Endothelium-derived relaxing factor [EDRF; identified as NO or as a closely related substance (1,2)] and prostacyclin are dilators released from endothelial cells in response to various stimuli including exposure to acetylcholine (3-5). In addition, endothelial cells release a yet-unidentified factor (endothelium-derived hyperpolarizing factor, EDHF; 6-10), which causes membrane hyperpolarization by opening K+ channels of the vascular smooth-muscle cells. EDHF may play an important role under conditions of impaired NO synthesis. In carotid arteries from hypercholesterolemic rabbits, although the contribution of NO is reduced, a normal relaxation to acetylcholine has been observed, which has been attributed to an enhanced release of EDHF (11). In contrast, in canine and porcine coronary arteries, sustained NO production through the induction of NOS II, induced by endotoxemia, virtually abolishes the component of the endothelium-dependent relaxation attributed to EDHF (12,13), suggesting that NO regulates the synthesis of EDHF (14). In endothelial cells, cross-talks between NO and the cyclooxygenase pathway have also been described (15,16). In rats, long-term in vivo inhibition of NOS provokes the induction of cyclooxygenase-2 pathway (17), and in dogs, the upregulation of the cyclooxygenase-1 pathway (18,19).
The purpose of our study was to determine in two different tissues and species [canine coronary artery, which demonstrates an upregulation of the cyclooxygenase pathway after long-term NOS inhibition (18), and guinea-pig carotid artery, the model preferentially studied in our laboratory] whether long-term NOS inhibition could affect the responses attributed to the EDHF pathway and whether endothelium-dependent hyperpolarization can act as a back-up mechanism for endothelium-dependent relaxations when NO production has been inhibited over the long term.
MATERIALS AND METHODS
Long-term treatment with nitric oxide synthase inhibitor. Studies were performed in 16 male mongrel dogs weighing 16-30 kg. Eight of these dogs underwent surgical preparation (seven dogs from the treated group and one dog from the control group), as described previously (18). In the seven treated dogs, the inhibitor of NOS, Nω-nitro-L-arginine (LNA), was infused slowly into the pulmonary artery, at the initial dose of 30 mg/kg on the first day and subsequently at the dose of 20 mg/kg (i.v.) on the 7 following days. One of the control dogs was sham-operated to verify that the surgical procedure did not affect the responses. The results obtained in this dog were not different from those obtained in control non-operated animals.
Male Hartley guinea-pigs (300-400 g) were treated for 1-2 weeks, by addition of L-nitro-arginine methyl ester (L-NAME) in the drinking water, as this inhibitor is more water soluble than LNA. The calculated dose ingested by the guinea pigs averaged 60 mg/kg/day.
Electrophysiologic experiments. Male dogs were anesthetized with sodium pentobarbital (30 mg/kg) and artificially ventilated. The heart was rapidly excised from the dogs, and small arteries from left descending coronary artery were dissected free. Guinea pigs were anesthetized with pentobarbital (250 mg/kg, intraperitoneally), and the carotid arteries dissected free.
Segments of artery (guinea pig carotid artery or dog coronary artery; 1 cm in length) were cleaned of adherent connective tissue and pinned down to the bottom of an organ chamber (3 ml) superfused continuously with modified Krebs-Ringer bicarbonate solution (37°C, aerated with a 95% O2/5% CO2 gas mixture; pH, 7.4) of the following composition (mM): NaCl, 118.3; KCl, 4.7; CaCl2, 2.5; MgSO4, 1.2; KH2PO4, 1.2; NaHCO3, 25; calcium-disodium EDTA, 0.026; and glucose, 11.1. In all experiments, care was taken to preserve the endothelium as intact as possible. Transmembrane potentials were recorded by using glass microelectrodes filled with KCl (3 M), with a tip resistance of 30-90 MΩ. The microelectrode was mounted on a sliding micromanipulator (Leitz, St. Gallen, Switzerland). The potential recorded was amplified by means of a recording preamplifier (Intra 767; WPI, New Haven, CT, U.S.A.) with capacitance-neutralization. The signal was monitored on an oscilloscope (3091; Nicolet, Madison, WI, U.S.A.), recorded on paper (Gould, Valley View, OH, U.S.A.) and on video recorder (TEAC XR310, Tokyo, Japan); the latter allowed replay for further analysis. Impalements were not accepted as valid unless they were signaled by a sudden change in voltage and were maintained for ≥3 min, at which point the membrane potential had stabilized. The impalements were performed from the adventitial side. The incubation time with the various inhibitors studied was ≥20 min. Acetylcholine was infused for ≤5 min to avoid desensitization of the preparation.
