By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Author for correspondence: Rhian M. Touyz, Kidney Research Centre, University of Ottawa/Ottawa Health Research Institute, 451 Smyth Rd, Ottawa, ON, KIH 8M5, Canada (fax 613 562 5487, e-mail firstname.lastname@example.org).
Abstract: The healthy endothelium prevents platelet aggregation and leucocyte adhesion, controls permeability to plasma components and maintains vascular integrity. Damage to the endothelium promotes endothelial dysfunction characterized by: altered endothelium-mediated vasodilation, increased vascular reactivity, platelet aggregation, thrombus formation, increased permeability, leucocyte adhesion and monocyte migration. Molecular processes contributing to these phenomena include increased expression of adhesion molecules, synthesis of pro-inflammatory and pro-thrombotic factors and increased endothelin-1 secretion. Decreased nitric oxide bioavailability and increased generation of reactive oxygen species (ROS) are among the major molecular changes associated with endothelial dysfunction. A critical source of endothelial ROS is a family of non-phagocytic nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, including the prototypic Nox2-based NADPH oxidases, Nox1, Nox4 and Nox5. Other possible sources include mitochondrial electron transport enzymes, xanthine oxidase, cyclooxygenase, lipoxygenase and uncoupled nitric oxide synthase (NOS). Cross-talk between ROS-generating enzymes, such as mitochondrial oxidases and Noxs, is increasingly implicated in cellular ROS production. The present review discusses the importance of endothelial ROS in health and disease and focuses on the major ROS-generating systems in the endothelium, namely uncoupled endothelial nitric oxide synthase and NADPH oxidases.
The endothelium comprises a simple squamous layer of cells that lines the inner surface of all blood vessels from the heart to the smallest capillary. It forms an interface between circulating blood and the vascular wall. The healthy endothelium prevents platelet aggregation and leucocyte adhesion and controls vascular permeability [1,2]. It is highly responsive to mechanical stimuli (stretch, shear stress, pressure), humoural agents [angiotensin II (Ang II), endothelin-1 (ET-1), aldosterone, bradykinin, thromoxane] and chemical factors [glucose, homocysteine, reactive oxygen species (ROS)] by releasing endothelial-derived mediators, such as nitric oxide, prostacyclin (PGI2), platelet-activating factor (PAF), C-type atrial natriuretic peptide and ET-1 to maintain vascular tone and structural integrity [3–6]. Nitric oxide, produced from endothelial nitric oxide synthase (eNOS) has potent vasodilatory, anti-inflammatory and anti-thrombotic characteristics. Endothelial cells also control the coagulation pathway through the production of molecules with anti-coagulant activities such as tissue plasminogen activator, urokinase plasminogen activator, tissue factor pathway inhibitor and thrombomodulin .
Under pathological conditions, the endothelium becomes decompensated such that protective mechanisms are overwhelmed by injurious processes leading to impaired endothelium-mediated vasodilation, increased vascular reactivity, platelet activation, thrombus formation, increased permeability, leucocyte adhesion and monocyte migration into the vascular wall [8,9]. These events underlie endothelial impairment, which by definition, is a functional and reversible alteration of endothelial cell function. Molecular mechanisms contributing to this include increased expression of adhesion molecules, increased synthesis of pro-inflammatory and pro-atherosclerotic factors, activation of the local renin-angiotensin system and increased ET-1 secretion [8–11]. Increased ROS bioavailability and dysregulated redox signalling (oxidative stress) together with decreased nitric oxide production because of reduced eNOS activity and increased nitric oxide consumption by ROS contribute to many of the molecular events underlying endothelial injury [8–12].
