NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Félétou M. The Endothelium: Part 2: EDHF-Mediated Responses “The Classical Pathway”. San Rafael (CA): Morgan & Claypool Life Sciences Publisher; 2011.

Cover of The Endothelium

The Endothelium: Part 2: EDHF-Mediated Responses “The Classical Pathway”.

Show details

Chapter 1Endothelium-Dependent Hyperpolarizations: The Classical “EDHF” Pathway

A lot of confusion has arisen in the field of endothelium-dependent hyperpolarizations due to the use of the acronym “EDHF” which implies that a single diffusible substance underlies this novel mechanism of relaxation of the vascular wall. In fact, as has been summarized in Part I Multiple Functions of the Endothelial Cells: Focus on Endothelium-Derived Vasoactive Mediators, numerous endothelium-derived factors, including nitric oxide and prostacyclin, can produce such hyperpolarization of the underlying smooth muscle cells, this effect contributing in varying degrees to the relaxing effect of these different substances. Endothelium-dependent relaxations, which are independent of the production of NO and prostacyclin, can be elicited by these diffusible factors of various origin (arachidonic acid metabolites, gaseous mediators, reactive oxygen species, peptides, etc.) and very often involve the hyperpolarization of the vascular smooth muscle cells. Confusingly, each of these factors, at one point or another, has been termed “endothelium-derived hyperpolarizing factor” or “EDHF.”

An additional pathway, which does not per se involve a diffusible factor, also plays an important role in the endothelium-dependent control of vascular tone. This pathway requires the activation of endothelial potassium channels and the hyperpolarization of the endothelial cells. This pathway should be referred to as “endothelium-dependent hyperpolarizations” or “EDHF-mediated responses” as there is, at present, no better name to characterize it [376,430,432].


Very early in the saga of the endothelium-dependent relaxations, pharmacological studies using various inhibitors of the metabolism of arachidonic acid suggested that at least three different pathways were involved [304]. At that time, only one of these was clearly identified, the cyclooxygenase-dependent production of prostacyclin [1071], but the others were unknown.

De Mey et al. [304] first proposed the existence of a third pathway because, in canine femoral arteries, the endothelium-dependent relaxations to acetylcholine obtained in the presence of indomethacin (excluding the involvement of prostacyclin) were abolished by mepacrine whereas the responses to thrombin and ATP were unaffected. Similarly, in the rabbit aorta, mepacrine inhibited acetylcholine-induced endothelium-dependent relaxation but did not affect that elicited by the calcium ionophore A 23187 [1434]. Furthermore, in canine arteries under various experimental conditions, the different susceptibility to antioxidants or to inhibitors of the arachidonic acid cascade strongly suggested the involvement of different relaxing factors [1224,1314,1318], which at that time were attributed to different arachidonic acid metabolites. However, the inhibitors used were very poorly specific, interfering, for instance, with intracellular calcium handling and/or potassium channel activation [430].

Independently, in the late 1970s, the group of Kuriyama [816,860,861], when measuring both changes in tension and the membrane potential of vascular smooth muscle cells, has observed that acetylcholine could produce a contraction in guinea-pig and rabbit coronary and mesenteric arteries with, paradoxically, a simultaneous hyperpolarization of the vascular smooth muscle cells. In guinea-pig basilar and coronary arteries, the hyperpolarization of the vascular smooth muscle cells was attributed to an increase in the membrane potassium conductance [778,816]. However, the modulatory role of the endothelial cells on vascular tone was then unforeseen. Later, in the guinea-pig mesenteric artery, Bolton et al. [123] demonstrated that acetylcholine not only caused direct contraction of the smooth muscle cells but also a concomitant endothelium-dependent relaxation which was accompanied by the endothelium-dependent hyperpolarization of the cell membrane of vascular smooth muscle. This initial observation has been confirmed in various blood vessels from different species [9395,121,227,427,428,683,836838,1529]. In most of these studies, it was suggested that the endothelium-dependent hyperpolarization of the smooth muscle cells could not be attributed to the release of the EDRF discovered by Robert Furchgott. For instance, in the guinea-pig mesenteric artery, acetylcholine-induced endothelium-dependent relaxation was associated with the endothelium-dependent hyperpolarization of the smooth muscle cells and was poorly inhibited by the presence of hemoglobin, while that elicited by substance P was not associated with hyperpolarization and was very sensitive to the presence of hemoglobin, indicating the involvement of different endothelium-derived “substances” [121]. In the rabbit saphenous artery, the endothelium-dependent relaxations and the endothelium-dependent hyperpolarizations to acetylcholine had a different time course, the former being sustained and the latter transient, and had different sensitivity to muscarinic antagonists [837,838]. Unfortunately, the major drawback of some of these latter studies was that the experiments were not performed in the presence of inhibitors of cyclooxygenases, preventing definitive conclusions for the occurrence of a third endothelial pathway. However, in canine coronary artery, the same difference in time course between endothelium-dependent relaxations and hyperpolarizations was reported, but since these experiments were performed in the presence of a cyclooxygenase inhibitor, the occurrence of a “factor” different from the EDRF described by Furchgott became extremely likely. Furthermore, in this blood vessel, a different susceptibility of the two responses to ouabain was observed, reinforcing the hypothesis of a third endothelial pathway besides the prostacyclin and the Furchgott's EDRF. Finally, in this artery, the NO-donor, sodium nitroprusside, which in many blood vessels was known to mimic at least partially EDRF responses, did not hyperpolarize the vascular smooth muscle cells. Taken together, these observations very strongly suggested that the occurrence of an alternative pathway, besides cyclooxygenases and the EDRF pathways, was contributing to the endothelial regulation of vascular tone [418,427,428,430]. Similarly, in rat aorta and pulmonary artery, the endothelium-dependent hyperpolarizations were insensitive to indomethacin but also hemoglobin and methylene blue (known EDRF inhibitors) and independent of cGMP accumulation [227,1529].

When EDRF was identified as NO, it was clearly shown that NO was not involved in most of the observed endothelium-dependent hyperpolarizations. Indeed, in the canine mesenteric artery and in the rabbit femoral artery, it was confirmed that neither authentic NO nor sodium nitroprusside could hyperpolarize the vascular smooth muscle cells, while in both tissues acetylcholine produced endothelium-dependent hyperpolarizations [683,836]. When L-arginine analogues became available as specific inhibitors of the production of NO [1287,1288], it became obvious that most endothelium-dependent hyperpolarizations and the associated relaxations were indeed resistant to the combined inhibition of cyclooxygenases and NO-synthases and occurred without any major increase in intracellular levels of cyclic nucleotides (cyclic GMP and cyclic AMP) in the smooth muscle cells. The production of an additional endothelial substance, which was neither a cyclooxygenase derivative nor the EDRF discovered by Dr Furchgott, was therefore suggested. By analogy with EDRF, this non-characterized endothelial factor was termed EDHF for endothelium-derived hyperpolarizing factor [168,376,380,430,432,1530] (Figure 1).

FIGURE 1. Membrane potential recording in an isolated blood vessel with an intracellular microelectrode.


Membrane potential recording in an isolated blood vessel with an intracellular microelectrode. An isolated strip or ring of a blood vessel is studied in an organ chamber filled with thermostated and oxygenated physiological solution. A sharp glass microelectrode (more...)


Initially, four major lines of evidence indicated that the endothelium-dependent hyperpolarization resulted from the opening of K+ channels in the vascular smooth muscle cells. Firstly, the amplitude of the hyperpolarization was inversely related to the extracellular concentration in K+ and abolished by K+ concentrations higher than 25 mM [224,225,273,1096] (Figure 2). Secondly, agonists that produced endothelium-dependent hyperpolarizations also stimulated the efflux of 42K (or 86Rb) from pre-loaded arteries [227,1529]. Thirdly, a decrease in membrane resistance of the vascular smooth muscle cells was observed during endothelium-dependent hyperpolarizations [123,224,225], suggesting that the hyperpolarization was due to the opening of a K+ channel rather than the closing of a chloride or a non-specific cationic channel. Fourthly, they were blocked by non-selective inhibitors of potassium channels such as tetraethylammonium or tetrabutylammonium [228,1096,1589].

FIGURE 2. Identification of the potassium channels involved in EDHF-mediated responses.


Identification of the potassium channels involved in EDHF-mediated responses. Characteristics of endothelium-dependent hyperpolarizations in the isolated carotid artery of the guinea-pig (in the presence of inhibitors of NO-synthase and cyclooxygenase). (more...)

EDHF-mediated responses, like the relaxations to endothelium-derived NO, are inhibited when the extracellular concentration of calcium is decreased [226,1202]. Endothelium-dependent hyperpolarization, in response to agonists that stimulate G-protein-coupled receptors, is associated with an increase in endothelial [Ca2+]i [749,953,954]. Finally, substances that increase endothelial [Ca2+]i in a receptor-independent manner (calcium ionophore, thapsigargin, cyclopiazonic acid) also produce endothelium-dependent hyperpolarization of the smooth muscle cells [482,1095,1097,1228], including human coronary arteries [1100].

Collectively, these observations suggested that the endothelium-dependent hyperpolarization of the smooth muscle cells should be attributed to an undetermined endothelial mediator (EDHF), released following an increase in endothelial [Ca2+]i in a similar manner to NO or prostacyclin, that would produce the hyperpolarization and the relaxation of the underlying smooth muscle cells by activating K+ channel(s), in essence an endogenous endothelium-derived potassium channel opener. However, in many arteries, this early interpretation was clearly wrong and had to be modified in the light of new experimental evidence [168].