Tension measurements. Experiments were performed in canine left circumflex coronary artery and guinea-pig aortas. The arteries were excised, immersed in cold (4°C) physiologic salt solution, and cleaned of adherent connective tissue. Arteries were cut into rings (4 mm long) and mounted in an organ bath for isometric tension recording. The tissues were bathed in modified Krebs-Ringer solution, connected to a UC2 force transducer (Gould, Longjumeau, France), and changes in tension were recorded on a polygraph (Gould).
The rings were stretched step by step (6-8 g of passive tension for the dog coronary artery; 1-2 g for the guinea pig aorta) until optimal and reproducible contraction to KCl (40 mM) was achieved.
Then a reference contraction was produced with a concentration of KCl (60 mM) that gave the maximal contraction to the depolarizing solution in those tissues. Repeated rinses were followed by an equilibration period of 45 min. Inhibitors of cyclooxygenase and NOS (i.e., indomethacin and LNA) were given 45 min before the contraction of the arteries with U46619 and the cumulative addition of agonist (i.e., bradykinin or acetylcholine). Experiments were performed in parallel in rings from the same tissue. Only one concentration-response curve was performed on a single ring.
Calculation and statistical evaluation
Data are shown as means ± SEM; n represents the number of animals from which tissue was taken (organ-bath experiments) or the number of cells that were recorded (electrophysiology). Changes in tension are expressed in percentage of maximal relaxation produced by sodium nitroprusside (1 mM). For concentration-response curves, statistical evaluation was performed by a three-way analysis of variance (Treatment × Concentration × Tissue). When a significant interaction was observed (p < 0.05), a complementary analysis was undertaken (Newman-Keul's test) to identify differences among groups. Median effective concentration (ED50) calculations were performed by a linear regression within the two half-log concentrations surrounding the 50% value, and they were compared by using a paired or unpaired t test. For electrophysiologic experiments, the unpaired Student's t test was performed.
Canine coronary artery
Isolated rings of coronary arteries, taken from control and animals after long-term treatment were contracted with U46619 (1-3 × 10−8M) to achieve a similar level of tension in the various conditions tested (e.g., in presence or absence of indomethacin, LNA, or their combination). There was no statistically significant difference between control and in vivo treated animals. However, in the experiments involving the effect of acetylcholine in control animals, the level of tension achieved in the presence of LNA plus indomethacin was significantly higher than that in control conditions (Table 1). In canine coronary arteries from control dogs, acetylcholine and bradykinin induced concentration-dependent, endothelium-dependent relaxations that were not affected significantly by the in vitro treatment with an inhibitor of cyclooxygenase, indomethacin (5 × 10−6M). In vitro, the inhibitor of NOS, LNA (10−4M), induced a significant shift to the right of the concentration-response curves to both agonists. However, the maximal relaxations were not affected significantly. The combination of indomethacin and LNA added in vitro, induced further inhibition; the maximal response to bradykinin remained unaffected (Figs. 1 and 2, Table 2).
In dogs treated for the long term in vivo with LNA, relaxations of the isolated coronary rings to acetylcholine and bradykinin were shifted significantly to the right when compared with the relaxation observed in coronary artery rings taken from control animals. In these arteries, the endothelium-dependent relaxations to both mediators were not influenced significantly by additional in vitro treatment with indomethacin (5 × 10−6M) or LNA (10−4M) or both (Figs. 1 and 2, Table 2).
In isolated coronary arteries taken from control dogs, in the presence of LNA (10−4M) and indomethacin (5 × 10−6M), the resting membrane potential of the smooth-muscle cells averaged −55.3 ± 1.5 mV (n = 31). The resting membrane potential of the smooth-muscle cells of coronary arteries taken from dogs treated in vivo over the long term was not significantly different from that observed in arteries from control dogs (−55.5 ± 1.4 mV; n = 35). In both groups, acetylcholine (10−6M) and bradykinin (10−8 and 10−7M) induced endothelium-dependent hyperpolarizations. When compared with control animals, the hyperpolarization induced by acetylcholine was increased significantly. In contrast, the hyperpolarizations induced by bradykinin were reduced significantly in animals after long-term treatment (Figs. 3 and 4).
