NF)-κB pathways (Kumar et al., 2006; Taglieri et al., 2014). For example, cytokine tumor necrosis factor α (TNFα) can mediate chemokine and adhesion molecule VCAM-1 (vascular cell adhesion molecule 1) expression in endothelial cells via the proinflammatory transcription factor NF-κB in a NOX/Rac1-dependent way (Chen et al., 2004; Marinkovic et al., 2014b) and AP1/ATF2 via JNK signaling (Marinkovic et al., 2013, 2014b). Attracting leukocytes is the first step of an inflammatory response, but leukocyte migration through the endothelial layer requires cell adhesion, which is the second step. Rac1 is critically involved in the formation of so called “docking structures” or “transmigratory cups” (Barreiro et al., 2002; Carman and Springer, 2004; van Buul et al., 2007), which are membrane protrusions that capture adherent and transmigrating leukocytes and depend on cytoskeletal remodeling (van Rijssel et al., 2012b). The main Rho-GEF involved in Rac1-induced docking structures in endothelial cells is TRIO (van Rijssel et al., 2012b). Whereas RhoG is a more potent substrate for TRIO-GEF1 than Rac1 (van Rijssel et al., 2012a; Jaiswal et al., 2013), RhoG is not activated upon TNFα stimulation, whereas Rac1 is (van Rijssel et al., 2013). This shows that TRIO can control Rac1 activation upon receiving stimuli that do not necessarily activate RhoG in endothelial cells. Cross-linking or clustering of intercellular adhesion molecule (ICAM)-1 recruits TRIO and triggers local activity of Rac1 (van Rijssel et al., 2012b). Rac1 can also be activated by VCAM-1 clustering (van Wetering et al., 2002; Marchese et al., 2012). TNFα-induced VCAM-1 expression involves post-translational modification by methylation of the C-terminal prenylcysteine in Rac1 to target Rac1 to the cell membrane (Ahmad et al., 2002; Papaharalambus et al., 2005). Interestingly, VCAM-1 expression is regulated in a Rac1-dependent fashion, whereas ICAM-1 is not (Ahmad et al., 2002; Ng et al., 2002; Marinkovic et al., 2014b). This exclusiveness of VCAM-1 regulation has also been observed with transcription factor Ets-2 in endothelial cells (Cheng et al., 2011), although Rac1 involvement has not been determined here.

Statins, which have been shown to inhibit Rac1-mediated NOX activation, are also reported to be protective against endothelial cell activation (Palinski and Napoli, 2002). In the presence of proinflammatory cytokines, such as TNFα, statins can suppress Rac1 activation and downstream ROS generation, transcription factor NF-κB activity, and expression of adhesion molecules such as ICAM-1 and VCAM-1, with extracellular signal-regulated kinase 5 (ERK5) as a target of statin therapy (Wu et al., 2013b). The abrogation of Rac1 activation by statins and ERK5 is transcription factor KLF2–dependent, which is the shear stress–sensitive factor responsible for healthy endothelium (Komaravolu et al., 2015).

ICAM-1 expression is not changed by Rac1 activation, but its functionality after clustering is Rac1-dependent and it is severely hampered by Rac1 inhibition (Marinkovic et al., 2014b). Anti–ICAM-1 antibody–coated beads are shown to efficiently cross-link ICAM-1, thereby inducing Rac1 activation and transmigratory cup formation in endothelial cells (Marinkovic et al., 2014b). This can be inhibited by 6-mercaptopurine, a metabolite of azathioprine, which has potent Rac1 inhibitory capacity. 6-Mercaptopurine, which is further converted to 6-thio-GTP, interferes with Rac1 function (Marinkovic et al., 2014b). Upon angiotensin-II infusion in mice, leukocyte transmigration into the aortic wall was inhibited when mice were treated with azathioprine, with endothelial cells showing decreased p-JNK activation (Marinkovic et al., 2013, 2014b).

