J Pharmacol Sci 118, 521 – 530 (2012) Journal of Pharmacological Sciences © The Japanese Pharmacological Society Full Paper Involvement of and K+ Channels in Relaxation Evoked by Ferulate in Rat Aorta Artery

Zhan-Qing Wang1,*, Jing-Feng Xu1, Jin-Ping Wang1, Wei-Juan Zhao1, and Ming Zeng1 1Department of Pharmacology, General Hospital of Beijing Military Command, Beijing 100700, China

Received September 29, 2011; Accepted February 21, 2012

Abstract. Vasorelaxant properties of N-2-(ferulamidoethyl)-nitrate (ferulate nitrate, FLNT), a newly synthesized nitrate, were compared with those of dinitrate, , nitro- glycerin, and 8-bromoguanosine 3,5-cyclic monophosphate (8-Br-cGMP) in rat aorta pre-contracted by phenylephrine. FLNT produced vasorelaxation in a concentration-dependent manner (0.1 – 100 μM). The degree of relaxation induced by FLNT was similar to that induced by . In addition, removal of endothelium did not affect the relaxant effect of FLNT. FLNT caused a rightward shift of the cumulative concentration–response curves of phenylephrine and reduced the maximal efficacy of contraction. 1H-[1,2,4]Oxadiazolo-[4,3-a]quinoxalin-1-one (ODQ, 10 μM) + and K -channel blockers charybdotoxin (CHT, 0.1 μM) and BaCl2 (1 μM) reduced the relaxant effect of FLNT in the endothelium-denuded arteries, whereas glibenclamide (1 μM) and 4-amino- pyridine (1 mM) failed to influence FLNT-induced vasorelaxation. Furthermore, in the presence

of ODQ, both CHT (0.1 μM) and BaCl2 (1 μM) still significantly reduced the relaxation evoked by FLNT. Pretreatment of vessels with hydroxocobalamin, a scavenger, abolished the FLNT effect. These findings demonstrate that FLNT induces relaxation of the rat aorta rings en- dothelium-independently. Furthermore, we demonstrated that FLNT-induced vasorelaxation is related to its stimulation of soluble guanylate cyclase and activation of K+ channels.

Keywords: organic nitrate, vasorelaxation, guanylate cyclase, K+ channel

Introduction (ALDH-2), and possibly esterases in vitro and in vivo (5, 6). Nitric oxide (NO) is an important regulator of vascular A few evidences indicate that vascular smooth muscle smooth muscle. It stimulates is responsible for denitration and release of NO from (sGC), which catalyzes the conversion of guanosine nitrate which in turn activates sGC (7). However, evi- 5-trisphosphate (GTP) to cyclic guanosine monophos- dences have accumulated to support the endothelium- phate (cGMP), and cGMP may induce dependent mechanism of nitrate-induced vasorelaxation through activation of myosin light chain phosphatase (1). (8, 9). Moreover, nitrate tolerance has been reported to NO also causes membrane hyperpolarization through be associated with effects of nitrate on the vascular + 2+ K -channel activation, decreasing Ca entry and leads to endothelium leading to NO synthase activation and L- vasodilatation (2, 3). NO is produced either through the depletion (8, 9). In addition, the tolerance was endogenous L-arginine–citrulline–NO pathway or from a also associated with sGC desensitization to NO because pharmacological NO donor. The organic are the of increased reactive oxygen species production and S- most commonly used NO donors and mimic endogenous of sGC (10 – 12). However, slight or no NO (3, 4). They are suggested to bio-transform into NO tolerance was developed against a nitrate, like nicorandil, by such as glutathione S-transferase, cyto- with the property of opening K+ channels (13, 14). chrome P-450, mitochondrial aldehyde dehydrogenase Ferulate nitrate (FLNT), N-2-(ferulamidoethyl)-nitrate is a newly synthesized ferulate derivative. Ferulate is an *Corresponding author. [email protected] antioxidant and protects vascular endothelium cells from Published online in J-STAGE being injured by hyperlipidemic serum. We have previ- doi: 10.1254/jphs.11179FP ously demonstrated that FLNT was an orally efficacious

521 522 Z-Q Wang et al agent (15). FLNT possesses a nitrate moiety g, Certification No. SCHK-JING-2000-0010), purchased in its chemical structure (Fig. 1). So we hypothesize that from Experimental Animal Breeding Center of Chinese FLNT will also possess potential vasorelaxant activity. Medical Institute (Beijing, China), were sedated via an- The aim of the present study was to determine whether esthetic overdose and killed by cervical dislocation and FLNT induces vasorelaxation in rat aorta artery and to exsanguination. Thoracic aortas were dissected free and address whether sGC and different K+ channels are re- surrounding connective tissues were carefully removed. sponsible for the vasorelaxation induced by FLNT. Four aortic rings (2 mm in length) were prepared from each rat and bathed in 10-ml organ baths containing Materials and Methods Krebs solution of the following composition: 116.3 mM