Endothelial dysfunction is a hallmark for vascular diseases and is associated with various pathologies including hypertension, diabetes, atherosclerosis, pulmonary hypertension, ischaemic heart disease and chronic kidney disease [13–16]. Impaired endothelial function may be an important predictor of cardiovascular outcomes and is an independent predictor of future events in patients with cardiovascular risk factors [17–19]. Endothelial impairment may be reversible with certain treatments. In patients with mild to moderate hypertension, angiotensin converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARB) and calcium channel blockers improved endothelial vasodilatory function [20–23]. These effects appeared to be independent of blood pressure lowering, because atenolol, a β-blocker, which reduced blood pressure did not influence endothelial function in hypertensive patients. Improvement in endothelial status in hypertensive patients is associated with a more favourable prognosis. This review highlights some molecular mechanisms contributing to endothelial dysfunction, focusing specifically on NOS/nitric oxide, Noxs, ROS and oxidative stress.
Nitric oxide synthase, nitric oxide and endothelial function
Nitric oxide, a potent endothelium-derived relaxing factor, is synthesized from the oxidation of l-arginine, mediated in a two-step process by NOS through the generation of N-hydroxyl l-arginine . Of the three characterized NOS isoforms, neuronal NOS, inducible NOS and eNOS, it is the endothelial isoform that is localized primarily in endothelial cells, closely associated with caveolae/lipid rafts, and critically involved in endothelial-derived nitric oxide [25,26]. Nitric oxide diffuses locally within endothelial cells, to the luminal surface of the endothelium, and into the smooth muscle cells of the vascular wall, where it signals through numerous downstream pathways via guanylate cyclase to generate cGMP and by direct S-nitrosylation of cysteine residues in proteins, to activate signalling molecules such as transcription factors NF-κB and AP-1 and to inhibit Ca2+-mediated vasoconstriction .
Multiple factors influence eNOS activity, including calcium/calmodulin, depalmitoylation of eNOS, displacement of caveolin-1 and release of eNOS from caveolae [28,29]. Changes in eNOS activity are mediated through Akt-dependent phosphorylation at Ser1177, which activates eNOS in response to stimuli, while phosphorylation at Thr459 decreases activation . Physiologically many stimuli promote eNOS expression to maintain nitric oxide production, whereas pathologically, eNOS protein levels may be normal or even increased, despite endothelial dysfunction and reduced nitric oxide generation. These findings have led to the concept of ‘eNOS uncoupling’, characterized by the discrepancy between eNOS protein levels and nitric oxide production, with a switch in the enzymatic activity of eNOS to generate superoxide () rather than nitric oxide . The role of NOS and nitric oxide in endothelial biology has been extensively reviewed [25–29], and only an overview is presented here with a focus on the role of uncoupled NOS and ROS generation.
Uncoupling of eNOS
Under physiological conditions, NOS, in the presence of co-factors l-arginine and tetrahydrobiopterin (BH4), produces nitric oxide. In the vascular wall, BH4 bioavailability is regulated by the enzyme GTP-cyclohydrolase I (GTPCH I), through a salvage pathway from the synthetic pterin, sepiapterin and via oxidative degradation of BH4 to BH2, which is inactive for eNOS co-factor function . BH4 is an essential co-factor for all three NOS isoforms, and all are capable of ‘uncoupling’. Basal NOS activity correlates with the amount of BH4 bound to the enzyme. BH4 regulates NOS at multiple levels: it shifts the NOS haem iron to a high-spin state, promoting arginine binding; BH4 bound to NOS acts as a redox-sensitive co-factor; BH4 increases substrate affinity of NOS; BH4 participates in electron transfer as it converts to BH3, and importantly, it stabilizes NOS dimer formation, which is critical for functional NOS activity [32–34]. When BH4 levels are reduced, the dimer structure is altered such that the oxidase domain yields molecular uncoupling and the catalytic activity becomes functionally ‘uncoupled’. Enzymatic reduction of molecular oxygen by eNOS no longer couples to l-arginine, resulting in the generation of injurious rather than protective nitric oxide (fig. 1). This eNOS uncoupling contributes to increased ROS production and decreased nitric oxide formation and consequent endothelial dysfunction . Although there is extensive evidence demonstrating that uncoupled NOS generates , these findings are based on in vitro studies, and to date, there is still no definitive proof that such phenomena also occur in vivo.
