Endothelium-dependent hyperpolarization. Beyond nitric oxide and cyclic GMP.

Abstract

HomeCirculationVol. 92, No. 11Endothelium-Dependent Hyperpolarization Free AccessResearch ArticleDownload EPUBAboutView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessResearch ArticleDownload EPUBEndothelium-Dependent Hyperpolarization Beyond Nitric Oxide and Cyclic GMP Richard A. Cohen and Paul M. Vanhoutte Richard A. CohenRichard A. Cohen From the Vascular Biology Unit (R.A.C.), Vascular Medicine Section, Robert Dawson Evans Department of Clinical Research, Boston University School of Medicine, Boston, Mass; and Institut de Recherche Servier (P.M.V.), Suresnes, France. Search for more papers by this author and Paul M. VanhouttePaul M. Vanhoutte From the Vascular Biology Unit (R.A.C.), Vascular Medicine Section, Robert Dawson Evans Department of Clinical Research, Boston University School of Medicine, Boston, Mass; and Institut de Recherche Servier (P.M.V.), Suresnes, France. Search for more papers by this author Originally published1 Dec 1995https://doi.org/10.1161/01.CIR.92.11.3337Circulation. 1995;92:3337–3349Blood flow and blood pressure are determined by an integration of reflex, humoral, and local vascular control mechanisms. Knowledge of these mechanisms has mushroomed over the past 15 years, particularly in the area of local endothelium-dependent vasomotor control. This has stemmed from the pioneering report in 1980 by Robert Furchgott1 demonstrating that the endothelium releases a vasodilator substance in response to acetylcholine. This concept has been expanded with knowledge that the endothelium releases a variety of relaxing and contracting factors that regulate the underlying smooth muscle.2 The most widely known endothelium-derived relaxing factor, nitric oxide, is released from endothelial cells in response to shear stress or stimulation of different receptors for a variety of neurohumoral mediators on the endothelial cell surface.3 The increase in endothelial cell calcium initiated by these stimuli increases the activity of a constitutively expressed enzyme, nitric oxide synthase, which converts l-arginine to nitric oxide and citrulline.4567 Nitric oxide thus formed diffuses to and inhibits contraction of the underlying vascular smooth muscle.3 The physiological significance of the production of endothelial nitric oxide is suggested by the vasoconstriction observed in most vascular beds8910111213 and the increase in systemic arterial blood pressure,141516 which occurs on infusion of inhibitors of nitric oxide synthase. This observation further implies that under normal conditions, endothelial cells are locally liberating nitric oxide which effectively inhibits vasoconstriction arising by other mechanisms. Thus, normal vascular homeostasis depends in the periphery on a balance between neurally and humorally mediated vasoconstriction in skeletal muscle, mesentery, and the kidney, and local endothelium-dependent vasodilatation. In the truly vital areas of the heart, brain, and genitalia, vasodilator neural mechanisms reinforce the vasodilator influence of the endothelium.17All is not known regarding the vasodilator function of the endothelium. In the present article, we elaborate that endothelium-dependent vasodilatation can best be explained by the participation of at least two vasodilator substances, nitric oxide and a substance whose identity is unknown. The identification of nitric oxide as an endothelium-derived relaxing factor rests on the similarities in physical, chemical, and physiological characteristics between the endogenous substance and authentic nitric oxide, as well as evidence derived from the use of arginine analogues as competitive inhibitors of nitric oxide synthesis.1819 Thus, although the precise chemical nature of the nitric oxide–like endothelium-derived vasodilator remains debatable,20 the evidence appears incontrovertible that nitric oxide, in one chemical form or another, participates in endothelium-dependent relaxation. On the other hand, it has long been known that acetylcholine, the prototypical endothelium-dependent vasodilator, causes hyperpolarization of smooth muscle,2122232425 and the endothelium dependence of the response was first described by Bolton and colleagues.26 Hyperpolarization of smooth muscle is an action that has not been generally associated with nitric oxide.272829 In addition, in many studies of isolated blood vessels and the intact circulation, the action of endothelium-dependent vasodilators is, at least in part, resistant to inhibitors of nitric oxide.9303132333435363738394041424344 Although the number of observations in human blood vessels is limited, endothelium-dependent hyperpolarization of human coronary arteries in response to bradykinin also is resistant to nitric oxide synthase inhibitors.43 Therefore, the existence of two endothelium-derived mediators has been proposed to account for endothelium-dependent relaxation—one