Classical EDHF-mediated responses are endothelium-dependent relaxations resistant to inhibitors of NO-synthases and cyclooxygenases and insensitive to specific blockers of KATP (Figure 2), BKCa and KV channels, but inhibited by non-selective blockers of KCa including tetraethylammonium or tetrabutylammonium [228,272,1589,1827], which rules out the involvement of the diverse endothelial-derived mediators mentioned in Part I of these two volumes dedicated to “The Endothelium” (residual or stored NO, metabolites of arachidonic acid, CNP, H2O2, and so on) [430]. Waldron and Garland [1632] first showed that EDHF-mediated responses were unaffected by either apamin (a specific blocker of SKCa) or by charybdotoxin (a non-specific inhibitor of BKCa, IKCa and some KV) alone but were totally abolished by a combination of the two toxins. Thereafter, it was shown that in most blood vessels, EDHF-mediated responses involve the activation of SKCa, which is blocked by apamin, scyllatoxin, or UCL 1684, and/or IKCa, which is blocked by charybdotoxin, TRAM-39 or TRAM-34 [37,210,272,331,551,1212,1739,1826]. Conversely, first-generation activators of SKCa and/or IKCa (but not of BKCa), such as 1-ethyl-2-benzimidazolinone (1-EBIO), NS-309 and SKA-31 produced endothelium-dependent hyperpolarizations, which were inhibited by the above-mentioned potassium channel blockers [161,176,188,269,377,379,902,983,1347,1481,1633]. Taken into conjunction, these results indicate that, in most blood vessels, the activation of SKCa and/or IKCa is required in order to observe endothelium-dependent hyperpolarizations (Figure 2).

Then, Edwards et al. [373], first and very elegantly, demonstrated that the two potassium channels activated during EDHF-mediated responses were located on the endothelial cells, and not on the vascular smooth muscle cells as had been generally assumed. Indeed, in healthy vascular smooth muscle cells from various species, IKCa and SKCa are not or only very poorly expressed, whereas functional BKCa are constitutively present [1117,1264]. In contrast, both SKCa (and especially the SK3 α subunit) and IKCa (IK1) are expressed in freshly isolated endothelial cells from various arteries (Figures 3 and 4). These two channels are activated by substances that elicit EDHF-mediated responses, and their activation produces the concomitant hyperpolarization of the endothelial cells [161,176,177,831,1422,1526]. Conversely, the blockade of these two channels inhibits the hyperpolarization of the endothelial cells produced, for instance, by acetylcholine or bradykinin [373,379,1145].

FIGURE 3. Intermediate conductance calcium-activated potassium channels in endothelial cells.


Intermediate conductance calcium-activated potassium channels in endothelial cells. (A) The various configurations of the patch clamp technique are shown. The patch-clamp technique is an electrophysiological technique that allows the study of ionic channels (more...)

FIGURE 4. SKCa in freshly dissociated porcine coronary endothelial cells (outside-out configuration of the patch-clamp technique).


SKCa in freshly dissociated porcine coronary endothelial cells (outside-out configuration of the patch-clamp technique). (A) A small-conductance channel, as observed in single-channel recordings (250 nM free Ca2+ pipette solution, in the presence of charybdotoxin, (more...)


Stimulation of G-protein-coupled receptors, calcium ionophore, thapsigargin and cyclopiazonic acid all increase endothelial [Ca2+]i, produce hyperpolarization of the endothelial cells and endothelium-dependent hyperpolarization of the smooth muscle cells [169,246,482,749,955,1097,1145,1197,1581,1623,1650]. The principal mechanism that sustains the opening of endothelial KCa channels, following stimulation, is the capacitive calcium entry elicited by the depletion of calcium stores [978,1121,1122,1388]. However, the endothelial hyperpolarization in turn favors the entry of calcium by increasing the driving force for this ion [773,953]. It is therefore legitimate to wonder whether the intracellular increase in calcium per se is the essential contributor in eliciting EDHF-mediated responses or whether it is the hyperpolarization of the endothelial cells. Indeed, increasing [Ca2+]i in endothelial cells not only activates calmodulin, which acts as the Ca2+-sensing subunit for both SKCa and IKCa [804,1376,1722], but also leads to the activation of various enzymes, for instance, NO-synthases and phospholipases with the subsequent metabolism of arachidonic acid by cyclooxygenases, cytochrome P450 epoxygenases and lipoxygenases.

Depolarization with KCl and inhibition of the two endothelial conductances (SKCa and IKCa) are two different maneuvers that fully prevent the hyperpolarization of the endothelial cells and the subsequent endothelium-dependent hyperpolarization of the vascular smooth muscle. However, these two maneuvers produce no or only partial inhibition in the sustained phase of the rise in endothelial intracellular calcium concentration [541,773,981,1739]. Conversely, 1-EBIO, the activator of endothelial KCa, evokes EDHF-mediated responses without increasing endothelial [Ca2+]i [983]. The primary role of the increase in the endothelial calcium in EDHF-mediated responses is therefore to stimulate endothelial KCa channels. Thus, activation of endothelial potassium channels and/or hyperpolarization of the endothelial cell, but not the rise in [Ca2+]i per se, is a key step in the generation of endothelium-dependent hyperpolarization of the underlying smooth muscle cells [1264].


1.5.1. Myoendothelial Gap Junctions

In the vascular wall, gap junctions link smooth muscle with other smooth muscle cells, endothelial with other endothelial cells and, in many blood vessels, smooth muscle with endothelial cells. The number of myoendothelial gap junctions increases with a reduction in the size of the artery [1343], a phenomenon that parallels the contribution of the EDHF-mediated responses to endothelium-dependent relaxations [691,1414]. Furthermore, the presence of these heterocellular gap junctions is associated with EDHF-mediated responses [339,1346]. Endothelium and smooth muscle cells can communicate via these myoendothelial gap junctions physically, as Ca2+ can diffuse from one cell type to another [335,1760], and electrically, since depolarization and hyperpolarization are conducted bi-directionally from one cell type to the other [96,270,391,392,979,1736,1737]. The Cx37, Cx40 and Cx43 are the predominant isoforms of gap junction proteins expressed in the vascular wall. Depending on the blood vessel, they can be involved in myoendothelial gap junction communication [581,600,872,994]. Blockers of gap junctions abolish or partially inhibit EDHF-like responses in many arteries in vitro [100,219,220,338,375,379,581] and possibly in vivo [315]. Additionally, in the rat mesenteric artery, antibodies directed against Cx40, when loaded selectively in the endothelial cells, block EDHF-mediated responses [995]. Furthermore, in mice, Cx40 is essential for the acetylcholine-activated regenerative endothelium-dependent vasodilatation [320,441]. Finally, the endothelial cell lining is a privileged low resistance pathway for conducted vasodilatation/hyperpolarization along the length of the vessel [391,392]. Taken into conjunction, the results of these in vitro experiments provide compelling evidence for a major contributing role of myoendothelial gap junction in EDHF-mediated responses [318] (Figure 5).

FIGURE 5. Mechanisms underlying EDHF-mediated responses.


Mechanisms underlying EDHF-mediated responses. Stimulation of neurohumoral receptors and the shear stress exerted by the flowing blood increase endothelial [Ca2+]i, which in turn activates endothelial calcium-activated potassium channels of small and (more...)

However, experiments performed in vivo generally failed to demonstrate such a significant role for myoendothelial gap junctions [315,318]. The origin of this discrepancy is unknown but may involve the type and size of arteries studied in vivo, the presence of shear stress, sympathetic innervation and circulating hormones as well as confounding factors such as the use of anesthetics which inhibit gap junctions [1342]. Additionally, a recent paper suggests that distinct EDHF-mediated responses occur in the same artery when studied under different experimental conditions, i.e., isometric versus isobaric conditions. In the former case, EDHF-mediated response involves a Cx40-dependent mechanism, while in the latter case, which mimics in vivo conditions, the EDHF-mediated responses are very potent but completely independent of myoendothelial gap junctions [115].

1.5.2. Regulation of Gap Junction Communication

Most connexin isoforms can be phosphorylated by protein kinases that lead to modifications in tyrosine, serine and threonine residues, affecting intercellular communication [1074]. EDHF-mediated responses were first thought to occur independently of changes in the levels of cyclic nucleotides. However, it has now been demonstrated that EDHF-mediated responses are associated with a small but significant early and transient endothelium-dependent increase in cAMP content of the smooth muscle [583,1009,1525]. Capacitive calcium influx promotes not only endothelial hyperpolarization by activating KCa but also stimulates the calcium-sensitive adenylyl cyclase isoform, resulting in an increase in the production of cAMP [582]. Inhibitors of adenylyl cyclase and of PKA inhibit both the production of cAMP and the endothelium-dependent hyperpolarization of the subintimal smooth muscle cells, while a phosphodiesterase inhibitor amplifies both phenomena [583,1525]. The cAMP-dependent activation of PKA may promote connexin transport to the cell membrane and their incorporation in gap junction plaques and could also directly induce connexin phosphorylation and enhance the channel opening, thereby facilitating the electrotonic transmission of the endothelial hyperpolarization to the smooth muscle layers [319,581].

Metabolites of arachidonic acid formed by cytochrome P450 monooxygenase are also involved in the regulation of gap junction communication. EETs are involved in the regulation of intracellular calcium homeostasis and activate endothelial adenylyl cyclase [461]. Endogenously generated EETs cause a biphasic change in gap junctional communication between endothelial cells [1239], suggesting that these arachidonic acid metabolites, provided they produce similar effects on myoendothelial gap junctions, could favor transmission of the endothelial hyperpolarization toward the smooth muscle cells.

Hydrogen peroxide potentiates EDHF-mediated responses by promoting endothelial calcium mobilization, but can also directly enhance cell–cell coupling by altering the phosphorylation/oxidation status of connexin residues [371,1310]. Conversely, oxidized phospholipids alter vascular connexin expression, phosphorylation and reduce heterocellular communication [711].

Finally, there is some evidence for a coordinated regulation between different connexins. Thus, the endothelial and smooth muscle cell distribution of Cx43 is dependent on the presence of Cx40, suggesting that interactions between Cx40 and Cx43 regulate communication between endothelial cells and perhaps between smooth muscle and endothelial cells as well [710].

1.5.3. Potassium Ion Accumulation in the Intercellular Space Physiological Evidence.

The activation of endothelial IKCa and SKCa causes an efflux of K+ ions from the intracellular compartment toward the extracellular space, and these K+ ions can accumulate in the intercellular space between endothelial and smooth muscle cells. A moderate increase in the extracellular K+ concentration can provoke relaxation of vascular smooth muscle cells [635] by activating the two “metabolic sensors,” KIR [1116] and the Na+/K+ pump [1252].