Factors that influence BH4 bioavailability and hence NOS uncoupling include oxidative stress, which decreases expression of GTPCH and depletes nicotinamide adenine dinucleotide phosphate (NADPH), Ang II, which decreases BH4 by down-regulating dihydrofolate reductase, an enzyme involved in BH4 biosynthesis from the salvage pathway, homocysteine, which reduces de novo synthesis of BH4, and folate and vitamin C, which increase BH4 bioavailability  (table 1).
Table 1. Factors that influence BH4 bioavailability.
Endothelial NOS uncoupling, which shifts the nitroso-redox balance with adverse consequences, has been demonstrated in experimental models of hypertension and is implicated in atherosclerosis and endothelial dysfunction in low-density lipoprotein receptor-deficient mice (LDLR−/−) fed with a high-salt and high-fat diet [33,34]. Uncoupled eNOS has also been demonstrated in essential hypertension, in patients with hypercholesterolaemia and in chronic smokers [35,36]. Reduced BH4 bioavailability and consequent NOS uncoupling and oxidative stress have been implicated in patients with diabetes, coronary artery disease, cardiac failure and ischaemia-reperfusion injury (table 2). In many of these conditions, exogenous BH4 has been shown to improve endothelial function and cardiovascular risk [32–35]. Accordingly, BH4 may have potential in the treatment of hypertension and associated cardiovascular diseases. While previously difficult to use clinically because of chemical and thermal instability and cost, newer methods to synthesize stable BH4 suggest its use as a therapeutic agent. The recent development of BH4 in the form of a thermostable and photostable compound (Biomarin, San Francisco, CA, USA) has facilitated large clinical trials that should yield important information on the therapeutic use of BH4 in cardiovascular disease. In addition to exogenous BH4 preventing uncoupling of NOS, some classical anti-hypertensive drugs, such as calcium channel blockers, ACE inhibitors, ARB and beta blockers, may induce beneficial effects by preventing eNOS uncoupling.
Table 2. Conditions associated with uncoupled nitric oxide synthase and decreased tetrahydrobiopterin [32–35,78–80].
Chronic kidney disease
Systolic cardiac failure
Radiation-induced endothelial dysfunction
ROS and vascular biology
Reactive oxygen species are generated during the reduction of oxygen and include unstable free radicals (species with an unpaired electron) such as and non-free radicals, such as hydrogen peroxide (H2O2). ROS, originally considered to induce damaging cellular effects, are now recognized to have important physiological actions such as the induction of host defence genes, activation of transcription factors and stimulation of ion transport systems . In the vascular system, ROS play a physiological role in controlling endothelial function, vascular tone and vascular integrity and a pathophysiological role in inflammation, hypertrophy, proliferation, apoptosis, constriction, migration, fibrosis, angiogenesis and rarefaction, important factors contributing to endothelial dysfunction, vascular contraction and arterial remodelling in cardiovascular diseases.
Molecular processes underlying ROS-induced cardiovascular injury involve activation of redox-sensitive signalling pathways. Superoxide anion and H2O2 stimulate mitogen-activated protein kinases (MAPK), tyrosine kinases, Rho kinase and transcription factors (NFκB, AP-1 and HIF-1) and inactivate protein tyrosine phosphatases . ROS also increase intracellular free Ca2+ concentration ([Ca2+]i) and up-regulate protooncogene and pro-inflammatory gene expression and activity. These phenomena occur through oxidative modification of proteins by altering pivotal amino acid residues, by inducing protein dimerization and by interacting with metal complexes such as Fe–S moieties . Changes in endothelial cell redox state through glutathione and thioredoxin systems may also influence intracellular signalling. Oxidative stress was originally defined as molecular damage because of an imbalance between pro-oxidants and antioxidants . More recently, the definition has expanded to include the concept that molecular damage is because of impaired redox signalling and equilibrium.
