Regulation of NO-dependent acetylcholine relaxation by K+ channels and the Na+–K+ ATPase pump in porcine internal mammary artery

European Journal of Pharmacology 641 (2010) 61–66

Contents lists available at ScienceDirect

European Journal of Pharmacology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e j p h a r

Cardiovascular Pharmacology

Regulation of NO-dependent acetylcholine relaxation by K+ channels and the Na+–K+ ATPase pump in porcine internal mammary artery Rosa María Pagán a, Dolores Prieto a, Medardo Hernández a, Carlos Correa b, Albino García-Sacristán a, Sara Benedito a,⁎, Ana Cristina Martínez a a b

Sección Departamental de Fisiología Animal, Facultad de Farmacia, Universidad Complutense de Madrid, Madrid, Spain Sección de Cirugía Experimental y Animalario, Hospital Universitario Ramón y Cajal, Madrid, Spain

a r t i c l e

i n f o

Article history: Received 16 November 2009 Received in revised form 10 March 2010 Accepted 6 May 2010 Available online 21 May 2010 Keywords: Porcine internal mammary artery Acetylcholine relaxation Endothelium Nitric oxide K+ channel Na-K+ ATPase pump

a b s t r a c t This study was designed to determine whether K+ channels play a role in nitric oxide (NO)-dependent acetylcholine relaxation in porcine internal mammary artery (IMA). IMA segments were isolated and mounted in organ baths to record isometric tension. Acetylcholine-elicited vasodilation was abolished by muscarinic receptor blockade with atropine (10-6 M). Incubation with indomethacin (3 × 10−6 M), superoxide dismutase (150 U/ml) and bosentan (10−5 M) did not modify the acetylcholine response ruling out the participation of cyclooxygenase-derivates, reactive oxygen species or endothelin. The relaxation response to acetylcholine was strongly diminished by NO synthase- or soluble guanylyl cyclase-inhibition using L-NOArg (10−4 M) or ODQ (3 × 10−6 M), respectively. The vasodilation induced by acetylcholine and a NO donor (NaNO2) was reduced when rings were contracted with an enriched K+ solution (30 mM), by voltagedependent K+ (Kv) channel blockade with 4-amynopiridine (4-AP; 10−4 M), by Ca2+-activated K+ (KCa) channel blockade with tetraethylammonium (TEA; 10−3 M), and by apamin (5 × 10−7 M) plus charybdotoxin (ChTx; 10−7 M) but not when these were added alone. In contrast, large conductance KCa (BKCa), ATPsensitive K+ (KATP) and inwardly rectifying K+ (Kir) channel blockade with iberiotoxin (IbTx; 10−7 M), glibenclamide (10−6 M) and BaCl2 (3 × 10−5 M), respectively, did not alter the concentration–response curves to acetylcholine and NaNO2. Na+−K+ ATPase pump inhibition with ouabain (10−5 M) practically abolished acetylcholine and NaNO2 relaxations. Our findings suggest that acetylcholine-induced relaxation is largely mediated through the NO-cGMP pathway, involving apamin plus ChTx-sensitive K+ and Kv channels, and Na+−K+-ATPase pump activation. © 2010 Elsevier B.V. All rights reserved.

1. Introduction The internal mammary artery (IMA) is the vessel of choice for coronary artery bypass grafting due to improved long-term patency rates compared with other arteries and saphenous vein (González Santos et al., 2005). However, arterial segments are inherently prone to vasospasm, which can lead to perioperative graft failure (Ding et al., 2008). Progress has been made in understanding the primary mechanism of the myogenic response in human IMA and other arterial segments, with increasing attention paid to the endothelium (Liu et al., 2002; Wei et al., 2007). The balance between endothelial-derived contractile and relaxant factors determines the tone and the physical state of vascular smooth muscle. Endothelium-dependent vascular relaxation is mediated by nitric oxide (NO), prostacyclin and endothelial-derived hyperpolariz⁎ Corresponding author. Sección Departamental de Fisiología Animal, Facultad de Farmacia, Universidad Complutense de Madrid, Plaza Ramón y Cajal s/n, 28040 Madrid, Spain. Tel./fax: + 34 913941696. E-mail address: [email protected] (S. Benedito). 0014-2999/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.ejphar.2010.05.004

ing factor (EDHF). The actions of these endothelium-derived relaxing substances often involve membrane hyperpolarization of vascular smooth muscle, and plasma membrane K+ channels are key molecules for producing membrane electrical events (Busse et al., 2002). Studies over the past 20 years have identified at least four different classes of K+ channels expressed by arterial smooth muscle cells. These include inward rectifier K+ (KIR), ATP-sensitive K+ (KATP), Ca2+-activated K+ (KCa), and voltage-dependent K+ (KV) channels (Busse et al., 2002; Feletou and VanHoutte, 2006). Several studies have tried to determine whether the higher patency of IMA as a bypass graft may be explained by a difference in its ability to release NO and EDHF. Thus, Hamilton et al. (1999) observed that endothelium-dependent relaxation in response to bradykinin in human IMA was strictly NO-dependent. Conversely, the group of He (He and Liu, 2001; Liu et al., 2000) reported the ability of IMA to release both NO and EDHF. More recently, Archer et al. (2003) identified the EDHF involved in human IMA and reported that acetylcholine and bradykinin induce the release of the cytochrome epoxygenase-derived 11,12-EET promoting hyperpolarization and relaxation through the activation of large conductance KCa channels (BKCa) located on smooth muscle cells.