Rac1 activation is necessary for the formation of membrane protrusions, which facilitate leukocyte transmigration, and it is also essential for closure of the micro-wounds in the endothelial cell layer, where leukocytes have passed through, to prevent leakage (Martinelli et al., 2013). In case of excessive inflammation, Rac1 inhibition can relieve the inflammatory burden by prevention of further leukocyte influx. Yet, micro-wounds created by leukocyte penetration will not be restored when Rac1 is inhibited, and leaky vessels could possibly be attributed to chronic use of Rac1 inhibitors. Thus, tight regulation of endothelial Rac1 is required to respond adequately to changes in the microenvironment.

Barrier Function.

In addition to Rac1-mediated cytoskeletal rearrangements to capture leukocytes, cytoskeletal changes are also involved in endothelial barrier function or cell migration. Maintaining the endothelial barrier is imperative for healthy blood vessels. Vascular leakage leading to edema or tissue inflammation is highly relevant in the microvasculature and plays a key role in cancer progression (Bid et al., 2013). Several reports show that Rac1 activation recovers or even stabilizes decreased endothelial barrier function after G protein–coupled receptor activation with vascular endothelial growth factor (VEGF), thrombin, histamine, or platelet-activating factor (Waschke et al., 2006; Tan et al., 2008; Baumer et al., 2009; Knezevic et al., 2009; Adamson et al., 2010; Hoang et al., 2011; Wang et al., 2011b; Bid et al., 2013; Maharjan et al., 2013). Interestingly, apart from proteins, certain lipids also modulate the barrier function. Two endogenous lipid-derived molecules, namely, the phospholipid sphingosine 1-phosphate (S1P) and bile acids (generated from the lipid cholesterol), are known to increase endothelial barrier function. They do so by enhancing Rac1 activity via activation of their respective receptors, S1P receptor type 1 (van Hooren et al., 2014) and the G protein–coupled bile acid receptor (Kida et al., 2014). The sugar, hyaluronan, a glycosaminoglycan that is abundantly present in the extracellular matrix, can be generated in a high-molecular-weight (HMW-HA) or low-molecular-weight (LMW-HA) form and also modulates the endothelial barrier. HMW-HA and LMW-HA bind to different isoforms of the cell surface receptor CD44, with opposing effects. HMW-HA enhances the barrier function via CD44s association with S1P receptor type 1 and Rac1 activation, whereas LMW-HA reduced the barrier function via CD44v10 association with S1P receptor type 3 and RhoA activation (Singleton et al., 2006). In light of the concept that sugars and lipids can modulate barrier function, the finding that the pseudo-sugar derivative of cholesterol Sac-1004 blocks stimulus-induced endothelial permeability is an interesting discovery (Maharjan et al., 2013). By improving Rac1 activity, Sac-1004 inhibits endothelial hyperpermeability, induced by VEGF, histamine, or thrombin, via stabilization of the cortical actin rings and adherens junction proteins at the cell-cell junction (Maharjan et al., 2013).

Rac1 is known to regulate cell-to-cell contact by modulation of junctional protein distribution, especially the adherens junctions (Bazzoni and Dejana, 2004). Adherens junctions are predominately composed of transmembrane VE-cadherin and intracellular catenins (Bazzoni and Dejana, 2004). Interference with VE-cadherin causes actin cytoskeleton reorganization (Hordijk et al., 1999), enhanced permeability (Hordijk et al., 1999; Timmerman et al., 2012), cell migration (Breviario et al., 1995), and angiogenesis (Carmeliet et al., 1999). Whereas Rac1 activity is essential for maintenance of junctional integrity, introduction of a constitutively active mutant of Rac1 (Tat-RacV12) into endothelial cells resulted in increased endothelial monolayer permeability by redistribution of the VE-cadherin complex (van Wetering et al., 2002). Rac1-dependent ROS production played a key role in this process, leading to phosphorylation of the VE-cadherin complex (van Wetering et al., 2002; Monaghan-Benson and Burridge, 2009). These studies show the importance of balanced Rac1 activity to maintain endothelial barrier function. Most probably, local Rac1 activity, induced through specific GEFs, may contribute to this delicate balance. However, how exactly this is regulated is intriguing and requires future research.