NaCl, 5.4 mM KCl, 1.8 mM CaCl2, 1.04 mM NaH2PO4, Chemicals 0.83 mM MgSO4, 19 mM NaHCO3, 5.5 mM glucose, pH FLNT (99.3% purity) was synthesized in our labo- 7.4. The bath solution was constantly gassed with a ratory; phenylephrine hydrochloride, acetylcholine mixture of 95% O2 and 5% CO2 and maintained at 37°C. chloride, 1H-[1,2,4]oxadiazolo-[4,3-a]quinoxalin-1-one Before the experiments were carried out, the aortic rings (ODQ), nicorandil, 8-bromoguanosine 3,5-cyclic mono- were allowed to equilibrate under a resting tension of 1.5 phosphate (8-Br-cGMP), hydroxocobalamin hydrochlo- g for 1 h, during which time Krebs solution was replaced ride, glibenclamide (GLM), charybdotoxin (CHT), every 15 min and muscle tensions were readjusted to 1.5

4-aminopiridine (4-AP), and barium chloride (BaCl2) g. Contractions were recorded isometrically by means of were purchased from Superior Chemical & Instrument four-channel JZ101 force transducers (Xin-Hang Instru- Co. (Beijing, China). All drugs were dissolved in Krebs- ment, Gao-beidian, China) and stored on computer using Henseleit solution except for FLNT, ODQ, nicorandil, Medlab-U-4CS software (Mei-Yi Instrument, Nan-jing, and GLM, which were dissolved in dimethyl sulfoxide. China). The endothelium-denuded rings were made by Dimethyl sulfoxide at a final concentration of 0.2% (v/v) rubbing the lumen of the artery with a steel wire. The did not affect phenylephrine-induced contraction. All removal of the endothelium was deemed successful if no drugs were added directly to the bath in a volume of 10 relaxation was induced in response to acetylcholine (10 mL and the concentrations given are the calculated final μM), and if the relaxation response was more than 80%, concentrations in the bath solution. the ring was deemed endothelium-intact, as described previously (16). Preparation of aortic rings All experiments reported here abided by the Guide for Phenylephrine contractions and FLNT relaxation the Care and Use of Laboratory Animals published by In the first set of experiments, the aortic rings were the US National Institutes of Health (NIH publication contracted with cumulative concentration of phenyleph- NO. 85-23, revised 1996). The experiments were con- rine (0.001 – 100 μM). The contraction concentration– ducted with the approval of the Ethics Review Committee response curves of phenylephrine were then conducted. for Animal Experimentation of the General Hospital of Based on the integrated concentration–response curve, Being Military Command. Male Wistar rats (250 – 300 the 80% maximally effective concentration of phenyleph- rine (1 μM) was selected as the concentration used for pre-contraction of the vessel rings in all the subsequent experiments. Cumulative relaxation responses to FLNT were then studied in the endothelium-intact and -denuded rings, with a 2-min exposure to each FLNT concentra- tion. The relaxation induced by FLNT was also compared to those induced by typical vasodilators: isosorbide dini- trate, nicorandil, , and 8-Br-cGMP in en- dothelium-intact rings. In the second group of experiments, endothelium-intact aorta arteries were pre-incubated for 5 min with FLNT (1, 10, and 100 μM). The effect of FLNT on the contrac- tion concentration–response curves of phenylephrine was studied. During the experiments, aortic rings were washed by changing the Krebs solution four times at 15-min intervals and then allowed to stabilize between Fig. 1. Chemical structure of FLNT. each FLNT concentration pre-incubation. Vasorelaxation Evoked by FLNT 523

+ Studies with K -channel blockers and the drug concentration exhibiting 50% of the Emax As described previously (6, 17 – 20), experiments (EC50) were calculated from the fitted concentration– were performed with CHT (0.1 μM), a selective Ca2+- response curves for each ring. Data were expressed as the + activated K -channel (KCa) blocker, as well as the inward mean ± S.E.M.; n refers to the number of rats from which + rectifier K -channel (KIR) blocker BaCl2 (1 μM). Experi- the vessel ring segments were taken. When the data were ments were also performed with the ATP-sensitive K+- distributed normally, statistical analysis was performed channel (KATP) blocker GLM (1 μM) and voltage-depen- using repeated-measures of analysis of variance, fol- + dent K -channel (KV) blocker 4-AP (1 mM). Tissues lowed by Dunnett’s test to analyze the difference between were pre-incubated with the K+-channel blockers 45 min the control group and experiment groups. The difference before contractions with 1 μM phenylephrine, and then was considered to be statistically significant when cumulative relaxations to FLNT and nicorandil were P < 0.05, determined using the statistical package SPSS studied. (Standard Version 13.0; SPSS Inc., Chicago, IL, USA).

Studies with sGC inhibitor Results To study the involvement of cGMP, the effect of ODQ (10 μM), a selective and well-established sGC inhibitor, Role of endothelium in FLNT-induced relaxation was studied. Tissues were exposed to 10 μM ODQ (45 In endothelium-intact and -denuded rings pre-contract- min) for pre-incubation. Tissues were then contracted ed by phenylephrine (1 μM), cumulative relaxation con- with 1 μM phenylephrine, and FLNT and nicorandil cu- centration–response curves for FLNT showed that FLNT mulative relaxations were studied. Control aortic rings (0.1 – 100 μM) produced concentration-dependent relax- from the same rat were used without ODQ pretreatment, ation compared with the control group (the mean value as described previously (4, 21, 22). of relaxation lower than 5%). There was no significant