62

R.M. Pagán et al. / European Journal of Pharmacology 641 (2010) 61–66

Endothelium-dependent vasodilation can be assessed using receptor-operated stimuli such as acetylcholine. This approach is commonly considered a marker of endothelial function. The differences revealed by the above-mentioned reports could be the outcome of variable IMA behaviour depending on the cardiovascular risk factors of the patients included in these studies. Cardiovascular risk factors such as diabetes, hypercholesterolemia, hypertension, obesity or smoking may be linked to endothelial dysfunction, atherosclerosis or coronary heart disease (Smith, 2007). Further, surgical manipulation or vasodilatory drugs administered during coronary artery bypass grafting may also affect the endothelial responses of human IMA (Gao et al., 2003; González Santos et al., 2005). Accordingly, data obtained using isolated human IMA segments may differ to those observed in healthy vessels. Thus, further investigations on the IMA require the use of an experimental model such as the pig (Pagán et al., 2009; Pesic et al., 1999) to eliminate confounding cardiovascular risk factors present in patients undergoing bypass surgery. Indeed, knowledge of endothelial function in this vessel could be essential for understanding the good long-term patency of the IMA. This study was designed to examine whether K+ channels and Na+– + K ATPase pump activation play important roles in NO-dependent acetylcholine relaxation in porcine IMA. 2. Materials and methods 2.1. Tissue preparation, dissection and mounting IMA segments from 37 cross-breed male pigs (weight 35–45 kg) were obtained from the Experimental Surgery Department of the Hospital Universitario Ramón y Cajal (Madrid, Spain) shortly after the animals were euthanized. Animals were handled according to European Union regulation (86/609/EEC) and the Spanish Normative for the Care and Use of Laboratory Animals (RD 1201/2005). The study protocol was approved by the Ethics Committee for Animal Welfare of the Hospital Universitario Ramón y Cajal. After dissection, the tissue was transported to the laboratory in cooled (4 °C) physiological saline solution (PSS). Arterial segments were cleaned from adhering connective tissue and cut into 2 mm length rings. Vessel rings were transferred to 5 ml organ baths containing PSS at 37 °C and gassed with a mixture of 95% O2 and 5% CO2 to maintain the pH at 7.4. Rings (external diameter = 3.18 ± 0.06 mm and internal diameter = 2.08 ± 0.04 mm; n = 37) were mounted between two parallel Lshaped stainless steel wires. One wire was fixed to a displacement unit allowing fine adjustment of tension while the other was attached to a force transducer (Grass FT03C). Special care was taken to avoid damage to the endothelium. The isometric tension of the vessel wall was displayed and recorded using a PowerLab data acquisition system and Chart v5.5 software. 2.2. Experimental procedure Each ring was stretched in a stepwise manner to its optimal resting tension (≈20 mN). This tension was determined previously in lengthactive tension relationship experiments. The contractile capacity of the preparation was tested by exposing the arterial rings to 124 mM K+ (K-PSS, 17.4 ± 0.3 mN, n = 37). Concentration–response curves for acetylcholine and a nitric oxide donor (NaNO2) were constructed by adding increasing concentrations of the agonist into the organ bath. Acetylcholine- and NaNO2-induced relaxations were examined in preparations contracted with noradrenaline (10−7–3 ×10−7 M). These concentrations of noradrenaline produced a stable contraction, corresponding to 60–70% of the response induced by K-PSS and of sufficient duration to permit the analysis of agonist responses. Previous experiments showed that two consecutive acetylcholine or NaNO2 concentration–response curves were not reproducible. Thus, the following experiments were conducted using consecutive

segments from the same animal, with one segment acting as the control of the other. This meant that only one acetylcholine or NaNO2 concentration–response curve per arterial ring could be obtained. We assessed the involvement of muscarinic receptors, NO-cGMP or cyclooxygenase pathways, K+ channels, Na+–K+ ATPase pump, reactive oxygen species and endothelin receptors in the acetylcholineor NaNO2-responses by adding to the organ bath atropine, L-NOArg, ODQ, indomethacin, different K+ channels blockers (4-aminopyridine (4-AP), apamin, barium chloride (BaCl2), charybdotoxin (ChTx), glibenclamide, iberiotoxin (IbTx), tetraethylammonium (TEA)), ouabain, superoxide dismutase (SOD) or bosentan, respectively, 30 min before the construction of the concentration–response curve. To test the participation of a hyperpolarizing component in acetylcholine-evoked relaxation, concentration–response curves for acetylcholine and NaNO2 were constructed on contractions elicited by a high concentration K+ solution (30 mM) and compared to those produced by noradrenaline. The stable contractions induced by 30 mM KCl (14.3 ± 2.6 mN; n = 16) and noradrenaline were effectively very similar (14.2 ± 1.7 mN; n = 16). 2.3. Drugs and solutions The following drugs were used: acetylcholine, 4-aminopyridine, apamin, atropine, barium chloride, charybdotoxin, glibenclamide, iberiotoxin, indomethacin, Nω-nitro-L-arginine (L-NOArg), noradrenaline, ouabain, 1H-[1,2,4] oxadiazol [4,3,-α]quinaxolin-1-one (ODQ), sodium nitrite, superoxide dismutase, tetraethylammonium (all from SigmaAldrich, St. Louis, MO, U.S.A.) and bosentan (a gift from Hoffmann-La Roche, Inc., USA). All drugs were dissolved in distilled water except: indomethacin, which was prepared in ethanol (96%), glibenclamide and ODQ, which were dissolved in dimethylsulphoxide, and NaNO2 , which required an acidified solution. In prior experiments, these solvents had no effect on the preparations. The concentrations of agents are expressed as their final concentration in the organ bath. The composition of PSS was (mM); NaCl 119, KCl 4.7, CaCl2 1.5, MgSO4 1.2, NaHCO3 25, glucose 11, KH2PO4 1.2 and ethylenediaminetetraacetic acid (EDTA) 0.027. K-PSS was identical to PSS except that NaCl was replaced on an equimolar basis. We also prepared a high K+ concentration solution using normal PSS in which 30 mM NaCl was exchanged isotonically with KCl. 2.4. Statistical analysis Relaxing responses to acetylcholine and NaNO2 observed in contracted arteries were expressed as the percentage inhibition of the vascular contraction induced by noradrenaline. Emax refers to the maximum response achieved. The effects of different blockers on basal tension were expressed as percentages of K-PSS contraction. For each concentration–response curve, the agonist concentration eliciting the half-maximal response (EC50) was estimated by computerized nonlinear regression analysis (GraphPad Software, U.S.A.). The sensitivity of the drugs is expressed in terms of their pD2, which is defined as the negative logarithm of the EC50 for the agonist used (pD2 = −log EC50). Results were expressed as the mean ± standard error of the mean (S.E.M.) of n animals. Statistical determinations were performed using the Student's t-test for unpaired data. A P value of less than 5% was taken to denote a significant difference. 3. Results Acetylcholine (10−9–10−5 M) caused the concentration-dependent relaxation of noradrenaline-contracted arterial rings (pD2 = 7.29± 0.09 and Emax = 75.4 ± 3.1%; n = 37) (Fig. 1A). This relaxant effect was

