Physiol Rev Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Physiol. Rev. 78: 53-97, 1998;
0031-9333/98 $15.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by FARACI, F. M.
Right arrow Articles by HEISTAD, D. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by FARACI, F. M.
Right arrow Articles by HEISTAD, D. D.

PHYSIOLOGICAL REVIEWS   Vol. 78 No. 1 January 1998, pp. 53-97
Copyright ©1998 by the American Physiological Society

Regulation of the Cerebral Circulation: Role of Endothelium and Potassium Channels

FRANK M. FARACI AND DONALD D. HEISTAD

Departments of Internal Medicine and Pharmacology, Cardiovascular Center and Center on Aging, University of Iowa College of Medicine, Iowa City, Iowa

I. INTRODUCTION
II. DISTINCTIVE CHARACTERISTICS OF CEREBRAL BLOOD VESSELS
    A. Humoral Stimuli
    B. Neural Stimuli
    C. Metabolic Stimuli
    D. Hypercapnia and Hypoxia
    E. Autoregulation
    F. Flow-Mediated Vasodilatation
III. ENDOTHELIUM-DERIVED VASOACTIVE FACTORS
    A. Nitric Oxide
    B. Prostacyclin
    C. Endothelium-Derived Hyperpolarizing Factor
    D. Endothelin
    E. Other Endothelium-Derived Vasoactive Factors
IV. INDUCIBLE NITRIC OXIDE SYNTHASE
V. POTASSIUM CHANNELS
    A. ATP-Sensitive Potassium Channels
    B. Calcium-Dependent Potassium Channels
    C. Voltage-Dependent Potassium Channels
    D. Inward-Rectifier Potassium Channels
VI. HYPERTENSION
    A. Acute Hypertension
    B. Chronic Hypertension
VII. HYPERCHOLESTEROLEMIA AND ATHEROSCLEROSIS
VIII. DIABETES
IX. AGING
X. ISCHEMIA
XI. SUBARACHNOID HEMORRHAGE
    A. Endothelium-Dependent Relaxation
    B. Endothelin
    C. Potassium Channels
XII. MENINGITIS
XIII. CONCLUSIONS AND FUTURE DIRECTIONS
REFERENCES

    ABSTRACT
Top
References

Faraci, Frank M., and Donald D. Heistad. Regulation of the Cerebral Circulation: Role of Endothelium and Potassium Channels. Physiol. Rev. 78: 53-97, 1998. --- Several new concepts have emerged in relation to mechanisms that contribute to regulation of the cerebral circulation. This review focuses on some physiological mechanisms of cerebral vasodilatation and alteration of these mechanisms by disease states. One mechanism involves release of vasoactive factors by the endothelium that affect underlying vascular muscle. These factors include endothelium-derived relaxing factor (nitric oxide), prostacyclin, and endothelium-derived hyperpolarizing factor(s). The normal vasodilator influence of endothelium is impaired by some disease states. Under pathophysiological conditions, endothelium may produce potent contracting factors such as endothelin. Another major mechanism of regulation of cerebral vascular tone relates to potassium channels. Activation of potassium channels appears to mediate relaxation of cerebral vessels to diverse stimuli including receptor-mediated agonists, intracellular second messengers, and hypoxia. Endothelial- and potassium channel-based mechanisms are related because several endothelium-derived factors produce relaxation by activation of potassium channels. The influence of potassium channels may be altered by disease states including chronic hypertension, subarachnoid hemorrhage, and diabetes.

    I. INTRODUCTION
Top
Next
References

Vascular tone in the cerebral circulation is regulated by several major mechanisms. One mechanism that has been under intense investigation involves endothelial factors. Endothelium produces and releases potent relaxing and contracting factors that regulate tone of underlying vascular muscle and may also influence vascular growth. A second major area of research involves regulation of cerebral vascular tone by activity of potassium channels. Activation of potassium channels appears to mediate relaxation of cerebral blood vessels in response to a diverse group of important stimuli.

This review summarizes concepts concerning the role of endothelium-derived vasoactive factors and potassium channels in the cerebral circulation. These two mechanisms are related because several endothelium-derived factors produce relaxation by activation of potassium channels. We discuss the functional importance of these mechanisms and, when data are available, review molecular mechanisms that contribute to regulation of cerebral vascular biology. Some abnormalities that occur in cerebral blood vessels under pathophysiological conditions are also reviewed. We focus on these topics because they are areas of intense investigation and because abnormalities in these mechanisms appear to play a key role in brain vascular pathophysiology.

    II. DISTINCTIVE CHARACTERISTICS OF CEREBRAL BLOOD VESSELS
Top
Previous
Next
References

There are basic principles of regulation of blood flow that apply in general to all vascular beds. There also, however, are some major differences between cerebral blood vessels and vessels in other organs. We first briefly review some distinctive characteristics of the cerebral circulation.

A. Humoral Stimuli

The endothelial blood-brain barrier limits access of many humoral stimuli to smooth muscle of cerebral blood vessels. It has been assumed that failure of humoral stimuli to alter cerebral blood flow was the result of absence of vasomotor effects of these stimuli on the blood vessels, because the endothelial blood-brain barrier prevents access to vascular muscle. In some cases, however, humoral stimuli can selectively alter resistance of large cerebral arteries, without altering blood flow, because small vessels compensate (presumably via an autoregulatory response) (204). In contrast, humoral stimuli can have major effects on blood flow to regions of the brain such as the choroid plexus, which lack a blood-brain barrier (203, 347, 479). For example, increases in plasma levels of vasopressin to concentrations observed under pathophysiological conditions produce marked reductions in blood flow to choroid plexus (203).

Thus the concept that the blood-brain barrier reduces vascular responses to humoral stimuli is sound. A more contemporary view, however, is that some humoral stimuli produce opposing vascular responses in large and small vessels, which result in failure of the stimuli to alter net blood flow. In addition, humoral stimuli may have significant effects on blood flow to circumventricular organs.

B. Neural Stimuli

Cerebral blood vessels have dense innervation from autonomic and sensory fibers. The sources of this innervation include sympathetic nerves (originating predominately in the superior cervical ganglion), parasympathetic fibers (originating primarily in the sphenopalatine, otic, and internal carotid ganglia), and the trigeminal nerve (originating in the trigeminal ganglion) (266). Although perivascular innervation of cerebral vessels is relatively abundant (63), the functional significance of much of this innervation is poorly defined.

Under normal conditions, sympathetic stimulation has little effect on cerebral blood flow, in contrast to far greater effects in other vascular beds (49, 281). Sympathetic stimulation constricts large cerebral arteries, but small vessels downstream dilate (probably an autoregulatory response to a decrease in intravascular pressure), so that blood flow does not decrease (49). Although sympathetic stimuli have little effect under normal conditions, they have profound and physiologically important effects during acute hypertension, because neural stimuli attenuate increases in cerebral blood flow (282). The functional significance of parasympathetic innervation is less clear, although it is known that electrical stimulation of fibers that originate in the sphenopalatine ganglion increase cerebral blood flow (555). Sensory fibers originating in the trigeminal ganglion appear to modulate constrictor responses of cerebral blood vessels (170, 516, 558) and contribute to increases in cerebral blood flow that occur during meningitis, cortical spreading depression, seizures, and reactive hyperemia (134, 557, 699, 798, 807).