Within the Rho subfamily, RhoA and Rac1 often have opposing roles and keep each other in balance. Knockdown of focal adhesion kinase, a mediator of signal transduction by integrins and growth factor receptors in endothelial cells, results in increased RhoA activity and decreased Rac1 activity. This caused a severe dysfunction of adherens junctions and subsequently compromised endothelial barrier function (Schmidt et al., 2013). However, this effect was rescued by a RhoA inhibitor that allowed Rac1 to become activated again. Regulation of specific GAPs seemed to contribute in the balancing act between RhoA and Rac1 in fibroblasts. RhoA was shown to inhibit Rac1 activity by antagonism of FILGAP (Ohta et al., 2006). In addition, Rac1-mediated ROS production resulted in inhibition of the low-molecular-weight protein tyrosine phosphatase (LMW-PTP), followed by increased phosphorylation and activation of its target, p190Rho-GAP, which in turn inhibited RhoA activation (Nimnual et al., 2003). A similar influence of RhoA or Rac1 on GAPs could be responsible for the opposing effects observed in endothelial cells, which requires future studies.

Migration.

Endothelial cell migration depends on lamellipodia formation. Lamellipodia are membrane protrusions that are induced through local (specific area within the cell) Rac1 activity and downstream activation of the actin-related protein 2/3 complexes to induce actin polymerization (Krause and Gautreau, 2014). Several studies have shown that local activation of Rac1 is essential for directional migration (Mayor and Carmona-Fontaine, 2010). Kraynov et al. (2000) were the first to show local Rac1 activity at the leading edge of migrating fibroblasts using a fluorescent resonance energy transfer–based biosensor for Rac1 activity. We have confirmed these data in endothelial cells (Fig. 1). The actin adapter protein cortactin plays an important role in the rearrangement of the actin cytoskeleton to generate and stabilize Rac1-dependent lamellipodia (Krause and Gautreau, 2014). Recently, the involvement of G protein–coupled receptor-2–interacting protein-1 (GIT1) became apparent in this process, since in GIT1-depleted endothelial cells, cortactin localization and, as a consequence, lamellipodia formation were strongly reduced. GIT1 promotes the assembly of signaling complexes to the membrane of the leading edge, and thereby activates Rac1 locally, resulting in the formation of lamellipodia (Majumder et al., 2014).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Local Rac1 activity at the leading edge of a migrating endothelial cell. An endothelial cell is transfected with an Rac1 FRET (fluorescent resonance energy transfer) biosensor and recorded in time to study spatio-temporal Rac1 activity during endogenous protrusion formation. Arrowheads show high FRET, i.e., active Rac1 regions, that coincide with protrusive activity. Scale bar, 20 μm. DIC, differential interference contrast.

Rac1-mediated ROS also play a role in vascular endothelial cell motility (Moldovan et al., 2006). Generation of ROS has been found at the leading edge of endothelial lamellipodia, exactly at the spots where local Rac1 activity is detected. Scavenging ROS reduced the speed of migration and directionality (Moldovan et al., 2006). The source of ROS that regulate endothelial cell migration may be NOX or mitochondria derived (Wang et al., 2011b). Rac1 and ROS are involved in phosphorylation of actin filament cross-linking protein myristoylated alanine-rich C-kinase substrate (MARCKS). This protein is essential for actin assembly and directional cell movement of endothelial cells (Kalwa and Michel, 2011; Kalwa et al., 2012). Small interfering RNA (siRNA)–mediated Rac1 knockdown blocked angiotensin-II–stimulated MARCKS phosphorylation, and siRNA-mediated knockdown of Rac1 or MARCKS disrupted actin polymerization and migration (Kalwa et al., 2012). In line with these data, we observed similar results using 6-mercaptopurine and 6-thio-GDP as Rac1 inhibitors in a model for endothelial cell migration (Fig. 2). Our data indicate that these metabolites of immunosuppressive drug azathioprine inhibit wound-induced Rac1 activity and additionally reduce the length and directionality of endothelial cell migration toward the wound.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Rac1-mediated migration is inhibited by 6-mercaptopurine (6-MP) and 6-thio-GDP (6-T-GDP). Rac1 activity is induced in human umbilical vein endothelial cells (HUVEC) by TNFα, and inhibited by 6-MP, 6-T-GDP, 6-T-GTP, and TRIO inhibitor ITX-3 (2-[(2,5-dimethyl-1-phenyl-1H-pyrrol-3-yl)methylene]-thiazolo[3,2-a]benzimidazol-3(2H)-one). In a scratch-wound assay, HUVEC migration was tracked in six cells per condition. Path length and directional migration was inhibited by 6-MP and 6-T-GDP after 32 hours.