R.M. Pagán et al. / European Journal of Pharmacology 641 (2010) 61–66

Fig. 1. Concentration–response relaxation curves to (A) acetylcholine and (B) NaNO2 on porcine endothelium-intact IMA segments contracted with noradrenaline in the absence and presence of atropine (10−6 M), L-NOArg (10−4 M) and ODQ (3× 10−6 M) and on endothelium-denuded IMA segments. Each point represents the means± S.E.M. of results obtained in 8–9 animals.

abolished by the muscarinic receptor antagonist, atropine (10−6 M) and endothelium removal (Fig. 1A). The possibility of contractile factors reducing the acetylcholineinduced relaxation was addressed by determining the effects of cyclooxygenase inhibition, reactive oxygen species removal and endothelin receptor blockade. The cyclooxygenase inhibitor, indomethacin (3 × 10−6 M), failed to modify the acetylcholine concentration–relaxation curve (pD2 = 7.29 ± 0.09 and Emax = 75.4 ± 3.1% for controls; pD2 = 7.10 ± 0.15 and Emax = 77.5 ± 5.6% for indomethacin; n = 9). The reactive oxygen species scavenger, SOD (150 U/mL), had no effect on the relaxation induced by acetylcholine (pD2 = 7.46 ± 0.13 and Emax = 76.5 ± 5.3 for controls; pD2 = 7.20± 0.20 and Emax = 76.6 ± 7.5% for SOD; n = 8). Pretreatment with the non-specific endothelin receptor antagonist, bosentan (10−5 M), did not modify the acetylcholine-elicited relaxation (pD2 = 6.58 ± 0.22 and Emax = 74.5 ± 8.0% for controls; pD2 = 6.82 ± 0.16 and Emax = 62.8 ± 8.9% for bosentan; n = 7). We also explored the role played by the NO-cGMP pathway in the acetylcholine-elicited relaxation response. Acetylcholine-induced relaxation was strongly inhibited by the nitric oxide synthase (NOS) inhibitor L-NOArg (10−4 M), and by the soluble guanylyl cyclase (sGC) inhibitor ODQ (3 × 10−6 M) (pD2 = 7.20 ± 0.17 and Emax = 64.1 ± 5.6% for controls, n = 9; Emax = 13.6 ± 3.7% for L-NOArg, P b 0.001, n = 9; Emax = 14.4 ± 3.3% for ODQ, P b 0.001; n = 9) (Fig. 1A). Endothelium-dependent relaxation due to acetylcholine is mainly related to NO synthesis. We therefore examined the effect of a NO donor such as NaNO2. Cumulative concentrations of NaNO2 (10−6–10−3 M) induced a dose-dependent vasodilation of porcine IMA segments contracted with noradrenaline (pD2 = 4.41 ± 0.04 and Emax = 77.5 ± 2.0%; n = 26) (Fig. 1B). Relaxations produced in response to NaNO2 were similar in preparations in which the endothelial cells had been mechanically removed (pD2 = 4.38 ± 0.12 and Emax = 75.2 ± 2.3%; n = 8) (Fig. 1B). NaNO2-evoked vasodilation was unmodified by −4 L-NOArg (10 M) (pD2 = 4.25 ± 0.15 and Emax = 78.9 ± 5.4%; n = 9), but practically abolished by ODQ (3× 10−6 M) (Emax = 16.4 ± 5.2%, P b 0.001; n = 9) (Fig. 1B). Resting tone was increased in segments incubated with L-NOArg (42.4 ± 6.4%, n = 15) and ODQ (51.2 ± 10.2%, n = 19). However, the final precontraction reached before constructing the agonist concentration–response curve was comparable to that obtained for controls. We also assessed the possibility that acetylcholine induces hyperpolarization, by comparing relaxant responses in preparations contracted with noradrenaline to those contracted using a high K+ solution (30 mM). The latter was used to abolish the driving force of K+ efflux and subsequent membrane hyperpolarization. In these experiments,