C. Metabolic Stimuli

Cerebral circulation, like most other vascular beds (e.g., coronary, mesenteric, and skeletal muscle), but in contrast to some other vascular beds (renal and cutaneous), is characterized by "coupling" of changes in metabolism and blood flow (281, 311). There may be multiple mechanisms by which changes in neuronal activity and metabolism produce a corresponding change in blood flow. For example, changes in tissue concentration of adenosine, lactate, and tissue PO2 , PCO2 , and pH may contribute to increases in blood flow during increases in cerebral metabolism. Recent findings strongly support the concept that release of nitric oxide by neurons plays a critical role in producing increases in blood flow during metabolic stimuli (neuronal activation) (183, 190, 191, 193, 195, 312, 315, 530, 596, 640).

D. Hypercapnia and Hypoxia

In some vascular beds (e.g., renal, cutaneous, and skeletal muscle), moderately severe levels of hypercapnia and hypoxia have relatively small effects on blood flow (278). Although hypercapnia and hypoxia have a direct dilator effect in these vascular beds, neurohumoral stimuli (especially the chemoreflex) produce vasoconstriction so that blood flow usually fails to increase (278).

In contrast, hypercapnia and hypoxia are extremely potent vasodilators in the cerebral circulation. The vasoconstrictor response to the chemoreflex is minimal in the cerebral circulation so that the direct vasodilator effects of hypercapnia and hypoxia typically produce large increases in cerebral blood flow. Recent studies suggest that cerebral vasodilatation in response to hypercapnia is dependent on formation of nitric oxide (184, 192, 310, 313, 316, 321, 330, 523, 801). Although formation of nitric oxide may contribute to mechanisms that mediate increases in cerebral blood flow during hypoxia (28, 36, 643, 675, 749), additional mechanisms involving formation of adenosine and activation of potassium channels may also be important (28, 226, 412, 675, 676, 711, 750, 811).

E. Autoregulation

Changes in perfusion pressure produce marked changes in cerebrovascular resistance and, therefore, contribute to maintenance of relatively constant levels of blood flow over a wide range of pressures. Autoregulation seems to be particularly effective in brain, renal, and mesenteric vessels and less effective in cutaneous and perhaps coronary vessels (281). Mechanisms that mediate autoregulation of cerebral blood vessels may include myogenic responses, metabolic factors, neural mechanisms, and activation of potassium channels (91, 443, 638).

F. Flow-Mediated Vasodilatation

Large arteries play a surprisingly large role in regulation of vascular resistance in the brain (199). Large arteries play a key role in regulation of the cerebral circulation, a moderate role in the coronary circulation, and little role in mesenteric and skeletal muscle circulation.

Constriction and dilatation of large arteries regulate cerebral vascular resistance and, perhaps most importantly, cerebral microvascular pressure. This effect may be a protective mechanism that attenuates changes in pressure in thin-walled intracranial blood vessels. There is, however, an important "cost" of the large proportion of resistance in large cerebral arteries: the vascular bed is particularly vulnerable to a vascular "steal." This concept has been summarized previously (199).

A major mechanism that likely protects the cerebral circulation against a vascular steal is flow-mediated vasodilatation. The concept is that focal increases in blood flow in one region of the brain produce flow-mediated dilatation of large arteries upstream, and thereby protect against a vascular steal. The concept of flow-mediated vasodilatation has been challenged in other vascular beds (519) but appears to be especially large in magnitude and importance in the cerebral circulation (229).

    III. ENDOTHELIUM-DERIVED VASOACTIVE FACTORS
Top
Previous
Next
References

A. Nitric Oxide

1. Nitric oxide synthases

Production and release of potent vasoactive factors by endothelium plays a major role in regulation of vascular tone (Fig. 1). Perhaps, the most important of these substances in relation to regulation of vascular tone is endothelium-derived relaxing factor (EDRF), the labile substance which was initially described by Furchgott and Zawadski (235) in rabbit aorta. Several lines of evidence obtained in subsequent studies have led to the concept that EDRF is nitric oxide (NO) (211, 622) or a closely related compound (429, 485, 566, 684). Although production of NO was first described in vascular endothelium, it is now clear that NO is produced in many cells types by a family of isoenzymes known as NO synthases (81, 224, 225, 547-549, 575, 709).


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 1.   Some mechanisms of endothelium-dependent relaxation of cerebral vascular muscle. Nitric oxide (NO) is produced by NO synthase (NOS) from amino acid L-arginine (L-Arg). NO diffuses to vascular muscle where it activates soluble guanylate cyclase, causing increased production of guanosine 3',5'-cyclic monophosphate (cGMP), which results in relaxation. Prostacyclin (PGI2) is normally produced by cyclooxygenase-1 (COX-1) from arachidonic acid (AA). PGI2 diffuses to vascular muscle where it activates adenylate cyclase, causing increased production of adenosine 3',5'-cyclic monophosphate (cAMP), which results in relaxation. Endothelium-derived hyperpolarizing factor (EDHF) is probably a product of AA metabolism. EDHF diffuses to vascular muscle where it activates potassium (K+) channels. Increased activity of potassium channels produces hyperpolarization and relaxation of vascular muscle. ACh, acetylcholine.

Nitric oxide is a potent vasodilator that produces relaxation of cerebral blood vessels both in vitro and in vivo (84, 363, 484, 845). After release by endothelium (or other cells), NO stimulates soluble guanylate cyclase in vascular muscle, resulting in an increase in the intracellular concentration of guanosine 3',5'-cyclic monophosphate (cGMP) and relaxation (Fig. 1) (707, 800, 804). Organic nitrovasodilators such as nitroglycerin also produce relaxation of cerebral vascular muscle by activation of guanylate cyclase after biotransformation to NO (5, 454, 800). In addition to this cGMP-dependent mechanism, NO can produce relaxation of some blood vessels from some species via activation of potassium channels (24, 71, 130).

All NO synthase enzymes catalyze a five-electron oxidation of a guanidino nitrogen of the basic amino acid L-arginine to NO and L-citrulline (Fig. 1) (395, 448, 621). This process requires the presence of L-arginine (the substrate) and molecular oxygen, NADPH, tetrahydrobiopterin, flavin adenine dinucleotide, and flavin mononucleotide (225, 395, 491, 549). Activity of NO synthases can be inhibited with analogs of L-arginine including N G-monomethyl-L-arginine (L-NMMA), N G-nitro-L-arginine (L-NNA), N G-nitro-L-arginine methyl ester, and N G,N G-dimethyl-L-arginine (ADMA) (395, 472, 549, 599, 674, 782). Of particular interest, ADMA is a potent inhibitor of NO synthase (196) that is produced endogenously in brain and other tissue (380, 381, 421, 777).

Inhibitors of NO synthase have been very useful in providing insight into the role of NO in the cerebral circulation. Inhibitors of soluble guanylate cyclase can also be used to examine this pathway, although inhibitors such as methylene blue may have nonspecific effects (485). A more recently developed inhibitor of soluble guanylate cyclase, 1H-[1,2,4]oxadiazolo[4,3,-a]quinoxalin-1-one, appears to be more selective that methylene blue (728).

