The mechanism of gentisic acid-induced relaxation of the guinea pig isolated trachea : the role of potassium channels and vasoactive intestinal peptide receptors

We examined some of the mechanisms by which the aspirin metabolite and the naturally occurring metabolite gentisic acid induced relaxation of the guinea pig trachea in vitro. In preparations with or without epithelium and contracted by histamine, gentisic acid caused concentration-dependent and reproducible relaxation, with mean EC50 values of 18 μM and Emax of 100% (N = 10) or 20 μM and Emax of 92% (N = 10), respectively. The relaxation caused by gentisic acid was of slow onset in comparison to that caused by norepinephrine, theophylline or vasoactive intestinal peptide (VIP). The relative rank order of potency was: salbutamol 7.9 > VIP 7.0 > gentisic acid 4.7 > theophylline 3.7. Gentisic acid-induced relaxation was markedly reduced (24 ± 7.0, 43 ± 3.9 and 78 ± 5.6%) in preparations with elevated potassium concentration in the medium (20, 40 or 80 mM, respectively). Tetraethylammonium (100 μM), a nonselective blocker of the potassium channels, partially inhibited the relaxation response to gentisic acid, while 4-AP (10 μM), a blocker of the voltage potassium channel, inhibited gentisic acid-induced relaxation by 41 ± 12%. Glibenclamide (1 or 3 μM), at a concentration which markedly inhibited the relaxation induced by the opener of ATP-sensitive K+ channels, levcromakalim, had no effect on the relaxation induced by gentisic acid. Charybdotoxin (0.1 or 0.3 μM), a selective blocker of the largeconductance Ca2+-activated K+ channels, caused rightward shifts (6and 7-fold) of the gentisic acid concentration-relaxation curve. L-NGnitroarginine (100 μM), a NO synthase inhibitor, had no effect on the relaxant effect of gentisic acid, and caused a slight displacement to the right in the relaxant effect of the gentisic acid curve at 300 μM, while methylene blue (10 or 30 μM) or ODQ (1 μM), the inhibitors of soluble guanylate cyclase, all failed to affect gentisic acid-induced relaxation. D-P-Cl-Phe,Leu[VIP] (0.1 μM), a VIP receptor antagonist, significantly inhibited (37 ± 7%) relaxation induced by gentisic acid, whereas CGRP (8-37) (0.1 μM), a CGRP antagonist, only slightly enhanced the action of gentisic acid. Taken together, these results provide functional evidence for the direct activation of voltage and large-conductance Ca2+-activated K+ channels, or indirect modulation of potassium channels induced by VIP receptors and accounts for the predominant relaxation response caused by gentisic acid in the guinea pig trachea. Correspondence

The mechanism of gentisic acid-induced relaxation of the guinea pig isolated trachea: the role of potassium channels and vasoactive intestinal peptide receptors Introduction 2,5-Dihydroxybenzoic acid or gentisic acid is an active metabolite of salicylic acid degradation (1).There is an increasing amount of evidence indicating that gentisic acid has a broad spectrum of biological activity such as anti-inflammatory, antirheumatic and antioxidant properties, independent of the action of salicylic acid (2,3).Thus, gentisic acid has been demonstrated to inhibit both the cyclooxygenase and 12-lipoxygenase enzymes (4) and has also been found to decrease the production of the leukotrienes C4 on the stomach wall during the gastric lesion induced by ethanol (5).
Several potentially active compounds including some naturally occurring substances capable of directly activating potassium channels have been recently reported (6,7).The new classes of potassium channel openers might be of potential clinical interest because they are able to induce vascular (8) and nonvascular (9) smooth muscle relaxation, including in the respiratory tract (10,11).
Despite the considerable number of reported studies for gentisic acid, so far its effect on nonvascular smooth muscle has not been investigated.The aims of the present study were to examine whether gentisic acid causes relaxation in the guinea pig trachea in vitro and if so, to characterize some of the pharmacological mechanisms responsible for such effects using selective receptor antagonists or ion channel blockers.We present evidence that the relaxation caused by gentisic acid in the guinea pig trachea is largely mediated by activation of potassium channels, especially the highconductance calcium-activated potassium channels, and by vasoactive intestinal peptide (VIP)-receptor interaction.