nitric oxide, and the other an endothelium-derived hyperpolarizing factor (EDHF).454647484950 It is important to note that hyperpolarization is, thus far, a phenomenon that has been measured only in vitro. Thus, the existence of a hyperpolarizing factor in vivo has been deduced to play a role in vasodilatation by the noted failure of nitric oxide synthase inhibitors to completely block endothelium-dependent vasodilation. Furthermore, the picture is complicated by the fact that at least two different modes of signal transduction may account for endothelium-dependent vasodilation, even for a single endothelium-derived relaxing factor. The first, described for nitric oxide, depends on the stimulation of guanylate cyclase and production of its product, cGMP. Reliance of endothelium-dependent vasodilation on this mechanism is based on the parallel drawn between the increase in cGMP content of arterial tissue caused by endothelium-dependent vasodilators and nitrovasodilators, whose action is based on releasing nitric oxide.5152535455 In addition, endothelium-dependent relaxation to nitric oxide may be reduced by hemoglobin or methylene blue, which antagonize the rise in cGMP either by inhibiting guanylate cyclase or by scavenging nitric oxide and preventing its stimulation of the enzyme.53545657585960 The cyclic nucleotide causes smooth muscle relaxation chiefly via phosphorylation of, and multiple actions mediated by cGMP-dependent kinase(s), and thus to a decrease in intracellular calcium and a decrease in calcium sensitivity of the contractile apparatus.61626364656667The second mechanism is largely independent of cGMP and may be mediated by hyperpolarization of the smooth muscle cell membrane. Having demonstrated efficacious inhibition of the rise in arterial cGMP by arginine analogue inhibitors of nitric oxide synthesis or by methylene blue or hemoglobin, potent relaxation303244 and hyperpolarization3768 to bradykinin or acetylcholine persist in the porcine coronary artery,3044 the carotid artery and abdominal aorta of the rabbit,32 and the aorta and pulmonary artery of the rat.37 These observations attest to cGMP-independent mechanisms of endothelium-dependent relaxation. A cGMP-independent mechanism of relaxation may also apply to nitric oxide and nitrovasodilators.6970 Thus, there is evidence for more than one endothelium-derived mediator, one or all of which may mediate relaxation by more than one mechanism. It is also possible that the two mediators are interrelated, either by sharing common mechanisms or by reinforcing or otherwise modulating the action of the other (Figs 1 and 2). The purpose of this review is to summarize the current understanding, gaps in knowledge, and potential directions for research on endothelium-dependent vasodilation as it relates to the question of the existence, identification, and mode of action of a second, as-yet-unidentified mediator that behaves principally as an EDHF. Mechanism of Endothelium-Dependent Hyperpolarization Endothelial Signal Vascular smooth muscle cell hyperpolarization may be initiated by several autacoids and hormones in addition to acetylcholine, including histamine,71 bradykinin,4368 substance P,7273 calcitonin gene–related peptide,73 ADP,74 endothelin,757677 and vasoactive intestinal polypeptide.72 The hyperpolarization of the smooth muscle usually precedes and is more transient than the accompanying relaxation induced by endothelium-dependent vasodilators.28377178 The mediator is believed to be a soluble transferable factor because of the success of some investigators in passing a hyperpolarizing substance between a vessel with endothelium to one that has been denuded of endothelium.7980 The release of EDHF, like that of nitric oxide, is believed to be initiated by an increase in free calcium concentration in the endothelial cell. This conclusion is based on the ability of the calcium ionophore A23187 to cause endothelium-dependent hyperpolarization and of calcium-free solutions to make the hyperpolarizations to acetylcholine more transient and prevent the action of A23187.81 The transient response to acetylcholine in calcium-free solutions can be attributed to the release of intracellular endothelial calcium stores, whereas the entry of extracellular calcium into the endothelial cell prolongs the response to acetylcholine or bradykinin and accounts entirely for the response to A23187. Like the release of nitric oxide, endothelium-dependent hyperpolarization apparently depends on calcium calmodulin, because inhibitors of this enzyme, calmidazolium and fendiline, decrease the endothelium-dependent hyperpolarization and relaxations of the porcine coronary artery to bradykinin, which are resistant to inhibitors of nitric oxide.8283Electrical Conduction An important issue that remains unresolved is a potential electrical conduction of hyperpolarization between endothelial and smooth muscle cells,8485 which might explain the