Nitric oxide can also be inactivated without affecting guanylate cyclase, by reacting with superoxide anion (362, 548, 693). The reaction of NO with superoxide anion results in the formation of peroxynitrite (53) (Fig. 2), a potent oxidant that may produce cytotoxicity, including nitrosylation of proteins and damage to DNA (51-53, 64, 556). Nitric oxide reacts with superoxide anion at a rate three times faster than the dismutation of superoxide anion by superoxide dismutase (51, 53, 64, 150, 556). Because of the efficiency of this reaction, the local concentration of superoxide dismutase may be an important determinant of activity (the biological half-life) of NO.


View larger version (14K):
[in this window]
[in a new window]
  		FIG. 2.   Schematic representation of interaction of NO and superoxide anion (O-2·). NO is produced from L-arginine (L-Arg) by NO synthases. NO can activate guanylate cyclase resulting in increased cGMP production. NO can also rapidly react with superoxide anion to form peroxynitrite (ONOO-). Peroxynitrite can form hydroxyl radical (OH·), which produces vascular injury. Superoxide anion can then be formed from metabolism of arachidonic acid in cyclooxygenase pathway or by other sources including NADH/NADPH oxidase and leukocytes. Superoxide anion can also be formed by NO synthase in absence of L-arginine. Superoxide anion is dismuted by superoxide dismutase (SOD) to hydrogen peroxide (H2O2). Hydrogen peroxide is degraded by catalase (CAT) but can also be catalyzed to hydroxyl radical via Haber-Weiss reaction.

Although formation of peroxynitrite (from interaction of NO and superoxide) has the potential to be cytotoxic, the inactivation of superoxide by NO also appears to be protective under some conditions, particularly in vivo if subsequent degradation products are nontoxic (64). Formation of peroxynitrite is not always toxic. For example, under some physiological conditions, peroxynitrite inhibits platelet aggregation and leukocyte adhesion to endothelium (protective effects) with no evidence of cell injury (64, 444). Thus the consequences of interaction of NO and superoxide appear to be dependent in part on factors such as the presence of plasma proteins, reduced thiols, and the ratio of NO to superoxide (64). Although peroxynitrite can also cause relaxation of blood vessels, the concentrations of peroxynitrite needed to produce this effect are 50- to 1,000-fold higher than NO (53). Thus peroxynitrite is only modestly effective as a vasodilator. Based on these biochemical characteristics, the concept has emerged that NO is a protective molecule in part because of its ability to react with and inactivate superoxide anion. Relatively little is known regarding the importance of this mechanism in vivo.

Of the three isoforms of superoxide dismutase (Mn-superoxide dismutase, Cu/Zn-superoxide dismutase, and extracellular superoxide dismutase), extracellular superoxide dismutase may be the most important in blood vessels, accounting for up to 70% of total activity of superoxide dismutase in some blood vessels (617, 618, 738). The distribution of extracellular superoxide dismutase in the vessel wall seems ideal for protection from superoxide anion, because NO diffuses from endothelium through vascular muscle (617, 618). Inactivation of NO by superoxide anion may contribute to impaired NO-mediated dilatation of cerebral blood vessels under pathophysiological conditions in which reactive oxygen species are produced. Production of superoxide anion has been reported to occur in brain in response to acute hypertension (814), seizures (30, 47), fluid-percussion injury (359, 410, 765), and perivascular blood (535), as well as during meningitis (446, 520), ischemia with reperfusion (29, 409, 578), and during asphyxia with reventilation (664, 665).

Three isoforms of NO synthase, which are products of separate genes, have been identified. Based on the cell type from which they were initially identified, these three isoforms are frequently described as neuronal, inducible (initially identified in macrophages), and endothelial NO synthase (81, 225, 548, 549, 832). Numerical nomenclature, based on the historical order of enzyme purification and gene cloning, is also used. Nitric oxide synthase I (NOS I), NOS II, and NOS III are sometimes used to describe isoforms initially isolated from neurons, macrophages, and endothelium, respectively (111, 225, 245, 395, 469, 484, 549, 588).

2. Nitric oxide synthase in endothelium

Expression of NO synthase in endothelium (NOS III) is controlled by a single gene (96, 225, 248, 341, 433, 484, 862). The promoter for endothelial NO synthase, like that of other housekeeping genes which are expressed constitutively, does not contain a TATA-like element (225, 540). Endothelial NO synthase also contains a shear stress response element in its promoter region (225, 540, 588, 789). Messenger RNA and protein for NO synthase are present in normal cerebral endothelium (82, 108, 159, 234, 253, 330, 483, 660, 661). Protein for NO synthase can be detected in cerebral endothelium relatively early in gestation (543, 594).

Immunocytochemistry and measurements of enzyme activity indicate that NO synthase is found predominantly in the particulate fraction of endothelium (225, 234, 705). Myristoylation and palmitoylation of the enzyme are essential for localization of NO synthase to plasmalemmal caveolae (218, 239, 712). Caveolae are plasmalemmal microdomains where signaling molecules appear to be concentrated. Localization of endothelial NO synthase in caveolae may be critical for optimal enzyme function, responsiveness to changes in shear stress, and extracellular release of NO (705, 710, 712).

Several lines of evidence suggest that constitutive levels of expression of NO synthase in endothelium are sufficient to influence tone in cerebral blood vessels under basal conditions. Basal levels of cGMP are much greater in cerebral arteries with endothelium than in vessels without endothelium (143, 171, 378, 743, 767). Inhibitors of NO synthase decrease basal levels of cGMP and produce contraction of cerebral arteries in vitro that is endothelium dependent (17, 143, 171, 541, 771). Inhibitors of NO synthase also produce constriction of cerebral blood vessels and decrease cerebral blood flow under basal conditions in several species including nonhuman primates (25, 32, 151, 156, 158, 160, 161, 183, 186, 187, 190, 191, 196, 198, 200, 258, 310, 312, 313, 321, 331, 382, 422, 423, 476, 523-526, 620, 642, 673, 687, 708, 746, 754, 764, 768, 783). This same effect has been observed in midgestation, indicating that the influence on NO in the cerebral circulation may begin early in development (403, 595). Reductions in cerebral blood flow in response to L-NNA are absent in mice that are deficient in expression of the gene for endothelial NO synthase (38, 470, 471), suggesting that endothelium is the primary source of NO that influences basal tone.

In contrast to constriction of cerebral blood vessels in response to inhibitors of NO synthase, exogenous administration of L-arginine, the substrate for NO synthase, has been frequently found to have no effect on cerebral vascular tone (194). Even at relatively high concentrations, L-arginine does not affect tone of cerebral arteries or arterioles in vitro (17, 382, 583), cerebral arterioles (32, 46, 187, 673, 815) and the basilar artery in vivo (186, 187, 386, 505), or cerebral blood flow (197, 198, 258, 275, 422). These findings suggest that availability of L-arginine is not rate limiting for activity of NO synthase in cerebral endothelium. This finding is not surprising because the Michaelis constant (half-saturating concentration of L-arginine) value for L-arginine for endothelial NO synthase is ~3 µM (224, 225), and levels of L-arginine in plasma and endothelium are in the range of 100-2,200 µM (68, 225, 277, 342, 822). In contrast to these studies, which indicate that L-arginine has no significant effect on tone of cerebral vessels, some studies have reported very small to moderate levels of vasodilatation in brain in response to L-arginine (94, 261, 351, 352, 552, 553, 620, 678, 687).