Guinea pig arteries
Aortic rings taken from control guinea pigs were contracted with U46619 (3 × 10−8M). In the presence of indomethacin (5 × 10−6M), acetylcholine induced either no or minor relaxations (data not shown). However, bradykinin (10−7M) induced consistent endothelium-dependent relaxations, which reached −45.5 ± 8.2% of the established tone (n = 6). In isolated aorta taken from animals treated in vivo over the long term with L-NAME, bradykinin (10−7M) did not induce significant changes in tension (+2.7 ± 5.4%; n = 10).
In the presence of inhibitors of LNA (10−4M) and indomethacin (5 × 10−6M), the value of the resting membrane potential of the isolated carotid artery smooth-muscle cells was not significantly different in arteries from the control and L-NAME-treated guinea pigs (−56.8 ± 2.0 mV; n = 10; and −60.0 ± 1.8 mV; n = 14 in control and L-NAME-treated animals, respectively). Acetylcholine (10−6M) induced an endothelium-dependent hyperpolarization, which was not affected by the long-term in vivo treatment with L-NAME (Fig. 5). In the guinea pig isolated carotid artery, bradykinin did not produce hyperpolarization in arteries obtained from either control or treated animals (data not shown).
These findings indicate that endothelium-dependent hyperpolarizations are preserved in animals treated over the long term with an NOS inhibitor and act as a back-up mechanism to elicit endothelium-dependent relaxations.
In both animal species studied (dogs and guinea pigs), long-term treatment with an NOS inhibitor was effective. In canine coronary arteries from animals after long-term treatment, the endothelium-dependent relaxations to both acetylcholine and bradykinin were not influenced further by additional in vitro treatment with LNA. Similarly, in the guinea pig isolated aorta, the endothelium-dependent relaxation to bradykinin, which in the presence of indomethacin can be attributed exclusively to NO release, was fully abolished by long-term L-NAME treatment. This was further confirmed in both species, as cyclic guanosine monophosphate (cGMP) accumulation in vascular tissue, an index of NOS activity (20), was reduced significantly by long-term in vivo treatment with NOS inhibitors in both canine coronary artery (18) and guinea pig aorta (data not shown).
In canine coronary arteries, the endothelium-dependent relaxations resistant to the inhibition of NOS and cyclooxygenase were not influenced significantly by long-term in vivo NOS inhibition. The endothelium-dependent hyperpolarizations were also preserved after long-term in vivo NOS inhibition, suggesting that these hyperpolarizations were the cause of the relaxations (9). Previous studies showed an increase in the production of prostacyclin in canine coronary arteries and rat mesenteric arteries taken from animals after long-term treatment with inhibitors of NOS. This upregulation of the cyclooxygenase pathway can act as a back-up mechanism to elicit endothelium-dependent relaxations (17-19). The involvement of the EDHF or the cyclooxygenase pathway or both as back-up mechanisms may depend on the experimental conditions used. In our study, the relaxations induced by acetylcholine and bradykinin observed in isolated canine coronary rings taken from animals after long-term treatment did not seem to involve the production of prostacyclin. The additional in vitro treatment with indomethacin did not modify the responses to either acetylcholine or bradykinin, suggesting that the component of the relaxation that persisted might be attributed to EDHF. However, by increasing the concentration of the stable analog of thromboxane A2 (U46619), the agonist that produced contraction of the coronary rings, this cyclooxygenase-dependent component is unmasked, and the EDHF component suppressed (18). The inhibition of the EDHF component by high concentrations of U46619 can be explained by two different mechanisms. First, U46619 directly inhibits endothelium-dependent hyperpolarizations (21), possibly by inhibiting potassium channels (22); and second, the endothelium-dependent relaxations attributed to the EDHF pathway are dependent on the level of tension developed by the tissue (23; i.e., these relaxations are inhibited when the tissue is maximally contracted). These two back-up mechanisms can be observed in vivo in dogs after long-term treatment, as the coronary vasodilator effects of acetylcholine and bradykinin are resistant to NOS and cyclooxygenase inhibition, although reactive hyperemia is sensitive to cyclooxygenase inhibition (18).