63

the sensitivity and the maximum response to acetylcholine were reduced when preparations were contracted with a high K+ solution (30 mM) (pD2 = 6.82 ± 0.14, P b 0.01 and Emax = 38.7 ± 5.6%, P b 0.01; n = 7) compared to noradrenaline (pD2 = 7.70 ± 0.19 and Emax = 84.5 ± 2.8%). The hyperpolarizing component of the vasodilation induced by acetylcholine can normally be attributed to activation of K+ channels. Blockade of KCa channels with TEA (10-3 M) significantly diminished relaxation induced by acetylcholine (pD 2 = 7.20 ± 0.15 and Emax = 66.1 ± 5.3% for controls; pD2 = 6.40 ± 0.20, P b 0.01 and Emax = 59.6 ± 5.7% for TEA; n = 8) (Fig. 2A). Treatment with IbTx (10-7 M), a large conductance KCa (BKCa) channel blocker, did not alter the relaxations elicited by acetylcholine (pD2 = 7.12 ± 0.19 and Emax = 77.3 ± 5.8% for controls; pD2 = 6.94 ± 0.13 and Emax = 67.6 ± 5.7% for IbTx; n = 7) (Fig. 2C). Similarly, the acetylcholine response (pD2 = 7.07 ± 0.17 and Emax = 73.3 ± 6.2%; n = 7) was not significantly affected by ChTx (10−7 M), a large- and intermediate-conductance KCa (IKCa) channel blocker (pD2 = 7.18 ± 0.07 and Emax = 72.5 ± 7.2%; n = 7), or by apamin (5 × 10−7 M), a small-conductance KCa (SKCa) channel blocker

Fig. 2. Concentration–response relaxation curves to (A, C, E) acetylcholine and (B, D, F) NaNO2 on porcine IMA segments contracted with noradrenaline in the absence and presence of (A, B) tetraetylammonium (TEA, 10−3 M), (C, D) iberiotoxin (IbTx, 10−7 M), (E, F) charibdotoxin (ChTx, 10−7 M), apamin (5 × 10−7 M) or ChTx plus apamin. Each point represents the means± S.E.M. of results obtained in 7–8 animals.

64

R.M. Pagán et al. / European Journal of Pharmacology 641 (2010) 61–66

(pD2 = 7.18 ± 0.13 and Emax = 67.9 ± 8.8%; n = 7) (Fig. 2E). However, ChTx plus apamin reduced the relaxation response to acetylcholine (pD2 = 6.95 ± 0.16 and Emax = 48.3 ± 4.9% for ChTx plus apamin, P b 0.01; n = 7) (Fig. 2E). 4-AP (10−4 M), a Kv channels inhibitor, reduced acetylcholineinduced relaxations (pD2 = 7.50 ± 0.09 and Emax = 77.6 ± 5.3% for controls; pD2 = 7.17 ± 0.13, P b 0.01and Emax = 73.3 ± 3.6% for 4-AP; n = 9) (Fig. 3A). In contrast, the concentration–response curves for acetylcholine (pD2 = 7.25 ± 0.11 and Emax = 82.7 ± 4.1%; n = 9) were unaffected by incubation with glibenclamide (10−6 M), a KATP channel blocker (pD2 = 7.20 ± 0.11 and Emax = 81.7 ± 6.9%; n = 9) (Fig. 3C). Blockade of Kir channels with BaCl2 (3 × 10−5 M) did not alter the relaxations elicited by acetylcholine (pD 2 = 7.15 ± 0.12 and Emax = 79.3 ± 4.0% for controls; pD2 = 7.10 ± 0.16 and Emax = 80.4 ± 8.5% for BaCl2; n = 7) (Fig. 3E). Conversely, ouabain (10−5 M), a Na+– K+ ATPase pump inhibitor, practically abolished the acetylcholine responses (Emax = 4.3 ± 1.4%, P b 0.001; n = 7) (Fig. 3E). Ouabain plus BaCl2 induced a similar effect to that observed when ouabain was added separately (Emax = 4.5 ± 1.3%, P b 0.001; n = 7) (Fig. 3E).

To test whether NO released from the endothelium is able to modulate K+ channel activity, we examined the effects of the same treatments on the response to NaNO2. Concentration–relaxation curves for NaNO2 in vessels contracted with 30 mM KCl (pD2 = 3.92 ± 0.13, P b 0.05 and Emax = 54.4 ± 5.8%, P b 0.01; n = 9) were inhibited with respect to those contracted by noradrenaline (pD2 = 4.33 ± 0.13 and Emax = 75.9 ± 3.0%). Similarly to the results yielded by acetylcholine, NaNO2-evoked relaxations (pD2 = 4.39 ± 0.12 and Emax = 82.2 ± 4.8%; n = 9) were reduced in the presence of TEA (pD2 = 3.98 ± 0.10, P b 0.05 and Emax = 63.3 ± 4.5%, P b 0.05; n = 9) (Fig. 2B) and there was no change in the presence of IbTx (pD2 = 4.25 ± 0.14 and Emax = 73.7 ± 6.4% for controls; pD2 = 4.25 ± 0.17 and Emax = 67.6 ± 6.9% for IbTx; n = 9) (Fig. 2D). NaNO2 vasodilation (pD2 = 4.37 ± 0.07 and Emax = 70.9 ± 3.3; n = 9) did not differ significantly in the presence of ChTx (pD 2 = 4.49 ± 0.16 and E max = 68.5 ± 6.3%; n = 9) or apamin (pD2 = 4.50 ± 0.07 and Emax = 68.4 ± 4.1%; n = 9) (Fig. 2F). However, treatment with ChTx plus apamin inhibited relaxations produced in response to the NO donor (pD2 = 4.06± 0.09, P b 0.05 and Emax = 61.4 ± 3.7, P b 0.05; n = 9) (Fig. 2F). Moreover, 4-AP also attenuated NaNO2-induced relaxations (pD2 = 4.46 ± 0.07 and Emax = 85.7 ± 4.4% for controls; pD2 = 4.21 ± 0.06, P b 0.05 and Emax = 56.6 ± 5.9%, P b 0.01 for 4-AP; n = 9) (Fig. 3B). In contrast, glibenclamide failed to modify the NaNO2 concentration– relaxation curve (pD2 = 4.27 ± 0.08 and Emax = 67.7 ± 9.2% for controls; pD2 = 4.35 ± 0.07 and Emax = 66.6 ± 6.8% for glibenclamide; n = 7) (Fig. 3D). BaCl2 showed no effect on the NaNO2-elicited relaxations (pD2 = 4.63 ± 0.16 and Emax = 69.1 ± 5.5% for controls; pD2 = 4.53 ± 0.22 and Emax = 66.0 ± 6.1% for BaCl2; n = 7) (Fig. 3F). However, the relaxation provoked by NaNO2 was almost abolished by ouabain (Emax = 4.3 ± 1.4%, P b 0.001; n = 7) and by ouabain plus BaCl2 (Emax = 4.5 ± 1.3%, P b 0.001; n = 7) (Fig. 3F). Basal tone was increased by the addition of TEA (11.3 ± 5.3%, n = 17), 4-AP (43.0 ± 6.37%, n = 18), and ouabain (74.8 ± 10.3, n = 14) to the organ bath. Nevertheless, the contraction achieved in these treated segments was similar to that obtained in control conditions. 4. Discussion