VEGF is well known for its capacity to induce angiogenesis through endothelial cell migration to form new, yet leaky, vessels. Interestingly, Rac1 activation reduces VEGF-induced vascular leakage (Hoang et al., 2011), yet Rac1 is essential for VEGF-induced migration (Wang et al., 2011a). Again, local activation of Rac1 very probably determines these different processes. Recently, Rac1 was also demonstrated to regulate the production of VEGF itself by endothelial cells. An increase in activity of Rac1 and also Cdc42 resulted in tumor suppressor p53 ubiquitination, reducing p53 protein levels, which in turn induced expression of VEGF in endothelial cells, and thereby promoting angiogenesis as a feed forward mechanism (Ma et al., 2013). The prominent role of VEGF in angiogenesis makes it a prime target in cancer therapy. Since cell migration is largely Rac1-dependent, a strong focus now is on development of specific Rac1 inhibitors to reduce cancer cell migration and angiogenesis (Bid et al., 2013; Majumder et al., 2014). These inhibitors could be of importance potentially for other angiogenic disorders, such in diabetes, in which vascular leakage and angiogenesis in the retina causes injury. Yet, in diabetic patients, decreased vascularization of extremities (such as feet) is unwanted with Rac1 inhibitors. Thus, local Rac1 activity determines the final phenotypic outcome in many processes, which calls for a more specific strategy to target Rac1 activity as a therapeutic reagent. Nevertheless, it is clear that blocking Rac1 activity shows strong potential to fight inflammatory diseases, while perturbed endothelial monolayer integrity is a downside of Rac1 inhibition.

Proliferation and Migration.

In VSMC, redox-dependent migration, and proliferation are central events in the physiology and the development of vascular pathology. NOX1 is the prominent Rac1-dependent NOX that contributes to these processes, which are activated through a variety of different stimuli (Janiszewski et al., 2005; Sturrock et al., 2006; Meng et al., 2008; Fernandes et al., 2009; Pescatore et al., 2012). NOX1 deficiency [Nox1(y/–)] inhibited wire injury–induced VSMC hyperplasia in mice, yet VSMC-specific overexpression of NOX1 did not influence hyperplasia (Lee et al., 2009). NOX1-deficient VSMC derived from the mice exhibited a decreased proliferation and migration response to platelet-derived growth factor (PDGF) stimulation. Detailed analysis showed that this phenotype was caused by reduced PAK1 activation and increased cofilin phosphorylation, leading to reduced actin de-polymerization. An essential role for PAK1 in VSMC hyperplasia is shown in a rat restenosis model (Wang et al., 2009; Hinoki et al., 2010). Thus NOX1 activation causes Rac1/PAK1-dependent actin polymerization in VSMC (Lee et al., 2009). Kalirin-9 (Kalrn) is an arterial Rho-GEF with high homology to the Rho-GEF TRIO. Interestingly, Kalrn contains two GEF domains for both Rac1 and RhoA. Neointimal hyperplasia, induced by carotid endothelial denudation, was significantly reduced in VSMC-specific Kalrn^–/+ mice compared with control mice. Reduced Kalrn function gave normal RhoA activation but diminished Rac1 activation in serum-, endothelin-1–, or PDGF-stimulated VSMC, as assessed by reduced Rac1-GTP levels, PAK1 activation, and reduced VSMC proliferation and migration (Wu et al., 2013a). These data suggest that the Rac-GEF domain is dominant over the Rho-GEF domain in Kalrn under these conditions.