inability to assay a soluble EDHF in arteries such as the porcine coronary artery.8687 Although simultaneous hyperpolarization of endothelial and smooth muscle cells occurs in the porcine coronary artery in response to bradykinin, it could not be demonstrated by current injections that the two cell types are electrically connected or by lucifer yellow injections that gap junctions exist between the two cell types.858889 These studies likely exclude, at least in this artery, the possibility that hyperpolarization is conducted between the endothelial and smooth muscle cells. The possibility of electrical coupling between endothelial and smooth muscle cells may be more likely in the microcirculation, where contacts between the two cell types are much more intimate.9091Vasoconstrictors such as phenylephrine, U46619, and endothelin have been reported to depolarize the overlying endothelial cells.92 Hyperpolarization by acetylcholine reduces the endothelial depolarization induced by these vasoconstrictors. The depolarization of the endothelial cell by the contractile agonist does not occur if the endothelium is separated from the smooth muscle, even if the smooth muscle and endothelium are reapposed, suggesting electrical transmission or an extremely short-lived mediator.92 Direct electrical signaling between endothelial and smooth muscle cells, if it is shown to occur in some arteries, could explain the difficulty in identifying a distinct diffusable EDHF. Endothelial Cell Membrane Potential Membrane potential is as important a determinant of endothelial cell function as it is for smooth muscle cells. Endothelial cells also are hyperpolarized by agents that cause endothelium-dependent hyperpolarization,85929394 an event that has been attributed to the opening of calcium-dependent potassium channels on the endothelial cell membrane in response to the increase in cell calcium.9596979899100101102103104 Because the endothelial cell lacks typical voltage-dependent calcium channels,97105106 the influx of calcium into the hyperpolarized cell is dependent on and enhanced by the greater electrochemical gradient for the cation107 (Fig 1). Although it is not known if shear stress regulates EDHF release, shear stress–induced release of nitric oxide from endothelial cells is inhibited by calcium-dependent potassium channel blockers, emphasizing the importance of these potassium channels102 to the release mechanism. Thus, endothelial cell hyperpolarization may enhance release of all endothelial cell autacoids whose liberation depends on calcium, including EDHF.93108109 Because endothelial cells release bradykinin,110 adenine nucleotides,111112 and endothelin,113 these autocoids could be involved in autocrine regulation of nitric oxide and EDHF release. It is even possible that EDHF acts in an autocrine fashion, hyperpolarizing the endothelial cell, and that a significant physiological role of EDHF is to modulate the release of nitric oxide and other autacoids, such as prostaglandins. This might occur by potassium channel or calcium channel activation. For example, an endothelial cell cytochrome P-450 product that is a potential EDHF, 5,6-epoxyeicosatrienoic acid, increases endothelial cell calcium influx.114 Cytochrome P-450 inhibitors diminish the sustained endothelial cell calcium rise caused by histamine, suggesting that endogenous epoxyeicosatrienoic acids synthesized by endothelial cells promote calcium influx and therefore stimulate autacoid release.114Response of Smooth Muscle Cells to EDHF How Does Hyperpolarization of the Smooth Muscle Occur? The hyperpolarization of the smooth muscle membrane mediated by the endothelium has been attributed to an increase in conductance to potassium ions. The evidence for this rests on the facts that (1) the cell membrane conductance of smooth muscle measured electrically during hyperpolarization is increased242671 ; (2) the magnitude of endothelium-dependent hyperpolarization is decreased in a concentration-dependent manner by extracellular potassium71 ; (3) radioactive rubidium efflux is increased during hyperpolarization37115 ; (4) the endothelium-dependent hyperpolarizations in certain cases have been prevented by potassium channel blockers2780116 ; and (5) many parallels can be drawn between the action of EDHF and pharmacological potassium channel openers, which cause relaxation primarily by causing hyperpolarization.117The role of potassium channels has been proposed also because these channels, particularly large conductance calcium-dependent potassium channels, account for the passage of large amounts of ion current across the membrane, far greater than, for example, the Na+,K+-ATPase, and because the membrane potassium gradient accounts to the greatest extent for the resting membrane potential.118 It has been estimated that only a small increase in the opening rate of large conductance potassium channels is required to nearly maximally hyperpolarize the cell