This hypothesis was demonstrated successfully in the hepatic and mesenteric arteries of the rat [373]. Edwards et al. not only provided evidence that the targets of apamin and charybdotoxin are located on the endothelial cells but also that potassium ions accumulate in the intercellular space between endothelial and smooth muscle cells and that the potassium efflux associated with the activation of endothelial IKCa and SKCa channels produces hyperpolarization by activating both KIR and the Na+/K+ pump on the smooth muscle cells [373]. A role for the accumulation of K+ ion in the intercellular space was then confirmed in many different arteries [97,171173,337,377,1020,1114,1279] (Figure 5). Potassium Cloud Theory.

Nevertheless, this hypothesis has been highly controversial, possibly because the experimental conditions, and especially the presence or not of contractile agents, can markedly influence the contribution of K+ ion in the overall mechanisms underlying endothelial-dependent hyperpolarizations. Contractile agonists stimulate receptors on vascular smooth muscle cells, increase intracellular calcium concentration and depolarize the cells. The rise in intracellular calcium activates BKCa, while the depolarization activates KV as braking mechanisms. K+ ion accumulation in the intercellular space was first described by Bolton and Clapp [120] and was thought to originate from the potassium channels involved in myocyte repolarization. More recently, the existence of this cloud has been measured using K+-selective electrodes placed on the adventitial surface of a contracting vessel. Not only does the extracellular potassium ion concentration rise by a mean value of 11 mM, the cloud itself is almost abolished by iberiotoxin, indicating a predominant role of BKCa in its formation [1667]. This “potassium cloud” activates KIR and Na+/K+-ATPase, preventing the effects of any subsequent rise in potassium during endothelium-dependent hyperpolarization [1292] (Figure 5).

However, in some blood vessels, K+ does not evoke, or inconsistently produces relaxations and hyperpolarizations [270,333,375,379,1263]. Therefore, in these blood vessels the contribution of K+ ions in EDHF-mediated responses must, if anything, be minimal. The involvement of gap junctions and K+ ions are not necessarily mutually exclusive. The relative proportion of each mechanism almost certainly depends on numerous parameters including the state of activation of the vascular smooth muscle, the density of myoendothelial gap junctions and the level of the expression of the appropriate isoforms of Na+/K+-ATPase and/or KIR on the vascular smooth muscle [430,432].


The involvement of two populations of endothelial KCa channels in EDHF-mediated responses has been puzzling for a long time but the discrete role of each channel is now better appreciated [283]. The IKCa and SKCa channels are constitutively expressed in endothelial cells but show a very different spatial distribution (Figure 6).

FIGURE 6. Sublocalization of calcium-activated potassium channels in the vascular wall.


Sublocalization of calcium-activated potassium channels in the vascular wall. The three subtypes of calcium-activated potassium channels of large (BKCa), intermediate (IKCa) and small conductance (SKCa) are present in the vascular wall but with very specific (more...)

1.6.1. SKCa Localization and Postulated Function

SKCa are diffusely distributed over the plasma membrane with preferential locations at sites of homocellular endothelial gap junctions with various connexins (Cx37, Cx40 and Cx43) and in caveolin-rich domains along with certain other components of the classical EDHF pathway such as ligand-specific receptors like the bradykinin B2, IP3 receptors, as well as the calcium-permeable TRPC1 and TRPV4 channels [4,336,941,1335]. Additionally, SKCa could preferentially be associated with the other potassium channel KIR. Indeed, the opening of endothelial SKCa channels (by the preferential SKCa-selective opener, CyPPA) [675] is specifically abolished not only by apamin but also by a KIR-selective blocking concentration of barium [1670]. These data suggest that endothelial SKCa opening hyperpolarizes the endothelial cells; the associated K+ efflux increases the concentration of K+ ions within the intracaveolar spaces and is able to relieve the endogenous block of KIR channels. Their opening probably functions to amplify endothelial cell hyperpolarizations [1192,1440] generated by the initial calcium-dependent opening of SKCa, following a receptor-mediated or shear stress-dependent activation of the endothelial cells [1670]. In quiescent arteries (in the absence of vasoconstrictor stimulation), EDHF-mediated responses are associated with the preferential activation of SKCa and the contribution of myoendothelial gap junctions [336,551,887,1342].

1.6.2. IKCa Localization and Postulated Function

In contrast to SKCa, IKCa channels are non-caveolar [4,1668] and are localized preferentially within the endothelial projections through the internal elastic lamina at the sites of myoendothelial gap junctions [4,336,887,1342,1344]. In endothelial projections, IKCa channels are co-located with Cx40, calcium-sensing receptors, the non-selective, calcium-permeable cation channel TRPA1 and sections of endoplasmic reticulum densely expressing IP3 receptors [4,336,361,712,887,1342,1344,1668]. Repetitive localized calcium events (pulsars), triggered by IP3 and/or Ca2+ ions (calcium-induced calcium release), originate from these endothelial endoplasmic reticulum calcium stores. IP3 can be generated endogenously by the endothelial cell itself (for instance, following acetylcholine stimulation) or be produced by the adjacent smooth muscle cells (for instance, following phenylephrine stimulation). In the latter case, IP3 would diffuse toward the endothelial cells (possibly with Ca2+ ions) through the myoendothelial gap junctions [336,712,777,887]. The closely situated IKCa channels are activated by these calcium pulsars and the resultant endothelial hyperpolarization can be transmitted to the smooth muscle cells either via the myoendothelial gap junctions, as described previously, or be elicited by the K+ ions accumulating in the restricted extracellular space surrounding these endothelial projections. The activation of endothelial IKCa by calcium pulsars is likely to influence tone in resistance vessels under resting conditions [135]. In the rat mesenteric artery, the opening of IKCa channels and the subsequent K+ ion accumulation in the intercellular space preferentially activates the smooth muscle Na+/K+ pump [336,1292,1670] (Figure 7).

FIGURE 7. Myoendothelial projections.


Myoendothelial projections. Stimulation of vascular smooth muscle cells by contractile agonists, for instance, phenylephrine (PE) increases the intracellular Ca2+ concentration via the release of calcium from internal stores, through the production of (more...)

The precise role of the calcium-sensing receptor at the site of these endothelial projections is not completely understood. Stimulation of the calcium-sensing receptor results in selective IKCa-dependent endothelial hyperpolarization and endothelium-dependent vascular smooth muscle hyperpolarization [1437,1667,1668]. It has been hypothesized that in quiescent arteries (in the absence of vasoconstrictor stimulation), the calcium-sensing receptor would be fully stimulated by the extracellular concentration of calcium bathing the endothelial cells and that this prolonged stimulation of the calcium-sensing receptor would lead to the inactivation of IKCa [336]. Stimulation of smooth muscle (for instance, by phenylephrine) opens CaV and, subsequently to calcium entry, localized calcium depletion would be created in the small extracellular space surrounding myoendothelial projections. The endothelial calcium-sensing receptor would detect the changes in extracellular calcium and allow the recruitment of endothelial IKCa and their calcium-dependent activation. The subsequent endothelium-dependent hyperpolarization would restrain excessive activation of the smooth muscle [336]. Indeed, reducing the extracellular calcium concentration enables IKCa activation by acetylcholine [336,551]. Endothelial projections and myoendothelial gap junctions are therefore key structures intrinsically associated with extracellular and intracellular calcium homeostasis in both endothelial and vascular smooth muscle cells [336,887,1342] (Figure 7).

The close association between SKCa and KIR and between IKCa and Na+/K+-ATPases could be vessel-specific. Thus, in rat cerebral arteries, barium alone abolished the EDHF response to a TRPA1 activator, despite the co-localization of the TRP channel with IKCa and not with SKCa, as in the mesenteric artery [361].

1.6.3. Regulation of IKCa and SKCa Activity

SKCa and IKCa are complexes of four alpha pore-forming subunits (in endothelial cells SK3 and IK1, respectively) each bound by calmodulin and regulated by calcium. However, at present, very little is known with regards to the regulation of the activity of the endothelial channels formed by SK3 and IK1 [410]. Proteomic analysis indicates that the SK2 channel isoform also binds protein kinase CK2 and protein phosphatase 2A [26]. Furthermore, a putative PKA/PKG consensus phosphorylation site (Ser334) has been identified in IK1 [1369] and can be regulated by PKC [1714]. Finally, oxidative stress may regulate the activity of IKCa by oxidazing sulfhydril groups involved in the gating of the channel [183].


1.7.1. NOS and COX Genes

In eNOS knock-out mice, EDHF-mediated responses play a compensatory role for the absence of endothelial NO in both agonist- and flow-induced endothelium-dependent vasodilatation as well as in the basal regulation of myogenic tone [142,332,681,682,1381,1631] (Figure 8), this adaptation to NO-synthase deletion being gender specific [1713]. Similarly, in resistance arteries from female double knock-out mice for eNOS and cyclooxygenase-1 (COX-1), endothelium-dependent relaxations are preserved by an EDHF-mediated mechanism, while in arteries from the males, the endothelium-dependent relaxations are severely impaired. In these female mice, the double deletion of eNOS and COX-1 does not affect mean arterial blood pressure, while the corresponding males are hypertensive [1381].

FIGURE 8. Compensatory EDHF-mediated responses in eNOS knockout mice.


Compensatory EDHF-mediated responses in eNOS knockout mice. (A) eNOS knock-out mice are hypertensive. (B) Acetylcholine (ACh) produces an endothelium-dependent relaxation in the isolated aortic rings of wild type mice [eNOS(+/+)] but no longer in that (more...)

1.7.2. TRP Channel Genes

Calcium-permeable channels of the TRP family are essential molecular components for calcium entry into endothelial cells [79,1759,1791].

In TRPV4 knockout mice, flow-mediated and agonist-evoked vasodilatations are markedly inhibited [616,832,947,1030] and, in these deficient mice, EDHF-mediated responses are reduced to a greater extent than NO-mediated relaxations [366,1335,1792]. The close vicinity between TRPV4 and SKCa channels in caveolae together with connexins supports a role for these microdomains in the generation and propagation of EDHF-mediated responses [1284]. TRPV4-deficient mice are normotensive but the hypertension caused by inhibitors of NO-synthase is greater in the TRPV4-/- mice than in the wild-type controls [366]. Additionally, endothelial TRPV1 and TRPV3 channels may also play a role in EDHF-mediated responses [79].