To determine whether the effect of FLNT and nic- difference in the EC50 value between the endothelium- orandil on K+ channels is sGC/cGMP pathway–dependent intact group and endothelium-denuded group (5.0 ± 0.1 or –independent, we studied the effect of BaCl2 (1 μM), vs. 5.2 ± 0.2 μM) (Fig. 2). CHT or GLM and ODQ (10 μM) in combination on the rat aorta rings, as described previously (19). Comparison of the concentration–response curves for FLNT and the typical vasodilators Studies with NO scavenger FLNT vasorelaxations were compared with the typical Because FLNT, isosorbide dinitrate, and nicorandil are structurally similar nitrates and were thought to produce their vasorelaxation over the same concentration range (0.3 – 300 μM) via the generation of NO, a comparative study was carried out to study the effect of hydroxocoba- lamin, a NO scavenger, on the relaxations induced by FLNT, isosorbide dinitrate, and nicorandil. Aortic rings were exposed to 1 mM hydroxocobalamin for 45 min, as described previously (23 – 25). The tissues were then contracted with 1 μM phenylephrine, and FLNT, isosor- bide dinitrate, and nicorandil cumulative relaxations were studied. Control aortic rings from the same rat were used without hydroxocobalamin pretreatment.

Statistical analyses FLNT-induced vasorelaxation responses were ex- pressed as a percentage of relaxation relative to the level of pre-contraction (before FLNT administration). The developed vessel tension responses were expressed as a percentage of the baseline values (before phenylephrine application), as described previously (26). The concen- tration–response curves were generated with non-linear Fig. 2. Cumulative relaxation concentration–response curves for FLNT in aorta rings with endothelium-intact (closed squares) or en- curve fitting by means of a logistic equation. The maxi- dothelium-denuded (closed triangles) pre-contracted with phenyleph- mal efficacies of concentration–response curves (Emax) rine (1 μM). Data are presented as the mean ± S.E.M. (n = 10). 524 Z-Q Wang et al

nitrate drugs isosorbide dinitrate nitroglycerin, nicorandil, response curve with FLNT EC50 values significantly in- and 8-Br-cGMP. In endothelium-intact aorta arteries creasing from a control value of 5.5 ± 0.1 to 8.3 ± 0.5 pre-contracted by phenylephrine, FLNT, isosorbide dini- and 9.5 ± 0.3 μM, respectively (P < 0.01 for both). The trate, nicorandil, nitroglycerin, and 8-Br-cGMP induced Emax of FLNT-induced vasorelaxation decreased from the nearly full relaxations (> 80%) at concentrations of 100, 100, 100, 5, and 300 μM, respectively. FLNT and the typical vasodilators produced relaxation concentration- dependently. The EC50 values were 5.2 ± 0.2, 5.3 ± 0.1, 10.5 ± 0.8, 0.3 ± 0.02, and 30.6 ± 1.2 μM, respectively. The extent of relaxation induced by FLNT is similar to that by isosorbide dinitrate (5.2 ± 0.2 vs. 5.3 ± 0.1 μM, P > 0.05), while it is less than that by nitroglycerin (5.2 ± 0.2 vs. 0.3 ± 0.02 μM, P < 0.01) (Fig. 3).

Effect of FLNT on the contraction concentration– response curve of phenylephrine In endothelium-intact aorta arteries pre-contracted by phenylephrine, pre-incubation with different concentra- tions of FLNT (1, 10, and 100 μM) caused a rightward shift of the cumulative contraction concentration– response curves of phenylephrine. And the Emax of the concentration–response curves of phenylephrine contrac- tions without FLNT pre-incubation decreased from the control value of 260.2% ± 3.5% to 226.7% ± 2.8%, 200.2% ± 4.5%, and 160.5% ± 2.6%, respectively (P < Fig. 4. Effect of FLNT on the cumulative contraction concentration– 0.01 for all), but the different concentration of FLNT response curves of phenylephrine. Endothelium-intact rings were pre-incubation had no effect on EC50 values from the pre-incubated for 5 min with FLNT 1 μM (open triangles), 10 μM control EC50 value of 0.1 ± 0.02 to 0.1 ± 0.01, 0.1 ± 0.02, (closed triangles), and 100 μM (closed squares), respectively. After and 0.1 ± 0.01 μM, respectively (Fig. 4). incubation, cumulative concentrations of phenylephrine (0.001 – 100 μM) were added, and concentration–response curves were generated Effect of K+-channel blockers on the relaxation response for phenylephrine. The contraction responses to phenylephrine are expressed as percentages of baseline values (before phenylephrine to FLNT administration), which were set at 100% for the sake of clarity. Con- Both CHT (0.1 μM) and BaCl2 (1 μM) caused inhibi- trol group (open circles): rings without FLNT pre-incubation (n = 8). tions through the entire range of the FLNT concentration– Data are presented as the mean ± S.E.M. (n = 8).

