Nitric oxide: Orchestrator of endothelium-dependent responses

Abstract

The present review first summarizes the complex chain of events, in endothelial and vascular smooth muscle cells, that leads to endothelium-dependent relaxations (vasodilatations) due to the generation of nitric oxide (NO) by endothelial nitric oxide synthase (eNOS) and how therapeutic interventions may improve the bioavailability of NO and thus prevent/cure endothelial dysfunction. Then, the role of other endothelium-derived mediators (endothelium-derived hyperpolarizing (EDHF) and contracting (EDCF) factors, endothelin-1) and signals (myoendothelial coupling) is summarized also, with special emphasis on their interaction(s) with the NO pathway, which make the latter not only a major mediator but also a key regulator of endothelium-dependent responses.

Key words::

• Atherosclerosis
• cyclo-oxygenase
• cytochrome P450 mono-oxygenase
• diabetes
• endothelin
• hypertension
• lipoxygenase
• nitric oxide synthase
• oxidative stress
• potassium channels
• soluble epoxide hydrolase
• tetrahydrobiopterin

Abbreviations
ADMA	=	asymmetric dimethyl-L-arginine
ADP	=	adenosine diphosphate
AT1	=	angiotensin receptor type 1
AT2	=	angiotensin receptor type 2
BH2	=	dihydrobiopterin
BH4	=	tetrahydrobiopterin
BKCa	=	calcium-activated potassium channels of large conductance
cAMP	=	cyclic adenosine monophosphate
CaV	=	voltage-activated calcium channel
cGKI	=	cGMP-dependent protein kinase I
cGMP	=	cyclic guanosine monophosphate
COX	=	cyclo-oxygenase
DDAH-2	=	dimethylarginine dimethylaminohydrolase-2
DHET	=	dihydroxyeicosatrienoic acids
DHFR	=	dihydrofolate reductase
DOCA-salt	=	deoxycorticosterone acetate-salt-induced rats hypertensive rats
ECE	=	endothelin-converting enzyme
EDCF	=	endothelium-derived contracting factor
EDHF	=	endothelium-derived hyperpolarizing factor
EDRF	=	endothelium-derived relaxing factor
EETs	=	epoxyeicosatrienoic acids
ET-1	=	endothelin-1
ETA	=	endothelin-receptors subtype A
ETB	=	endothelin-receptors subtype B
GTP	=	guanosine triphosphate
GTPCH	=	guanosine triphosphate cyclohydrolase
H2O2	=	hydrogen peroxide
H2S	=	hydrogen sulfide
HETE	=	hydroxyeicosatetraenoic acids
IKCa	=	calcium-activated potassium channels of intermediate conductance
IP	=	prostacyclin receptor
KCa	=	calcium-activated potassium channels
KIR	=	inwardly rectifying potassium channel
KLF-2	=	Krupper-like factor-2
L-NMMA	=	N(G)-monomethyl L-arginine
LOX	=	lipoxygenase
M3	=	muscarinic receptor type 3
NADPH	=	nicotinamide adenine dinucleotide phosphate
NO	=	nitric oxide
NOX	=	NAD(P)H oxidases
eNOS	=	endothelial nitric oxide synthase
iNOS	=	inducible nitric oxide synthase
nNOS	=	neuronal nitric oxide synthase
NO2-	=	nitrite ion
NO3-	=	nitrate ion
O2.-	=	superoxide anion
ONOO-	=	peroxynitrite
P450	=	cytochrome P450 mono-oxygenases
PDE-5	=	phosphodiesterase-5
PGD2	=	prostaglandin D2
PGE2	=	prostaglandin E2
PGF2α	=	prostaglandin F2α
PGI2	=	prostaglandin I2 or prostacyclin
PGH2	=	prostaglandin H2, endoperoxide
PKG	=	protein kinase G
PLA2	=	phospholipase A2
PPAR	=	peroxisome proliferator-activated receptor
PRMT-I	=	protein arginine N-methyltransferase type I
SERCA	=	sarcoplasmic reticulum calcium ATPase
SHR	=	spontaneously hypertensive rat
SHRSP	=	stroke-prone spontaneously hypertensive rat
ROS	=	reactive oxygen species
sEH	=	soluble epoxide hydrolase
sGC	=	soluble guanylyl cyclase
SKCa	=	calcium-activated potassium channels of small conductance
THETA	=	trihydroxyeicosatrienoic acids
TP	=	thromboxane/endoperoxide receptor
TRP	=	transient receptor potential channel
TXA2	=	thromboxane A2

Key messages

• NO, an endothelium-derived powerful vasodilator and a potent anti-thrombotic agent is also an important regulator of the other endothelium-mediated responses (EDHF, EDCF, endothelin).

• Multiple potential therapeutic targets have been identified along the L-arginine–NOS pathway, in particular at the levels of the substrate (L-arginine), the NO-generating enzyme (eNOS), the soluble guanylyl cyclase (the main target of NO), and cGMP (the main effector of its action).

• Additional therapeutic opportunities have been identified in the regulation of the other endothelium-mediated responses in various cardiovascular diseases such as hypertension, diabetes, or atherothrombosis.

Introduction

The endothelium plays a major role as local regulator of vascular tone by releasing vasodilator (endothelium-derived relaxing factors (EDRF)) and vasoconstrictor (endothelium-derived contracting factors (EDCF)) that diffuse to the underlying vascular smooth muscle to cause moment-to-moment changes in their degree of contraction. Of those endothelium-derived mediators, nitric oxide (NO) is the key player as it not only induces vasodilatation but also greatly protects the vascular wall by inhibiting platelet aggregation (in synergy with prostacyclin), expression of adhesion molecules (and thus penetration of white cells in the intima and media), and oxidation of lipids. These actions converge in healthy arteries to prevent the unwanted coagulation of the blood and the inflammatory response that leads to the atherosclerotic process. When the endothelial cells lose the ability to release enough NO or when the bioavailability of the gaseous mediator is reduced, endothelial dysfunction ensues, a hallmark of vascular disease, and indeed a prognostic marker of severe cardiovascular events (1–5). Besides its pivotal role in modulating the function of blood components and vascular smooth muscle directly, NO also appears to interact closely with the other EDRFs (endothelium-derived hyperpolarizing factors (EDHFs)) and to control the production/action of endothelium-derived vasoconstrictors (endothelin-1 (ET-1) and vasoconstrictor prostanoids (EDCFs)), making it the gate-keeper of endothelial function in terms of vasomotor control (6–9). This review summarizes the current knowledge concerning the role of NO in vascular homeostasis, focusing on its intrinsic effect but also discussing its modulatory role in controlling the other endothelium-derived mediators. Because of space limitations, we have selected to refer only to key or recent original articles and to reviews by others and us, to which the reader is referred for a more complete coverage of the literature.

Nitric oxide

Furchgott and Zawadzki (10) first showed that the presence of the endothelial lining is required to observe the relaxation of isolated strips of rabbit aorta in response to acetylcholine. They attributed the relaxation to the calcium-dependent release of an endothelial factor that was later identified as NO (4,11–13). Subsequently, Moncada and colleagues demonstrated that the enzymatic origin of NO production in the endothelial cells involves the L-arginine–NO-synthase pathway (14). NO is not only a powerful vasodilator but is also a potent anti-thrombotic agent, and it plays important biological roles in the gastrointestinal, respiratory, nervous, and immune systems (15,16).