In the endothelial cells of TRPC4-deficient mice animals, the agonist-induced calcium entry is reduced and is associated with a substantial reduction in endothelium-dependent relaxations [475]. In contrast, the deletion of TRPC1 is associated with a specific amplification of EDHF-mediated vasodilatations without alteration in NO-dependent relaxations. TPRC1-deficient mice have lower systolic blood pressure than the wild-type controls [1362]. TRPC1 is thus acting as a specific negative regulator of endothelial KCa channel-mediated vasodilatations, which appears paradoxical, since this channel is thought to be involved in endothelial calcium entry [1759]. However, TRPC1 can form heteromeric complexes with TRPV4, which are negatively modulated by protein kinase G [967]. The deletion of TRPC1 may possibly remove a constitutive NO-dependent inhibition of EDHF-mediated responses.

1.7.3. Potassium Channel Genes

Transgenic mice harboring genetically targeted alleles for the SK3 channel, in which SK3 gene expression can be experimentally controlled by dietary doxycycline, have been engineered [125]. In these mice, the level of expression of SK3 channels on the endothelial cells is correlated with the value of the cell membrane potential of both endothelial and vascular smooth muscle cells, the tone of isolated mesenteric arteries, the diameter of these arteries in situ and the arterial blood pressure of the animals [1526]. The functional importance of SK3 is supported by the observation that the deletion of the SK2 gene is without effect on mouse blood pressure [135,1526]. Genetic deficiency of SK3 strongly attenuates arteriolar vasodilatations to skeletal muscular contraction but not those in response to acetylcholine, suggesting that activation of endothelial SK3 contributes to exercise hyperemia [1040].

Disruption of the IK1 gene reduces the hyperpolarization of endothelial and smooth muscle cells in response to acetylcholine. The associated vasodilatations in the carotid artery and in resistance vessels, including skeletal muscle arterioles, are also diminished because of a substantial reduction in EDHF-mediated responses. Moreover, the IK1 deletion also leads to a significant increase in arterial blood pressure and to mild left ventricular hypertrophy [1040,1422].

These results confirm that, in mice, the endothelial SKCa and IKCa are fundamental determinants of endothelial hyperpolarization and EDHF signaling and indicate that they actively control vascular tone and overall regulation of the circulation. The differential impact of SK3- and IK1-deficiency on dilations to distinct stimuli suggests stimulus-dependent activation of these endothelial channels [1040]. The relative importance of IKCa and SKCa channels to the EDHF response could be tissue- and species-dependent. However, gene-knockout studies in mice show that IKCa is important in conduit arteries and in resistance arterioles, but that loss of IK1 function cannot be compensated for by up-regulation of SK3 [135,1422,1693]. Loss of IK1 reduces the EDHF response to acetylcholine by about 75% and increases blood pressure. Although loss of SK3 further reduces the EDHF response, there is no additional depressor effect on blood pressure in the conscious, unrestrained IK1 knockout mice [135].

In mice that lack Kir2.1, but not in those deleted for Kir2.2 genes, Kir currents are absent, and stimulation with moderate increases in K+ concentration does not produce relaxation, indicating that Kir2.1 gene expression is required for Kir currents and K+-induced dilatation. Kir2.1(-/-) animals die 8 to 12 hours after birth, preventing the acquisition of more information on cardiovascular parameters such as arterial blood pressure and amplification of EDHF-mediated responses [134,1786].

1.7.4. Connexin Genes

C37 and Cx40 are the predominant gap junction proteins in the endothelial cells of the mouse [1429]. Cx40 proteins are involved in endothelial homocellular gap junctions and also in heterocellular gap junctions linking endothelial cells not only to smooth muscle cells but also to renin-producing juxtaglomerular cells. Global deletion of Cx40 caused local dissociation of renin-producing cells from endothelial cells, renin hypersecretion and hypertension. Mice deficient for Cx40 in renin-producing cells were hypertensive, and these cells were ectopically localized. In contrast, in mice lacking endothelial Cx40, the blood pressure, renin-producing cell distribution and control of renin secretion were similar to wild-type mice. Thus, endothelial Cx40 is not essential for the control of blood pressure and for renin expression and/or release but Cx40 is required in renin-producing cells to ensure their correct positioning in the juxtaglomerular area and the control of renin secretion by pressure. [1626,1627].

In skeletal muscle arterioles, genetic deficiency of Cx40, like the deficiency of SK3, attenuates dilations induced by muscular contractions, but not those in response to acetylcholine. However, the arterioles of Cx40-deficient mice exhibit a reduced spread of dilatation in response to endothelium-dependent vasodilators and irregular arteriolar vasomotion [320,321,443,1429]. Therefore, the conducted hyperpolarization/vasodilatation ascending the vascular tree, in response to either acetylcholine stimulation or exercise hyperemia, is orchestrated by Cx40-forming gap junctions [1040].

Cx37 is also highly expressed in endothelial cells and in juxtaglomerular cells but does not appear to play an essential role in the control of renin secretion [1628]. The deletion of Cx37 does not alter blood pressure and, in these mice, the acetylcholine-induced vasodilatation and the ascending vasodilatation are unaffected [441]. However, gene polymorphism of Cx37 has been associated with arterial stenosis and myocardial infarction. Indeed, Cx40, Cx37 and eNOS are tightly associated, and Cx37 specifically interacts with eNOS. The deletion of Cx40 markedly decreases the expression of Cx37 and that of eNOS at the site of endothelial intercellular junctions [28,1214]. In the cerebral circulation, Cx37 along with Cx40 may be involved in endothelium-dependent synchronized vasomotion, where the intracellular calcium waves generated within the smooth muscle cells are coordinated by electrical coupling via myoendothelial gap junctions [600]. Double Cx37-Cx40 knockout mice show severe vascular abnormalities and die perinatally, providing direct evidence for a critical role of endothelial gap junction-mediated communication in the development and/or functional maintenance of mouse vasculature [1428], but did not allow further investigation on EDHF-mediated responses.

Mice subjected to specific deletion of endothelial Cx43 do not present major alterations in arterial blood pressure [916,1544], possibly because this connexin is not the major one expressed in murine endothelial cells.

These results show that deletion of each key molecular component of EDHF-mediated responses is associated with hemodynamic alterations, suggesting that this endothelial pathway contributes to the overall regulation of arterial blood pressure.


1.8.1. EDHF-Mediated Responses and Arterial Blood Pressure

Arterial blood pressure and flow to the tissues are determined mainly by the degree of opening of resistance arteries with an internal diameter smaller than 300 μM. The presence of an EDHF-component to endothelium-dependent relaxations (dilatations) becomes preponderant in resistance arteries [1414] and is more pronounced in arteries than in veins [418,1797].

Experiments performed in genetically modified animals (see above) strongly suggest that EDHF-mediated responses contribute to the regulation of arterial blood pressure [580]. Furthermore, activation of endothelial KCa channel, for instance, by the riluzole derivative SKA-31, decreases arterial blood pressure in both normotensive and hypertensive mice [135,1347]. However, the generation of NO may contribute to this hypotensive effect [289]. Besides, EDHF-mediated responses can counteract the hypertension caused by deficiency in other vasodilator systems, i.e., NO and/or prostacyclin [142,1381].

In contrast, in anesthetized rats, acute administration of the two KCa blockers, charybdotoxin plus apamin, does not affect basal arterial blood pressure and does not modify basal conductance in the mesenteric artery or the hindlimb, although the combination of these toxins inhibits the vasodilatation in both vascular beds as well as the decrease in blood pressure induced by the administration of neurohumoral substances such as acetylcholine [310,1193,1415]. In the same species, inhibition of gap junction communication by the intra-renal administration of elevated concentrations of Gap peptide [219] produces a decrease in basal renal blood flow and an increase in arterial blood pressure [315].

These results indicate that the alteration in EDHF-mediated responses can affect mean arterial blood pressure. It appears likely that they contribute tonically to the homeostatic regulation of arterial blood pressure. The issue to be addressed specifically is the role of EDHF-mediated responses in the control of peripheral resistance in conscious animals, when the activity of the NO-synthase is either inhibited or preserved. Indeed, anesthetic drugs may inhibit some EDHF-mediated responses [662,1342].

1.8.2. Flow-mediated Vasodilatation

Efficient regulation of blood flow requires coordinated changes in the diameter of both arterioles and the feeding arteries that supply them. This involves flow-induced vasodilatation and a mechanism that is independent of flow or changes in internal pressure, spreading vasodilator responses. Flow-induced vasodilatation is an important physiological stimulus involved in the control of vascular tone as the blood vessel continuously adapts its diameter to minimize the shear forces exerted by the flowing blood on the endothelial surface.

Shear stress activates mechanosensors on the endothelial cell surface, promoting calcium influx via cationic channels, TRPV4 being a major player in this phenomenon [1791], and the release of calcium from intracellular stores [137,676]. This is a key step not only for the activation of the L-arginine–NO-synthase pathway but also for the opening of endothelial potassium channels, the subsequent hyperpolarization of the endothelial cell membrane [676] and consequently EDHF-mediated responses [168]. Long-term exposure of human endothelial cells to laminar shear stress up-regulates the expression and the function of IKCa [139], suggesting that this potassium channel, one of the key players in EDHF-mediated responses, contributes to the endothelium-dependent adaptation to hemodynamic changes.

1.8.3. Conducted Vasodilatation

Dilatation along and among branches of a resistance network governs blood flow distribution and magnitude. Vasodilatation must be coordinated among daughter and parent arterioles in order to obtain adequate changes in blood flow [54]. The transmission of vasomotor signals along the axis of an arteriole, i.e., conducted vasodilatation, is one of the pathways, besides flow-mediated and metabolic vasodilatation, by which a stimulus, sensed by a localized portion of the artery, is communicated to remote upstream and downstream regions. Local stimulation of endothelial cells with acetylcholine evokes dilatation that can spread along the length of the arteriole [351,1389]. This phenomenon is endothelium-dependent, but even if NO is involved in some arteries, it alone cannot explain the phenomenon [154,1201]. Thus, EDHF-mediated responses are strongly involved in conducted vasodilatation.