membrane.118 Nevertheless, other potential explanations exist that could contribute to endothelium-dependent hyperpolarization, including activation of the Na+,K+-ATPase79 or inactivation of chloride channels.119120 For example, taking the latter as a hypothetical example, vasoconstrictors increase smooth muscle cell intracellular calcium, which can lead to depolarization, in part due to the activation of calcium-activated chloride channels. Nitric oxide released from the endothelium decreases intracellular calcium in the smooth muscle.121 Thus, the decrease in calcium caused by nitric oxide could deactivate chloride channels and account for the hyperpolarizing effects of nitric oxide. This example points out that reduction in intracellular calcium induced by nitric oxide, or by another EDHF, could cause secondary changes in membrane potential that may contribute to the regulation of contractile tone. Thus, although experiments with nitric oxide synthase inhibitors suggest that endothelial factors other than nitric oxide can cause hyperpolarization, it is difficult to exclude membrane potential changes due to endothelium-derived nitric oxide when its synthesis is not blocked. There are several important additional caveats of which to be aware when considering a potential regulation of membrane potential in mediating endothelium-dependent relaxation. First, there is only a certain range of membrane potentials over which an EDHF that operates by opening potassium channels might be expected to contribute significantly to the regulation of membrane potential. On the one hand, very negative membrane potentials are close to the equilibrium potential for potassium, and therefore opening of potassium channels would have little further effect. On the other hand, at depolarized potentials the conductance for potassium is already high, so the effect of further increasing the opening of channels may be negligible. Thus, just as changes in intracellular calcium could cause secondary changes in membrane potential, differences in resting membrane potential could explain differences in experimental results that have otherwise been attributed to the altered release or the absence of EDHF. Second, caution must be exercised in interpreting experiments in which inhibitors of hyperpolarizing mechanisms are used, such as ouabain to block the Na+,K+-ATPase, glibenclamide to inhibit ATP-dependent potassium channels, or tetraethylammonium and charybdotoxin to block calcium-dependent potassium channels. Although these inhibitors may very well inhibit their intended target, they also may, simply by depolarizing the membrane, nonspecifically inhibit a response initiated by the mechanism that is the actual target of EDHF in the smooth muscle. In addition, many of the inhibitors have been used in concentrations that greatly exceed the range over which they are considered specific. For example, studies have been performed using glibenclamide to inhibit ATP-dependent potassium channels at concentrations exceeding the micromolar range that is more than 10 000 times higher than the concentration that inhibits the ATP-dependent channels in pancreatic β cells.122 At micromolar concentrations, glibenclamide can inhibit responses mediated by calcium-activated potassium channels.123124 Another factor that complicates interpretation of experiments is that both endothelial cells and smooth muscle cells possess a large complement of ion channels. None of the inhibitors are specific for one cell type or the other, making it difficult to determine if a pharmacological agent is affecting EDHF release from endothelial cells, its action on endothelial cells, or its action on the smooth muscle. Third, contractile agents can have various independent effects on membrane potential or intracellular calcium, both of which can influence the response to an EDHF. To avoid movement of the preparation during contraction and the associated difficulty in maintaining cell impalements, many measurements of membrane potential in response to endothelium-dependent agonists have been made in the absence of contractile agents. Different membrane potential responses to endothelium-dependent agents have been recorded in the presence and absence of contractile agents,125126 so caution is necessary in extrapolating data obtained in resting vascular smooth muscle as to how it relates to the mechanisms by which contracted smooth muscle relaxes. These factors emphasize the importance of future studies to determine the role of EDHF in vivo independent of exogenous contractile agents and at physiological membrane potentials. How Does Hyperpolarization Cause Relaxation? The mechanism by which hyperpolarization causes relaxation is controversial. The most direct and obvious explanation is that hyperpolarization of the smooth muscle cell membrane inhibits the opening of voltage-dependent calcium channels, allows calcium sequestration and removal mechanisms to lower