Fig. 3. Concentration–response relaxation curves to (A, C, E) acetylcholine and (B, D, F) NaNO2 on porcine IMA segments contracted with noradrenaline in the absence and presence of (A, B) 4-aminopyridine (4-AP, 10−4 M), (C, D) glibenclamide (10−6 M), (E, F) barium chloride (barium, 3 × 10−5 M), ouabain (10−5 M), or barium plus ouabain. Each point represents the mean ± S.E.M. of results obtained in 7–9 animals.

The results of this study indicate the involvement of apamin plus ChTx-sensitive K+ and KV channels, and Na+–K+ ATPase pump activation in acetylcholine-induced relaxation, mainly mediated by NO. Our findings obtained in porcine IMA segments reveal that the acetylcholine-elicited vasodilatory effect is achieved by endothelial muscarinic receptor activation. Pesic et al. (1999, 2001) characterized muscarinic receptor subtypes in human and porcine IMA. Although some studies have addressed the mechanisms underlying the acetylcholine relaxant response in human IMA, results have been controversial (Liu et al., 2002; Wei et al., 2007). Thus, it seems the acetylcholine vasodilatory response varies depending on the level of endothelial dysfunction induced by cardiovascular risk factors. Acetylcholine-evoked vasodilatation is thought to depend on NO synthesized by endothelial NO synthase and prostanoids, as well as a non-NO/non-prostanoid EDHF (Busse et al., 2002). A healthy endothelium, but especially a dysfunctioning endothelium, may be a source of other substances and mediators that are detrimental to the arterial wall, including endothelin-1, thromboxane A2, prostaglandin H2 and reactive oxygen species (Taddei et al., 2003) due to their vasoconstrictive actions. The synthesis of arachidonic-acid derivates may contribute to acetylcholine-induced relaxation. In our experiments, the lack of any effect of the cyclooxygenase inhibitor indomethacin precludes a role

R.M. Pagán et al. / European Journal of Pharmacology 641 (2010) 61–66

for prostanoids in the acetylcholine-evoked response. Liu et al. (2002) reported that the addition of indomethacin increases vasodilation in response to acetylcholine in human IMA segments. In contrast, Wei et al. (2007) observed no effects of indomethacin. These discrepancies could be due to differences in the cardiovascular risk factors of patients undergoing coronary artery bypass graft surgery. Given that our data were obtained in porcine IMA rings, in healthy segments with an intact endothelium, cyclooxygenase-derived prostanoids may not be involved in acetylcholine-elicited relaxation. Huraux et al. (1999) were unable to attribute reduced acetylcholine-elicited relaxant responses to increased reactive oxygen species production in human IMA grafts. In the present study, SOD exerted no effect on acetylcholine-induced relaxation, ruling out the role of reactive oxygen species in the response to acetylcholine in control conditions. Similarities between our results and those of Huraux et al. (1999) could be explained if the oxidative stress present in the patients of that study was insufficient to modify the acetylcholineinduced response. In addition to reactive oxygen species, the endothelium may produce vasoconstrictive factors such as endothelin-1. Notably, this vasoactive mediator is produced in abundance during coronary artery bypass grafting (Danová et al., 2007) and has been implicated in the genesis of vasospasm (Lockowandt et al., 2004). Endothelin-1 usually acts as a vasoconstrictor through the activation of endothelin type-A receptors on smooth muscle cells. In contrast, endothelin type-B receptors are abundant on endothelial cells and mediate a relaxation response thus inhibiting vasoconstriction (Penna et al., 2006). Both receptor subtypes have been characterized in human IMA grafts and specific endothelin type-A receptor blockade has also been proposed as a useful tool for preventing IMA vasospasm (Lockowandt et al., 2004). In our experiments, bosentan failed to modify acetylcholineinduced relaxation, indicating this vasodilatory response is unaffected by endothelin release. Although endothelin-1 production in the IMA could be offset by increased NO production, under disease conditions the endothelin-1/NO ratio can be augmented and potentially contribute to impaired IMA graft function. This rationale could support the findings of Verma et al. (2001), who suggested that endothelin-1 receptor blockade by bosentan improves acetylcholine-induced relaxation in human IMA grafts. There is a large body of evidence demonstrating that NO is the main mediator of acetylcholine-induced relaxation in different vascular beds (Martínez et al., 2005; Segarra et al., 2000; Wei et al., 2007). In the IMA, NO production is greater than in other vessels used for revascularization and this has been associated with a low incidence of perioperative myocardial ischemia (Liu et al., 2000; Segarra et al., 2000). Accordingly, NO donors and phosphodiesterase inhibitors are used in coronary artery bypass surgery to deal with vasospasm (Ding et al., 2008). Our experiments reveal that acetylcholine-induced relaxation is chiefly mediated through the NO-cGMP pathway, since this response was strongly diminished by L-NOArg and ODQ. These data are in agreement with previous findings in human IMA (Liu et al., 2002; Wei et al., 2007). Acetylcholine-evoked relaxation was reduced after contracting porcine IMA with a depolarizing concentration of K+. The hyperpolarizing component of the response to vasoactive substances can normally be attributed to activation of different K+ channels (Busse et al., 2002; Feletou and VanHoutte, 2006). Indeed, KCa channels may counteract vasopressin-induced contractions in the human IMA (Novella et al., 2007), and KCa and Kv have been shown to counterbalance adrenergic contractions in porcine radial artery (Pagán et al., 2009). KATP channel openers are thought to be beneficial against vasospasm since they seem to contribute to IMA relaxation (Ding et al., 2008). Kir channels might also be involved in the CGRP relaxation of human radial artery (Zulli et al., 2008). In porcine IMA, the partial reduction produced by a concentration of TEA selective for KCa channels suggests the participation of these channels in the