Fig. 3. Cumulative relaxation concen- tration–response curves for FLNT (open circles), isosorbide dinitrate (closed trian- gles), and nitroglycerin (closed squares) (A) and FLNT (open circles), nicorandil (open triangles), and 8-Br-cGMP (open squares) (B) in endothelium-intact aorta rings. Data are presented as the mean ± S.E.M. (n = 8 – 10). Vasorelaxation Evoked by FLNT 525

control value of 98.1% ± 8.7% to 61.5% ± 5.5% and Effects of GC inhibitor ODQ, ODQ + BaCl2, and 23.4% ± 3.5% (P < 0.01, both), respectively (Fig. 5A). ODQ + CHT on the FLNT-induced relaxation CHT (0.1 μM) had no effect on resting tension or phen- Pre-incubation with ODQ (10 μM) caused inhibition ylephrine (1 μM) contraction, whereas BaCl2 caused through the entire range of FLNT relaxation concentra- approximately 30% to 40% increase in resting tension tion–response curves, with FLNT EC50 values signifi- but did not increase the size of the phenylephrine con- cantly increasing from a control value of 5.6 ± 0.1 to traction (data not shown). In contrast to the KCa and KIR 9.5 ± 0.2 μM (P < 0.01). The Emax of FLNT-induced blockers, the KATP blocker GLM 1 μM and KV blocker vasorelaxation decreased from the control value of 4-AP 1 mM did not have any significant effect on FLNT 95.6% ± 8.5% to 62.5% ± 4.2% (P < 0.01). ODQ had no relaxation (Fig. 5B). In addition, only GLM caused sig- effect on resting tension or phenylephrine contraction nificant inhibition through the entire range of the nic- (Fig. 6A). + orandil concentration–response curve, with EC50 values Pre-incubation of ODQ in combination with the K - significantly increasing from a control value of 10.8 ± 0.9 channel blocker CHT (0.1 μM) [(ODQ + CHT)] or BaCl2 to 14.4 ± 1.9 μM (P < 0.01) and Emax decreasing from (1 μM) [(ODQ + BaCl2)] also caused inhibition through 90.1% ± 6.7% to 66.3% ± 3.5% (P < 0.01) (Fig. 5D). the entire range of FLNT relaxation concentration–re-

sponse curves, with FLNT EC50 values significantly in-

Fig. 5. Effect of incubation with charyb- dotoxin (0.1 μM, open triangles), 4-amin- opyridine (1 mM, closed triangles), glib-

enclamide (1 μM, closed circles), or BaCl2 (1 μM, closed squares) on the cumulative relaxation concentration–response curves for FLNT (A, B) and nicorandil (C, D) in endothelium-denuded aortic rings pre- contracted with phenylephrine (1 μM). Rings were pre-incubated with K+-channel blockers for 45 min. After the pre-incuba- tion, tissues were contracted with phen- ylephrine (1 μM) in the presence of blockers. Control group (open circles): rings that were not pre-incubated with K+- channel blockers. Data are presented as the mean ± S.E.M. (n = 8 – 9). 526 Z-Q Wang et al

Fig. 6. Effect of ODQ (10 μM, closed triangles) and charybdotoxin (0.1 μM,

open triangles), BaCl2 (1 μM, open circles), or glibenclamide (1 μM, open squares) with ODQ (10 μM) in combination on FLNT (A, B) (n = 7 – 9) and nicorandil (C, D) relaxation concentration–response curves in rat endothelium-denuded aorta rings. Rings were pre-incubated for 45

min with ODQ and CHT, BaCl2, or GLM with ODQ in combination (ODQ + CHT,

ODQ + BaCl2, ODQ + GLM). After the pre-incubation, tissues were contracted with phenylephrine (1 μM) in the presence of ODQ, ODQ plus CHT, or ODQ plus

BaCl2. Control group (closed squares): rings that were not pre-incubated with ODQ or K+-channel blockers. Data are presented as the mean ± S.E.M. (n = 8).

creasing from a control value of 5.3 ± 0.2 to 11.2 ± 0.2 Effects of NO scavenger hydroxocobalamin on the (ODQ + CHT, P < 0.01) and 11.9 ± 0.2 μM (ODQ + FLNT-induced relaxation

BaCl2, P < 0.01). The Emax of FLNT-induced vasorelax- Pretreatment of vessels with 1 mM hydroxocobalamin ation decreased from the control value of 95.3% ± 8.2% caused significant inhibition of FLNT relaxation, with to 35.2% ± 2.8% (ODQ + CHT, P < 0.01) and 26.4% ± EC50 values significantly increasing from 5.4 ± 0.5 to 1.7% (ODQ + BaCl2, P < 0.01) (Fig. 6B). 6.1 ± 0.5 μM (P < 0.05). The Emax significantly decreased Similarly, Pre-incubation with ODQ (10 μM) also from the control value of 93.5% ± 7.8% to 16.2% ± 1.8% caused inhibition of nicorandil relaxation, with EC50 (P < 0.01). Similarly, hydroxocobalamin almost abol- values significantly increasing from 10.2 ± 0.9 to 18.5 ± ished the response to isosorbide dinitrate. However, the

1.2 μM (P < 0.01). The Emax decreased from the control nicorandil relaxation concentration–response curve was value of 88.6% ± 6.4% to 36.2% ± 3.4% (P < 0.01) (Fig. only slightly shifted to the right after hydroxocobalamin