Nitric oxide synthases

The production of NO is associated with three different subtypes of biosynthetic enzymes. These were cloned, characterized, and termed according to the chronology of their characterization: NO-synthase I (or neuronal, nNOS, NOS-1) (17,18), NO-synthase II (or inducible, iNOS, NOS-2) (19,20), and NO-synthase III (or endothelial, eNOS, NOS-3) (21,22). The two constitutive isoforms (nNOS and eNOS) are stimulated by an increase in intracellular calcium concentration, while iNOS is a calcium-independent enzyme (23). In fact, the expression of the various NOS isoforms is much more diverse than their name would suggest. For instance, the three NOS can be present in the vascular wall. Thus, nNOS is expressed not only in perivascular nerves but can also be detected in endothelial and smooth muscle cells (24–26). The expression of iNOS has been documented in all nucleated cells in the cardiovascular system. In the blood vessel wall, this includes endothelial and smooth muscle cells as well as fibroblasts, leukocytes, and mast cells (27). Furthermore, eNOS is expressed not only in endothelial cells but also in cardiac myocytes and platelets (28,29). Nevertheless, the preponderant NOS isoform in healthy endothelial cells is the calcium-dependent eNOS, and eNOS is predominantly expressed in endothelial cells (30).

The three NOS enzymes contain a reductase and an oxygenase domain. In the presence of molecular oxygen, they catalyze a five-electron oxidation of one of the guanido nitrogens of L-arginine. They all require the co-factors NADPH, flavin adenine dinucleotide (FAD), flavin mononucleotide (FMN), heme, and tetrahydrobiopterin (BH₄) in order to catalyze this reaction (31). An essential feature of NOS, despite the ability of the reductase or the oxygenase domain to function independently, is that the NO synthase activity requires dimerization. The heme group of the enzyme plays an essential role in the formation of the homodimer, and BH₄ stabilizes this dimer once formed. The activation of eNOS requires also the calcium-dependent binding of calmodulin (32) (Figure 1).

Figure 1. NO biosynthesis and vascular smooth muscle targets. In endothelial cells, the L-arginine–endothelial nitric oxide synthase (eNOS) system and the synthesis via the de-novo synthesis and salvage pathways of tetrahydrobiopterin (BH₄), an essential co-factor of eNOS, are shown. ([Ca²⁺]_i = intracellular calcium concentration; Ca²⁺-sensitivity = sensitivity to calcium of contractile proteins; cGMP = cyclic guanosine monophosphate; cGKI = cGMP-dependent protein kinase I; Cyt-C = cytochrome C; DHFR = dihydrofolate reductase; FAD = flavin adenine dinucleotide; FMN = flavin mononucleotide; GTP = guanosine triphosphate; NADPH = nicotinamide adenine dinucleotide phosphate; NO₂^- = nitrite ion; NO₃^- = nitrate ion; O₂.^- = superoxide anion; ONOO^- = peroxynitrite; P450 = cytochrome P450 mono-oxygenases; SERCA = sarcoplasmic reticulum calcium ATPase; sGC = soluble guanylyl cyclase; XO = xanthine oxidase).

Transcriptional and post-translational regulation of eNOS in endothelial cells

Since NO is a gaseous freely diffusible radical with a high biological reactivity, its synthesis has to be tightly controlled in order to be delivered at the right time, the right place, and in the right amount (23).

The transcription of eNOS is under epigenetic control and can be regulated by various physical, chemical, and pharmacological stimuli. For instance, laminar shear stress is a major stimulus not only for the transcription of eNOS but also for the co-ordinated expression of many other genes that not only control vascular tone but also promote an atheroprotective, anti-thrombotic, and anti-inflammatory phenotype. This phenomenon involves the up-regulation of an endothelial transcription factor KLF-2 (33–35). Since eNOS is only one component of a large signaling complex that, besides calmodulin and caveolin-1, involves many other different proteins, its activity can be altered by several overlapping post-translational regulatory modifications, including lipidation, phosphorylations (at various serine, threonine, and tyrosine residues), S-nitrosylations, acetylation, and protein–protein interactions (34,36,37). The subcellular location of eNOS is critical for its state of activation, and the trafficking of the enzyme is under multiple levels of regulation (38).

NO storage and NO-synthase-independent production of nitric oxide

NO is a reactive radical with a half-life of a few seconds in tissues and physiological fluids, since it reacts with superoxide anion to form peroxynitrite or is rapidly inactivated by oxyhemoglobin to form nitrate and methemoglobin (39–41). However, NO can also exert remote and long-lasting effects. Indeed, it can be either transported by protein carriers or stored locally. S-nitrosylation/denitrosylation of protein cysteine residues is a dynamic process than can affect hundreds of proteins in intact cells and albumin and hemoglobin in the blood. Not only can S-nitrosylation affect the activity of these proteins, but the generation of S-nitrosoglutathione allows the transfer of NO activity at distance from the site where it has been generated (42). Besides S-nitrosylation of local proteins, NO can also be stored by forming protein-bound dinitrosyl iron complexes to be displaced later by low-molecular weight thiols (43) (Figure 1).

In addition, NO can also be oxidized into nitrite, either spontaneously in the presence of oxygen or enzymatically by the copper storage plasma enzyme, ceruloplasmin (44,45). Furthermore, nitrite can originate directly from ingested food or indirectly from the reduction of nitrates either by reductase enzymes from commensal bacteria colonizing the mouth and the digestive track or by xanthine oxidase (46). When nitrite is present, deoxyhemoglobin in particular and heme-globins in general, as well as xanthine oxidase, cytochrome P450 mono-oxygenase, cytochrome C, or eNOS itself, can act as nitrite reductases and produce NO (46–48) (Figure 1).

Multiple targets of nitric oxide

Most of the physiological effects of NO are associated with the activation of soluble guanylyl cyclase (49,50). In addition, NO forms S-nitrosylated proteins (see above) (42) and can interact with the heme iron moiety of various proteins. For instance, the reduction in the intracellular concentration of calcium, which contributes to the inhibition of platelet activation and initiates vascular relaxation, involves the uptake of calcium by sarcoplasmic reticulum ATPase (SERCA) (51,52). This guanylyl cyclase-independent effect is attributed to the redox regulation of SERCA via the NO-dependent S-glutathiolation of a reactive cysteine (53,54). NO binds reversibly to cytochrome C oxidase and competes with oxygen. The NO-mediated inhibition of cytochrome C oxidase activity plays a fundamental role in the regulation of mitochondrial signaling and the cellular response to hypoxia (55,56) (Figure 1).

Soluble guanylyl cyclase is a heterodimer, consisting of an α- and a β-subunit, the latter containing the prosthetic ferrous heme group (Fe²⁺) that catalyzes the conversion of guanosine-5′ triphosphate (GTP) to the second messenger cyclic 3′,5′-guanosine monophosphate (cGMP) (57). The latter predominantly activates cGMP-dependent protein kinases, cGKI being the preferential isoform expressed in vascular smooth muscle cells, platelets, and to a lesser extent endothelial cells (50,58). Furthermore, cGMP interacts with phosphodiesterases and regulates the concentration of cyclic 3′,5′-adenosine monophosphate (cAMP) (59), and activates cyclic nucleotide-gated cation channels (60). In vascular smooth muscle cells, activation of cGKI produces relaxation by multiple mechanisms, which either reduce the intracellular calcium concentration or affect the calcium sensitivity of the contractile proteins (50) (Figure 1).

L-arginine–NOS pathway and potential therapeutic targets in cardiovascular diseases

Various potential therapeutic targets have been identified along the L-arginine–NOS pathway, in particular at the levels of the substrate (L-arginine), the NO-generating enzyme (eNOS), soluble guanylyl cyclase (the main target of NO), and cGMP (the main effector of its action).

Availability of the substrate L-arginine. When L-arginine is deficient, eNOS can generate both superoxide anions and NO, leading to the detrimental production of peroxynitrite (61) (Figures 1 and 2). A few clinical studies have shown that L-arginine supplementation could bring some benefit under pathophysiological conditions, including hypercholesterolemia and hypertension, although the elevated doses required are not appropriate for a standard therapy.