65

acetylcholine response. In an attempt to elucidate which KCa channel subtypes were involved, we assessed the effects of IbTx (BKCa channel blocker), ChTx (BKCa and IKCa channel blocker) and apamin (SKCa channel blocker) on relaxations produced in response to acetylcholine. Concentration–response curves for acetylcholine were unaltered when these blockers were added separately. However, combined treatment with ChTx and apamin reduced the acetylcholine-evoked relaxation. These results are similar to those obtained in rat mesenteric artery in which simultaneous blockade of IKCa and SKCa channels revealed an inhibitory effect on acetylcholine relaxation not produced by the separate actions of ChTx and apamin (Stankevicius et al., 2006). Inhibition of relaxation by this combination of substances has become widely regarded as a hallmark of the EDHF phenomenon (Edwards et al., 1998). In isolated guinea-pig carotid artery, Gluais et al. (2005) hypothesized these results could be explained if apamin increases ChTx binding to a heteromultimer channel with two binding sites, rather than these two blockers acting independently blocking specific apamin- or ChTx-sensitive channels. This theory does not seem plausible since, in our experiments, no effects were observed when apamin was applied on its own. The fact that ChTx added alone was ineffective, while 4-AP, a specific Kv blocker, significantly diminished the acetylcholine-elicited relaxation, precludes a non-specific effect of ChTx on Kv (Ko et al., 2008) and suggests a role for Kv channels. Moreover, the lack of effect of IbTx, a selective blocker of BKCa, rules out the role for BKCa channels in the response to acetylcholine. Despite functional KATP and Kir channels having been characterized in human IMA (Karkanis et al., 2003; Luu et al., 1997), the present findings revealed no inhibitory effects of glibenclamide or BaCl2 on acetylcholine relaxation, suggesting that these types of K+ channel are not involved. Ouabain has been used to characterize the contribution of this pump to EDHF-induced responses (Edwards et al., 1998). A role for Na+–K+ ATPase has generated renewed interest after K+ was suggested as an EDHF in the rat hepatic artery and it was observed that this response could be inhibited by a combination of Ba2+ and ouabain (Bény and Schaad, 2000; Edwards et al., 1998). In the present study, acetylcholineevoked vasodilation was abolished by Na+–K+ ATPase pump inhibition. Previous reports have advocated the involvement of different K+ channels and subsequent Na+–K+ ATPase pump activation in the mechanisms underlying acetylcholine-elicited relaxation in rat femoral arteries (Savage et al., 2003), the rabbit aorta (Ferrer et al., 1999), and horse penile small arteries (Prieto et al., 1998). In fact, it has been suggested that KCa (Prieto et al., 1998), Kir (Savage et al., 2003) or Kv (Ferrer et al., 1999) channel opening can activate the Na+–K+ ATPase pump to induce hyperpolarization. Given our results preclude a role of KATP or Kir channels in the response to acetylcholine, Na+–K+ ATPase pump activation is probably mediated by K+, effluxing through K+ channels located in the porcine IMA wall. Having established that NO was the most important factor mediating acetylcholine relaxation, we decided to assess whether NO was responsible for opening these ion channels (Martínez et al., 2005) and for Na+–K+ ATPase pump activation, as has been reported for other blood vessels (Palacios et al., 2004). Similar data to those obtained for acetylcholine were recorded for NaNO2, indicating that NO mediates relaxation by apamin- and ChTx-sensitive K+ and Kv channel opening and Na+–K+ ATPase pump activation. In smooth muscle, different mechanisms have been proposed for NO relaxation through K+ channel opening and Na+–K+ ATPase pump activation. In fact, it has been demonstrated that NO may follow both cGMP-dependent (Palacios et al., 2004; Archer et al., 1994) and -independent (Bolotina et al., 1994; Gupta et al., 1995) pathways. In the present study, ODQ practically abolished both acetylcholine- and NaNO2-induced vasodilations. Thus, it is highly likely that in porcine IMA, acetylcholine-evoked activation of K+ channels and the exchanger pump is mediated by NO, which in turn