intracellular calcium, and leads to relaxation.127 Although this mechanism may operate in some blood vessels under some conditions, it does not fully explain the mechanism of relaxation. As has been pointed out previously,117 the isolated rabbit thoracic aorta is not depolarized during contractions induced by norepinephrine, and those contractions are poorly inhibited by blockers of voltage-dependent calcium channels such as nifedipine. Nevertheless, potassium channel activators, like chromakalim, whose relaxations are attributed solely to opening of potassium channels and hyperpolarization, induce potent relaxations of that preparation. Taken together, these observations suggest that potassium channel openers may relax by mechanisms other than inhibition of voltage-dependent calcium channels.117This point was made more evident in an investigation of the bradykinin-induced relaxations of the porcine coronary artery that persist in the presence of nitric oxide synthesis inhibitors. The artery was treated with nifedipine before contracting the artery with the thromboxane mimetic, U46619.128 Despite the fact that potassium could no longer contract the artery in the presence of nifedipine (indicating that voltage-dependent calcium channels were blocked), the potassium-induced depolarization inhibited the relaxations to bradykinin. This is consistent with the relaxations being mediated by a hyperpolarizing mechanism. However, in the absence of potassium, the bradykinin-induced relaxations of arteries treated with nifedipine were similar to those observed by others in the absence of nifedipine.3068129 This suggests that inhibition of voltage-dependent channels, at least of the L-type blocked by nifedipine, is not important in mediating relaxations associated with hyperpolarization that occur in the presence of nitric oxide synthase inhibitors and makes it likely that EDHF can induce relaxations by other mechanisms. Potassium channel openers may relax by inhibiting vasoconstrictor-induced, phospholipase C–mediated inositol triphosphate production that would decrease intracellular calcium release.130131132 Also, possibly as a result of a reduction in diacylglycerol production by phospholipase C, there is a reduction in calcium sensitivity mediated by protein kinase C.133134135136 These actions of potassium channel openers can all be attributed to hyperpolarization because they are prevented by elevated extracellular potassium and are blocked by glibenclamide.117 There are several lines of additional evidence that the potassium channel openers also may relax by mechanisms independent of the changes in membrane potential that they produce,117 possibly by enhancing the refilling of intracellular calcium stores.137138 These observations regarding potassium channel openers have important implications for mechanisms attributed to an EDHF. First, relaxations caused by EDHF may differ depending on the mechanism by which the contractile agonist causes contraction. Contractions mediated by agonists that depolarize and activate voltage-dependent calcium channels may be affected differently by a hyperpolarizing stimulus than are those that do not. Thus, rather than indicating that different endothelial factors relax an artery contracted by two different agents, it may be that the difference is explained by the differences in the mechanisms of contraction.139 Second, one should be aware that an EDHF could potentially relax by mechanisms other than those dependent on membrane hyperpolarization, and its effect on membrane potential may be an additional mechanism of relaxation or even an epiphenomenon. Nature of EDHF Two Known EDHFs: Nitric Oxide and Prostacyclin Nitric oxide and prostacyclin are two known endothelial products that can be released from the endothelium in sufficient quantities to cause relaxation and hyperpolarization at least in some blood vessels and thus fulfill the criteria to be called EDHFs. Nitric oxide and prostacylin are distinguishable from a putative distinct EDHF by having their hyperpolarizing and relaxing actions annulled by competitive inhibitors of nitric oxide synthase or inhibitors of cyclooxygenase, respectively. Prostacyclin Prostacyclin contributes to endothelium-dependent relaxation of several isolated blood vessels and vasodilation of perfused organs.140141142143144145146147 The vasodilation of the rabbit coronary circulation by prostacyclin is blocked by glibenclamide,148 suggesting an underlying opening of potassium channels as the mechanism. Prostacyclin causes relaxation by increasing cAMP in smooth muscle cells. Prostacylin has also been shown to hyperpolarize cultured canine carotid artery smooth muscle cells,149 and a prostacylin analogue hyperpolarizes the isolated guinea pig coronary artery.150 In the latter study, cAMP also hyperpolarized the smooth muscle, suggesting that hyperpolarization may occur via cyclic nucleotide-dependent protein kinases, which are