ROS – basic concepts
Reactive oxygen species are produced as intermediates in reduction-oxidation (redox) reactions leading from O2 to H2O. Of the ROS generated in endothelial cells, and H2O2 appear to be particularly important. In biological systems, is short-lived and unstable owing to its rapid reduction to H2O2 by superoxide dismutase (SOD), of which there are three mammalian isoforms, copper/zinc SOD (SOD1), mitochondrial SOD (SOD2) and extracellular SOD (SOD3) . The charge on the superoxide anion makes it unable to cross cellular membranes except possibly through ion channels. H2O2 has a longer lifespan than , which is relatively stable and is easily diffusible within and between cells. The main source of H2O2 in vascular tissue is the dismutation of . This reaction can be spontaneous or it can be catalysed by SOD.
ROS production in the endothelium
Reactive oxygen species are products of normal cellular metabolism and derive from many sources in different cellular compartments. Enzymatic sources of ROS in endothelial cells are uncoupled NOS, xanthine oxidoreductase, mitochondrial respiratory enzymes and NADPH oxidase [42–45]. Of these, Nox family NADPH oxidases are particularly important and are the focus of the present review.
Nox family NADPH oxidases
Nicotinamide adenine dinucleotide phosphate oxidase has as its primary function the formation of ROS and accordingly is termed a ‘professional’ ROS producer. NADPH oxidase was originally considered to be expressed only in phagocytic cells. It is now evident that there is a family of NADPH oxidases, based on homologues of the catalytic subunit gp91phox that are functionally active in non-phagocytic cells. The new homologues, along with gp91phox, are designated the Nox family of NADPH oxidases  and contribute to endothelial cell ROS production (fig. 2). The prototypical gp91phox-containing phagocytic NADPH oxidase (now termed Nox2) comprises five subunits: p47phox (‘phox’ stands for phagocyte oxidase), p67phox, p40phox, p22phox and the catalytic subunit gp91phox (90). In basal conditions, p47phox, p67phox and p40phox exist in the cytosol, whereas p22phox and gp91phox are in the membrane, where they occur as a heterodimeric flavoprotein (cytochrome b558). Upon stimulation, p47phox and p67phox form a complex that translocates to the membrane, where it associates with cytochrome b558 to assemble the active oxidase, which transfers electrons from the substrate to O2 forming . Activation also requires Rac 2 (or Rac 1) and Rap 1A.
The mammalian Nox family comprises seven members: Nox1, Nox2, Nox3, Nox4, Nox5, Duox1 and Duox2  that are encoded by separate genes. All are transmembrane proteins that have conserved structural properties and that transport electrons across biological membranes to reduce O2 to . Nox1, Nox2, Nox4 and Nox5 have been identified in endothelial cells (fig. 2).
Relative to other endothelial Noxs, Nox1 is expressed at low levels. Nox1 requires p22phox, p47phox [or its homologue NoxO1 (Nox organizer 1)], p67phox [or its homologue NoxA1 (Nox activator 1)] and Rac1 for its activity . Novel Nox1 regulators have been identified, tyrosine kinase substrates 4 and 5 (Tks4 and Tks5), which resemble p47phox and NoxO1 and which interact with NoxA1. Nox1 localizes with p22phox in caveolae/lipid rafts and is expressed at low levels in physiological conditions. Nox1-derived is increased in a stimulus-dependent manner, involving interactions between NoxO1, NoxA1 and p22phox. It is also regulated by the redox chaperone protein disulphide isomerase in vascular smooth muscle cells. In cultured endothelial cells, Nox1 is up-regulated by mechanical factors (shear stress), vasoactive agents (Ang II, aldosterone) and growth factors (EGF, PDGF) . Ang II-induced induction of Nox1 may involve mitochondria, suggesting an interaction between Nox1 and mitochondria, possibly through a Ca2+-dependent mechanism.