Figure 2. L-arginine–NO-synthase–soluble guanylyl cyclase pathway and potential sites of therapeutic intervention. The circled numbers (from 1 to 14) indicate potential sites of therapeutic intervention. 1) L-arginine supplementation. 2) Inhibition of protein arginine N-methyltransferase type I (PRMT-I) to prevent the formation of asymmetric dimethyl-L-arginine (ADMA). 3) Increasing the expression and/or the activity of dimethylarginine dimethylaminohydrolase-2 (DDAH-2) to facilitate ADMA catabolism. 4) Inhibition of arginase-2 to prevent L-arginine metabolism. 5) Increasing the expression and/or activity of endothelial nitric oxide synthase (eNOS). 6) Designing drugs that stimulate endothelium-derived nitric oxide release. 7) Enhancing the expression and/or activity of guanosine triphosphate cyclohydrolase (GTPCH), the rate-limiting enzyme for tetrahydrobiopterin synthesis (BH₄), or direct supplementation with BH₄, or with its precursor sepiapterin. 8) Enhancing the expression and/or activity of dihydrofolate reductase (DHFR), involved in BH₄ regeneration. 9) Scavengers of reactive oxygen species (ROS), anti-oxidants. 10) Inhibition of the activity and/or expression of enzymes that generate ROS such as NAD(P)H oxidases (NOX), cyclo-oxygenases (COX), lipoxygenases (LOX), or cytochrome P450 mono-oxygenases (P450). 11) Enhancing the expression and/or activity of enzymes that metabolized ROS such as superoxide dismutase (SOD) or catalase (or alternatively synthesis of mimetics). 12) Stimulators of soluble guanylyl cyclase (sGC). 13) Activators of sGC. 14) Inhibitors of phosphodiesterase-5 (PDE-5). (BH₂ = dihydrobiopterin; CAT-1 = cationic amino acid transporters; CaV = voltage-activated calcium channel; cGMP = cyclic guanosine monophosphate; EC = endothelial cell; FAD = flavin adenine dinucleotide; FMN = flavin mononucleotide; O₂.^- = superoxide anion; ONOO^- = peroxynitrite; PKG = protein kinase G; VSMC = vascular smooth muscle cell).

In vivo, true L-arginine deficiency is unlikely to occur (62). However, L-arginine can be locally submitted to competition with endogenously produced analogs such as symmetric dimethyl-L-arginine, asymmetric dimethyl-L-arginine (ADMA), and N(G)-monomethyl L-arginine (L-NMMA). L-arginine is transported through the plasma membrane by the system Y+, a member of the family of cationic amino acid transporters, and is in competition with other amino acids such as L-lysine but also with the endogenous analogs mentioned above. Furthermore, ADMA and L-NMMA directly compete with L-arginine at the NOS site and are inhibitors of the enzyme (63). ADMA represents a novel independent predictor for all causes of cardiovascular mortality (62). Free dimethylarginines are the products of proteolytic degradation of arginine-methylated proteins following the preferential action of protein arginine N-methyltransferase type I (PRMT-I). ADMA is mainly metabolized by dimethylarginine dimethylaminohydrolase (DDAH), whereby the DDAH-2 isoform is the preponderant enzyme expressed in endothelial cells. Endothelial cells also express and contain arginases (mainly arginase-II), which metabolize L-arginine to L-ornithine and urea. Arginase-II competes with eNOS for substrate, and its expression and/or activity are enhanced in cardiovascular diseases. Therefore, PRMT-1, arginase II (inhibition), and DDAH-2 (activation) are promising therapeutic targets for the treatment of cardiovascular disease (64–66) (Figure 2).

Proper expression and functionality of eNOS. A reduced expression of eNOS could be responsible for a decreased NO production. However, in most situations associated with endothelial dysfunction the expression of eNOS is increased paradoxically. The augmented endothelial level of reactive oxygen species, in particular those of hydrogen peroxide, which increases the expression of eNOS at the transcriptional and translational levels, likely explains this paradox. An endothelial dysfunction associated with an increased expression of eNOS shows that the ability to generate NO is reduced or/and that its bioavailability is decreased. The reduction in NO generation can be attributed to the eNOS uncoupling phenomenon whereby NOS itself becomes a source of superoxide anions exacerbating endothelial dysfunction. The decrease in NO bioavailability is associated with an oxidizing environment, i.e. superoxide anion generation from different sources, and the formation of peroxynitrite (67) (Figures 1 and 2).

The two independent enzymatic steps required for the eNOS-dependent conversion of L-arginine to NO necessitate the presence of tetrahydrobiopterin (BH₄). Decreased endothelial levels of this essential pteridine co-factor are not only responsible for a reduction in NO production but also for the uncoupling of eNOS. In such an uncoupled state, electrons normally flowing from the reductase domain of one subunit to the oxygenase domain of the other subunit are diverted to molecular oxygen rather than to L-arginine, resulting in production of superoxide anion rather than NO (68). Homocysteine is a potent risk factor for endothelial dysfunction and atherosclerosis, and some of its effects may be due to inhibition of BH₄ de-novo synthesis (69) (Figure 2).

Direct supplementation with BH₄, or with its precursor sepiapterin, improves endothelial function in numerous animal models of cardiovascular disease. Controversial findings have been published concerning the therapeutic efficacy of BH₄, although some positive observations have been reported in patients with hypertension, diabetes, hypercholesterolemia, coronary insufficiency and in smokers (69). Enhancing the expression or the activity of GTP-cyclohydrolase (GTPCH-I), the first and rate-limiting step of de-novo BH₄ synthesis, or that of dihydrofolate reductase (DHFR), involved in BH₄ regeneration, may prevent the occurrence of endothelial dysfunction. Alternative strategies include the administration of folic acid, which enhances the binding affinity of BH₄ to NOS, stabilizes BH₄ chemically, and by stimulating DHFR enhances its regeneration. Preventing peroxynitrite formation and therefore BH₄ oxidation or facilitating the recycling of BH₄ regeneration with vitamin C may also be considered (67,70) (Figure 2).

In various animal models, inhibition or genetic deletion of eNOS causes accelerated atherosclerosis, abdominal aortic aneurysms, and ischemic heart disease (71–74), indicating that endogenous eNOS plays an important anti-atherosclerotic role. However, apoE-/- mice over-expressing eNOS develop paradoxically larger atherosclerotic lesions than control apoE-/- mice (75). This paradox is explained by the fact that in the presence of oxidative stress, reactive oxygen species from various sources (kindling radicals) and the formation of peroxynitrite will lead to eNOS uncoupling and the subsequent eNOS contribution to superoxide production (bonfire radicals). Therefore, up-regulation of eNOS in itself is a double-edged sword strategy. However, if the up-regulation of eNOS is associated with an increased availability of essential co-factors such as BH₄, eNOS uncoupling can be prevented. Some compounds, such as AVE9488 and AVE3085, which enhance eNOS promoter activity, possess such a co-ordinated activity and validate the concept that enhanced transcriptional expression of eNOS can be beneficial in atherosclerosis and possibly in other cardiovascular diseases (76) (Figure 2).

Furthermore, statins (3-hydroxy-3-methylglutaryl-CoA reductase inhibitors), by their ‘pleiotropic’ (independent of their lipid-lowering action) effects, and the polyphenol, resveratrol, which activates SIRT-1, increase the expression of the transcription factor KLF-2 (77,78), which is also stimulated by laminar shear stress (34). KLF-2 is described as a master regulator of the transcription of anti-inflammatory, anti-thrombotic, and vasoactive genes, including eNOS, leading to a vasoprotective phenotype (35). Additionally, statins and polyphenols as well as drugs such as fibrates, metformin, and peroxisome proliferator-activated receptor-gamma (PPARγ) agonists activate AMP-kinase. This could lead to an increased expression of eNOS and to post-translational modifications, in particular phosphorylations, that enhance the protein expression and the enzymatic activity of eNOS (79–81) (Figure 2).