known to modulate potassium channels. Likewise, in the canine saphenous vein denuded of endothelium, other activators of adenylate cyclase induce hyperpolarization by opening ATP-dependent potassium channels.151Nitric OxideNitric oxide in solution or nitrovasodilators can hyperpolarize smooth muscle of several different isolated blood vessels, including the small mesenteric artery125 and aorta152 of the rat, the uterine126 and coronary artery of the guinea pig,150 and the basilar artery of the rabbit.78 Nitric oxide in concentrations exceeding 5×10−7 mol/L hyperpolarized all vessels studied, including various arteries of rat, rabbit, dog, and guinea pig.126 In some arteries, nitric oxide caused relaxation at lower concentrations without hyperpolarization, but hyperpolarization was always observed at higher concentrations. In the present study, it was found to be necessary to depolarize and contract the arteries with phenylephrine to observe hyperpolarization to nitric oxide. The failure to observe a hyperpolarizing response to nitric oxide in resting arteries was attributed to a highly negative resting membrane potential. Stretching the guinea pig coronary artery was found to be necessary to observe nitric oxide−induced hyperpolarizations.150 In contrast, nitric oxide hyperpolarized the small mesenteric artery of the rat only in the absence of contractile force.125 These differences may, as discussed previously, depend on differences in resting membrane potential in the various preparations or on other experimental conditions. Although the concentrations of nitric oxide in vivo are not known, the concentrations that caused hyperpolarization in several of the above studies125126150 were in the micromolar range, which is equal to the concentration of nitric oxide measured in rabbit aortic medial smooth muscle after stimulation of the endothelium by acetylcholine.153It was also demonstrated that the hyperpolarization to acetylcholine in the guinea pig artery was reduced by 41% by a relatively low concentration of NG-monomethyl-l-arginine.126 Another nitric oxide synthase inhibitor, NG-l-arginine methyl ester, as well as NG-monomethyl-l-arginine, had no effect on peak acetylcholine-induced hyperpolarization in the rat aorta but made the response more transient, reducing the more sustained hyperpolarization by about 50%.152 These studies suggest that at least under some circumstances, the release of nitric oxide accounts for part of the hyperpolarizing response to endothelium-dependent vasodilators.Even in the absence of an effect of inhibitors of nitric oxide on the hyperpolarization stimulated by endothelium-dependent vasodilators, hemoglobin,2937 methylene blue,71126 or nitric oxide synthase inhibitors152154 can depolarize resting arteries. This suggests that nitric oxide tonically released from the endothelium can under certain circumstances hyperpolarize the smooth muscle cell. Despite these findings, removal of the endothelium does not usually depolarize the underlying smooth muscle cells, as one would expect by removing a source of hyperpolarizing factor.2728294371778081151 This is perhaps due to the physical trauma to the blood vessel associated with removal of the endothelium or to the simultaneous removal of depolarizing factors released by the endothelium. One mechanism by which nitric oxide could hyperpolarize smooth muscle is by increasing cGMP, which in turn causes protein kinase–dependent activation of calcium-dependent potassium channels.67155156 In addition, ATP-dependent potassium channels appear to be activated by nitric oxide, through a cGMP-dependent mechanism.157 All endothelium-dependent relaxations and hyperpolarizations have been attributed to cGMP-dependent effects on potassium channels.158 This is clearly not the case, as demonstrated by endothelium-dependent relaxations that persist after blocking the rise in cGMP.303237 Some investigators have relied on the ability of methylene blue to block cGMP,27287174126158159 without accounting for its ability to scavenge nitric oxide,57 as well as to inhibit nitric oxide synthesis by the endothelium.160Nitric oxide in concentrations from 5×10−7 to 10−5 mol/L increases the activity of single calcium-dependent potassium channels in isolated membrane patches of rabbit aortic smooth muscle cells.70 Because the membrane patches in these experiments were devoid of all nucleotides and cofactors necessary for cGMP generation, this observation provides a mechanism whereby nitric oxide could hyperpolarize smooth muscle cells independent of cGMP. Relaxations to nitric oxide of the rabbit aorta were incompletely blocked by methylene blue, which effectively prevented the increase in cGMP, and the residual relaxations were blocked by charybdotoxin.70 This suggests that cGMP-independent potassium channel activation by nitric oxide can cause relaxation. These effects of exogenous nitric oxide were reproduced by native EDRF. In this case, the