Conducted vasodilatation in response to acetylcholine, involves a rise in endothelial intracellular calcium, hyperpolarization of the endothelial cells and endothelium-dependent hyperpolarization of the underlying smooth muscle cells at the site of stimulation. The hyperpolarization of the endothelial cells and the endothelium-dependent hyperpolarization of the smooth muscle cells are transmitted, but not the rise in intracellular calcium [340,358,1506,1761]. Thus, conducted vasodilatations evoked by these endothelium-dependent agonists are inhibited by blockers of SKCa and IKCa [662,1575,1760]. In response to acetylcholine, IKCa is the predominant endothelial potassium channel involved in conducted vasodilatation [1693], while endothelial SKCa activation contributes to exercise hyperemia [1040].

Intercellular communication is essential in order to observe conducted vasodilatation. The hyperpolarization is conducted via the endothelial cells, which are coupled by homocellular gap junctions and aligned following the axis of the arterioles, providing a predominant low resistance cellular pathway for conducting vasoactive signals along the vessel wall. Endothelial cells are also coupled to smooth muscle cells by myoendothelial gap junctions, but if a direct coupling between these two cell layers has consistently been reported in vitro, the apparent lack of myoendothelial coupling found in vivo indicates that heterocellular coupling does not explain all the observed conducted vasodilatations and that in the absence of such coupling, respective cell layers provide parallel paths for conduction [54]. Nevertheless, mice deficient for CX40 show an impaired conduction of vasodilatation [320].

Cell activation by acetylcholine appears to facilitate communication through myoendothelial gap junctions. Indeed, the conducted vasodilatation observed in arteries exposed to the muscarinic agonist is more pronounced than the one elicited by injection of negative currents either in the smooth muscle or the endothelial cells [391,392]. This suggests the involvement of an undefined active membrane process in conducted hyperpolarization [284]. One possibility is that interstitial cells of Cajal-like cells, which have been observed in some blood vessels, could act as pacemaker cells [613,1243,1255]. Additionally, the release and accumulation of potassium ions in the intercellular myoendothelial space could play a critical role in conducted vasodilatation by activating endothelial and/or smooth muscle KIR [54,1294,1440].

1.8.4. Vasomotion

Vasomotion is the rhythmic change in arterial or venous diameter occurring spontaneously or upon stimulation with an agonist. The physiological role of this universal phenomenon is not completely understood, although it is likely to contribute to the harmonious and efficient distribution of blood flow.

Vasomotion is generated by cyclical contractions and relaxations of the vascular smooth muscle cells directly associated with oscillatory changes in [Ca2+]i and in membrane potential. In many arteries, this phenomenon relies on the presence of endothelial cells where synchronous calcium waves can also be observed [1127]. NO release and the subsequent elevation of cyclic-GMP in the smooth muscle cells contribute to the mechanism of underlying vasomotion [1206]. In rat mesenteric arteries, the vasomotion evoked by the stimulation of adrenergic receptors is critically dependent on the activation of endothelial IKCa and SKCa, indicating that the EDHF-mediated response is essential for the initiation and maintenance of the phenomenon [1012,1149]. Furthermore, mice deficient for Cx40 show an irregular vasomotion [320,321,443]. These findings support the hypothesis that myoendothelial and/or endothelial to endothelial cell coupling contributes to synchronous vasomotion, conducted vasodilatation and to the regulation of peripheral resistance [945,946].

1.8.5. Hypoxia

Acute and chronic hypoxia affect the caliber of arteries and veins by altering the release of endothelial relaxing and contracting factors, and by acting directly on the vascular smooth muscle cells. Endothelium-dependent relaxations are inhibited by hypoxia [496]. This is understandable since molecular oxygen, along with L-arginine, is a required substrate for NO-synthases [1180]. However, in blood vessels that exhibit EDHF-mediated responses (e.g., pulmonary, cerebral and coronary arteries), the EDHF- but not the NO-component of endothelium-dependent relaxations is preserved under severe hypoxia [417,1213,1411]. Whether or not EDHF-mediated responses blunt acute hypoxic pulmonary vasconstriction remains controversial [197,619,1077].

Thus, EDHF-mediated responses can prevent vasospasm and in some instances compensate for the disappearance of eNOS-dependent relaxation under hypoxic conditions. This, along with various other mechanisms such as the release of prostacyclin, adenosine, EETs and the formation of NO from nitrite by enzymes possessing nitrite reductase activity, as well as the direct activation of potassium channels and the inhibition of CaV in the smooth muscle cells [364,365,367,1438,1592], contribute to the maintenance of a normal caliber of the blood vessel and the preservation of blood flow.

1.8.6. EDHF-Mediated Responses and Gender Sex Hormones.

Men and post-menopausal women have a greater incidence of hypertension and coronary artery disease when compared to pre-menopausal women. The vascular protection observed in the latter has been attributed to a beneficial effect of the female sex hormones, especially estrogens, and justified the now contested hormone replacement therapy. Sex hormones interact with both cytosolic, nuclear, and plasmalemnal receptors in the vasculature, producing both genomic and non-genomic effects. The acute administration of sex hormones, including testosterone, produces rapid vascular relaxation or vasodilation, which can be either endothelium-dependent or -independent [1166,1348]. However, the physiological relevance of these acute effects is uncertain. The genomic effects of estrogen influence the expression of proteins involved in multiple endothelial pathways including the NOS, cytochrome P450 epoxygenase and EDHF pathways [1029,1166,1615].

In isolated mesenteric, femoral and tail arteries as well as in the perfused renal vascular bed taken from sexually mature rats, the contribution of the EDHF-mediated responses is greater in females than in males [960,1018,1178,1656,1674]. This sex difference was absent in estrogen receptor β knockout mice [960]. In resistance arteries of eNOS/COX-1 double knock-out female mice, an EDHF-mediated mechanism maintains normal endothelium-dependent relaxations, while these relaxations are markedly reduced in the same arteries taken from the corresponding males [1381].

In reproductive (uterine) and non-reproductive (mesenteric) arteries of the female rat, EDHF-mediated responses are reduced in ovariectomized animals while estrogen supplementation restores them [158,932,933,1333]. In the mesenteric artery, ovariectomy induces reciprocal changes in the respective contribution of EDHF- and NO-mediated responses in the overall endothelium-dependent relaxations, so that the disappearance of the EDHF-mediated response is fully compensated by the increase in NO production, a phenomenon associated with a decrease in the expression of the negative regulator of eNOS function, caveolin-1 [1110]. Even short-term alteration in blood levels of estrogens affects the amplitude of the endothelium-dependent hyperpolarizations since, in young female rats at the diestrus phase of the oestrus cycle, the amplitude of the hyperpolarizations is reduced significantly [932]. In the gracilis muscle of ovariectomized rats exposed to inhibitors of NO-synthase, flow-induced dilatation depends on the activation of cyclooxygenase, but becomes, after estrogen supplementation, fully mediated by EDHF-mediated responses [681]. The variation in the amplitude of the EDHF-mediated response produced by ovariectomy and estrogen supplementation is paralleled by variations in the expression of Cx40 and Cx43 in the endothelial layer of the mesenteric arterial wall, suggesting that estrogens affect EDHF-mediated responses by controlling the level of cell coupling by gap junctions [933,1110]. In addition, estrogen can enhance endothelium-dependent hyperpolarizations by increasing the expression of SK3 [725].

By contrast, in the middle cerebral artery of the rat, the endothelium-dependent relaxation is fully NO-mediated in the females but involves both NO and an EDHF-mediated component in males [559]. In the middle cerebral artery and pial arterioles of ovariectomized females, the NO-mediated relaxation disappears and an EDHF-mediated component is revealed. Estrogen supplementation reverses the balance between NO and EDHF to what is observed in control non-ovariectomized females [559,1728]. These changes are not associated with modifications in calcium homeostasis in the endothelial cells [560], but rather with changes in the expression of Cx43 [1728]. They are unlikely to be explained by NO affecting EDHF-function since manipulations of NO do not alter agonist-induced EDHF-mediated vasodilatations [1358]. The inverse modulation by estrogen of the contribution of NO and EDHF-Cx43 in endothelium-dependent relaxations of cerebral and peripheral blood vessels is puzzling. The observation that in the middle cerebral artery of orchietomized male rats, treatment with testosterone suppresses EDHF-mediated relaxation [567] does not help to clarify the role of sex hormone in the regulation of cerebral vascular tone.

In the forearm of post-menopausal women, short-term estrogen treatment improves both the NO-dependent and the EDHF-mediated vasodilatations induced by acetylcholine and substance P, respectively [1502]. Taken in conjunction, these data show that sex hormones exert a prominent role in modulating vascular tone and that the targets affected by sex hormones include EDHF-mediated responses. Pregnancy.

A successful pregnancy outcome relies on extensive maternal cardiovascular adaptation, including enhanced uteroplacental vasodilator mechanisms. In pregnant rats, the endothelium-dependent relaxations not only of uterine but also of other peripheral arteries such as the mesenteric artery are enhanced. This involves an increased production of NO as well as augmented EDHF-mediated responses, associated with the activation of the two endothelial potassium channels SKCa and IKCa [286,287,486,538,558]. In pregnant women, similar conclusions have been reached in omental and subcutaneous blood vessels [39,545,1195]. In the myometrium and in subcutaneous arterioles from healthy pregnant women, these EDHF-mediated responses involve gap-junctional communication [799,872,957,958].


Endothelial dysfunctions are observed in various cardiovascular diseases and are often associated with a decrease in NO synthesis and/or a loss of its biological activities. However, alterations in the EDHF pathway can also contribute to these endothelial dysfunctions or conversely compensate for the loss of NO bioavailability. EDHF-mediated responses are modified in various pathological conditions (hypertension, atherosclerosis, hypercholesterolemia, heart failure, ischemia–reperfusion, angioplasty, eclampsia, diabetes, sepsis) and with aging [429431].

1.9.1. Impairment of EDHF-Mediated Responses Hypertension.

Hypertension per se does not produce a consistent depression of the EDHF-mediated responses but has been observed in various animal models and in humans. Genetic Models of Hypertension.