66

R.M. Pagán et al. / European Journal of Pharmacology 641 (2010) 61–66

activates a cGMP-dependent pathway. Nevertheless, since a residual relaxant response to acetylcholine- and NaNO2 remains after adding L-NOArg or ODQ to the organ bath, we do not exclude the possibility that K+ channels or the Na+–K+ ATPase pump could also be activated by other NO- or cGMP-independent ways. 5. Conclusions Our findings suggest that acetylcholine-induced relaxation is largely mediated through the NO-cGMP pathway, involving apamin plus ChTx-sensitive K+ and Kv channels as well as Na+–K+-ATPase pump activation. These results provide functional evidence for understanding the mechanisms underlying the acetylcholine-evoked relaxant response observed in healthy IMA segments. Acknowledgments This study was supported by grant PI 031257 from the FIS, Spanish Ministry of Health. The authors thank Dr. Luis Orensanz for his kind cooperation in this study and Ms. Consolación de la Calle for technical assistance. References Archer, S.L., Huang, J.M., Hampl, V., Nelson, D.P., Shultz, P.J., Weir, E.K., 1994. Nitric oxide and cGMP cause vasorelaxation by activation of a charybdotoxin-sensitive K+ channel by cGMP-dependent protein kinase. Proc. Natl Acad. Sci. USA 91, 7583–7587. Archer, S.L., Gragasin, F.S., Wu, X., Wang, S., McMurtry, S., Kim, D.H., Platonov, M., Koshal, A., Hashimoto, K., Campbell, W.B., Falck, J.R., Michelakis, E.D., 2003. Endothelium-derived hyperpolarizing factor in human internal mammary artery is 11, 12-epoxyeicosatrienoic acid and causes relaxation by activating smooth muscle BKCa channels. Circulation 107, 769–776. Bény, J.L., Schaad, O., 2000. An evaluation of potassium ions as endothelium-derived hyperpolarizing factor in porcine coronary arteries. Br. J. Pharmacol. 131, 965–973. Bolotina, V.M., Najibi, S., Palacino, J.J., Pagano, P.J., Cohen, R.A., 1994. Nitric oxide directly activates calcium-dependent potassium channels in vascular smooth muscle. Nature 368, 850–853. Busse, R., Edwards, G., Feletou, M., Fleming, I., Vanhoutte, P.M., Weston, A.H., 2002. EDHF: bringing the concepts together. Trends Pharmacol. Sci. 23, 374–380. Danová, K., Pechán, I., Olejárová, I., Líska, B., 2007. Natriuretic peptides and endothelin1 in patients undergoing coronary artery bypass grafting. Gen. Physiol. Biophys. 26, 194–199. Ding, R., Feng, W., Li, H., Wang, L., Li, D., Cheng, Z., Guo, J., Hu, D., 2008. A comparative study on in vitro and in vivo effects of topical vasodilators in human internal mammary, radial artery and great saphenous vein. Eur. J. Cardiothorac. Surg. 34, 536–541. Edwards, G., Dora, K.A., Gardener, M.J., Garland, C.J., Weston, A.H., 1998. K+ is an endothelium-derived hyperpolarizing factor in rat arteries. Nature 396, 269–272. Feletou, M., VanHoutte, P.M., 2006. Endothelium-derived hyperpolarizing factor. Where are we now? Arterioscler. Thromb. Vasc. Biol. 26, 1215–1225. Ferrer, M., Marín, J., Encabo, A., Alonso, M.J., Balfagón, G., 1999. Role of K+ channels and sodium pump in the vasodilation induced by acetylcholine, nitric oxide and cyclic GMP in the rabbit aorta. Gen. Pharmacol. 33, 35–41. Gao, Y.J., Yang, H., Teoh, K., Lee, R.M., 2003. Detrimental effects of papaverine on the human internal thoracic artery. J. Thorac. Cardiovasc. Surg. 126, 179–185. Gluais, P., Edwards, G., Weston, A.H., Falck, J.R., Vanhoutte, P.M., Félétou, M., 2005. Role of SKCa and IKCa in endothelium-dependent hyperpolarizations of the guinea-pig isolated carotid artery. Br. J. Pharmacol. 144, 477–485. González Santos, J.M., López Rodríguez, J., Dalmau Sorlí, M.J., 2005. Arterial grafts in coronary surgery. Treatment for everyone? Rev. Esp. Cardiol. 58, 1207–1223. Gupta, S., Moreland, R.B., Munarriz, R., Daley, J., Goldstein, I., Saenz de Tejada, I., 1995. Possible role of Na+-K+-ATPase in the regulation of human corpus cavernosum smooth muscle contractility by nitric oxide. Br. J. Pharmacol. 116, 2201–2206. Hamilton, C.A., Williams, R., Pathi, V., Berg, G., McArthur, K., McPhaden, A.R., Reid, J.L., Dominiczak, A.F., 1999. Pharmacological characterization of endothelium-dependent relaxation in human radial artery: comparison with internal thoracic artery. Cardiovasc. Res. 42, 214–223.