Activity of NO synthase in endothelium, and NO synthase in neurons (NOS I), is dependent on the presence of calcium (395, 549). Basal activity of NO synthase can be further stimulated by increases in intracellular calcium in response to receptor-mediated agonists (Fig. 1) (337). Lee (442) first demonstrated that acetylcholine produces endothelium-dependent relaxation of cerebral arteries. It is now known that many substances, in addition to acetylcholine, produce endothelium-dependent relaxation of cerebral vessels (188), which is dependent on production of NO (194). Relaxation of cerebral vessels in response to acetylcholine (17, 46, 135, 176, 186, 187, 190, 191, 196, 200, 222, 494, 583, 631, 671, 687, 728, 729, 815), bradykinin (254, 363, 494, 541, 614), arginine vasopressin (143, 361, 364, 620, 746), oxytocin (620), substance P (92, 606, 614, 648, 688), histamine (37, 166, 340, 502, 780), sodium fluoride (a G protein activator) (541), endothelin (via activation of endothelin-B receptors) (390, 391, 396, 702), ADP (370, 404, 405, 498), ATP (337), UK-14304 (an alpha 2-adrenoceptor agonist) (87, 88), UTP (539), and prostaglandin F2alpha (370) are dependent on production of NO. Basic fibroblast growth factor (348, 681) and some opioids (156) also produce NO-dependent dilatation of cerebral vessels in vivo that is presumably endothelium dependent. In addition to these receptor-mediated agonists, relaxation of cerebral blood vessels in response to A-23187 (606), increases in shear stress (583), a product released by cultured astrocytes (564), and 4-hydroxynonenal (a product of lipid peroxidation) is dependent on production of NO (487). Recent studies indicate that endothelium-dependent relaxation that is mediated by NO is present in human cerebral arteries (15, 112, 340, 486, 487, 614, 648) as well as in vessels from experimental animals.

A murine model, which is deficient in expression of the endothelial NO synthase gene, has been developed (307, 713). As might be expected, these mice are chronically hypertensive (307, 713) and exhibit impaired endothelium-dependent relaxation in response to acetylcholine in the aorta (307). In contrast, dilatation of cerebral arterioles in response to acetylcholine is normal in endothelial NO synthase mutants (529). After targeted deletion of a specific gene, it is not uncommon for mutant animals to display no or minimal differences in phenotype, suggesting the presence of redundant or compensatory mechanisms (306, 471). Such a compensatory mechanism in cerebral vascular responses probably explains the observation that responses to acetylcholine are normal in cerebral arterioles (529).

A recent study using antisense oligonucleotides suggested that in addition to expression of endothelial NO synthase, endothelium of cerebral vessels may also express neuronal NO synthase (NOS I) (685). The importance of this finding is unclear, however, because other studies (including the use of the relatively selective inhibitor of neuronal NO synthase 7-nitroindazole) suggest that neuronal NO synthase is not expressed in cerebral endothelium or contribute to endothelium-dependent relaxation (195, 287, 315, 330, 453, 801, 841, 847, 860).

An important property of endothelium is that it provides an antithrombogenic surface for blood vessels (Fig. 1). Aggregation of platelets is inhibited normally by luminal release of NO by endothelium (Fig. 1) (462, 547, 686). The antithrombogenic property of endothelium is due in part to synergistic effects of NO with prostacyclin (another relaxing factor produced by endothelium) (462, 829). Nitric oxide also inhibits neutrophil aggregation, adhesion of leukocytes to endothelium, and proliferation of vascular muscle (428, 462, 829). Nitric oxide inhibits expression of endothelial-leukocyte adhesion molecules such as vascular cell adhesion molecule-1 (VCAM-1, an endothelial-leukocyte adhesion molecule) (153, 374, 450). This effect on expression of adhesion molecules appears to be mediated by inhibitory effects of NO on the transcription factor nuclear factor kappa B (95, 374). Nitric oxide may be an important regulator of expression of redox-sensitive genes such as VCAM-1 in endothelium (374). Inhibition of endogenous production of NO in cerebral endothelium increases expression of VCAM-1 (56). Thus, in addition to influencing vascular tone, release of NO protects cerebral endothelium by inhibiting aggregation and adherence of platelets and leukocytes (455).

3. Regulation of endothelial nitric oxide synthase gene expression

Although endothelial NO synthase is frequently referred to as a constitutive isoform of NO synthase, the level of gene expression for NO synthase in endothelium may be increased or decreased in response to several stimuli and under pathophysiological conditions. Gene expression of endothelial NO synthase (NOS III) may be upregulated by increasing shear stress (540, 588, 769, 776), estrogen (293), cGMP (672), basic fibroblast growth factor (420), transforming growth factor-beta 1 (328), and in the presence of atherosclerosis (356), cirrhosis (551), and pregnancy (827). Immunocytochemistry suggests that expression of endothelial NO synthase increases in cerebral vessels after ischemia (863) and during experimental allergic encephalomyelitis (864). In contrast, expression of endothelial NO synthase may be downregulated by oxidized low-density lipoprotein (452), hypoxia (225, 527, 654, 865), tumor necrosis factor-alpha , and lipopolysaccharide (via decreased stability of mRNA) (457, 464, 849) and during heart failure (725). Nitric oxide itself may be involved in regulation of levels of gene expression for endothelial NO synthase via a negative-feedback mechanism (677).

B. Prostacyclin

Prostacyclin is a metabolite of arachidonic acid that is produced, with other prostaglandins and thromboxanes, by the two rate-limiting cyclooxygenase (COX) enzymes [also called prostaglandin H (PGH) synthases] (537, 828). Expression of the two isoforms of cyclooxygenase (COX-1 and COX-2) is determined by separate genes (821, 828). Although COX-1 is expressed constitutively (at varying levels) in most cells including endothelium (726) (Fig. 1), regulatory elements in the promoter region suggest that expression of COX-1 can be modulated by several factors including shear stress (828). The promoter for the gene which encodes COX-1 (like endothelial NO synthase) does not contain a TATA-like element and thus has one characteristic of a housekeeping gene (726).

Cyclooxygenase-2 is an inducible isoform that is expressed primarily in macrophages, fibroblasts, endothelium, and vascular muscle (327, 726, 828). Although it is characterized as an inducible isoform (and is undetectable in most mammalian tissues under normal conditions), COX-2 is expressed constitutively in some organs including brain (726). Many of the same stimuli that result in expression of inducible NO synthase also stimulate expression of COX-2 (458, 537, 821, 828). These stimuli include lipopolysaccharide, adenosine 3',5'-cyclic monophosphate (cAMP), hypoxia, and some cytokines, growth factors, and hormones (80, 327, 706, 828). Expression of COX-2 is involved in inflammation (537, 828) and may be the predominate source of prostaglandins under these conditions (458). Under basal conditions in adult rats, COX-2 is undetectable using in situ hybridization in cerebral vessels (104). In contrast, under normal conditions in newborn pigs, the predominant source of prostaglandins in cerebral vessels appears to be COX-2 (644). Messenger RNA levels for COX-1 and COX-2 increase in brain in response to ischemia (133, 296, 593, 607, 656), after systemic administration of lipopolysaccharide (80), and in response to Borna disease virus (554). Expression of COX-2 is also increased in brain for prolonged periods of time after repeated episodes of cortical spreading depression (99).