Increasing the levels of NO and/or its bioavailability. Besides preventing eNOS uncoupling (see above), different therapeutic strategies can be designed to restore proper NO levels. These include drugs that stimulate the release of NO by endothelial eNOS, NO donor drugs, anti-oxidant compounds, drugs that boost the anti-oxidant defense mechanisms, and inhibitors of the enzyme(s) involved in the generation of reactive oxygen species (Figure 2).

Some anti-hypertensive drugs can stimulate the release of endothelial NO. This is the case for nebivolol, a third-generation β-blocker (82,83), and various dihydropyridines, by a mechanism that is unrelated to the calcium channel-blocking properties of these compounds (84–86). Inhibitors of angiotensin-converting enzyme indirectly produce endothelium-dependent relaxations in vitro or vasodilatations, in vivo, by preventing the degradation of endogenous bradykinin, a potent stimulator of eNOS (87–89). Independently of the metabolism of bradykinin, inhibitors of angiotensin-converting enzyme can also enhance endothelium-dependent relaxations by regulating the sequestration and the sensitivity of the bradykinin B₂ receptors (90,91). Similarly, in various arteries, including human coronary arteries, antagonists of the angiotensin AT1 receptor unmask endothelium-dependent, NO-dependent vasodilator effects due to the preferential activation of AT2 receptor subtype by the high levels of circulating angiotensin-II (92–94). Additionally, chronic treatments with inhibitors of angiotensin-converting enzyme or antagonists of the angiotensin AT1 receptor increase eNOS expression in both animals with high blood pressure and in hypertensive humans (95,96). The limited information available in animals with the renin inhibitor, aliskiren, suggests that this way of inhibiting the renin–angiotensin system is also likely to improve endothelium-dependent relaxations and to increase eNOS expression (97,98).

Besides uncoupled eNOS, potential sources of superoxide anions in the vascular wall include NAD(P)H oxidases, xanthine oxidase, cyclo-oxygenases, lipoxygenases, cytochrome P450 mono-oxygenases, and the mitochondrial respiratory chain. The production of reactive oxygen species is increased in patients with essential hypertension, renovascular hypertension, malignant hypertension, and pre-eclampsia. In essential hypertension, an increase in the levels and/or activity of vascular NAD(P)H oxidase is a predominant source of excess reactive oxygen species (99). Superoxide anions not only decrease NO bioavailability, but the peroxynitrite generated exacerbates vascular injury by oxidation of sulfhydryl groups as well as nitration and hydroxylation of aromatic groups, including tyrosine, tryptophan, and guanine (1,100). Thus, the scavenging of reactive oxygen species with anti-oxidants and inactivation of superoxide anions with superoxide dismutase mimetics seem obvious ways of increasing NO bioavailability (Figure 2). However, if such a strategy has been successful in various animal models of hypertension, large clinical trials in hypertensive patients showed that chronic anti-oxidant therapy does not produce a major reduction in arterial blood pressure and does not improve the associated morbidity and mortality, with the possible exception of the long-term intake of polyphenols that are present in red wine, fruits, and vegetables, i.e. the so-called beneficial effect of a Mediterranean diet (101,102). The reasons underlying these failures are still not completely understood.

Direct inhibition of the generation of reactive oxygen species might be a better alternative. Major efforts toward the synthesis of NADPH oxidase inhibitors have been performed in academic institutions and pharmaceutical companies. However, it is not yet clear which isoform(s) of NADPH should be inhibited, and the inhibitors available are still not specific, not very potent, and possess poor pharmacokinetics properties (103) (Figure 2).

The activation of the renin–angiotensin system and the resulting stimulation of AT1 receptors are a major stimulus for NAD(P)H oxidase and reactive oxygen species (ROS) production, also in hypertensive humans (99,104). Preventing the generation of reactive oxygen species, for instance by deleting a subunit of the NADPH oxidase, causes resistance to angiotensin II-induced hypertension and markedly reduces the associated endothelial generation of superoxide anions (105). Statins, by another of their ‘pleiotropic’ actions, decrease the expression of NADPH subunits, improving, therefore, the balance between NO and reactive oxygen species. These endothelial effects may contribute to the therapeutic benefice exerted by these compounds, although no clinical data exist demonstrating that the reversal of endothelial dysfunction is associated with a reduction in cardiovascular events (95,106–109).

Substituting NO with NO donor drugs started empirically in the late nineteenth century with the nitrate compound nitroglycerin to alleviate the symptoms of angina. However, the development of nitrate tolerance, a complex multifactorial phenomenon, limits their continuous clinical use and is associated with the generation of oxidative stress and endothelial dysfunction (110). Different chemical classes of NO donors exist. Some of those release NO spontaneously and may be of therapeutic interest beyond coronary disease and heart failure (111). S-nitrosothiols are not prone to tolerance and are not cross-tolerant with nitrates. They possess vasodilator and anti-aggregating properties and may boost the anti-oxidant capacity of the plasma and the vascular wall (111).

Alternatively, hybrid compounds possessing dual activities may be of therapeutic interest to treat endothelial dysfunction in hypertension and the ensuing end-organ damage. Earlier compounds such as nicorandil, an organic nitrate that also opens potassium channels, or nipradilol, a non-selective adrenoceptor blocker with weak β-adrenoceptor-antagonistic and NO-releasing properties are still predominantly prescribed for angina and glaucoma, respectively (107,111). However, among the new hybrid compounds that have been designed, i.e. NO-releasing anti-adrenergic drugs, NO-releasing-dihydropyridines, NO-releasing statins, NO-releasing antagonists of the angiotensin AT1 receptor, or NO-releasing inhibitors of angiotensin-converting enzyme, some may present a therapeutic interest in hypertension and other cardiovascular diseases (112). For instance, a NO-releasing derivative of the angiotensin-converting enzyme inhibitor enalapril, NCX-899, has demonstrated a superior activity when compared to enalapril alone in animal models of heart failure and hypertension (113,114). Proper clinical trials must be undertaken in order to demonstrate whether or not this strategy brings a therapeutic benefit when compared to the original therapeutic entity.

Soluble guanylyl cyclase and cGMP production. When NO binds to the heme moiety of soluble guanylyl cyclase, the activity of the enzyme, leading to the production of the signaling molecule cGMP, is enhanced 400-fold. Thus, regulation of NO bioactivity is achieved not only by modulation of eNOS and curtailing oxidative stress but also by affecting soluble guanylyl cyclase (115) or the half-life of cGMP, which is readily hydrolyzed by phosphodiesterases (116) (Figure 2).

NO cannot activate soluble guanylyl cyclase when the enzyme is heme-free (117) or when the heme moiety is oxidized (Fe³⁺), following oxidative stress (118,119). However, some pharmacological agents can interact with soluble guanylyl cyclase either under the native form or when oxidized. They generate cGMP independently of NO and are potent vasodilators and inhibitors of platelet aggregation (120,121). Some of these compounds are undergoing clinical trials in acute decompensated heart failure, peripheral arterial occlusive diseases, and pulmonary hypertension. In addition, the potential therapeutic indications of these agents may include hypertension, myocardial ischemia, erectile dysfunction, atherosclerosis, and thrombosis (120,121) (Figure 2).