In the mesenteric arterial bed of SHR, the endothelium-dysfunction can be attributed to a marked attenuation of the EDHF-mediated component and a concomitant production of cyclooxygenase-derived contractile prostanoids (EDCF) with no or little alteration in the production of NO [477,478,689,963,977,1599]. This decrease in EDHF-mediated response has been associated, but not yet causally linked, to a change in the expression profile of gap junctions in endothelial cells. Indeed, Cx37 and Cx40 are lower in the artery of the SHR than in that of the WKY, while the reverse is observed for Cx43 [776,1322]. The endothelial SKCa activation (but not that of IKCa) and the amplification mediated by KIR opening are compromised in SHR mesenteric artery, and this is associated with a decreased expression of both SK3 and Kir2.1 [1670]. Plasma levels of ADMA, an endogenous inhibitor of eNOS, are elevated in hypertension, and this arginine analogue decreases the expression of SK3 mRNA [908]. Furthermore, a reduced dimerization of caveolin-1 can modify the interactions between SKCa and caveolins and contribute to the pronounced decrease in SKCa function [4,513,1670]. The conducted vasodilatation is reduced in the mesenteric artery of the SHR, a phenomenon also observed in arteries of the retractor muscle of hypertensive hamster [572,862]. In the WKY, such conducted vasodilatation is facilitated by the activation of KIR, a phenomenon that disappears in SHR [572]. Additionally, in the SHR, activation of a fast endothelium-dependent depolarization attributed to activation of a calcium-activated chloride conductance counteracts the EDHF-mediated relaxation [272,570].

In the isolated renal artery of the WKY, the EDHF-mediated response relies on the activation of endothelial potassium channels and the release of potassium ions. In this artery, a total and selective disappearance of this response occurred in aged SHR, while in young ones, a compensatory mechanism involving a cytochrome P450-dependent pathway preserved the endothelium-dependent hyperpolarization [172]. However, in the renal artery of SHR subjected to a high-salt diet, the release of EDCF and NO is increased and the alternative pathway is reduced [759].

In the stroke-prone SHR (SHRSP), endothelial dysfunction in the mesenteric artery is more pronounced than in the SHR, and both EDHF-mediated responses and NO activity are decreased [776,1475]. Again, reduced SK3 expression has been observed, but a parallel increase in IK1 expression, attributed to a decreased expression of the repressor element-1 silencing transcription factor (REST), partially compensates the deficient activity of SKCa [543].

In the isolated mesenteric artery of the salt-sensitive Dahl rat, the endothelium-dependent hyperpolarizations are reduced [1161]. Conversely, in the Munich Wistar Fromter, a genetically hypertensive rat with spontaneous albuminuria, and in the Lyon hypertensive rat strain, the responses of the mesenteric artery are unaltered. However, in the coronary artery of the former model, both EDHF- and NO-mediated responses are reduced [476,591]. Induced Hypertension.

In mesenteric arteries, but not necessarily in other vascular beds, from DOCA-salt and two-kidney one-clip hypertensive rat models, EDHF-mediated responses are impaired [8,625,975,976,1161,1393,1712]. In the angiotensin II-infused hypertensive rats, the alteration of the EDHF-mediated response is associated with an increase in plasma ADMA, reduced endothelial expression of both SK3 mRNA and protein and decreased expression of vascular Cx37, Cx40 and Cx43 [288,647,1635].

In rats with renal mass reduction subjected to a high-salt diet, both the NO- and the EDHF-mediated responses are reduced [806], while in a severe model of subtotal nephrectomy leading to chronic renal failure, only the EDHF-mediated response is markedly impaired [769]. In this latter model, the reduced EDHF-mediated response appears to be related to the severity of the renal dysfunction [1613]. Human Hypertension and Eclampsia.

In humans with essential hypertension, the mechanism underlying endothelial dysfunction is linked to oxidative stress, which reduces availability of NO, and to the activation of cyclooxygenase-derived constricting factors. Few studies have combined inhibitors of both cyclooxygenases and NO-synthases in order to assess the role of EDHF-mediated responses in the endothelial dysfunction characteristic of human hypertension. In the forearm vascular bed, the presence of such an alternative pathway, possibly a cytochrome P450-dependent mechanism, compensates the decreased NO bioavailability in order to sustain endothelium-dependent vasodilatation [1487,1491,1500]. This compensatory pathway can be depressed by additional aggravating factors such as hyperhomocysteinemia [1497]. Indeed, in rats with chronic hyperhomocysteinemia, induced by a high-methionine/low vitamin-B diet, a severely reduced EDHF-mediated vasodilatation is observed in the renal vasculature [313,630]. In patients with acromegaly, the prevalence of hypertension and diabetes is increased. Subcutaneous arteries from these patients undergo hypertrophic remodeling and present endothelial dysfunction involving both a reduction in NO- and EDHF-mediated responses [1177].

Eclampsia is a pregnancy-specific disorder characterized by hypertension, proteinuria and alterations in endothelial cell function. Human omental arterioles taken from preeclamptic women exhibit a deficit in the endothelial production of prostacyclin [1479]. In addition, in myometrial arterioles from preeclamptic mothers, the up-regulation of the EDHF-mediated responses observed in normal pregnancy does not occur [799,800]. A decrease in myoendothelial communications has been associated with this endothelial dysfunction [957959] and may contribute to the clinical features of the disease. Aging.

In the aging normotensive WKY rat, a progressive and slow decrease in the release of NO is observed together with a reduction in EDHF-mediated responses and an increase in the production of EDCF [477,703,977,1101]. The endothelial dysfunction phenotype is similar to that observed |in the SHR, and one could consider this hypertensive model as a model of premature aging.

There is a significant inhibition of endothelium-dependent hyperpolarizations with aging in human coronary and peripheral arteries, although in the human, it is difficult to separate what can be ascribed to aging per se rather than to other risk factors such as atherosclerosis, hyperlipidemia or smoking [1100,1579]. The measurement of forearm blood flow in vivo suggests that aging and hypertension produce the same endothelial dysfunction or, in other words, that hypertension causes premature aging of the endothelium. With aging, oxidative stress and the activation of the metabolism of arachidonic acid by cyclooxygenase reduce the availability of NO which is, at least initially, compensated for by the presence of a cytochrome P450-related mechanism [1494,1499]. Diabetes Insulin-Dependent Diabetes.

Most of the studies designed to assess EDHF-mediated responses in diabetes have been performed in streptozotocin-treated rodents. In this model, a decrease in EDHF-mediated response has been generally reported in various vascular beds including, mesenteric, carotid, renal and coronary arteries, in arterioles supplying the sciatic nerves, retina, and kidneys as well as in the corpus carvernosum [4042,280,314,452,484,770,797,898,973,974,1005,1104,1531,1678].

These dysfunctional EDHF-mediated responses have been observed with or without a concomitant decrease in the NO-mediated responses and have been attributed to various parameters and include alterations at the level of the two endothelial potassium channels and at the level of myoendothelial gap junction communications. Indeed, the endothelium-dependent hyperpolarization in response to 1-EBIO, an opener of SKCa and IKCa, is also inhibited [1678]. However, a parallel decreased expression of SK3 and/or IK1 is not consistently observed in these diabetic vascular beds [898,1810]. Additionally, the down-regulation of the expression of Cx40 [974], a reduction in the cAMP-dependent facilitation of gap junction communication [583] following an increase in phosphodiesterase-3 activity and a decreased expression and activity of protein kinase A has also been reported [1002,1009].

In patients with type I diabetes under good glycemic control and without albuminuria, endothelial function appears normal. This has been demonstrated by various clinical investigations and confirmed in isolated subcutaneous arteries where both the NO- and the EDHF-mediated responses were preserved [39,314]. However, in patients with microalbuminuria, impairment of the endothelium-dependent vasodilatation is observed [314]. In these patients, it is not known whether or not the various components of the endothelium-dependent vasodilatation are differentially affected by the disease. Nevertheless, the experimental data suggest that the impaired EDHF-mediated responses could be involved in diabetic macro- and microangiopathy, i.e., diabetic neuropathy, retinopathy and nephropathy as well as in erectile dysfunction and impotence. Insulin-Independent Diabetes.

Type II diabetes, in its early phase before pancreatic beta cell failure, is characterized by insulin-resistance, hyperinsulinemia, moderate hyperglycemia and dyslipidemia, and is often associated with hypertension (syndrome X, metabolic syndrome). This cluster of abnormalities contributes to the development of atherosclerosis [1107,1286]. In these diabetic patients, endothelial dysfunction has also been reported [314].

Various genetic and non-genetic animal models have been created to study type II diabetes, and in virtually all of them, the EDHF-mediated responses are impaired while the NO-mediated relaxations are not necessarily affected. These models include the fructose-fed rat [783,784,1041,1042], the cafeteria diet-induced obese rat [204,601], the leptin-deficient, genetically obese and mildly hypertensive Zucker rat [50,1712,1774], the more severe model of Zucker diabetic fatty rat (ZDF rat) [150,162,1667], the Otsuka Long–Evans Tokushima fatty rat, a strain that spontaneously develops hyperglycemia, hyperinsulinemia, insulin resistance and mild obesity (OLEFT rat) [761,1003,1006,1007,1048], the Goto-Kakizati rat, an inbred model of lean type 2 diabetes expressing glucose intolerance [1163] and the nicotinamide and streptozotocin-treated rats, a model of partial pancreatic beta cell destruction that appears closer to type II than to type I diabetes [1010].

In the OLEFT rat, reduced cAMP/PKA signaling and a decrease in endothelial potassium channel activity contribute to the impaired EDHF-mediated response [1003]. A reduction in SKCa but not IKCa channel activity, without reduction in SK3 expression, was also reported in the ZDF rat [150,162]. In the Zucker fatty rat, the altered EDHF-mediated response has been associated with decreased expression of Cx40 [1774]. In the mesenteric artery of the Sprague–Dawley rat model of cafeteria diet-induced obesity associated with mild hyperglycemia and hyperinsulinemia, the EDHF-mediated response is also blunted. A decrease in SKCa activity and a concomitant up-regulation of IKCa activity, associated with an enhanced expression of IK1, was observed. In the mesenteric artery of these obese rats, the relaxation associated with these EDHF-mediated responses involves predominantly the activation of Na+/K+-ATPase, while the Kir function is significantly reduced [601].