Prostacyclin is a powerful inhibitor of platelet aggregation (Fig. 1) (829). In addition, prostacyclin and stable prostacyclin analogs produce relaxation of cerebral arteries in vitro and cerebral arterioles in vivo (25, 178, 226, 371, 612, 613, 688, 766, 779, 799) (Fig. 1). Relaxation of cerebral vessels in response to prostacyclin is generally thought to be endothelium independent (371, 688, 766) but may be mediated by more than one mechanism. Relaxation of cerebral vessels in response to prostacyclin may be mediated by activation of adenylate cyclase with accumulation of cAMP (626, 762) (Fig. 1) and activation of potassium channels (141, 226). In contrast, recent findings in the newborn pig suggest that dilatation of cerebral arterioles in response to prostacyclin is partially dependent on formation of NO (25).

Factors that regulate COX gene expression and production of prostacyclin in cerebral endothelium are not well defined. Several stimuli, including bradykinin, thrombin, and A-23187, increase prostacyclin production in cerebral arteries and cultured cerebral endothelium (513, 550, 610, 834). Endothelium-dependent relaxation of human cerebral arteries in response to 4-hydroxynonenal, and acetylcholine-induced relaxation of the vertebral artery from newborn humans, is inhibited by indomethacin and thus may be mediated by prostacyclin (112, 487).

Expression of COX-2 in cerebral vessels may influence vascular tone. Messenger RNA for COX-2 is expressed in cerebral microvessels in response to lipopolysaccharide (104) and interleukin-1beta (105), and interleukin-1 increases production of prostacyclin in cerebral endothelium (513). Interleukin-1 produces dilatation of cerebral arterioles in newborn pigs that is inhibited by indomethacin (715) and presumably mediated by activation of COX-2.

C. Endothelium-Derived Hyperpolarizing Factor

In addition to production and release of NO and prostacyclin, endothelium may also produce relaxation of underlying vascular muscle by release of endothelium-derived hyperpolarizing factor(s) (EDHF) (55, 130, 241, 574) (Fig. 1). In large cerebral arteries, for example, acetylcholine, substance P, and bradykinin produce endothelium-dependent hyperpolarization and relaxation of vascular muscle (77, 142, 546, 590, 648, 740), which appears to be mediated, in part, by an EDHF. Although EDHF is more difficult to bioassay than EDRF, recent evidence indicates that EDHF is a diffusible (transferable) factor that causes relaxation by hyperpolarizing underlying vascular muscle (118, 130, 546, 662). The importance of EDHF as a mediator of endothelium-dependent relaxation has been reported to increase as vessel size decreases (216, 718). For example, expression of endothelial NO synthase and the importance of NO have been reported to decrease, and the importance of EDHF increase, as vessels become smaller in the mesenteric circulation (718).

The identity of EDHF remains a subject of debate and investigation. In some extracranial arteries, NO (sometimes only at relatively high concentrations) produces hyperpolarization of vascular muscle and thus may function as an EDHF (130). Nitric oxide and donors of NO like 3-morpholinosydnonimine (SIN-1) produce marked relaxation but have little or no effect on membrane potential in large cerebral arteries (77, 563, 658, 659). The absence of hyperpolarization suggests that NO is not an EDHF in these blood vessels. In contrast, NO produces glibenclamide-sensitive hyperpolarization of the guinea pig carotid artery, suggesting that the response is mediated by ATP-sensitive potassium channels (141). Because relaxation of cerebral blood vessels in response to prostacyclin is antagonized by inhibitors of potassium channels in some cerebral blood vessels (41, 141, 226), prostacyclin may also act as an EDHF in the cerebral circulation.

In many arteries, however, it seems clear that EDHF is not NO or a prostanoid, because inhibitors of NO synthase and COX do not attenuate endothelium-dependent hyperpolarization or relaxation of vascular muscle (118, 130, 142, 574). A substantial body of evidence, obtained primarily from studies of coronary blood vessels, suggests that an EDHF is a product of cytochrome P-450 monooxygenase metabolism of arachidonic acid (44, 102, 116, 267, 276, 373, 662). The products of arachidonate that mediate this effect appear to be epoxyeicosatrienoic acids (102, 267). Epoxyeicosatrienoic acids such as 11,12-epoxyeicosatrienoic acid produce hyperpolarization and relaxation of coronary arteries in vitro (102).

Epoxyeicosatrienoic acids are produced in brain (19, 177, 244, 267) and by astrocytes (16), and some epoxyeicosatrienoic acids produce relaxation of cerebral blood vessels (19, 177, 244). For example, both 5,6-epoxyeicosatrienoic acid and 11,12-epoxyeicosatrienoic acid produce relaxation of the middle cerebral artery in vitro (244). Consistent with a possible function as EDHFs, 11,12-epoxyeicosatrienoic acid produces relaxation of the middle cerebral artery, which is inhibited by a high concentration of tetraethylammonium (TEA) ion (244), and 14,15-epoxyeicosatrienoic acid enhances an outward potassium current in smooth muscle isolated from cerebral microvessels (16). In contrast, 5,6-epoxyeicosatrienoic acid (but not 11,12-epoxyeicosatrienoic acid) is a potent dilator of cerebral arterioles in vivo (19, 177). Interestingly, effects of 5,6-epoxyeicosatrienoic acid on cerebral arterioles were inhibited by indomethacin or by superoxide dismutase plus catalase (177), suggesting 5,6-epoxyeicosatrienoic acid does not directly produce relaxation in cerebral arterioles. It is known that 5,6-epoxyeicosatrienoic acid can be metabolized by COX, resulting in formation of reactive oxygen species that mediate vasodilatation (177).

Although it seems clear that some epoxyeicosatrienoic acids produce relaxation of cerebral blood vessels, little is known whether these substances are produced by cerebral endothelium and function as EDHFs in the cerebral microcirculation. A recent study suggests that release of cytochrome P-450 products by endothelium contributes to dilatation of cerebral microvessels during hypoxia in newborn pigs (445). In the carotid artery, acetylcholine causes release of an EDHF that may be a product of cytochrome P-450 monooxygenase metabolism (44, 456). Interestingly, bioassay studies suggest that NO inhibits the formation and/or release of EDHF in this artery (45). The implication of this finding is that when activity of NO is reduced, such as under pathophysiological conditions, synthesis or release of EDHF may increase as a compensatory mechanism. In contrast to these findings, endothelium-dependent hyperpolarization in the carotid artery of the guinea pig does not appear to be a product of cytochrome P-450 monooxygenase, COX, or lipoxgenase and thus may not be a product of arachidonic acid metabolism (142).

Studies of coronary blood vessels that have implicated cytochrome P-450 monooxygenase metabolites as potential EDHFs have frequently used arachidonic acid as a stimulus to produce endothelium-dependent relaxation. In contrast to coronary arteries (102), dilatation of cerebral arterioles in several species in response to arachidonic acid is mediated by COX-dependent generation of reactive oxygen species. Cerebral microvascular responses to arachidonate are inhibited by indomethacin (90, 93, 177, 416, 810, 845) and by scavengers of reactive oxygen species (177, 209, 415, 417, 844). Thus arachidonic acid does not appear to dilate cerebral arterioles by generation of P-450 metabolites unless these metabolites subsequently act as substrate for COX as described for 5,6-epoxyeicosatrienoic acid (177).