Phosphodiesterases are a diverse family of enzymes that hydrolyze cyclic nucleotides. In the cardiovascular system, phosphodiesterase 5 (PDE5) was the first identified selective cGMP esterase and is the major isoform involved in the hydrolysis of the cGMP pools generated by the activation of soluble guanylyl cyclase (122,123). Specific PDE5 inhibitors are currently approved for erectile dysfunction but, in the future, their therapeutic indications may also include pulmonary hypertension, heart failure, and essential hypertension (124–126) (Figure 2).

In conclusion: The NO pathway is a major regulator of cardiovascular homeostasis. A decrease in NO bioavailability or function is associated with most of endothelial dysfunctions observed in virtually all cardiovascular diseases. However, the mechanisms underlying the deficiencies in the NO pathway are multiple and thus so are the new potential therapeutic targets. Only oncoming clinical trials will define which of these new targets and in which indication(s) they will fulfill their therapeutic potential.

NO and EDHF

Nature and action of EDHF

Endothelium-dependent vasodilatations resistant to inhibitors of NO and prostacyclin synthesis are mediated by hyperpolarization leading to the closure of voltage-gated calcium channels in the vascular smooth muscle cells (127,128). The resulting decrease in intracellular calcium concentration leads to relaxation. This hyperpolarization has been proposed initially to be caused by a factor, of at that time unknown chemical nature, released by the endothelium and termed endothelium-derived hyperpolarizing factor (EDHF). Structurally very different factors including eicosanoids (129,130), hydrogen peroxide (131), C-type natriuretic peptide (132), potassium ions released by endothelial KCa channels (133), NO itself (134), and H₂S (135) have been proposed to act as diffusible EDHFs that cause opening of K⁺ channels of the underlying smooth muscle (for extensive review and debate see (8,136)). However, it soon appeared that the term EDHF may be misleading because endothelium-dependent hyperpolarization can also be caused by non-chemical, electrotonical coupling of endothelial and vascular smooth muscle cells via myoendothelial gap junctions without the need of the generation of a chemical diffusible mediator (137). In aggregate, EDHF(s) pathways are presumably multiform, and their roles differ between species and within the vascular tree (Figure 3). In any case, EDHF-signaling is a powerful vasodilator system, particularly in small-caliber arteries and arterioles in which NO-signaling is not the predominant pathway (138–140). However, in several cardiovascular pathologies (hypertension, hypercholesterolemia, atherosclerosis, and diabetes) disturbances in EDHF-signaling contribute to endothelial dysfunction and thus arterial malfunction (for in-depth review see (128,141)). This section introduces the major candidate molecules and pathways leading to endothelium-dependent hyperpolarizations and their interactions with NO. From a more medical perspective, it focuses on how EDHF(s) systems fit into the complex setting of arterial malfunction and discusses whether or not drug-targetable EDHF(s) systems can theoretically be exploited for novel endothelium-specific treatments of cardiovascular disease involving endothelial dysfunction.

Figure 3. NO and EDHF. Endothelium-dependent relaxations resistant to inhibitors of NO synthases and cyclo-oxygenases, leading to the hyperpolarization of the underlying smooth muscle cells, are associated with different mechanisms: the metabolism of arachidonic acid via the lipoxygenases (LOX) or cytochrome P450 mono-oxygenases (P450) pathways and the non-chemical EDHF pathways, which, following the calcium-calmodulin-dependent activation of endothelial potassium channels, involve either the direct transfer of negative charges (e⁻) via the myoendothelial gap junction or the accumulation of potassium ions in the intercellular space and the subsequent activation of the smooth muscle Na⁺/K⁺-ATPase and/or the inwardly rectifying potassium channel. The circled numbers (from 1 to 7) indicate the principal steps in the development of endothelium-dependent hyperpolarizations and the potential sites of therapeutic intervention. The privileged therapeutic options, inhibitors of soluble epoxide hydrolase (sEH) and activators of endothelial potassium channels, are indicated by a broad arrow. 1) Agonists interacting with their receptor and physical forces such as shear stress increase endothelial intracellular calcium concentration. 2) Activation of transient receptor potential channels (TRP) contributes to calcium entry. 3) Activation of endothelial calcium-activated potassium channels of small (SK_Ca or KCa2.3) and intermediate conductance (IK_Ca or KCa3.1); the endothelial cell hyperpolarization increases the driving force for calcium entry. 4) Calcium-dependent activation of the endothelial nitric oxide synthase (eNOS); epoxyeicosatrienoic acids, P450 metabolites of arachidonic acid (AA), enhance NO production. 5) EETs and LOX metabolites, such as hydroxyeicosatetraenoic acids (HETE) and trihydroxyeicosatrienoic acids (THETA), can be released by endothelial cells and activate potassium channels of the smooth muscle cells to produce their hyperpolarization and relaxation. EETs are metabolized by sEH to form the generally less active dihydroxyeicosatrienoic acids (DHET). 6) Electrotonical coupling between endothelial and vascular smooth muscle cells via myoendothelial gap junctions clustered at specific endothelial projections toward the vascular smooth muscle cells. 7) Vascular smooth muscle hyperpolarization inhibits the activation of voltage-gated calcium channels (CaV) and decreases the intracellular calcium concentration.

Eicosanoids. Endothelial lipoxygenases and cytochrome P450 enzymes metabolize the polyunsaturated fatty acid, arachidonic acid, to eicosanoids which stimulate the activity of the large-conductance KCa channels (BK_Ca or maxi K) of vascular smooth muscle and thus could act as diffusible endothelium-derived hyperpolarizing factors (129,130).

Epoxygenases. The cytochrome P450 epoxygenases 2C and 2J and their subforms produce epoxyeicosatrienoic acids (EETs; in varying amounts 5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET), all of which open BK channels of vascular smooth muscle and cause relaxation (129) (Figure 3). They are considered to contribute predominantly to endothelium-dependent dilatations attributed to EDHF in porcine and bovine coronary arteries (142,143) and in the human forearm microcirculation (144). However, they are not equally effective in all human vascular beds (e.g. the cutaneous microcirculation (145)). In patients with essential hypertension (144) and primary hyperparathyroidism (146), the production of EETs appears to take over, at least partially, during endothelial dysfunction related to NO deficiency.

EETs may contribute to the systemic hemodynamic control. Indeed, a lower arterial blood pressure has been observed in male mice, which lack the EET-degrading enzyme soluble epoxide hydrolase (sEH) (147). Transgenic mice with endothelial over-expression of the human CYP2J2 and CYP2C8 epoxygenases, and resulting enhanced EET synthesis, exhibit augmented dilator responses to acetylcholine. Their arterial blood pressure tends to be lower than in wild-type animals and increases less upon administration of a NO synthase inhibitor (148), suggesting a particular relevance and efficacy of this system during systemic hypertension. Moreover, pharmacological inhibition of sEH ensuing higher EET levels leads to a reduction in arterial blood pressure in experimental angiotensin II hypertension (149,150). The specific inhibitor of sEH, AR9281 (149), when tested in clinical trials for efficacy in mild to moderate hypertension, failed to have blood pressure-lowering efficacy, presumably because of the low in-vivo potency of the compound at the dosage used. However, AR9281 and analogs with an improved profile (151) are effective in experimental pulmonary hypertension (152,153), atherosclerosis (154–156), and have additional vascular protective effects in brain (157,158) and heart (159,160) and in diabetes (161). Moreover, genetic sEH deficiency in mice and pharmacological inhibition of the enzyme with the improved sEH inhibitor t-AUCB prevent hyperglycemia, favor insulin secretion, and protect pancreatic islets against apoptosis (161). As to atheroprotective effects, both pharmacological inhibition and genetic sEH deficiency reduce neointima formation in hyperlipidemic mice (162). The former decreases neointima formation in the presence but not in the absence of the endothelium (163). In addition, t-AUCB protects against ischemia–reperfusion injury (164). In deoxycorticosterone acetate (DOCA)-salt hypertensive mice, genetic sEH deficiency and t-AUCB have also beneficial effects on the kidneys by reducing inflammation and injury (165), although they augment albuminuria in mice with progressive renal disease (166). Thus, these results highlight the efficacy of sEH inhibitors in renal hypertension and inflammation but perhaps question their therapeutic utility in cardiovascular disease associated with chronic renal failure. Besides the potential cardiovascular protective actions of sEH inhibitors, they may also be considered anti-inflammatory drugs because of their ability to augment the anti-inflammatory actions of aspirin and of inhibitors of synthesis of pro-inflammatory 5-lipoxygenase-derived leukotrienes (167) and to reduce inflammatory pain (151).