Nox1 expression/activity is increased in the vasculature in models of cardiovascular disease including hypertension, atherosclerosis, diabetes and hypercholesterolaemia . Nox1 is also implicated in ageing. In human endothelial cells, Nox1 is implicated in ROS-induced senescence . In aged rats, endothelial Nox1 is up-regulated and influences vascular function in an NADPH oxidase inhibitor-reversible manner . Endothelial Nox1 may also be important in angiogenesis. Nox1-deficient mice exhibit impaired angiogenesis, and Nox1 expression and activity are increased in mouse and human endothelial cells upon angiogenic stimulation . These processes engage Nox1 signalling through peroxisome proliferator-activated receptor (PPARα/NF-κB). In the lung, endothelial Nox1 may be important in ROS production and cell death of the alveolocapillary barrier during hyperoxia and acute lung injury . Despite extensive experimental data implicating Nox1 in cardiovascular disease, there is still little information in humans, although expression of Nox1 and NoxA1is increased in human atherosclerotic vessels .
Nox2 is the catalytic subunit of the respiratory burst oxidase in phagocytes, but is also expressed in endothelial cells . Nox2 is unstable without p22phox and requires p47 phox, p67phox and Rac1/2 for its full activation. In neutrophils, Nox2 localizes to intracellular and plasma membranes, and in endothelial cells, it also localizes with the cytoskeleton, lipid rafts/caveolae and perinuclear compartment. Endothelial Nox2 is activated by Ang II, shear stretch . Recently identified novel mechanisms in endothelial cell Nox2 activation involve peroxiredoxin 6 (Prdx6), a bifunctional enzyme with glutathione peroxidase and phospholipase A2 activities , and adenosine A2A receptor . Endothelial Nox2-derived ROS induces activation of MAP kinases, heat shock factor-1 and plasminogen activator inhibitor-1 , implicated in endothelial dysfunction and vascular remodelling.
Transgenic mice with endothelial-specific Nox2 over-expression exhibit increased blood pressure responses, enhanced endothelial impairment and exaggerated vascular remodelling in response to Ang II compared with wild-type counterparts , implicating the importance of endothelial-specific Nox2 in Ang II-induced hypertension . Nox2-deficiency protects against ischaemia in conditions of increased oxidative stress, through improved neovascularization, preserved activation of the vascular endothelial growth factor (VEGF)/nitric oxide angiogenic pathway and improved functional activities of endothelial progenitor cells . In humans, Nox2-based NADPH oxidase has been shown to play a role in endothelial function, as patients with chronic granulomatous disease (CGD), who have an X-linked Nox2 mutation, exhibit a significant increase in forearm-mediated vasodilation with increased nitric oxide bioavailability , suggesting that Nox2-based NADPH oxidase influences endothelial function and nitric oxide biology in humans. Patients with CGD also exhibit blunted endothelial ischaemia/reperfusion injury suggesting a role for NADPH oxidase-derived ROS in human ischaemia/reperfusion injury ,
Nox4, of which four splice variants have been identified (Nox4B, NoxC, Nox4D and Nox4E), is found in many vascular cell types and seems to be particularly important in endothelial cells . Nox4 has been identified in the endoplasmic reticulum, mitochondria and nucleus of vascular cells. Nox4 does not require p47phox, p67phox, p40phox or Rac for its activation, although Nox R1 and Poldip2, Nox4-binding protein, may be important. Unlike Nox1 and Nox2, Nox4 is constitutively active, producing primarily H2O2 rather than . It contributes to basal ROS production through its constitutive activity and to increased ROS generation, when stimulated by Ang II, glucose, TNFα and growth factors. The pathophysiological role of endothelial Nox4 remains unclear, because both damaging and protective effects have been attributed to this Nox.
Nox4-derived ROS has been implicated in hypertension, atherosclerosis and cardiovascular and renal complications of diabetes and in remodelling of pulmonary arteries important in hypoxia-dependent development of pulmonary hypertension. Nox4 has also been suggested to play a role in cellular senescence and ageing. This may occur, in part, through loss of replicative potential, a process that is partly independent of telomere attrition . In human umbilical vein endothelial cells (HUVECs), microRNA-146a (miR-146a) is involved in a premature senescence-like phenotype through direct targeting of Nox4 .