In the guinea pig carotid artery, in the presence of inhibitors of NOS and cyclooxygenase, the endothelium-dependent hyperpolarizations to acetylcholine were not affected by long-term in vivo treatment with L-NAME, whereas in dogs, the hyperpolarization to acetylcholine was enhanced, and the hyperpolarization to bradykinin reduced. This indicates that long-term inhibition of NO production does not evoke a consistent upregulation of the EDHF pathway. Hypertension affects endothelium-dependent relaxation and hyperpolarization (24). However, this cannot explain the alterations observed in dogs, as in these animals after long-term treatment with LNA, a transient increase in arterial blood pressure is observed only the first day of treatment, and thereafter, the arterial blood pressure returns to control value (18). In this species, although systemic vascular resistances are increased, the ability to decrease heart rate and cardiac output prevents major systemic hypertension (18). The changes observed in canine coronary arteries could be linked to alterations in neurohumoral control induced by the long-term in vivo treatment with NOS inhibitors. Indeed, the bradycardia induced by long-term LNA treatment is corrected by atropine, indicating that NOS inhibition may activate parasympathetic tone (18), which in turn could alter the absolute number of muscarinic receptors or their sensitivity. However, a similar phenomenon also is observed in the guinea pig and rat, in which NOS inhibition induces hypertension and moderate bradycardia mediated by the baroreceptor reflex (25,26). The origin of the alterations in the endothelium-dependent hyperpolarizations evoked by acetylcholine and bradykinin in isolated canine coronary artery is not known at present, but according to our results, it cannot be linked to direct regulation of the EDHF pathway by the inhibition of NO production. These findings are at variance with previous reports suggesting that NO regulates EDHF synthesis (11-14). NO has been shown directly to inhibit EDHF-induced relaxation and hyperpolarization (14). The sustained NO production due to endotoxemia-induced induction of NOS II produces a similar effect (12,13). However, this phenomenon may be species dependent, as it was not observed in the rat mesenteric artery (27). The study suggesting that, in carotid arteries from hypercholesterolemic rabbits, EDHF may play a more prominent role under conditions of impaired NO synthesis (11) should be interpreted with caution because in the rabbit carotid artery, NO production can fully account for endothelium-dependent relaxations and hyperpolarizations without the involvement of EDHF (28).
In conclusion, our study shows that endothelium-dependent hyperpolarizations are maintained during long-term NOS inhibition and act as a back-up mechanism to elicit endothelium-dependent relaxations. Although NO may inhibit the formation of EDHF, long-term inhibition of NO production does not evoke upregulation of the EDHF pathway.
Acknowledgment: We thank Dr. A. Rupin (Suresnes, France) for the cGMP dosage in the guinea pig isolated aorta.
1. Palmer RMJ, Ferridge AG, Moncada S. Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Nature
2. Myers PR, Minor RL Jr, Guerra R Jr, Bates JN, Harrison DG. Vasorelaxant properties of the endothelium-derived relaxing factor more closely resemble S
-nitrosocysteine than nitric oxide. Nature
3. Furchgott RF, Zawadzki JV. The obligatory role of the endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature
4. De Mey JG, Claeys K, Vanhoutte PM. Endothelium-dependent inhibitory effects of acetylcholine, adenosine triphosphate, thrombin and arachidonic acid in the canine femoral artery. J Pharmacol Exp Ther
5. Furchgott RF, Vanhoutte PM. Endothelium-derived relaxing and contracting factors. FASEB J
6. Feletou M, Vanhoutte PM. Endothelium-dependent hyperpolarization of canine coronary smooth muscle. Br J Pharmacol
7. Chen G, Suzuki H, Weston AH. Acetylcholine releases endothelium-derived hyperpolarizing factor and EDRF from rat blood vessels. Br J Pharmacol
8. Chen G, Yamamoto Y, Miwa K, Suzuki H. Hyperpolarization of arterial smooth muscle induced by endothelial humoral substances. Am J Physiol
9. Nagao T, Vanhoutte PM. Hyperpolarization as a mechanism for endothelium-dependent relaxations in the porcine coronary artery. J Physiol
10. Corriu C, Feletou M, Canet E, Vanhoutte PM. Endothelium-derived factors and hyperpolarization of the carotid artery of the guinea-pig. Br J Pharmacol
11. Najibi S, Cohen RA. Enhanced role of potassium channels in relaxation of hypercholesterolemic rabbit carotid artery to nitric oxide and sodium nitroprusside. Am J Physiol
12. Mombouli JV, Bissiriou I, Agboton VD, et al. Endotoxemia reduces the contribution of endothelium-derived hyperpolarizing factor, but augments that of prostanoids in canine coronary arteries. In: Vanhoutte PM, ed. EDHF.