6C). In addition, pre-incubation of ODQ in combination pretreatment, with EC50 increasing from 10.5 ± 0.7 to with the KATP blocker GLM 1 μM almost abolished the 11.1 ± 0.2 μM (P < 0.05) and the Emax significantly de- nicorandil relaxation (Fig. 6D). creasing from the control value of 88.3% ± 3.2% to 73.6% ± 6.8% (P < 0.01). Hydroxocobalamin (1 mM) Vasorelaxation Evoked by FLNT 527 had no effect on resting tension or phenylephrine (1 μM) that FLNT causes vasorelaxation concentration-depend- contraction (Fig. 7). ently in rings pre-contracted by phenylephrine (1 μM) and pre-incubation with different concentration of FLNT Discussion causes a rightward shift of the contraction concentration– response curve of phenylephrine. The findings indicate The present study shows that FLNT causes relaxation that FLNT-induced relaxation links to receptor-operated of artery pre-contracted by phenylephrine in a concentra- intracellular Ca2+ channels and a decrease in the Ca2+ tion-dependent manner (0.1 – 100 μM) and endothelium- sensitivity. independently. Our previous research on the pharma- The Ca2+ channel closure is partly attributable to K+- cokinetic analysis of FLNT in rats (3 mg/kg, p.o.) shows channel opening (3, 32) or sGC / cGMP / cGMP-depen- that FLNT is rapidly metabolized (T1/2 = 0.5 h) and the dent protein kinase (cGK-I) signal pathway stimulation + maximum concentration (Cmax) reached 74 μg/L. It indi- (4). So we firstly examine the effects of nicorandil (K - cates that the use of a physiological dose of FLNT may channel opener and sGC activator), isosorbide dinitrate cause relaxation of arteries. (sGC activator), and 8Br-cGMP (the cell permeable ana- The contraction of the vascular smooth muscle is pri- logue of cGMP) on pre-contracted aortic rings. Our data marily regulated by the intracellular Ca2+ signal. Contrac- show they all cause vasorelaxation concentration- tions elicited by phenylephrine involve α-adrenergic re- dependently in rat aortic rings pre-contracted by phen- ceptor–operated intracellular Ca2+ release, Ca2+ influx ylephrine, indicating that K+-channel opening or the through receptor-operated Ca2+ channels, and increase in sGC/cGMP/cGK-I signal pathway plays important roles the Ca2+ sensitivity (27, 28). Stimulation of in vasodilation. We therefore investigated the effect of α-adrenoceptors increases both the Ca2+ sensitivity of FLNT on the K+ channels and sGC. contractile elements and the cytosolic Ca2+ level (29). It Four distinct types of K+ channels have been identified is well established that cGMP plays the key role in me- in vascular smooth muscle: KV, KCa, KATP, and KIR (18, diating smooth muscle relaxation. Attenuating myofila- 33); they can be blocked selectively by a number of K+- ment calcium sensitivity plays a major role in the cGMP- channel inhibitors at a certain concentration. The KATP mediated relaxation (30, 31). The present study shows channel is selectively inhibited by micromolar concentra-

tions of GLM (20, 34); KCa channels can selectively be blocked by low concentrations of CHT (20, 35). The KV channels are preferentially blocked by millimolar con-

centrations of 4-AP (18) and the KIR channels, by low barium concentrations (31, 34).

The present study shows both BaCl2 (1 μM) and CHT (0.1 μM) pretreatment significantly reduce the relaxation

induced by FLNT. It indicates that KIR and KCa channels are involved in the regulation of relaxation induced by FLNT. Previous studies showed that NO donor or nitrate

activated KCa channels directly (36) or through increasing the activity of cGMP-dependent protein kinase (3, 37). Another study showed that a NO donor could activate

KIR by increasing the KIR current (25). Furthermore, the results in our study also show that neither GLM nor 4-AP pretreatment has significant ef- fect on FLNT-induced relaxation. It suggests that neither

KATP nor KV channels are involved in the pathway by which FLNT produces a relaxation of the rat aorta artery. Fig. 7. Effect of hydroxocobalamin on FLNT (open circles), isosor- These findings are in agreement with the previous report bide dinitrate (open triangles), and nicorandil (open squares) relax- that 4-AP and GLM did not antagonize the vasorelaxant ation–concentration response curves in endothelium-denuded rat aorta effects induced by NO donors in rat aorta artery (3), but rings. Rings were pre-incubated for 45 min with hydroxocobalamin. not in agreement with other reports that NO or NO donor After the pre-incubation, tissues were contracted with phenylephrine has been found to activate KATP channels or KV channels (1 μM) in the presence of hydroxocobalamin, and then cumulative in rabbit mesenteric arteries (38), human umbilical artery relaxations to FLNT (closed circles), isosorbide dinitrate (closed tri- angles), or nicorandil (closed squares) were studied. Data are pre- (39), or rat small mesenteric arteries (19); these discrep- + sented as the mean ± S.E.M. (n = 8 – 12). ancies may be due to the variations in the K channel 528 Z-Q Wang et al distribution or receptor sensitivity in smooth muscle cells the nicorandil relaxation in rat aorta artery. This observa- among the different species or tissues (40 – 42). In addi- tion is in agreement with a previous study showing that tion, comparative studies show that relaxation responses nicorandil relaxed rat aorta via activation of sGC inde- to nicorandil are inhibited significantly by the KATP pendent of NO (49). In fact, sGC activation independent blocker GLM, but not by the other K+-channel blocker. of NO is not unique to nicorandil; (BAY58- These findings are in agreement with previous reports 2667) and YC-1 have also been shown to be an activator (43, 44). or stimulator of sGC independent of NO, and they both The present findings that FLNT-induced relaxation is increase sGC activity independent of NO, causing vaso- reduced by a sGC inhibitor, ODQ, indicate that FLNT- relaxation (50, 51). In addition, nitroglycerin has also induced relaxation involves the sGC/cGMP/cGK-I signal been shown to activate the sGC/cGMP/cGK-I pathway pathway mechanism. These findings were supported by and produce vasodilation independently of NO in vasore- previous reports about vasorelaxation induced by NO or laxant concentrations (0.01 – 1 μM), while nitroglycerin NO donors (3, 45). Furthermore, the present study also in suprapharmacological concentrations (10 – 1000 μM) shows that the presence of ODQ, CHT, or BaCl2 (1 μM) and isosorbide dinitrate exert their activity via an NO- further inhibited relaxations to FLNT, indicating that dependent activation of vascular sGC (52).