Tissue preparations
Guinea pigs (250-400 g) of both sexes were killed by cervical dislocation.The trachea was rapidly removed, freed from connective tissue, and cut into six transverse rings (3-4 mm wide), each containing 3 cartilages as described previously (12,13).The rings were opened (usually 6 strips of 8-10 mm in length were obtained from the same animal) and suspended in individual 10-ml jacketed organ baths containing Krebs-Henseleit solution maintained at 37 o C, pH 7.2, gassed with a mixture of 95% O 2 and 5% CO 2 .The Krebs solution was of the following composition: 118.0 mM NaCl, 4.4 mM KCl, 1.1 mM MgSO 4 , 2.5 mM CaCl 2 , 25.0 mM NaHCO 3 , 1.2 mM KH 2 PO 4 , and 11.0 mM glucose.Tissues were allowed to equilibrate for at least 60 min before drug additions under a resting tension of 1 g, during which the buffer solution was renewed every 15 min.Isometric responses were recorded on a polygraph (Letica Scientific Instruments, Barcelona, Spain) by means of TRI-201 force displacement transducers.In some experiments, the epithelial layer of the trachea was gently removed with a cotton-tipped applicator.The integrity of the epithelium was assessed by the ability of bradykinin to induce relaxation (12).

Experimental procedure
After an equilibration period of at least 60 min, the preparations with or without epithelium were precontracted with histamine (1 to 3 µM), and once the tonic contraction became stable, gentisic acid (0.3 µM to 1 mM) was added to the bath by the cumulative method (14).Two to five complete cumulative concentration-response curves were obtained for gentisic acid in each preparation, at 60-min intervals between curves.

Drugs
The following drugs were supplied by Sigma Chemical Co., St. Louis, MO, USA: gentisic acid (standard), histamine, TEA, 4-AP, glibenclamide, charybdotoxin, L-NOARG, methylene blue, ODQ, D-P -Cl-Phe 6 ,Leu 17 [VIP], porcine VIP, CGRP .The other chemical products and salts were supplied by Merck (Darmstadt, Germany) and were of high grade analytic purity.The stock solutions of these drugs were prepared and stored at -20 o C. The bath concentration of ethanol did not exceed 0.03%, which was shown to have no effect per se on the basal tonus of the preparations or on the agonistmediated contraction or relaxation.

Statistical analysis
Responses are reported as percentage of the maximal effect of control relaxation caused by gentisic acid.Data were analyzed statistically by the unpaired Student t-test, with the level of significance set at P<0.05.The EC 50 (i.e., the molar concentration of gentisic acid and of the reference drugs causing half-maximum responses) values were determined from individual experiments for the complete agonist concentration-response curves by graphic methods.In some experiments, the pD 2 values (i.e., log of the EC 50 ) were calculated.The EC 50 and pD 2 values are reported as geometric means accompanied by their respective 95% confidence limits.All other results are reported as means ± SEM.
The relaxation caused by gentisic acid was of slow onset when compared to that caused by VIP, norepinephrine or theophylline.In epithelium-denuded and histamineprecontracted preparations the relative rank order of relaxant potency (pD 2 ) was: salbu- In order to characterize the role of potassium channels in the relaxation caused by gentisic acid, various potassium channelblocking agents were tested.TEA (100 µM) caused a partial displacement to the right of the relaxation-concentration cumulative curve induced by gentisic acid (Figure 2B), while 4-AP (10 µM) inhibited gentisic acidinduced relaxation by 41 ± 12% (Figure 2C).Glibenclamide (1 or 3 µM), at a concentration which markedly inhibited the relaxation induced by the opener of ATP-sensitive K + channels, levcromakalim (10 nM-10 µM) (data not shown), did not modify the relaxation induced by gentisic acid (Figure 3A), but preincubation of the preparations with charybdotoxin (0.1 or 0.3 µM) caused a 6and 7-fold shift to the right of the concentration-relaxation curve for gentisic acid, respectively (Figure 3B).
The nitric oxide and cGMP pathways were also investigated.L-NOARG (100 µM), methylene blue (10 or 30 µM) and ODQ (1 µM) all failed to affect gentisic acid-induced relaxation, but an increase in L-NOARG concentration to 300 µM produced a slight rightward displacement of the relaxation curve for gentisic acid (Figure 4A, B and C, respectively).D-P -Cl-Phe 6 ,Leu 17 [VIP] (0.1 µM), a VIP receptor antagonist, also significantly inhibited (37 ± 7%) gentisic acidinduced relaxation (Figure 5A), while CGRP When the experiments were performed in Krebs-Henseleit solution containing 20 mM of KCl, the relaxation caused by gentisic acid was slightly reduced (24 ± 7%).In preparations contracted with 40 mM KCl, the relaxation caused by gentisic acid was reduced by 43 ± 4%, while in preparations contracted with 80 mM KCl, a concentration that is presumed to induce maximal depolarization of the cell membrane, the response to gentisic acid was markedly attenuated (78 ± 5.6%), although a weak relaxation response was still evident between 30 µM and 1 mM (Figure 2A).In addition, gentisic acid-induced relaxation was markedly antagonized when a higher K + concentration was present in the medium, further confirming the requirement of the integrity of the cell membrane potential for its action.Also relevant are the results showing that the relaxation caused by gentisic acid in the guinea pig trachea was largely antagonized by charybdotoxin, a neurotoxin known to antagonize selectively high-conductance calcium-activated potassium channels (18).However, glibenclamide, a selective blocker of ATP-sensitive K + channels, at concentrations at which it completely antagonized the relaxation caused by the opener of ATPsensitive potassium channels levcromakalim in the guinea pig trachea, failed to affect gentisic acid-mediated relaxation in this preparation, suggesting that ATP-sensitive potassium channels have no major effect on gentisic acid relaxation.
There is an increasing amount of experimental evidence suggesting that nitric oxide or nitric oxide-related substances present in the airways, including the guinea pig trachea (20,21), are able to modulate the relaxation response induced by several drugs and neurotransmitters.Our results, however, cannot fully rule out an action of nitric oxide on the relaxation produced by gentisic acid in the (8-37) (0.1 µM), a CGRP antagonist, only slightly enhanced the gentisic acid-induced relaxation (Figure 5B).