In the penile artery of humans with type II diabetes, the impairment of the endothelium-dependent relaxations involves both the NO and the EDHF-mediated components [40]. As has been demonstrated in numerous animal studies [430], diabetes can also affect vascular smooth muscle potassium channels and alter the intrinsic ability of vascular smooth muscle cells to hyperpolarize [1059]. Similarly, in human arteries, in vitro exposure to elevated glucose concentrations can decrease EDHF-mediated responses and alter vascular smooth muscle potassium channel activity [810].

Insulin is a physiological vasodilator which involves the L-arginine-nitric oxide pathway in healthy subjects but possibly the EDHF-pathway in hypertensive patients [1500]. However, hyperinsulinemia can also exert a deleterious effect on endothelium-dependent vasodilatation. Thus, acute incubation of the isolated mesenteric artery of the rat with insulin selectively depresses the acetylcholine-induced EDHF-mediated relaxation [808].

In conclusion, in most of the animal models of type I or type II diabetes as well as in human diabetic patients, EDHF-mediated responses are altered. However, a direct link between hyperglycemia and impairment of the endothelium-dependent hyperpolarizations has not been established. Angioplasty.

Under physiological conditions, endothelial cells are normally quiescent and are among the most genetically stable cells of the body. They replicate at a very slow rate since their turnover time is over hundreds of days. However, during angiogenesis, pathological situations (hyperlipidemia, hypertension) and surgical interventions (angioplasty), endothelial cells can proliferate rapidly with a turnover time of less than 5 days. Nevertheless, the capacity of endothelial cells to divide is limited and ultimately the cells enter a state of growth arrest, i.e., senescence. Senescent endothelial cells are altered morphologically, have an increased generation of reactive oxygen species, a decreased production of NO and are more sensitive to apoptotic stimuli [431]. Restenosis and altered vasomotion frequently occur after coronary angioplasty.

In the carotid artery of the rat after endothelial denudation by balloon angioplasty, the endothelium-dependent relaxations are impaired and both the NO- and EDHF-mediated responses are reduced. However, the alteration in NO-dependent responses is transient and is restored 4 weeks after the procedure, while the reduction in the EDHF-mediated responses is sustained [830,929]. This latter dysfunction is associated with decreased endothelial expression of both SKCa and IKCa [830].

In porcine coronary, changes in morphology and functionality are observed in the regenerated endothelial cells that re-colonize the site of angioplasty and the NO-mediated relaxations are impaired [128,431]. In segments of porcine coronary arteries with regenerated endothelium, the endothelium-dependent hyperpolarization in response to serotonin and substance P are reduced, while those in response to bradykinin are maintained or even enhanced [128,1413,1548,1549,1552]. These divergent effects of angioplasty could be explained by the difference in the mechanisms involved in these endothelium-dependent hyperpolarizations. The EDHF-mediated response evoked by substance P relies exclusively on the activation of endothelial KCa, while the endothelium-dependent hyperpolarization produced by bradykinin involves both the activation of endothelial KCa and the release of cytochrome P450 metabolites [379,1669]. Thus, the preserved hyperpolarization/relaxation in response to bradykinin may be attributed to a compensatory involvement of EETs. Transplantation.

Coronary graft vasculopathy develops in the majority of heart transplant recipients. The accelerated atherosclerosis of the coronary arteries is preceded by reduced endothelium-dependent relaxations. In a porcine model of heterotopic heart transplantation, without immunosuppression but with immunological typing in order to mimic the kinetics of slow low-grade rejection observed in human transplantation, obvious signs of rejection are present in the grafted heart 6 weeks after transplantation. In the isolated coronary arteries of these hearts, impairment of the NO- and the EDHF-mediated relaxations is observed [12091211].

Cardioplegic hyperkalemic solution used to produce heart arrest and organ preservation solutions such as the University of Wisconsin or the Euro-Collins solution produce an alteration in the EDHF-mediated responses in porcine coronary and pulmonary arteries [529,627,628,1796,1821]. Preservation of the organ in St Thomas's Hospital, in histidine–tryptophan–ketoglutarate or Celsior solution appears to be less damaging for EDHF-mediated responses [529,1710,1753]. Adding hyperpolarizing/relaxing substances such as 11,12-EETs, nicorandil, a KATP channel opener, adenosine or 1-EBIO, an activator of IKCa/SKCa, to the preservation solution attenuates the deleterious effect of these solutions [733,944,1751,1752,1820]. Sepsis.

The release of bacterial lipopolysaccharides (LPS) is in part responsible for vasoplegia and the hyperdynamic state of the circulatory system observed during sepsis. Sepsis and LPS produce vasodilatation by endothelium-dependent and -independent relaxations of the vascular smooth muscle [407,858,1357] and by promoting the expression of iNOS [457].

In various animal models, acute administration of LPS and sepsis, induced by cecal ligation and puncture, inhibits endothelium-dependent hyperpolarizations. This dysfunction can be prevented by inhibiting NO generation or by inhibiting iNOS or nNOS expression [762,802,844,917,1053,1464,1465]. The bacterial release of LPS during sepsis is associated with the generation of cytokines. In human omental arteries, tumor-necrosis factor-α (TNFα) impairs both NO- and EDHF-mediated relaxations [546]. However, in the rat mesenteric artery, TNFα inhibits only the NO-component [1687], but in this tissue, LPS administration does not affect EDHF-mediated responses [483].

By interacting with the heme moiety of cytochrome P450 and inhibiting its activity, NO can markedly affect endothelium-dependent relaxations which rely on EET generation [802]. Additionally, in theory, an increased NO generation could inhibit EDHF-mediated responses by reducing the permeability of Cx37-containing gap junctions [771], by depressing the expression of Cx43 [437] or by activating potassium channels in the vascular smooth muscle [232,233], which would prevent further hyperpolarization produced by an endothelial mechanism [210,1053]. These different mechanisms involved in NO- and prostacyclin-independent relaxations may explain the different susceptibility to LPS and/or sepsis.

In vivo, a treatment with apamin prevents LPS and TNF-induced mortality in mice and the associated hypotension, bradycardia and hypothermia [199,200]. The mechanism underlying this beneficial effect of apamin has not been elucidated.

1.9.2. Enhancement of EDHF-Mediated Responses?

In various cardiovascular pathologies, endothelial-dependent responses resistant to inhibition of NO-synthases and cyclooxygenases can be preserved or even enhanced, and thus are likely to play a compensatory role for the decrease in NO bioavailability. However, these responses are not necessarily attributable to “classical” EDHF-mediated responses. Hypercholesterolemia and Atherosclerosis.

Hypercholesterolemia is generally associated with a preserved or an enhanced contribution of EDHF-mediated responses that compensate for the decrease in NO-mediated relaxation [1390]. The resistance of endothelium-dependent hyperpolarizations to hypercholesterolemia has been demonstrated in arteries from pigs [1700], rabbits [14,141,1522], SHR [759], APOE-deficient mice [330,1075,1692] and dyslipidemic hApoB+/+ mice [847]. In some instances, such as in rabbit arteries, an increase in the generation of lipoxygenase- and/or cytochrome P450-derived products may explain this phenomenon [14,1522]. In the renal artery of the SHR, a high-cholesterol diet augments EDHF-like responses and reduces the responses attributed to NO [759].

However, hypercholesterolemia decreases KIR current in porcine endothelial cells leading to a reduction in flow-mediated vasodilatation [403]. Whether or not the EDHF-mediated responses that are involved in this phenomenon are also affected has not been determined.

In isolated gastroepiploic arteries from atherosclerotic patients, endothelium-dependent hyperpolarizations are inhibited [1579]. The prolonged duration of hypercholesterolemia and the severity of the atherosclerosic process in the human may contribute to the degree of dysfunction of the EDHF pathway [430]. Diabetes.

In contrast to most reports, it has also be shown that in the mesenteric vascular bed and in the femoral artery of streptzotocin-treated rats, the endothelium-dependent vasodilatation induced by histamine or acetylcholine shows a potentiation of the EDHF-mediated component, compensating for the decreased contribution of NO [1405,1775]. The reasons for the discrepancies between the results presented in these two studies and those discussed earlier are unknown. Additionally, the in vivo study of the endothelium-dependent regulation of the renal circulation in streptozotocin-treated rats shows that the EDHF-mediated response, sensitive to the combination of charybdotoxin plus apamin, is fully maintained, while the NO-mediated component is impaired [370]. This may be attributed to a specificity of the renal circulation (see below).

In the db/db-/- mouse, a genetic model of type II diabetes with a mutation on the leptin receptor, the NO-mediated relaxation of the mesenteric artery is reduced while the EDHF-component is preserved, though the mechanism involved in these responses differs in the diabetic and the db/db+/? mice. In control mice, the EDHF-mediated response involves “classical” EDHF-mediated responses (gap junctional communication and K+ ion accumulation), while in the diabetic mice an additional contribution of EETs is observed [1182,1183].

In the saphenous artery of the Sprague–Dawley rat model of cafeteria diet-induced obesity, in contrast to what was observed in the mesenteric artery, the endothelium-dependent relaxations are not affected when compared to lean controls. However, the NO-mediated relaxation, which fully explains the relaxation in control animals, was markedly impaired in the obese animals, but was fully compensated by an EDHF-mediated response. Overexpression of IK1 and increased IKCa function associated with up-regulation of myoendothelial gap junctions explain the maintenance of the endothelial vasodilator function in this artery from obese rats [204]. Ischemia-Reperfusion.

Ischemia-reperfusion is associated with endothelial dysfunction and decreased availability of NO [756,893].

In anesthetized dogs, 1 hour of occlusion of the left circumflex coronary artery followed by 2 hours of reperfusion does not induce major changes in NO-mediated relaxation of the ischemic coronary artery rings, but potentiates the contribution of NO-PGI2-independent dilatations [208]. In isolated perfused rat hearts, ischemia–reperfusion does not reduce these responses, while the availability of NO is decreased [1022]. Several pathways could be involved in these preserved responses resistant to inhibitors of NO-synthases and cyclooxygenases and they include the endothelial release of CNP [654] and the production of CO [1128]. In severe cases of ischemia-reperfusion, the production of peroxynitrite may impair EDHF-mediated responses by directly decreasing KCa activity [936].