Lipoxygenases. 12-HETE, a product of the lipoxygenase isoform 12-LO expressed in the arterial wall, is a vasodilator and enhances BK channel activity in vascular smooth muscle (130). Another isoform, 15-LO, generates HEETA and hydroxyepoxyeicosatrienoic acid (THETA) which activate a KCa-like channel (168). However, HETE, derived from lipoxygenase but also from cytochrome P450, and cysteinyl-leukotrienes can evoke contractions of vascular smooth muscle cells. Endothelium-dependent contractions involving a lipoxygenase pathway have been reported in various arteries and veins in response to neuro-humoral mediators, mechanical stretch, and ischemia (1–5). Therefore, their true contribution as vasodilator molecules in vivo is still uncertain but they are not likely to contribute to the control of basal arterial blood pressure since the administration of lipoxygenase inhibitor BW755C to rats and rabbits or deletion of the 12/15-LO gene in mice does not result in marked hemodynamic alterations (130). Nonetheless, the insertion of human 15-LO-1 gene in rabbits enhances the depressor responses to the classical endothelial agonist, acetylcholine (169). In angiotensin-II-hypertensive animals, lipoxygenase-inhibitors also lower arterial blood pressure due to a diminished production of vasoconstrictor HETE(s) (130). An enhanced production of 5-HETE, 12-HETE, and 15-HETE metabolites has been observed in placental tissue of pre-eclamptic women, suggesting either a compensatory or a pathophysiological role of these eicosanoids in pregnancy-induced hypertension (170). Moreover, the 12/15-LO inhibitor baicalein and its derivatives have favorable effects in animal models of ischemic stroke (130). In keeping with this on first sight Janus-faced vasoactive profile of HETE(s), it remains uncertain whether HETE(s) are beneficial or detrimental in hypertension (Figure 3).

Non-chemical EDHF pathways. Electrotonical coupling between endothelial and vascular smooth muscle cells is a simple but effective route for direct transfer of negative charges and may help to explain why EDHF-signaling is particularly prominent in smaller arteries (139,140,171), since their wall, consisting of only one or few layers of smooth muscle cells, represents a small electrical sink. Here endothelial hyperpolarization mediated by small/intermediate conductance Ca²⁺-activated potassium channels (KCa3.1 and KCa2.3) appears to be the starting-point (133,172,173). In addition, K⁺-release mediated by perhaps KCa3.1 (133), which is located at myoendothelial junctions (174,175), increases the local concentration of the ion in the inter-myoendothelial space and activates inwardly rectifying (K_IR) channels and the Na⁺/K⁺ -pump to elicit hyperpolarization of the smooth muscle (133). This EDHF-signaling involving electrotonical coupling—with, as initial trigger, endothelial hyperpolarization mediated by KCa3.1 and KCa2.3—is a Ca²⁺ -dependent process and involves calmodulin which serves as the Ca²⁺ -sensing channel subunit (141,176) (Figure 3). KCa3.1 channels sense preferentially Ca²⁺ originating from stores after receptor stimulation. By contrast, KCa2.3 channels sense Ca²⁺ influx through ‘osmo/mechanosensitive’ transient receptor potential (TRP) channels (177,178), as suggested by the reduced flow-induced vasodilatation observed in genetic deficiencies of both KCa2.3 (172) and TRPV4 (179,180), and appear required to produce hyperemia in working skeletal muscle in order to meet the metabolic demands (181). The role of KCa3.1 and KCa2.3 in systemic cardiovascular homeostasis is indicated by the higher arterial blood pressure in mice lacking one or both of these channels (172,182,183).

In several models of hypertension, hypercholesterolemia, atherosclerosis and diabetes as well as in some human cardiovascular pathologies, alterations of the KCa3.1/KCa2.3-EDHF system appear to contribute to endothelial dysfunction (184-187). For example, mesenteric arteries of spontaneously hypertensive mice exhibit compromised endothelium-dependent hyperpolarization and relaxations resulting from defective expression/activity of KCa2.3 and inwardly-rectifying K⁺ channels and KIR as well as from a change in caveolin-1 oligomers (188). By contrast, EDHF-mediated responses involving KCa3.1 are intact. Similar results are obtained in mesenteric arteries of stroke-prone spontaneously hypertensive rats (SHRSP), in which actually the expression of KCa3.1 is enhanced, possibly to compensate for the KCa2.3 deficiency (189).

Although the roles of KCa3.1 and KCa2.3 in human cardiovascular regulation are not properly defined yet, there is evidence for endothelial defects related to alterations of KCa3.1 functions in isolated coronary arterioles after cardioplegic arrest and reperfusion (184) and in the skeletal muscle microvasculature (185). The downstream electrotonic transfer of hyperpolarization via myoendothelial gap junctions may also be impaired in small myometrial arteries from women with pre-eclampsia (186).

Small molecule activators of these channels may provide a way to improve endothelial function. Indeed, NS309, an activator of KCa3.1, re-establishes EDHF-responses in mesenteric arteries from type 2 diabetic rats (190). Furthermore, the KCa3.1/KCa2.3 opener SKA-31, with a 5-fold higher potency on KCa3.1 over KCa2.3, lowers arterial blood pressure in angiotensin-II hypertensive mice and improves the defective endothelium-dependent vasodilatation in carotid arteries of eNOS-deficient mice (191–193). Moreover, inhibition of endothelial KCa promotes endothelium-dependent contractions (194), which further augments the potential of KCa3.1/KCa2.3 activators to counteract endothelial dysfunction. Thus, the KCa3.1/KCa2.3–EDHF-dilator system provides promising perspectives in terms of therapeutic improvement of endothelial function and/or arterial blood pressure. However, compounds like NS309 and SKA-31 are not optimized, and their bioavailability is still poor (192).

Modulation by NO

In a number of blood vessels, EDHF-mediated responses become obvious only when the synthesis of NO is inhibited (128,195). Thus, EDHF(s) are able to take over in the case of ‘classical’ endothelial dysfunction associated with a loss of NO synthesis (e.g. in arteries with regenerated endothelium (196) or in eNOS-deficient mice (197)), thus demonstrating strong compensatory efficiency of EDHF(s). Conversely, activation of EDHF pathways can impact on the production and/or action of endothelium-derived NO:

EETs. Pharmacological inhibition of sEH ensuing higher EET levels leads to a reduction of arterial blood pressure in experimental angiotensin-II hypertension (149,150). This blood pressure-lowering action depends on intact NO-synthesis as it is prevented by systemic administration of L-NAME (150), a finding that further supports the notion that, besides being EDHF(s), higher EET levels improve NO-formation, and thus that EET and NO are part of the same pathways and/or act synergistically.