Amsterdam: Harwood Academic, 1996:271-8.
13. Kristof AS, Noorhosseini H, Hussain NAS. Attenuation of endothelium-dependent hyperpolarizing factor by bacterial lipopolysaccharides. Eur J Pharmacol
14. Popp R, Bauersachs J, Sauer E, Hecker M, Fleming I, Busse R. The cytochrome P450 monooxygenase pathway and nitric oxide-independent relaxations. In: Vanhoutte PM, ed. EDHF.
Amsterdam: Harwood Academic, 1996:65-72.
15. Doni MG, Whittle B, Palmer RMJ, Moncada S. Actions of nitric oxide on the release of prostacyclin from bovine endothelial cells in culture. Eur J Pharmacol
16. Habib A, Creminon C, Frobert Y, Grassi J, Pradelles P. Maclouf J. Demonstration of an inducible cyclooxygenase in human endothelial cells using antibodies raised against the carboxyl-terminal region of the cyclooxygenase-2. J Biol Chem
17. Henrion D, Dechaux E, Dowell FJ, et al. Alteration of flow-induced dilatation in mesenteric resistance arteries of L-NAME treated rats and its partial association with induction of cyclo-oxygenase-2. Br J Pharmacol
18. Puybasset L, Béa M-L, Ghaleb B, Guidicelli J-F, Berdeaux A. Coronary and systemic hemodynamic effects of sustained inhibition of nitric oxide synthesis in conscious dogs. Circ Res
19. Beverelli F, Béa M-L, Puybasset L, Giudicelli J-F, Berdeaux A. Chronic inhibition of NO-synthase enhances the production of prostacyclin in coronary artery through up-regulation of cyclooxygenase type 1 isoform. Fundam Clin Pharmacol
20. Moncada S, Palmer RJM, Higgs EA. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev
21. Plane F, Garland CJ. Influence of contractile agonists on the mechanism of endothelium-dependent relaxation in rat isolated mesenteric artery. Br J Pharmacol
22. Scornik FS, Toro L. U46619, a thromboxane A2
agonist, inhibits KCa channels activity from pig coronary artery. Am J Physiol
23. Hatake K, Wakabayashi I, Hishida S. Endothelium-dependent relaxation resistant to NG
-nitro-L-arginine in rat aorta. Eur J Pharmacol
24. Nakashima M, Félétou M, Vanhoutte PM. Endothelium-dependent hyperpolarization and potassium channels. In: Godfraind T, Mancia G, Abbracchio MP, Aguilar-Bryan L, Govoni S, eds. Pharmacological control of calcium and potassium homeostasis: biological, therapeutical and clinical aspects.
Amsterdam: Kluwer Academic Publishers, 1995:73-84.
25. Aisaka K, Gross SS, Griffith OW, Levi R. NG
-methylarginine, an inhibitor of endothelium-derived nitric oxide synthesis, is a potent pressor agent in the guinea-pig: does nitric oxide regulate blood pressure in vivo? Biochem Biophys Res Commun
26. Küng CF, Moreau P, Takase H, Düscher TF. L-NAME hypertension alters endothelial and smooth muscle function in rat aorta: prevention by trandolapril and verapamil. Hypertension
27. Fukao M, Hattori Y, Kanno M, Sakuma I, Kitabatake A. Evidence against a role for cytochrome P450-derived arachidonic acid metabolites in endothelium-dependent hyperpolarization by acetylcholine in rat isolated mesenteric artery. Br J Pharmacol
28. Cohen RA, Plane F, Najibi S, Malinski T, Garland CJ. Nitric oxide is the mediator of both endothelium-dependent relaxation and hyperpolarization of the rabbit carotid artery. Proc Natl Acad Sci U S A