Hyperpolarization of vascular muscle in response to EDHF is most likely mediated by activation of potassium channels (102, 130, 241, 546, 662) (Fig. 1). For example, relaxation of cerebral vessels in response to acetylcholine appears to be mediated, in part, by production of an EDHF that activates ATP-sensitive potassium channels (77, 192, 740). Activation of calcium-dependent potassium channels may also mediate EDHF-induced relaxation in some arteries (117, 118, 130, 662), including the canine carotid artery (546). Charybdotoxin, an inhibitor of calcium-dependent potassium channels, produces partial inhibition of relaxation of the middle cerebral artery in response to acetylcholine (817). Endothelium-dependent hyperpolarization, which is not mediated by NO, has been described in human cerebral arteries in response to substance P (648).

The role of cytochrome P-450 monooxygenase metabolites as possible EDHFs has been supported, in part, by the use of cytochrome P-450 inhibitors. Although there is evidence that these inhibitors abolish release of EDHF from endothelium (662), high concentrations of these compounds may also exert nonspecific effects (173, 216). For example, some inhibitors of P-450 including clotrimazole have been reported to directly inhibit activation of potassium channels (independent of effects on activity of P-450 enzymes) (173, 871). Thus interpretation of results obtained with these agents may need to be made with caution (173).

In contrast to studies that implicate a role for EDHFs in regulation of vascular tone, several studies suggest that activation of potassium channels does not contribute to relaxation of cerebral vessels in response to endothelium-dependent vasodilators in vitro and in vivo (17, 78, 200, 631, 751, 817). Thus the overall functional importance of EDHF in cerebral vessels is not clear.

D. Endothelin

In addition to relaxing factors, endothelium can release substances that produce contraction of blood vessels. The endothelium-derived contracting factor (EDCF) that has received the most investigation is endothelin, a peptide originally isolated from aortic endothelium (840) (Fig. 3). There are three isopeptides of endothelin, all 21 amino acids, that are the products of separate genes [endothelin (ET)-1, ET-2, and ET-3] (692). All endothelins are synthesized as larger preproforms (~200 amino acids), which are converted by endopeptidases to propeptides (also known as big endothelins). For example, prepro-ET-1 is cleaved to big ET-1, which is then converted to ET-1 by endothelin-converting enzyme (255, 692) (Fig. 3). Endothelin-converting enzymes (ECE-1 and ECE-2) are metalloproteases that are associated with the membrane fraction of cells (255, 257, 785, 803).


View larger version (21K):
[in this window]
[in a new window]
  		FIG. 3.   Summary of synthesis of endothelin-1 (ET-1) in endothelium. ET-1 gene transcription results in formation of prepro-ET-1 mRNA (ET gene expression). Prepro-ET-1 is converted to big ET-1 by an endopeptidase. Big ET-1 is cleaved to ET-1 by endothelin-converting enzyme (ECE). ET-1 can activate both ETA and ETB receptors on vascular muscle and cause contraction. ET-1 can also activate ETB receptors on endothelium and cause activation of NO synthase and production of NO. NO and cGMP inhibit ET-1 gene expression.

Cerebral endothelium can produce endothelin under some conditions. Of the three isopeptides, only ET-1 is produced normally by endothelium (692), and mRNA for ET-1 can be detected in cerebral endothelium (848). Messenger RNA for ECE-1 has also been detected in cerebral endothelium using in situ hybridization (833). Endothelin gene expression can be enhanced by several factors including thrombin, transforming growth factor-beta , hemoglobin, and tumor necrosis factor-alpha (181, 605, 692) and can be inhibited by NO and cGMP (76, 165, 255, 692). A shear stress response element is present in the promoter of the ET-1 gene, but in contrast to the gene for endothelial NO synthase, activation of this element decreases transcription of the ET-1 gene (692).

Endothelins are known to mediate effects in blood vessels through activation of two receptors, endothelin-A (ETA) and endothelin-B (ETB) receptors, which have been cloned (23, 255, 700). In general, ETA receptors are expressed in vascular muscle and mediate contraction to endothelin (257) (Fig. 3).

The response to activation of ETB receptors on vascular tone depends on localization of the receptor. Endothelin-B receptors are expressed in smooth muscle in some blood vessels and mediate contraction (Fig. 3). In contrast, activation of ETB receptors (which are commonly expressed in endothelium) produces relaxation of blood vessels through release of prostacyclin or EDRF (257, 290, 692, 774) (Fig. 3). Some studies suggest there may be two distinct subtypes of ETB receptors (receptors that produce vasorelaxation or vasoconstriction) (806).

Endothelin-1 produces potent and long-lasting contraction of cerebral vessels both in vivo and in vitro (1, 2, 14, 31, 152, 185, 212, 217, 219, 265, 338, 390, 600, 625, 632, 680, 697, 698, 701, 757, 820). Endothelin-1 and ET-2 are much more potent than ET-3 in producing contraction (698, 701). Vasoconstriction in response to endothelin is dependent on extracellular calcium (152, 217, 219, 265, 338) and may be mediated by activation of protein kinase C (219, 565).

Endothelin-A receptors are expressed normally in cerebral blood vessels (286, 302, 586, 850), and the predominant mechanism of vasoconstriction in response to endothelin in these vessels is by activation of ETA receptors (1, 2, 154, 212, 217, 350, 359, 390, 586, 632, 647, 695, 701, 820, 869). In contrast to this contractile response, low concentrations of endothelin produce dilatation of cerebral arterioles in vivo that may be mediated by activation of ETB receptors (31, 185). Selective activation of ETB receptors produces marked relaxation of cerebral vessels that is mediated by NO (390, 391, 634, 702). Sarafotoxin 6c (a selective agonist for ETB receptors) produces relaxation of human cerebral arteries, suggesting that ETB-mediated relaxation can also be induced in human vessels (586). Application of inhibitors of ETA , or both ETA and ETB , receptors has no effect on tone of cerebral vessels in vivo, which suggests that production of endothelin does not contribute to basal tone in the cerebral circulation (249, 632-634, 869, 870).

E. Other Endothelium-Derived Vasoactive Factors

1. Reactive oxygen species

In addition to the factors described above, cerebral endothelium may produce several additional vasoactive substances including reactive oxygen species, carbon monoxide, and nonendothelin EDCFs. For example, although bradykinin produces NO-dependent relaxation of the basilar artery and the middle cerebral artery (254, 363, 417, 494, 541, 614), another mechanism mediates bradykinin-induced dilatation of pial arterioles that supply the cerebrum. In these microvessels, dilatation in response to bradykinin is endothelium dependent but mediated by reactive oxygen species (either hydroxyl radical or hydrogen peroxide depending on the species) (414, 415, 683, 844). Reactive oxygen species are dilators in the cerebral microcirculation of several species (409, 413, 417, 682, 778, 809, 812, 844, 845). As discussed in section VB, dilatation of cerebral microvessels in response to bradykinin appears to be mediated by activation of potassium channels (208).


View larger version (19K):
[in this window]
[in a new window]
  		FIG. 4.   Two mechanisms of endothelial dysfunction in cerebral vessels. A cyclooxygenase (COX)-derived endothelium-derived contracting factor (EDCF) may be formed from arachidonic acid (AA). This EDCF can diffuse to vascular muscle and cause contraction by activation of receptors for thromboxane A2/prostaglandin H2 receptors (TxA2/PGH2). NO is normally produced from L-Arg by NOS. Superoxide anion, produced in cyclooxygenase pathway or from other sources, can interact with and destroy NO. Thus endothelial dysfunction can result from either simultaneous release of an EDCF or impairment of relaxation by degradation of NO.