In addition to its direct EDHF actions on BK channels, 5,6-EET improves endothelial Ca²⁺ - signaling (198) and likely also NO synthesis in an autocrine fashion (150). In fact, 5,6-EET, 11,12-EET, 14,15-EET, and, surprisingly, their degradation products dihydroxyeicosatrienoic acids (DHETs), which have been suggested to have no biological activity whatsoever, cause vasodilatation by stimulating endothelial NO formation and NO-dependent vasodilatation (150) (Figure 3).

KCa3.1/KCa2.3 channels. Endothelium-dependent hyperpolarization, in response to agonists that stimulate G protein-coupled receptors, are associated with an increase in endothelial [Ca²⁺]_i. Furthermore, substances that increase endothelial [Ca²⁺]_i in a receptor-independent manner, such as calcium ionophore, also produce endothelium-dependent hyperpolarization of the smooth muscle cells. The principal mechanism that sustains the opening of endothelial K_Ca channels, following stimulation, is the capacitive calcium entry elicited by the depletion of calcium stores (7,129,180). KCa3.1/KCa2.3 activity may be required for proper eNOS activation or NO-signaling (172,199). This idea is based on the assumption that endothelial hyperpolarization mediated by the KCa3.1/KCa2.3 channels also increases the driving force for Ca²⁺ influx through TRP channels or, and perhaps also, through store-operated channels after receptor activation and Ca²⁺ release from the endoplasmic reticulum. This Ca²⁺ influx-promoting action of endothelial hyperpolarization thus also increases Ca²⁺ -dependent eNOS activation and NO formation (Figure 3). Likewise, prevention of any endothelial hyperpolarization by non-vasospasmogenic K⁺ levels impairs endothelium-dependent and NO-mediated vasodilatations (200). However, there are also other reports challenging the biophysical dogma according to which membrane hyperpolarization promotes Ca²⁺ influx in endothelial cells (199). Irrespectively, the KCa3.1 activator, NS309, not only elicits EDHF responses by producing strong endothelial and smooth muscle hyperpolarization but also increases NO-signaling in porcine retinal arterioles (199) and NO formation in a human endothelial cell line (201). Moreover, deficiency of KCa2.3 channels in mice appears to impair the large NO-mediated vasodilatation in carotid arteries in response to acetylcholine and to shear stress stimulation, while the EDHF-mediated vasodilatation to the muscarinic agonist is still intact in KCa2.3-deficient mice (172). Mechanistically, this role of KCa2.3 for proper NO synthesis can be explained by the coupled compartmentalization of KCa2.3, eNOS, and Ca²⁺ influx channels of the TRP-type in caveolae (small Ω-shaped membrane invaginations) and/or intracellular caveosomes which—when shuttled to the cell membrane—may constitute a functional signalosome for synergistic NO/EDHF vasodilatation (141) (Figure 3).

The combined improvement of NO- and EDHF-signaling by endothelial KCa-activators may be useful considering the regionally different and vessel size-dependent contribution of the two vasodilator systems. Where EDHF-signaling is weak and that of NO strong, i.e. in large conduit arteries, an improvement of NO-signaling could be advantageous and could also be relevant in terms of preventing platelet aggregation and thrombosis. Where EDHF is strong and NO less dominant, i.e. in the arterioles of the microcirculation, the synergistic activation of both systems may produce a decrease in total peripheral resistance and exert anti-thrombotic actions.

In conclusion: Endothelial key enzyme and ion channels driving EDHF-signaling also stimulate NO synthesis by amplifying intracellular Ca²⁺ -signaling and eNOS activation. Thus, EDHF signals, besides acting as vasodilator stimuli per se, converge at least in part at the level of the NO pathway. From a medical and pharmacological perspective, pharmacologic stimulators of EDHF(s) systems (e.g. channel modulators with an improved selectivity and bioavailability profile compared to the known KCa3.1/KCa2.3 activators NS309 and SKA-31) could be useful to improve endothelial dysfunction in cardiovascular disease, not only by virtue of their own vasodilator potential, but also by improving NO formation. However, whether this principle applies to the treatment of human endothelial dysfunction warrants further translational research.

NO and endothelin-1

Release and action of endothelin-1

Endothelial cells can produce the potent vasoconstrictor peptide endothelin-1 (ET-1), one of the most potent vasoconstrictor substances known, and the release of the peptide may contribute to the long-term regulation of arterial blood pressure (202–209). However, its contribution to moment-to-moment changes in vascular tone under physiological conditions is not established (6,210), except in arteries of aged rats (211). Although under normal physiological conditions the production and/or the action of ET-1 cannot proceed unmatched (6,9), its generation can be up-regulated by a number of factors believed to play a role in vascular disease (203,206,208,211–213). Thus, the peptide may become progressively more important with aging, contribute to the abnormalities of pulmonary hypertension, and play a role in the terminal stages of vascular disease (206,208,211,213,214).

Modulation by NO

Augmenting the production of NO and/or increasing the intracellular concentration of cGMP reduce the expression and the production of ET-1 (6,9,215,216). Both exogenous and endothelium-derived NO reduces, in a cGMP-dependent manner, the contraction of vascular smooth muscle cells in response to stimulation of ET_A (and ET_B) receptors by ET-1 (217–221). The peptide also activates endothelial ET_B receptors, linked to the activation of eNOS by G_i proteins (6,9,222–228). Hence, augmented production of ET-1 by endothelial cells under physiological conditions will be offset by an augmented release of NO, which not only curtails the further generation of the peptide but also prevents its vasoconstrictor effects (Figure 4). The unleashing of the production of ET-1 by the absence of NO is exemplified best by the sustained increase in arterial blood pressure caused by eNOS inhibitors, which is nearly prevented by antagonists of ET receptors (6,9). The feedback inhibition of the production and/or action of ET-1 is not limited to NO, as the peptide also increases the release of prostacyclin and EDHF, both of which further reduce any activation of the vascular smooth muscle that it may cause. Thus, the endothelial response to ET-1 provides an explanation for the minor contribution of the peptide to the moment-to-moment regulation of vascular tone under physiological conditions. As the endothelial cells age and/or regenerate and thus lose their G_i protein-mediated responsiveness, the moderation of the production and action of ET-1 declines, and the peptide contributes more and more to the dysfunction characteristic of the terminal stages of vascular disease (9,229) (Figure 4).

Figure 4. NO and endothelin. Release and actions of endothelin-1 (ET-1) in the vascular wall. (AA = arachidonic acid; AVP = arginine vasopressin; cAMP = cyclic AMP; cGMP = cyclic GMP; COX = cyclo-oxygenases; ECE = endothelin-converting enzyme; ET_A and ET_B = endothelin-receptors; NO = nitric oxide; NOS = nitric oxide synthase; PGI₂ = prostacyclin; R = cell membrane receptor) (from Feletou et al., 2008, with permission). (Reference 9)

NO and EDCF

Release and action of EDCF

Increases in intracellular calcium concentration in the endothelial cells not only evoke the release of relaxing (NO and EDHF) but also that of endothelium-derived contracting factors (EDCF). The mechanical consequence of these concomitant releases depends on the balance between the resulting tendencies toward relaxation or contraction. The EDCFs contributing to moment-to-moment changes in vascular tone are the products of endothelial cyclo-oxygenase, mainly its COX-1 isoform (230–232). Indeed, increases in the endothelial intracellular calcium concentration activate the enzyme phospholipase A₂, which liberates arachidonic acid from the cell membrane providing the substrate for COX-1 (233). The breakdown of arachidonic acid by cyclo-oxygenase generates prostanoids (mainly endoperoxides and prostacyclin, but in certain cases also thromboxane A₂ and prostaglandin F_2α) and reactive oxygen species (234–239). These mediators activate TP receptors of the vascular smooth muscle to evoke their contraction (Figure 5) (240–242). The release of EDCF, and the resulting activation of TP receptors, contributes to the vasoconstrictor responses to adenine nucleotides and ET-1 (243–246).