Similar results have been obtained in the renal [595] and cerebral circulation [984]. A decrease in the NO-component with an associated increase in EDHF-mediated responses has been observed in the middle cerebral artery of the rat [984]. In the middle cerebral artery and parenchymal arterioles, this compensation is attributed to the “classical” EDHF-mediated responses involving the activation of endothelial SKCa and IKCa, the expression of which remains unaffected by ischemia-reperfusion [253]. Besides, in a model of head injury, the endothelium-dependent NO-mediated relaxation of the middle cerebral artery is abolished while an increase in the EDHF-mediated response maintains endothelium-dependent relaxations [560]. These effects cannot be explained by the disappearance of a tonic repression of the EDHF component exerted by NO under control conditions [1358] but are due to an augmented intracellular endothelial calcium concentration [982]. Heart Failure.

Congestive heart failure, which results in reduced peripheral blood flows, is mainly attributed to dysfunction of the cardiac myocyte. However, endothelial dysfunction with a decreased bioavailability of NO contributes to the development of heart failure [345,928].

In coronary arteries from rats subjected to coronary ligation, which subsequently develop congestive heart failure, flow-induced dilation is not affected when compared to control animals because an EDHF-mediated response, involving KCa activation, compensates for the decrease in NO-mediated vasodilatation [1574]. In this model, similar observations have been made in the endothelium-dependent relaxation of the peripheral mesenteric artery. Additionally, in coronary arteries from a different rat model of right ventricular hypertrophy induced by pulmonary hypertension, the NO-mediated responses were also inhibited and a compensatory mechanism involving the activation of KCa channels, possibly by H2O2, was also reported [766].

A decrease of the NO-dependent component associated with a compensatory role of EDHF-mediated relaxations has also been described in the forearm vasculature of patients with congestive heart failure, in response to the administration of acetylcholine [791]. Hypertension.

Hypertension, as discussed above, is generally associated with an impaired EDHF-mediated response. However, in some models, especially those involving NOS inhibition or deletion, these responses are preserved or even enhanced. Additionally, most of the work showing an impairment of EDHF-mediated responses involves the study of the mesenteric vascular bed. It appears quite clearly now, that some vascular beds, and especially the renal circulation, behave very differently and appear to maintain EDHF-mediated responses, whereas they are severely impaired in other territories. Renal Circulation.

Thus, in SHR perfused kidneys, the EDHF-mediated response is not affected when compared to the vasodilatation observed in normotensive WKY [180,1431]. In aging SHR and in SHR chronically treated with a NOS inhibitor, a more severe model of hypertension, the renal vasodilatation is fully maintained, because the EDHF-mediated component has fully compensated the reduction of the NO-mediated vasodilatation. In both cases, this effect can be attributed to an increase in the functional contribution of the IKCa channel [1431,1432].

EDHF-mediated responses also plays a preponderant role in the murine kidneys [1625]. When receiving a high-salt diet plus a NOS inhibitor, double transgenic mice expressing the human angiotensinogen and renin genes, develop malignant hypertension and renal failure [1395]. In mice expressing the two transgenes, treated or not with the high-salt diet plus the NOS inhibitor, EDHF-mediated responses are maintained and even enhanced when compared to the wild-type animals. NS309 and SKA-31, two activators of endothelial SKCa and IKCa, produced long-lasting vasodilatations that were significantly more sustained in mice with an exacerbated renin–angiotensin system than in control animals [1625]. NOS Inhibition or Deletion.

Acute L-NAME treatment increases the arterial blood pressure and inhibits the fall in blood pressure produced by acetylcholine or bradykinin administration. However, under chronic L-NAME treatment, the blood pressure remains elevated but the hypotensive responses to these two mediators are restored and become more sensitive to blockade by the combination of apamin plus charibdotoxin, suggesting the development of a compensatory mechanism, most likely an EDHF-like-mediated response [310]. Indeed, in vitro, in the mesenteric and kidney arteries, the decrease in NO availability is generally associated with either maintained or enhanced EDHF-mediated responses [807,970,1321,1601,1802]. Similarly, in the guinea-pig, NOS inhibitors are also potent pressor agents [22], but chronic L-NAME treatment does not significantly affect endothelium-dependent hyperpolarizations [275]. In dogs, chronic NOS-inhibition provokes a transient increase in arterial blood pressure followed by a return to control value. In this species, although systemic vascular resistances are increased, the ability to decrease heart rate and cardiac output prevents the occurrence of systemic hypertension [1257]. But again, in the canine coronary arteries of these dogs, the endothelium-dependent hyperpolarizations are not or only very marginally affected [275].

eNOS knock-out mice are moderately hypertensive [685] and both prostacyclin and in particular EDHF-mediated responses appear to play a compensatory role both in vivo and in vitro [142,211,331,332,681,682,1380,1631]. In NOS-disrupted murine arteries, relaxations resistant to inhibitors of NO-synthases and cyclooxygenases are sensitive to catalase and have been attributed to the formation of H2O2 [999,1076]. The maintenance of these endothelium-dependent relaxations/hyperpolarizations in NOS-3 knockout mice was explained by the compensatory endothelial expression of other NOS genes, the production of H2O2 being preserved up to the total disruption of the three NOS genes [1505]. Nevertheless, whether or not the decrease in the EDHF-mediated responses is directly associated with the disruption of the NOS genes or is independently associated with the severe phenotype of these mice, especially with the one observed in the triple knockout mice (hypertension, dyslipidemia, myocardial infarction and nephrogenic diabetes insipidus) remains to be fully assessed.

In conclusion, these findings indicate that endothelium-dependent hyperpolarizations can be maintained during long-term inhibition of NO-synthase and probably act as a back-up mechanism to elicit endothelium-dependent relaxations. The precise role of H2O2 in these responses has yet to be clarified. Miscellaneous.

In the tail artery of young SHR, the endothelium-dependent relaxations are also maintained. This has been attributed to a preserved functional role for EDHF-mediated responses resulting from an increase in the heterocellular myoendothelial coupling, which compensates for other structural changes in the arterial media [1341]. In the weanling Sprague–Dawley rat, a high-salt diet also induces hypertension but, in contrast to the Dahl-salt sensitive rats, no apparent endothelial dysfunction is observed, because an increase in the EDHF-component compensates for the decrease in the NO-mediated relaxation [1445].

Taken in conjunction these results show that hypertension per se does not produce a consistent alteration of the EDHF-mediated responses. In models, where such an alteration is observed, the question remains as to whether this dysfunction could contribute to the genesis of the syndrome or is a consequence of the hypertensive process.


No drug is available designed to target EDHF-mediated responses. Nevertheless, therapeutic interventions with beneficial effects on the cardiovascular system such as angiotensin-converting enzyme inhibitors, antagonists of angiotensin receptors and phosphodiesterase-3 inhibitors [429,430,1004] can restore these responses, suggesting that the improvement in the EDHF pathway contributes to the observed beneficial effect. Similarly, various so-called non-pharmacological therapeutic strategies including exercise and supplementation with estrogens, omega-3 polyunsaturated fatty acids, polyphenol derivatives, potassium and/or calcium help to reverse endothelial dysfunction, including blunted EDHF-mediated responses [429,430].

The improvement or restoration of EDHF responses has not yet been the direct purpose of any pharmaceutical effort. SKA-31, a preferential activator of murine IKCa, potentiates EDHF-mediated responses in vitro in normotensive mice and in mice with eNOS deletion. This compound also lowers mean arterial blood pressure in normotensive and in angiotensin II-hypertensive mice [617,1347]. However, IKCa channels are also required for the differentiation of vascular smooth muscle cells, as well for their proliferation and migration [833,1117,1542,1543]. Selective blockade of IKCa [1717] prevents phenotypic changes in smooth muscle and coronary artery neointimal formation in two different models of post-angioplasty restenosis and the development of atherosclerosis in ApoE(-/-) mice [833,1543,1560]. IKCa are also involved in the proliferation of endothelial [579] and various cancerous cells [729,1655]. Therefore, activation of IKCa might be a double-edged sword and be associated with some unwanted detrimental effects. Alternatively, SKCa could substitute for IKCa function, as suggested by the partial restoration of endothelium-dependent vasodilatation in IK1 knockout mice overexpressing SK3 [135]. However, IK1 and SK3 are activated by different stimuli and serve different functions [135,1040,1693]. Furthermore, SK3 channels are widely expressed, especially in the central nervous system, and the result of a prolonged activation of this population of potassium channels is basically unknown [109,919,1716].

Additionally, activation of endothelial TRP channels, calcium sensing receptors, smooth muscle KIR and/or specific isoform(s) of Na+/K+-ATPase as well as facilitating myoendothelial communication and increasing the expression of appropriate connexins, channels and receptors may represent new potential targets. However, the precise role of these various molecular elements is far from being completely understood. For instance, TRPV4 channel could appear as a promising target in cardiovascular diseases since this cationic channel is involved in calcium entry following endothelial stimulation [1125]. Indeed, the arterial responses to shear stress critically depend on the activation of this endothelial channel and both the NO and the EDHF-mediated components of acetylcholine-induced vasodilatation are attenuated in TRPV4-deficient mice [616,1792]. However, GSK1016790A, a specific and potent agonist which, as expected, increases endothelial intracellular calcium concentration and produces endothelium-dependent relaxations, also causes endothelial failure, circulatory collapse and death [1553,1680].

Rotigaptide (ZP123), an antiarrhythmic peptide that prevents uncoupling of Cx43-mediated gap junction communication [818], has no effect on basal vascular tone and does not enhance endothelium-dependent or -independent vasodilatation in the forearm arterial circulation of healthy subjects [873]. Whether or not augmenting Cx43 communication would improve endothelial function in patients with vascular disease and whether or not Cx40, in humans, would be a more appropriate target than Cx43 remains to be determined.

Copyright © 2011 by Morgan & Claypool Life Sciences Publisher.
Bookshelf ID: NBK57107