2. Carbon monoxide

Recent evidence suggests that carbon monoxide might also function as an EDRF in some blood vessels (853). Carbon monoxide is produced from heme by two isoforms of heme oxygenase (HO-1 and HO-2) (478), and HO-2 is expressed constitutively in endothelium of cerebral arteries (853). The functional significance of carbon monoxide as a regulator of tone in cerebral vessels is unclear however. Carbon monoxide produces relaxation of some blood vessels, including the aorta, but does not produce direct relaxation of cerebral arteries (84). Messenger RNA for HO-1, the inducible form of heme oxygenase, is expressed in cerebral endothelium in response to hemin (a hemoglobin degradation product), but expression is not associated with increased levels of cGMP (790), which provides additional evidence that carbon monoxide may not function as a direct dilator in cerebral blood vessels. Although carbon monoxide may not directly relax cerebral vessels, recent evidence obtained in neurons suggests that carbon monoxide may modulate formation of cGMP in response to NO (326). Thus carbon monoxide may influence cerebral vascular tone indirectly via effects on production of cGMP in smooth muscle in response to NO.

3. Other endothelium-derived constricting factors

In some cerebral arteries such as the canine basilar artery, and in the presence of some pathophysiological conditions, contracting substances that are not endothelin may be produced by cerebral endothelium. These EDCFs are primarily metabolites of arachidonic acid produced through the COX pathway (366), which produce contraction by activation of PGH2-thromboxane A2 receptors (144) (Fig. 4). Superoxide anion, which is also produced via the COX pathway, produces contraction of the canine basilar artery, and thus may be an EDCF under some conditions (144), although recent evidence suggests contraction to superoxide anion is due to inhibition of basal effects of NO (362) (Fig. 4). Endothelium-dependent contraction may occur in response to A-23187 (144, 367, 369, 720), xanthine plus xanthine oxidase (which generates reactive oxygen species) (365), hydrogen peroxide (365), acetylcholine (344, 367, 379, 780), arachidonic acid (379, 780), anoxia (368), angiotensins (480), nicotine (719), and phospholipase A2 (571).

    IV. INDUCIBLE NITRIC OXIDE SYNTHASE
Top
Previous
Next
References

In contrast to endothelial NO synthase (NOS III), which is expressed constitutively, an inducible or "immunologic" isoform of NO synthase (NOS II) can be expressed in many cells types including vascular muscle and endothelium (113, 245, 335, 357, 358, 426, 434, 477, 548). Inducible NO synthase is the product of a separate gene (155, 225, 245, 395, 426, 597). Messenger RNA for inducible NO synthase is typically not present or is present in very low levels in normal cells under basal conditions (225, 556, 576). In response to inflammatory factors and other stimuli, expression of inducible NO synthase occurs over a period of hours (225, 426, 548, 556) (Fig. 5). Thus activity of inducible NO synthase is regulated primarily at the level of gene expression (548, 556, 575, 598).


View larger version (17K):
[in this window]
[in a new window]
  		FIG. 5.   Summary of synthesis of inducible NO synthase (iNOS, NOS II) in smooth muscle or other cells that express iNOS. NOS II gene transcription is activated by many stimuli including lipopolysaccharide (LPS), tumor necrosis factor-alpha (TNF-alpha ), and interleukin-1beta (IL-1beta ). NOS II gene expression can by inhibited by factors including interleukin-10 (IL-10) and transforming growth factor-beta (TGF-beta ). After translation, iNOS protein converts L-Arg to NO. NO activates soluble guanylate cyclase, resulting in increased production of cGMP and relaxation.

Inducible NO synthase produces much greater amounts of NO than the constitutive forms of NO synthase (endothelial and neuronal NO synthase) (395, 426, 548). Because high concentrations of NO can be toxic to cells, either alone or in combination with superoxide ion that produces peroxynitrite (53) (Fig. 2), inducible NO synthase is often referred to as the "pathological" form of NO synthase (426, 548, 576, 707). Once formed from NO and superoxide, peroxynitrite may be protonated to yield hydroxyl radical, which is extremely reactive (51, 53, 120, 150). Thus, in addition to its direct effects, formation of NO may contribute to formation of toxic reactive species such as hydroxyl radical (Fig. 2).

Stimuli that cause expression of inducible NO synthase and have received the most study are lipopolysaccharide (endotoxin) and some cytokines. Specific cytokines that cause expression of inducible NO synthase vary among species and between cell types but include interleukin-1beta , interferon-gamma , and tumor necrosis factor-alpha (225, 548, 575) (Fig. 5). Adenosine 3',5'-cyclic monophosphate enhances the expression of inducible NO synthase in response to cytokines in several types of cells including vascular muscle (324, 406). S100beta , a neurotrophic protein derived from glia, also causes expression of inducible NO synthase in cultured astrocytes (303).

The promoter of the gene for human inducible NO synthase contains a shear-stress response element (113, 225), which suggests that expression of this gene in endothelium might be modulated by blood flow. The 5'-flanking region of the inducible NO synthase gene also contains a hypoxia-responsive element, which suggests that inducible NO synthase is a hypoxia-inducible gene (528). Stimuli that induce expression of inducible NO synthase also induce expression of GTP-cyclohydrolase I, the rate-limiting enzyme in biosynthesis of tetrahydrobiopterin (491). Tetrahydrobiopterin is one of the cofactors required for enzyme activity (225, 395). Lipopolysaccharide also induces mRNA for arginase II (256). Activity of arginase II may play a major role in production of NO by influencing the availability of L-arginine, the substrate for NO synthases.

It is noteworthy that mRNA for interleukin-1beta converting enzyme and other genes that encode the interleukin-1 system appear to be expressed constitutively in cerebral microvessels (825). Expression of this "vascular interleukin-1 system" may contribute to regulation of expression of inducible NO synthase and COX-2.

Expression of inducible NO synthase can be inhibited by several factors including interleukin-4, interleukin-10, transforming growth factor-beta , basic fibroblast growth factor, aldosterone, heat shock protein 70, and insulin-like growth factor (50, 131, 213, 225, 323, 645, 704) (Fig. 5). Preliminary experiments in mice that lack gene expression for interleukin-10 suggest that endogenous interleukin-10 is a major regulator of expression of inducible NO synthase (206). Inhibitors of tyrosine kinase and glucocorticoids also block expression of inducible NO synthase (225, 575, 670). Expression of the inducible NO synthase in vascular muscle is dependent on binding of nuclear factor kappa B to the promoter (737) and can be blocked by inhibition of activation of nuclear factor kappa B (737, 831). As described previously for endothelial NO synthase, NO may modulate the level of gene expression for inducible NO synthase via a negative-feedback mechanism (132, 628, 629). For example, NO inhibits expression of inducible NO synthase in human microglia through a mechanism that may involve decreased availability of nuclear factor kappa B (132).