He, G.W., Liu, Z.G., 2001. Comparison of nitric oxide release and endothelium-derived hyperpolarizing factor-mediated hyperpolarization between human radial and internal mammary arteries. Circulation 104, I344–I349. Huraux, C., Makita, T., Kurz, S., Yamaguchi, K., Szlam, F., Tarpey, M.M., Wilcox, J.N., Harrison, D.G., Levy, J.H., 1999. Superoxide production, risk factors and endothelium-dependent relaxations in human internal mammary arteries. Circulation 99, 53–59. Karkanis, T., Li, S., Pickering, J.G., Sims, S.M., 2003. Plasticity of KIR channels in human smooth muscle cells from internal thoracic artery. Am. J. Physiol. Heart Circ. Physiol. 284, H2325–H2334. Ko, E.A., Han, J., Jung, I.D., Park, W.S., 2008. Physiological roles of K+ channels in vascular smooth muscle cells. J. Smooth Muscle Res. 44, 65–81. Liu, Z.G., Ge, Z.D., He, G.W., 2000. Difference in endothelium-derived hyperpolarizing factor-mediated hyperpolarization and nitric oxide release between human internal mammary artery and saphenous vein. Circulation 102, III296–301. Liu, M.H., Jin, H., Floten, H.S., Ren, Z., Yim, A.P., He, G.W., 2002. Vascular endothelial growth factor-mediated, endothelium-dependent relaxation in human internal mammary artery. Ann. Thorac. Surg. 73, 819–824. Lockowandt, U., Ritchie, A., Grossebener, M., Franco-Cereceda, A., 2004. Endothelin and effects of endothelin-receptor activation in the mammary and radial artery. Scand. Cardiovasc. J. 38, 240–244. Luu, T.N., Dashwood, M.R., Tadjkarimi, S., Chester, A.H., Yacoub, M.H., 1997. ATPsensitive potassium channels mediate vasodilatation produced by calcitonin generelated peptide in human internal mammary but not gastroepiploic arteries. Eur. J. Clin. Invest. 27, 960–966. Martínez, A.C., Prieto, D., Hernández, M., Rivera, L., Recio, P., García-Sacristán, A., Benedito, S., 2005. Endothelial mechanisms underlying responses to acetylcholine in the horse deep dorsal penile vein. Eur. J. Pharmacol. 515, 150–159. Novella, S., Martínez, A.C., Pagán, R.M., Hernández, M., García-Sacristán, A., GonzálezPinto, A., González-Santos, J.M., Benedito, S., 2007. Plasma levels and vascular effects of vasopressin in patients undergoing coronary artery bypass grafting. Eur. J. Cardiothorac. Surg. 32, 69–76. Pagán, R.M., Martínez, A.C., Martínez, M.P., Hernández, M., García-Sacristán, A., Correa, C., Prieto, D., Benedito, S., 2009. Endothelial and potassium channel dependent modulation of noradrenergic vasoconstriction in the pig radial artery. Eur. J. Pharmacol. 616, 166–174. Palacios, J., Marusic, E.T., Lopez, N.C., Gonzalez, M., Michea, L., 2004. Estradiol-induced expression of Na+–K+-ATPase catalytic isoforms in rat arteries: gender differences in activity mediated by nitric oxide donors. Am. J. Physiol. Heart Circ. Physiol. 286, H1793–H1800. Penna, C., Rastaldo, R., Mancardi, D., Cappello, S., Pagliaro, P., Westerhof, N., Losano, G., 2006. Effect of endothelins on the cardiovascular system. J. Cardiovasc. Med. 7, 645–652. Pesic, S., Grbovic, L., Jovanovic, A., Radenkovic, M., 1999. Acetylcholine-induced contractions in the porcine internal mammary artery: possible role of muscarinic receptors. Zentralbl. Veterinärmed. A 46, 509–515. Pesic, S., Jovanoviv, A., Grbovic, L., 2001. Muscarinic receptor subtypes mediating vasorelaxation of the perforating branch of the human internal mammary artery. Pharmacology 63, 185–190. Prieto, D., Simonsen, U., Hernández, M., García-Sacristán, A., 1998. Contribution of K+ channels and ouabain-sensitive mechanisms to the endothelium-dependent relaxations of horse penile small arteries. Br. J. Pharmacol. 123, 1609–1620. Savage, D., Perkins, J., Hong Lim, C., Bund, S.J., 2003. Functional evidence that K+ is the non-nitric oxide, non-prostanoid endothelium-derived relaxing factor in rat femoral arteries. Vasc. Pharmacol. 40, 23–28. Segarra, G., Medina, P., Vila, J.M., Martínez-León, J.B., Ballester, R.M., Lluch, P., Lluch, S., 2000. Contractile effects of arginine analogues on human internal thoracic and radial arteries. J. Thorac. Cardiovasc. Surg. 120, 729–736. Smith Jr., S.C., 2007. Multiple risk factors for cardiovascular disease and diabetes mellitus. Am. J. Med. 120, S3–S11. Stankevicius, E., Lopez-Valverde, V., Rivera, L., Hughes, A.D., Mulvany, M.J., Simonsen, U., 2006. Combination of Ca2+-activated K+ channel blockers inhibits acetylcholine-evoked nitric oxide release in rat superior mesenteric artery. Br. J. Pharmacol. 149, 560–572. Taddei, S., Ghiadoni, L., Virdis, A., Versari, D., Salvetti, A., 2003. Mechanisms of endothelial dysfunction: clinical significance and preventive non-pharmacological therapeutic strategies. Curr. Pharm. Des. 9, 2385–2402. Verma, S., Lovren, F., Dumont, A.S., Mather, K.J., Maitland, A., Kieser, T.M., Kidd, W., McNeill, J.H., Stewart, D.J., Triggle, C.R., Anderson, T.J., 2001. Endothelin receptor blockade improves endothelial function in human internal mammary arteries. Cardiovasc. Res. 49, 146–151. Wei, W., Chen, Z.W., Yang, Q., Jin, H., Furnary, A., Yao, X.Q., Yim, A.P., He, G.W., 2007. Vasorelaxation induced by vascular endothelial growth factor in the human internal mammary artery and radial artery. Vasc. Pharmacol. 46, 253–259. Zulli, A., Ye, B., Wookey, P.J., Buxton, B.F., Hare, D.L., 2008. Calcitonin gene-related peptide inhibits angiotensin II-mediated vasoconstriction in human radial arteries: role of the Kir channel. J. Thorac. Cardiovasc. Surg. 136, 370–375.