Figure 5. Endothelium-derived contracting factors (EDCF). Endothelium-dependent contraction is likely to be comprised of two components: generation of prostanoids and ROS. Each component depends on the activity of endothelial COX-1 and the stimulation of the TP receptors located on the smooth muscle to evoke contraction. In the SHR aorta, there is an increased expression of COX-1 and EP3 receptors, increased release of calcium, ROS, endoperoxides, and other prostanoids, which facilitates the greater occurrence of endothelium-dependent contraction in the hypertensive rat. The necessary increase in intracellular calcium can be triggered by receptor-dependent agonists, such as acetylcholine or ADP, or mimicked with calcium-increasing agents, such as the calcium ionophore A23187. The abnormal increase in intracellular ROS can be mimicked by the exogenous addition of H₂O₂ or the generation of extracellular ROS by incubation of xanthine with xanthine oxidase. The circled numbers indicate potential sites of therapeutic intervention: 1) inhibition of cyclo-oxygenases; and 2) antagonists of TP receptors. (AA = arachidonic acid; ACh = acetylcholine; ADP = adenosine diphosphate; H₂O₂ = hydrogen peroxide; m = muscarinic receptors; P2Y = purinergic receptor isoform 2Y; PGS, prostaglandin synthases; PGD₂ = prostaglandin D₂; PGE₂ = prostaglandin E₂; EP3, prostaglandin E₂ receptor subtype 3; PGF_2α = prostaglandin F_2α; PGI₂ = prostacyclin; PGIS = prostacyclin synthase; PLA₂ = phospholipase A₂; ROS = reactive oxygen species; TXA₂ = thromboxane A₂; TXAS = thromboxane synthase; X + XO = xanthine plus xanthine oxidase) (from Tang and Vanhoutte, 2009, with permission). (Reference 239)

In certain arteries, ROS produced by the endothelium diffuse to the underlying vascular smooth muscle cells and cause them to contract, presumably by activating the cyclo-oxygenase that they contain (247). However, in most cases, ROS mainly facilitate rather than induce EDCF-mediated responses (238). Obviously, since superoxide anions scavenge NO (40), resulting in the formation of peroxynitrite, oxidative stress impairs endothelium-dependent relaxations, which per se favors the occurrence of endothelium-dependent contractions (239,248–250).

Under physiological conditions, the releases of NO (and/or EDHF) predominates to maintain a normal vascular tone. However, with age and diseases such as hypertension, diabetes, and atherosclerosis, the balance between the endothelium-derived relaxing and contracting mediators is lost and the generation of EDCF becomes prominent (5,230,251–254). This preponderance becomes a hallmark of endothelial dysfunction, in particular in hypertension (Figure 6).

Figure 6. Endothelium-dependent effects of acetylcholine in rat aorta. Left: endothelium-dependent relaxations in normotensive rats. Right: cyclo-oxygenase-dependent, endothelium-dependent contractions to acetylcholine in SHR aorta. (AA = arachidonic acid; AC = adenylyl cyclase; COX₁ = cyclo-oxygenase 1; IP = PGI₂ receptor; M = muscarinic receptor; PGH₂ = endoperoxides; PGI₂ = prostacyclin; PGIS = prostacyclin synthase; PLA₂ = phospholipase A₂; R = receptor; sGC = soluble guanylyl cyclase; SR = sarcoplasmic reticulum; TP = TP receptor; + = activation; - = inhibition) (from Félétou and Vanhoutte 2006, with permission from the American Physiological Society). (Reference 1)

Modulation by NO

Acute inhibition. In the aorta of the spontaneously hypertensive rat (SHR), endothelium-dependent contractions, whether induced by acetylcholine or a calcium ionophore, are immediately potentiated by inhibitors of eNOS, intra- or extracellular NO scavengers, and selective inhibitors of soluble guanylyl cyclase (255,256). Since the combination of eNOS inhibitors or NO scavengers with inhibitors of soluble guanylyl cyclase has no additive potentiating effects on endothelium-dependent contractions (256), the inhibition of EDCF-mediated responses by NO must result from a functional antagonism at the level of the vascular smooth muscle cells. This acute functional antagonism by endothelium-derived NO explains why most studies on endothelium-dependent contractions are performed in the presence of eNOS inhibitors, NO scavengers, and/or inhibitors of soluble guanylyl cyclase to optimize the EDCF-mediated responses (Figure 6).

Long-term inhibition. In the SHR aorta, previous exposure to NO, as provided by an exogenous donor (sodium nitroprusside), causes a time- and concentration-dependent subsequent inhibition of EDCF-mediated contractions (obtained in the presence of an eNOS inhibitor), which persists long after the disappearance of NO from the biophase (257). Thus, NO not only exerts an acute functional antagonism on endothelium-dependent contractions but also exerts a long-term inhibition of EDCF-mediated responses (257). The mechanism underlying the prevention by previous exposure to NO of subsequent endothelium-dependent contractions is identical for acetylcholine (which activates M3 muscarinic receptors) and the calcium ionophore, A23187 (which opens pores in the cell membrane and permits the entry of extracellular calcium in cells). Indeed, previous exposure to the exogenous NO donor caused a comparable reduction in subsequent EDCF-mediated responses evoked by both agonists (257), suggesting that sodium nitroprusside interferes with an event downstream of the surge in calcium required to initiate endothelium-dependent contractions. The long-term inhibition by previous exposure to the NO donor of subsequent endothelium-dependent contractions is reduced by an inhibitor of soluble guanylyl cyclase, demonstrating the involvement of cyclic GMP (257). Since direct contractions of the vascular smooth muscle evoked by a specific TP receptor agonist were not blunted by a previous exposure to the NO donor (257), the cyclic GMP-dependent long-term inhibition exerted by NO takes place within the endothelial cells.

In conclusion: NO and EDCF behave acutely as functional antagonists at the level of vascular smooth muscle. In addition, NO exerts a long-term negative modulation of the bioavailability of EDCF and thus curtails the occurrence of endothelium-dependent contractions. Under conditions of reduced NO production, for example due to eNOS uncoupling or augmented release of the endogenous eNOS inhibitor asymmetric dimethylarginine (ADMA) (258), or in the case of a reduced NO bioavailability resulting from enhanced formation of superoxide anions (40), the emergence of EDCF-mediated responses becomes a prominent sign of endothelial dysfunction (9,254).

General conclusions

In conclusion, the available experimental evidence highlights endothelial NO as the major orchestrating vasodilator molecule on which other vasodilator systems (EDHFs) and also detrimental EDCFs converge in an integrative fashion. The therapeutic utility of pharmacologic manipulations of the synthesis and action of NO to treat hypertension and cardiovascular disease remains likely but is still puzzling. Indeed, it appears that an improvement of eNOS function cannot be simply achieved by supplementation of substrate and co-factors for the enzyme. This may be explained by the more complex nature of arterial malfunction in cardiovascular disease states, involving disintegration/inefficacies at the level of gene transcription, enzyme function, and fate of NO. Moreover, NO actions are likely to be curtailed at the smooth muscle level. From the pharmacotherapeutic perspective, the question whether endothelial dysfunction, malfunction of vascular smooth muscle, and the resulting pathological vessel remodeling can be reversed or at least improved is of on-going interest. Results from experimental pharmacological interventions targeting the complex interconnections between NO, EDHFs, and EDCFs systems provide new venues for improved NO-signaling and thus for novel treatments and are an incentive for further laboratory and clinical research and for pharmaceutical discovery.

Declaration of interest: The authors state no conflict of interest and have received no payment in preparation of this manuscript.