FLNT activates KCa and KIR independently of the sGC/ In the experiments studying the effect of FLNT on cGMP pathway. A few previous investigations also sGC, K+ channels, and the effect of hydroxocobalamin showed that NO or NO donors relax vessels indepen- on NO production, endothelium-denuded aortic rings dently of the sGC/cGMP pathway (34, 46). However, were used to escape the effect of factors located in the there is limited research showing that a NO donor induces endothelium, which will possibly interfere with the re- rat aorta relaxation through activation of KCa by a sGC/ sults of our investigation. cGMP pathway-dependent pathway (3). Similarly, com- In conclusion, this study has provided the first evidence parative studies also show GLM partially blocked the that the new agent FLNT induces relaxation of the rat nicorandil-induced relaxation and fully blocked the nic- aorta rings endothelium-independently and that NO-, orandil-induced effects when given in combination with sGC-, BaCl2- and CHT-sensitive KIR and KCa channels ODQ, suggesting that the effects of GLM and ODQ on located in vascular smooth muscle are involved in the relaxation to nicorandil are cumulative. These findings relaxation induced by FLNT. This is expected to provide show that FLNT and nicorandil exhibit a dual mechanism evidence for the research and development of new nitro- of action acting partly as a nitrate and partly as a K+- vasodilators with less tolerance development and less channel opener. Our findings of two pathways for relax- unwanted side effects; however, this observation needs ation by nicorandil are consistent with two previous re- further tolerance tests and electrophysiological evalu- ports (43, 44), while not consistent with other observations ation. using piglet mesenteric arteries pre-contracted by nora- drenaline (47). The reasons of the variable contribution Acknowledgments of KATP channel opening and stimulation of sGC on nic- The authors wish to thank Dr. C.L. Long for his excellent technical orandil-induced relaxation depend on the differences in assistance. This work was supported by the Fund of National New the animal species, tissues, size of vessels, and agonist Drug Research, China (No. 96-901-05-243). studied (48). (organic nitrates and NO donors) relax blood vessels primarily via activation of the sGC/ References cGMP/cGK-I pathway. Although the precise mechanism of sGC activation by nitrovasodilators in the vascular 1 Moncada S, Palmer RM, Higgs EA. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev. 1991;43: wall is unknown, the mediatory role of NO has been 109–142. postulated. In the present study, we found that hydroxo- 2 Sampson LJ, Plane F, Garland CJ. Involvement of cyclic GMP cobalamin, a NO scavenger (23, 24), caused the nearly and potassium channels in relaxation evoked by the nitric oxide complete inhibition of FLNT relaxation. It suggests that donor, diethylamine NO, in the rat small isolated mesenteric the vasorelaxation effect of FLNT is mediated by NO. artery. Naunyn Schmiedebergs Arch Pharmacol. 2001;364: This finding is in agreement with previous reports show- 220–225. ing that the vasorelaxation induced by nitrate involves 3 Lunardi CN, Vercesi JA, Silva RS, Bendhack LM. Vasorelaxation NO production (25, 36). In contrast, we found that nic- induced by the new nitric oxide donor cis-[Ru(Cl)(bpy)(2)(NO)] (PF(6)) is due to activation of K(Ca) by a cGMP-dependent orandil still retained most of its ability to cause relaxation pathway. Vascul Pharmacol. 2007;47:139–144. of the rat aorta artery in the presence of hydroxocobala- 4 Tseng CM, Tabrizi-Fard JE, Fung HL. Differential sensitivity min, which suggests that NO is not a primary factor for among nitric oxide donors toward ODQ-mediated inhibition of Vasorelaxation Evoked by FLNT 529