With the use of cells in culture, several lines of evidence indicate that glia (110, 140, 214, 215, 228, 232, 236, 237, 284, 285, 294, 407, 440, 441, 544, 545, 589), cerebral endothelium (57-59, 375, 561, 562, 616), cerebral vascular muscle (181, 741), and neurons (533, 534, 601) can be stimulated to express inducible NO synthase. Evidence of expression of inducible NO synthase in these cells includes analysis of mRNA for inducible NO synthase, Western blotting for protein, measurement of activity of inducible NO synthase (calcium-independent formation of L-citrulline), and measurement of nitrite (a breakdown product of NO). Expression of inducible NO synthase occurs in brain in situ in response to ischemia (180, 314, 317, 320, 322, 855). Inducible NO synthase can be detected in the presence of other pathophysiological conditions including meningitis (101), multiple sclerosis (39, 65), immunodeficiency viral and experimental allergic encephalitis (89, 301, 418, 435, 609), herpes simplex virus (418), Borna disease virus (301, 418, 554), encephalomyelitis (784), and rabies virus (301, 418). Inducible NO synthase is expressed in cerebral microvessels in Alzheimer's disease (162), after ischemia (317, 567, 568), and in brain tumors (175, 264) and endothelium of vessels supplying brain tumors (127). Protein for inducible NO synthase is present in cerebral vascular muscle and infiltrating leukocytes after traumatic brain injury (124). mRNA for inducible NO synthase can be detected in meninges and choroid plexus after systemic administration of lipopolysaccharide (826). Recent studies suggest that NO may modulate permeability of the blood-brain barrier (504), and expression of inducible NO synthase may contribute to increased permeability of the blood-brain barrier after exposure to lipopolysaccharide (69, 721).

Although expression of inducible NO synthase has generally been found to be associated with inflammatory or pathophysiological conditions, it appears that the gene may be active during development. For example, mRNA and protein for inducible NO synthase have been detected in parenchymal microvessels in brain during normal embryonic development and in newborn rats (238). Inducible NO synthase is not found in vessels in immature or adult animals under normal conditions (124, 238), and the significance of vascular expression of this isoform of NO synthase during early life is not clear. Because NO can influence vascular growth, we speculate that expression of inducible NO synthase early in ontogeny may contribute to vascular remodeling.


View larger version (25K):
[in this window]
[in a new window]
  		FIG. 6.   Summary of mechanisms by which ATP-sensitive potassium (K+) channels produce relaxation of vascular muscle. Potassium channels, which are present in cell membrane, can be activated by several stimuli including receptor-mediated agonists, increases in cAMP, hypoxia, and reductions in intracellular ATP. These channels can also be activated by synthetic compounds including cromakalim and aprikalim. ATP-sensitive potassium channels can be inhibited with glibenclamide. When potassium channels open, potassium exits the cell. This efflux of potassium ions hyperpolarizes the cell membrane, resulting in closure of voltage-gated calcium (Ca2+) channels, reductions in intracellular levels of calcium, and relaxation.

Recent evidence suggests that lipopolysaccharide and tumor necrosis factor-alpha cause expression of inducible NO synthase that affects cerebral vascular tone in vivo (83, 85, 86, 561, 593). Lipopolysaccharide causes marked, progressive dilatation of cerebral arterioles and increases in levels of cGMP in perivascular cerebrospinal fluid that are attenuated by inhibitors of NO synthase, including L-NMMA and aminoguanidine (85, 86). Aminoguanidine appears to be a relatively selective inhibitor of inducible NO synthase (536, 736, 824) and, at appropriate concentrations, does not inhibit cerebral vascular responses that are mediated by endothelial NO synthase (86). Increases in arteriolar diameter in response to these stimuli are also attenuated by dexamethasone (85), which inhibits expression of the inducible NO synthase gene (225, 670). In addition to these findings in cerebral arterioles, lipopolysaccharide causes similar, NO-dependent, dilatation of the basilar artery (755).

    V. POTASSIUM CHANNELS
Top
Previous
Next
References

Changes in activity of potassium channels represent a major mechanism that regulates vascular tone. Activation of potassium channels in vascular muscle produces hyperpolarization of the cell membrane, closure of voltage-dependent calcium channels, a decrease in intracellular calcium, and vascular relaxation (387, 579, 582) (Fig. 6). The membrane potential of cerebral vascular muscle measured in vitro has ranged widely from approximately -40 to -70 mV (77, 78, 240, 268, 338, 431, 522, 590, 648, 658, 659, 696, 867). The membrane potential of cerebral vascular muscle in vivo is not known. Changes in membrane potential of only a few millivolts are associated with substantial changes in vascular tone (394, 431, 582, 657). Activation of potassium channels mediates cerebral vascular response to several stimuli including some receptor-mediated agonists, second messengers, and hypoxia (207). Endothelium-derived hyperpolarizing factor produces hyperpolarization and relaxation of vascular muscle by activation of potassium channels (241, 387) (Figs. 6 and 7).


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 7.   Summary of mechanisms by which calcium-dependent potassium channels produce relaxation of vascular muscle. Calcium-dependent potassium channels can be activated by several stimuli including receptor-mediated agonists, increases in cAMP, increases in cGMP in response to NO, and increases in intracellular calcium. Whether NO or cGMP activates calcium-dependent potassium channels seems to vary among vascular beds, species, and perhaps size of vessels. These channels can also be activated by reactive oxygen species (ROS) including hydrogen peroxide. Calcium-dependent potassium channels can be inhibited with tetraethylammonium (TEA), iberiotoxin, or charybdotoxin. When potassium channels open, potassium exits cell. This efflux of potassium ions hyperpolarizes cell membrane, resulting in closure of voltage-gated calcium (Ca2+) channels, reductions in intracellular levels of calcium, and relaxation.

Both electrophysiological and pharmacologically based studies have provided evidence that several types of potassium channels are present in cerebral blood vessels (387, 582). Although the majority of studies have focused on potassium channels in vascular muscle, potassium channels can also be expressed in endothelium (115, 431, 465, 466). In endothelium, membrane hyperpolarization causes mobilization of calcium and may be an important mechanism for release of NO, prostacyclin, and EDHF (241, 431, 465).

A. ATP-Sensitive Potassium Channels

Adenosine 5'-triphosphate-sensitive potassium channels have been described in vascular muscle, including cerebral vessels (392, 582, 740). These potassium channels are defined based on sensitivity to intracellular ATP, which inhibits activity of the channel (582) (Fig. 6). Dissociation of ATP from the channel results in channel opening and hyperpolarization. In contrast to intracellular ATP, other factors including reductions in PO2 or pH open the channels and produce vasorelaxation (579, 582). These properties support the concept that activity of ATP-sensitive potassium channels may, in part, reflect the metabolic state of cells (579).

Electrophysiological and pharmacological studies of cerebral vascular muscle indicate that ATP-sensitive potassium channels are activated by synthetic compounds including cromakalim (or levcromakalim) and aprikalim (392, 569, 740) (Fig. 6). Activity of these channels is inhibited by sulfonylureas including glibenclamide (392, 582, 657, 740) (Fig. 6). Glibenclamide appears to be a selective inhibitor of ATP-sensitive potassium channels at the most commonly used concentrations (<3 µM) and has been used frequently to examine the role of ATP-sensitive potassium channels in intact cerebral vessels.

Activators of ATP-sensitive potassium channels produce hyperpolarization and relaxation of cerebral arteries (205, 392, 427, 489, 522, 569, 570, 630, 639, 658, 703, 749, 758, 817, 861), including human arteries (282, 648) in vitro. These activators of ATP-sensitive potassium channels also cause dilatation of the basilar artery (200, 389, 499, 569, 729, 732, 770) and cerebral arterioles in vivo (26, 40, 41, 201, 297, 332, 412, 443,