In addition, basic fibroblast growth factor (bFGF) induces NOX1/Rac1-dependent ROS generation and VSMC migration. bFGF-induced migration was attenuated in VSMC derived from NOX1-deficient mice, and reduced VSMC outgrowth from aortic segments of these mice was observed (Schroder et al., 2007). Transduction of NOX1 restored normal migration. bFGF activated JNK and enhanced phosphorylation of adaptor protein paxillin, which brings together regulators of actin organization. Paxillin is crucial for migration and secretion of matrix-metalloproteinases (MMPs) to facilitate movement (through matrix) (Schroder et al., 2007). Insulin-like growth factor-1 could also increase MMP-2 and MMP-9 activity and promote VSMC migration via Rac1 and (Rac1-independent) NOX4 activation (Meng et al., 2008). NOX4 knockdown inhibited VSMC migration owing to reduced ROS generation (Meng et al., 2008). In endothelial cells, it has been shown that NOX4 induced activation of redox-sensitive Nrf2 and its target genes (Schroder et al., 2012). In VSMC, PDGF promotes nuclear translocation of Nrf2 as an antioxidant response against excessive oxidative stress (Ashino et al., 2013). Nrf2 depletion by siRNA enhanced PDGF-promoted Rac1 activation and ROS production and phosphorylation of downstream ERK1/2. In line with these data, in vivo, Nrf2-deficient mice showed enhanced neointimal hyperplasia in a wire injury model (Ashino et al., 2013). Thus the Nrf2-mediated pathway seems a negative feedback mechanism to protect the cell, and may be NOX4-mediated.

Interestingly, 8-hydroxy-2-deoxyguanosine (8-OHdG), which is a product of DNA oxidation upon excessive oxidative stress, can decrease Rac1 activity. 8-OHdG has been shown to inhibit Rac1 by binding in the Rac1 GTP-active site, preventing Rac1 recycling. 8-OHdG decreased VSMC hyperplasia in a carotid artery ligation model probably by inhibition of Rac1 (Huh et al., 2012).

Cholesterol-lowering statins inhibit geranylgeranyl pyrophosphate generation and thereby limit the substrate for the enzyme GGTase-1 (Kou et al., 2009). The effect of decreased GGTase-1 function on VSMC reveals that inhibition of GGTase-1 decreased activation of RhoA and Rac1 as well as proliferation, and thereby decreased intimal hyperplasia after balloon injury in rats. TNFα- or angiotensin-II–induced activation of NOX and ROS was decreased by GGTase-1 inhibition. Whereas GGTase-1 inhibition in macrophages activated RhoA and Rac1, it reduced RhoA and Rac1 activity in VSMC, which may be part of the atheroprotective pleiotropic effect of statins (Zuckerbraun et al., 2005).

In conclusion, proliferation and migration of VSMC are largely regulated by NOX1/Rac1 and ROS, which can be beneficial in vascular repair but also detrimental in VSMC hyperplasia disorders.

Contraction.

NOX1 activation is known to enhance calcium (Ca²⁺) mobilization in VSMCs, which does not affect VSMC migration alone (Rossi et al., 2009; Li et al., 2011; Zimmerman et al., 2011; Duran-Prado et al., 2013). VSMC contraction is highly Ca²⁺-dependent, and the role of Rho GTPases therein is extensively discussed in a recent review by Loirand and Pacaud (2014). Whereas the role of GTPase RhoA in VSMC contraction is clear, inducing vasoconstriction upon activation resulting from the rise of cytosolic Ca²⁺, the role of Rac1 is still quite obscure. VSMC contraction requires phosphorylation of myosin light chain (MLC) by the Ca²⁺-calmodulin–dependent MLC-kinase (MLCK), which is activated by a rise in cytosolic Ca²⁺ concentration. MLCK is dephosphorylated again by the Ca²⁺-independent MLC phosphatase. Upon increased influx of Ca²⁺, activation of RhoA leads to activation of Rho-kinase (Rock), which phosphorylates MLC phosphatase, thereby inactivating it (Uehata et al., 1997). This inactivation prolongs the phosphorylated state of MLC and thus increased contractility of VSMC.