vascular relaxation. J Pharmacol Exp Ther. 2000;292:737–742. hydrolase activities in parallel with glucose-stimulated insulin 5 Torfgard KE, Ahlner J. Mechanisms of action of nitrates. Cardio- secretion. J Endocrinol. 2006;190:681–693. vasc Drugs Ther. 1994;8:701–717. 23 Kruszyna H, Magyar JS, Rochelle LG, Russell MA, Smith RP, 6 Chen Z, Foster MW, Zhang J, Mao L, Rockman HA, Kawamoto Wilcox DE. Spectroscopic studies of nitric oxide (NO) interac- T. An essential role for mitochondrial aldehyde dehydrogenase in tions with cobalamins: reaction of NO with superoxocobalamin(III) nitroglycerin bioactivation. Proc Natl Acad Sci U S A. 2005; likely accounts for cobalamin reversal of the biological effects of 102:12159–12164. NO. J Pharmacol Exp Ther. 1998;285:665–671. 7 Murad F. Cyclic guanosine monophosphate as a mediator of va- 24 Gerth K, Ehring T, Braendle M, Schelling P. Nitric oxide scav- sodilation. J Clin Invest. 1986;78:1–5. enging by hydroxocobalamin may account for its hemodynamic 8 Abou-Mohamed G, Kaesemeyer WH, Caldwell RB, Caldwell profile. Clin Toxicol. 2006;44:29–36. RW. Role of L-arginine in the vascular actions and development 25 Schubert R, Krien U, Wulfsen I, Schiemann D, Lehmann G, Ulfig of tolerance to nitroglycerin. Br J Pharmacol. 2000;130: N, et al. Nitric oxide donor dilates rat small 211–218. arteries by activation of inward rectifier potassium channels. 9 Fayers KE, Cummings MH, Shaw KM, Laight DW. Nitrate toler- Hypertension. 2004;43:891–896. ance and the links with endothelial dysfunction and oxidative 26 Löhn M, Steioff K, Bleich M, Busch AE, Ivashchenko Y. Inhibi- stress. Br J Clin Pharmacol. 2003;56:620–628. tion of Rho-kinase stimulates nitric oxide-independent vasore- 10 Daiber A, Oelze M, Sulyok S, Coldewey M, Schulz E, Treiber N, laxation. Eur J Pharmacol. 2005;507:179–186. et al. Heterozygous deficiency of manganese superoxide dis- 27 Thorin E, Huang PL, Fishman MC, Bevan JA. Nitric oxide in- mutase in mice (Mn-SOD+/−): a novel approach to assess the hibits alpha2-adrenoceptor-mediated endothelium-dependent role of oxidative stress for the development of nitrate tolerance. vasodilation. Circ Res. 1998;82:1323–1329. Mol Pharmacol. 2005;68:579–588. 28 Hirano K. Current topics in the regulatory mechanism underlying 11 Sorajja P, Cable DG, Hamner CE, Schaff HV. Prolonged expo- the Ca2+ sensitization of the contractile apparatus in vascular sure of canine coronary arteries to a nitric oxide donor desensi- smooth muscle. J Pharmacol Sci. 2007;104:109–115. tizes soluble guanylate cyclase. J Surg Res. 2005;123:82–88. 29 Karaki H. Ca2+ localization and sensitivity in vascular smooth 12 Sayed N, Kim DD, Fioramonti X, Iwahashi T, Durán WN, Beuve muscle. Trends Pharmacol Sci. 1989;10:320–325. A. Nitroglycerin-induced S-nitrosylation and desensitization of 30 Carvajal JA, Germain AM, Huidobro-Toro JP, Weiner CP. Mo- soluble guanylyl cyclase contribute to nitrate tolerance. Circ Res. lecular mechanism of cGMP-mediated smooth muscle relaxation. 2008;103:606–614. J Cell Physiol. 2000;184:409–420. 13 Kool MJ, Spek JJ, Struyker Boudier HA, Hoeks AP, Reneman 31 Nauli SM, Ally A, Zhang L, Gerthoffer WT, Pearce WJ. Matura- RS, van Herwaarden RH, et al. Acute and subacute effects of tion attenuates the effects of cGMP on contraction, [Ca2+]i and nicorandil and isosorbide dinitrate on vessel wall properties of Ca2+ sensitivity in ovine basilar arteries. Gen Pharmacol. 2000; large arteries and hemodynamics in healthy volunteers. Cardio- 35:107–118. vasc Drugs Ther. 1995;9:331–337. 32 Nelson MT, Quayle JM. Physiological roles and properties of 14 O’Rourke ST. KATP channel activation mediates nicorandil-in- potassium channels in arterial smooth muscle. Am J Physiol. duced relaxation of nitrate-tolerant coronary arteries. J Cardio- 1995;268:C799–C822. vasc Pharmacol. 1996;27:831–837. 33 Ko EA, Han J, Jung ID, Park WS. Physiological roles of K+ 15 Xu JF, Wang ZQ, Wang JP. Ferilnic nirate produces delayed pre- channels in vascular smooth muscle cells. J Smooth Muscle Res. conditioning against myocardial ischemia and reperfusion injury 2008;44:65–81. in rats. Arch Pharm Res. 2010;33:881–887. 34 Hempelmann RG, Seebeck J, Ziegler A, Mehdorn HM. Effects 16 Nomura H, Yamawaki H, Mukohda M, Okada M, Hara Y. Mecha- of potassium channel inhibitors on the relaxation induced by the nisms underlying pioglitazone-mediated relaxation in isolated nitric oxide donor diethylamine nitric oxide in isolated human blood vessel. J Pharmacol Sci. 2008;108:258–265. cerebral arteries. J Neurosurg. 2000;93:1048–1054. 17 Khan SA, Higdon NR, Meisheri KD. Coronary vasorelaxation 35 Chen SJ, Wu CC, Yen MH. Exercise training activates large- by nitroglycerin: involvement of plasmalemmal calcium-activat- conductance calcium-activated K(+) channels and enhances nitric ed K+ channels and intracellular Ca++ stores. J Pharmacol Exp oxide production in rat mesenteric artery and thoracic aorta. J Ther. 1998;284:838–846. Biomed Sci. 2001;8:248–255. 18 Standen NB, Quayle JM. K+ channel modulation in arterial 36 Bolotina VM, Najibi S, Palacino JJ, Pagano PJ, Cohen RA. Nitric smooth muscle. Acta Physiol Scand. 1998;164:549–557. oxide directly activates calcium-dependent potassium channels 19 Irvine JC, Favaloro JL, Kemp-Harper BK. NO- Activates soluble in vascular smooth muscle. Nature. 1994;368:850–853. guanylate cyclase and Kv channels to vasodilate resistance arter- 37 Bychkov R, Gollasch M, Steinke T, Ried C, Luft FC, Haller H. ies. Hypertension. 2003;41:1301–1307. Calcium-activated potassium channels and nitrate-induced vaso- 20 Erdei N, Papp Z, Pollesello P, Edes I, Bagi Z. The dilation in human coronary arteries. J Pharmacol Exp Ther. metabolite OR-1896 elicits vasodilation by activating the K 1998;285:293–298. (ATP) and BK (Ca) channels in rat isolated arterioles. Br J Phar- 38 Murphy ME, Brayden JE. Nitric oxide hyperpolarizes rabbit macol. 2006;148:696–702. mesenteric arteries via ATP-sensitive potassium channels. J 21 Zhao YD, Brandish PE, DiValentin M, Schelvis JP, Babcock GT, Physiol. 1995;486:47–58. Marletta MA. Inhibition of soluble guanylate cyclase by ODQ. 39 Lovren F, Triggle CR. Nitric oxide and sodium nitroprusside-in- Biochemistry. 2000;39:10848–10854. duced relaxation of the human umbilical artery. Br J Pharmacol. 22 Mosén H, Salehi A, Henningsson R, Lundquist I. Nitric oxide 2000;131:521–536. inhibits, and carbon monoxide activates, islet acid alfa-glucoside 40 Nagao T, Ibayashi S, Sadoshima S, Fujii K, Fujii K, Ohya Y, 530 Z-Q Wang et al