Recent studies suggest that Nox4 may have protective actions. In mice with a genetic deletion of Nox4 or a cardiomyocyte-targeted over-expression of Nox4, basal cardiac function was normal in both models, but Nox4-null animals developed exaggerated contractile dysfunction, hypertrophy and cardiac dilatation during exposure to chronic overload, whereas Nox4-transgenic mice were protected . In transgenic mice with endothelium-targeted Nox4 over-expression, blood pressure was lower and vasodilatory responses enhanced compared with wild-type counterparts . These beneficial effects of Nox4 were attributed to increased H2O2 production , but decreased ROS production may also be important, as long-term application of arterial laminar shear stress reduces endothelial formation and Nox4 expression .
Protective effects of Nox4-based NADPH oxidase are unclear, but may relate to potential antioxidant capacity of Nox4 and to the species of ROS generated by Nox4. The Nox4 promoter possesses an antioxidant response element (ARE-like/Oct-1) binding site  that may impart antioxidant activity. In brain endothelial cells, Nox4 initiates a cell survival mechanism by increasing production of a gaseous antioxidant mediator carbon monoxide (CO) by constitutive haem oxygenase-2 . Nox4-derived H2O2 has been considered to be an endothelium-derived hyperpolarizing factor that promotes vasodilation in some vascular beds .
Nox5 is the most recently identified of the Nox enzymes and has unique features compared with other family members. While all Noxs are present in mice, rats and man, the rodent genome does not contain the nox5 gene . Unlike other vascular Noxs, Nox5 possesses an amino-terminal calmodulin-like domain with four binding sites for Ca2+ (EF hands), and unique to Nox5 is its lack of requirement for p22phox or other subunits for its activation. Nox5 is directly regulated by intracellular Ca2+ ([Ca2+]i), the binding of which induces a conformational change leading to enhanced ROS formation . Five splice variants of Nox5 have been identified, Nox5α, β, δ and γ, and a truncated variant (Nox5-S or ε). All five Nox5 isoforms are expressed in endothelial cells and generate ROS in response to thrombin, ionomycin, PDGF, Ang II and ET-1 [73,74]. Endothelial Nox5 up-regulation promotes cellular proliferation and organization of endothelial cells into three-dimensional structures resembling capillary networks. Increased expression of Nox5 transgenes in endothelial cells inhibits the extracellular actions of nitric oxide and prevents endothelium-dependent relaxation. Endothelial Nox5 has also been shown to increase eNOS activity through eNOS/hsp90 binding, which may prevent eNOS uncoupling promoting increased nitric oxide generation . The biological significance of vascular Nox5 is unknown, although it has been implicated in cell proliferation, angiogenesis and migration and in oxidative damage in atherosclerosis .
Endothelial dysfunction, characterized by altered endothelium-mediated vasodilation, increased vascular reactivity, platelet activation, thrombus formation, increased permeability and leucocyte adhesion, is an early event in many cardiovascular diseases. These phenomena are due in large part to reduced endothelial cell nitric oxide bioavailability and oxidative stress. Factors contributing to such processes include decreased NOS expression/activity, NOS uncoupling and increased activation of Noxs. Of the seven Nox isoforms, Nox 1, 2, 4 and 5 are functionally active ROS-generating systems in the endothelium. Activation of vascular Noxs is complex and is regulated by phosphorylation and assembly of different oxidase subunits using cholesterol-rich microdomains and cytoskeletal proteins as scaffolding platforms. Whereas Nox1, 2 and 5 have been implicated in oxidative damage and endothelial dysfunction, Nox4 may be both injurious and protective. Despite the growing field of Nox biology, there is still confusion as to the exact function of vascular Noxs. With the recent development of genetically engineered mouse models of Nox isoforms  and the ability to knockdown Nox proteins with siRNA/shRNA, the molecular mechanisms and functional significance of these oxidases and their role in redox signalling and oxidative stress will be elucidated. As we learn more about the biology of Noxs, the interest in targeting specific isoforms as therapeutic strategies in cardiovascular disease may become a reality . However, to date, the clinical utility of such potential strategies still awaits confirmation.
Work from the author’s laboratory was supported by grants 44018 and 57886, both from the Canadian Institutes of Health Research (CIHR). RMT is supported through a Canada Research Chair/Canadian Foundation for Innovation award.