Loirand and Pacaud (2014) extensively describe Rac1 activation in VSMC as not that straightforward. Collectively, the context of Rac1 activation seems to determine if Rac1 activation leads to vasoconstriction or vasodilation. Rac1 activation can inactivate RhoA via PAK1-dependent inhibition of cGMP conversion to GMP, thus promoting cGMP-dependent protein kinase activity and subsequent RhoA phosphorylation and inactivation. Another Rac1/PAK1-mediated vasodilatory pathway affects endothelial eNOS and thus NO production, and NO increases cGMP generation and cGMP-dependent protein kinase activation. Rac1/PAK1 activation also inhibits MLCK, thus inhibiting phosphorylation of MLC and downstream contraction. In contrast, Rac1/PAK3 activation enhances phosphorylation of caldesmon, which induces vasoconstriction. In addition, Rac1-mediated NOX activation results in increased ROS generation, which also induces vasoconstriction. Clearly, further research on the functional relevance of the different contexts of Rac1 activation and its effect on VSMC contraction is necessary.

Fibrosis.

The importance of Rac1 in tissue repair is illustrated by the delayed wound healing response in mice with a specific deletion of Rac1 in fibroblasts. These mice revealed reduced collagen production and fibrosis (Liu et al., 2009). In cultured Rac1-deficient fibroblasts, adhesion, spreading, and migration were significantly inhibited (Liu et al., 2009). Rac1-deficient fibroblasts showed decreased myofibroblast formation by reduced α-smooth muscle actin (stress fiber) expression as well as matrix contraction (pulling force). In addition, in vivo and in vitro, Rac1-deficient fibroblasts showed reduced generation of ROS (Liu et al., 2009). In a hepatic fibrosis model, Rac1 caused NOX1 activation, NOX4 upregulation, ROS generation, and subsequent liver fibrosis, showing again the involvement of the NOX (Aoyama et al., 2012). Myofibroblast conversion is characterized by increased contractility of the cells, which coincides with decreased ciliation of the cell. Cilia are membrane protrusions that sense environmental mechanical cues. Rac1 enhances deciliation via transforming growth factor-β (TGFβ), NOX4, and ROS, and sustained phosphorylation of MLC, which promoted the contractile myofibroblast phenotype (Rozycki et al., 2014). There is an important role for integrin-linked kinase (ILK) in myofibroblast formation and wound repair, since fibroblast-restricted inactivation of ILK in mice resulted in decreased myofibroblast conversion and decreased repair (Blumbach et al., 2010). ILK is recruited to integrin β1 in focal adhesions, mediating the communication between the cell and the extracellular matrix. When this communication is disturbed in ILK-deficient fibroblasts, decreased α-smooth muscle actin, TGFβ, and Rac1 activity is observed, whereas RhoA activity is enhanced. Exogenous TGFβ could restore myofibroblast formation, cell shape, and directional migration (Blumbach et al., 2010).

Fibroblasts also display AT1R-mediated signaling, which is one of the main activators of NOX signaling. AT1R activation induces myofibroblast formation and fibrosis, which is strongly dependent on NOX and ROS (Adam et al., 2010). A prominent downstream gene of AT1R signaling is TGFβ, which induces connective tissue growth factor expression, which in turn induces matrix (such as collagen) production (Che et al., 2008). Rac1 is involved in the production of connective tissue growth factor in fibroblasts, which could be completely abolished by AT1R antagonist losartan (generic name) (Cotton et al., 2007; Tsai et al., 2008; Loirand and Pacaud, 2010). AT1R-mediated signaling in fibroblasts also induced a proinflammatory milieu by the production of cytokine interleukin-6 and chemokine monocyte chemoattractant protein-1 (Tieu et al., 2011), which will provoke an immune response that contributes to wound repair or, when excessive, harming the repair response. Rac1 activation may easily be involved in this proinflammatory gene transcription pattern upon AT1R-mediated signal transduction, since this has also been described for other cell types (Marinkovic et al., 2014a,b). Collectively, it shows that Rac1 is one of the central mediators of wound healing by promoting fibrosis.

Proliferation and Migration.