et al. Distribution and physiological roles of ATP-sensitive K+ new ruthenium complex. Vascul Pharmacol. 2007;46:215–222. channels in the vertebrobasilar system of the rabbit. Circ Res. 47 Pérez-Vizcaíno F, Cogolludo AL, Villamor E, Tamargo J. Role of 1996;78:238–243. K+ channel opening and stimulation of cyclic GMP in the vaso- 41 Belevych AE, Beck R, Tammaro P, Poston L, Smirnov SV. De- relaxant effects of nicorandil in isolated piglet pulmonary and velopmental changes in the functional characteristics and expres- mesenteric arteries: relative efficacy and interactions between sion of voltage-gated K+ channel currents in rat aortic myocytes. both pathways. Br J Pharmacol. 1998;123:847–854. Cardiovasc Res. 2002;54:152–161. 48 Akai K, Wang Y, Sato K, Sekiguchi N, Sugimura A, Kumagai T, 42 Archer SL, Wu XC, Thébaud B, Nsair A, Bonnet S, Tyrrell B, et al. Vasodilatory effect of nicorandil on coronary arterial mi- et al. Preferential expression and function of voltage-gated, O2− crovessels: its dependency on vessel size and the involvement of sensitive K+ channels in resistance pulmonary arteries explains the ATP-sensitive potassium channels. J Cardiovasc Pharmacol. regional heterogeneity in hypoxic pulmonary vasoconstriction: 1995;26:541–547. ionic diversity in smooth muscle cell. Circ Res. 2004;95: 49 Minamiyama Y, Takemura S, Hai S, Suehiro S, Okada S, Funae 308–318. Y. Nicorandil elevates tissue cGMP levels in a nitric-oxide-inde- 43 Hsieh GC, Kolasa T, Sullivan JP, Brioni JD. Dual mechanism of pendent manner. J Pharmacol Sci. 2007;103:33–39. action of nicorandil on rabbit corpus cavernosal smooth muscle 50 Chester M, Tourneux P, Seedorf G, Grover TR, Gien J, Abman tone. Int J Impot Res. 2001;13:240–246. SH. Cinaciguat, a soluble guanylate cyclase activator, causes 44 Davie CS, Kubo M, Standen NB. Potassium channel activation potent and sustained pulmonary vasodilation in the ovine fetus. and relaxation by nicorandil in rat small mesenteric arteries. Br J Am J Physiol Lung Cell Mol Physiol. 2009;297:L318–L325. Pharmacol. 1998;125:1715–1725. 51 Berkan O, Bagcivan I, Kaya T, Yildirim K, Yildirim S, Doğan K. 45 Nimmegeers S, Sips P, Buys E, Brouckaert P, Van de Voorde J. Investigation of the vasorelaxant effects of 3-(5′-hydroxymethyl- Functional role of the soluble guanylyl cyclase alpha(1) subunit 2′-furyl)-1-benzyl (YC-1) and diethylamine/nitric oxide in vascular smooth muscle relaxation. Cardiovasc Res. 2007; (DEA/NO) on the human radial artery used as coronary bypass 76:149–159. graft. Can J Physiol Pharmacol. 2007;85:521–526. 46 Bonaventura D, de Lima RG, Vercesi JA, da Silva RS, Bendhack 52 Kleschyov AL, Oelze M, Daiber A, Huang Y, Mollnau H, Schulz LM. Comparison of the mechanisms underlying the relaxation E, et al. Does nitric oxide mediate the vasodilator activity of ni- induced by two nitric oxide donors: sodium nitroprusside and a troglycerin? Circ Res. 2003;93:e104–e112.