Enhanced proliferation of adventitial fibroblasts is detrimental, such as in hypoxia-induced pulmonary hypertension. Excessive fibroblast proliferation under these conditions was caused by a Rac1/p38 signaling cascade (Carlin et al., 2007). Different statins inhibited adventitial fibroblast proliferation, and reduced p38 activation. The statin-mediated effect can be mimicked by inhibition of GGTase-1 or Rac1 (Carlin et al., 2007). Thus, as they did for VSMCs, statins inhibited proliferation via inhibition of GGTase-1 activity, probably preventing Rac1 activation. Rats infused with angiotensin-II (activating AT1R) displayed not only increased levels of active Rac1, collagen, and fibrosis but also presented enhanced proliferation and phosphorylation of signal transducer and activator of transcription 3 (STAT3), which was efficiently dampened by a AT1R-blocker or statin (Tsai et al., 2008; Tieu et al., 2011). Activation of AT1R-mediated JAK/STAT signaling was Rac1-dependent, and the statin blocked Rac1 localization to the membrane (Tsai et al., 2008). A Rac1-dependent STAT3-mediated proliferative response was also observed in our gut epithelial cell study, where cyclin D1 was ultimately increased (Marinkovic et al., 2014a).

Active Rac1 has been described as binding STAT3, which seems to target Rac1 to become phosphorylated and activated (Simon et al., 2000). This complex formation between Rac1 and STAT3 is not always found, however; RhoA, Rac1, and Cdc42 can all mediate phosphorylation and nuclear translocation of STAT3, independently of the interleukin-6 pathway (Debidda et al., 2005). In embryonic fibroblasts, tumor suppressor p53 stimulates actin cytoskeleton remodeling and limits lamellipodia formation, since in the absence of p53 these features were disturbed and migration was induced. P53 depletion induced STAT3, and inhibition of Rac1, actin-related protein 2/3, or NF-κB suppressed STAT3 and diminished lamellipodia formation (Guo et al., 2014). In addition, a role for STAT3 independent of its transcriptional activity was found. In rescue experiments with STAT3-deficient fibroblasts, introduction of STAT3 reduced the high activation level of Rac1 in these cells. Cytoplasmic STAT3 could bind to and thereby inhibit Rac1-specific GEF β-PIX, and normalize Rac1 activity (Teng et al., 2009). GEF β-PIX plays a key role in the negative regulation of focal adhesion maturation and promotion of lamellipodial protrusion formation, thus driving cell migration (Kuo et al., 2011). Collectively, it shows that Rac1/STAT3 activation is involved in fibroblast proliferation and migration, with different mechanisms of STAT3 regulation, and a negative feedback loop to inactivate Rac1.

Apart from angiotensin-II, bFGF is also known to induce proliferation and migration. Inhibition of Rac1 blocks bFGF-induced fibroblast migration, whereas inhibition of RhoA does not. Here, Rac1 activation was dependent on phosphoinositide 3-kinase activation and Rac1-induced downstream JNK activation. Inhibition of each step could inhibit fibroblast migration (Kanazawa et al., 2010). In endothelial cells and fibroblasts, Rac1-mediated JNK activation has been reported and, thus, also seems a common Rac1 activation pathway. In line with these data, high glucose could inhibit bFGF-induced Rac1/JNK-dependent fibroblast migration, which is presumably the cause of delayed wound repair in diabetic patients (Xuan et al., 2014). On the other hand, a neurotrophic factor called nerve growth factor (NGF) has been described as accelerating wound healing. NGF promoted the migration, but not proliferation, of dermal fibroblasts by inducing an increase in the activity of phosphoinositide 3-kinase/Akt/Rac1/JNK/ERK (Chen et al., 2014). Blockade with their specific inhibitors impaired the NGF-induced migratory response of fibroblasts.

Clearly, on the basis of the data presented above, targeting Rac1 for inhibition or activation in fibroblasts may depend on the type of pathology to be treated. Although Rac1-mediated fibrosis and migration/hyperplasia is necessary for wound healing, it is undesirable when there is excessive fibrosis or hyperplasia, which disturbs the normal function of the vasculature and/or organ perfusion.