Discussion
Potassium channels are important target proteins which are involved in the relaxation of numerous smooth muscle cells, including airway smooth muscle, where the large-conductance calcium-activated potassium channel (BK Ca ) has an important regulatory role in tissue responsiveness (15,16).Several types of potassium channels have been reported to produce smooth muscle relaxation.Among them are the delayed rectifier channels (17), the large-and low-conductance calcium-activated channels (15) and also the ATP-sensitive potassium channels (16).Potassium channels also are an important effector element for several relaxant agents in the guinea pig isolated trachea, like nitric oxide (18) and VIP (19).The results of the present study provide functional evidence indicating that gentisic acid, an active metabolite of the salicylic route of degradation (1), causes concentration-dependent relaxation of the guinea pig trachea in both the presence and absence of epithelium.
The gentisic acid relaxation response was largely mediated by the opening of K + chan- guinea pig trachea, since in the presence of high L-NOARG concentrations a slight displacement of the relaxant concentration response to gentisic acid was observed.The observation that the relaxation caused by gentisic acid was not affected by the absence of the epithelium in the preparations or by the guanylate cyclase inhibitors further confirms our hypothesis that nitric oxide seems to play only a minor role in the gentisic acid relaxation of the guinea pig trachea.A possible antioxidant property of gentisic acid cannot be fully excluded, since it has been reported that gentisic acid inhibits the reaction of nitric oxide with superoxide anions generated in the medium, thus enhancing the half-time of nitric oxide (22).Our results also show that the cGMP pathway seems unlikely to be involved in the relaxant effect of gentisic acid on the guinea pig trachea, since the inhibitors of soluble guanylate cyclase, methylene blue and ODQ ( 23), at concentrations known to inhibit the responses mediated by the cGMP-dependent mechanism, fully failed to affect gentisic acid-mediated relaxation.
The possible role of VIP and CGRP receptors in the gentisic acid-induced relaxation of the guinea pig trachea was also assessed by the use of selective antagonists of these receptors.Preincubation of the preparations with porcine D-P -Cl-Phe 6 ,Leu 17 [VIP] consistently antagonized gentisic acid-mediated relaxation, suggesting some modulatory role of VIP in the gentisic acid response.There is increasing evidence suggesting that the relaxation induced by VIP in the airway smooth muscles is largely mediated by activation of potassium channels including BK Ca , while ATP-sensitive potassium channels apparently play a minor role in the VIP response (19).On the other hand, human CGRP (8-37), another neuropeptide involved in the modulation of the responsiveness of airway smooth muscles (24), only slightly enhanced the action of gentisic acid, indicating that CGRP has apparently no major effect on gentisic acid-mediated relaxation.The precise mechanisms by which VIP mediated the relaxant response of gentisic acid in the guinea pig trachea are still unknown and additional studies are required to clarify this issue.
In summary, the present results have provided direct pharmacological evidence indicating that gentisic acid, an active metabolite of the salicylic acid route of degradation, produced concentration-dependent and reproducible relaxation in the guinea pig isolated trachea precontracted with histamine both in the presence and in the absence of epithelium.Its precise mechanism of action still remains partially undefined, but several mechanisms seem to account for its relaxant effect, such as its ability to activate largeconductance Ca 2+ -activated K + channels sensitive to charybdotoxin and also voltagesensitive K + channels highly sensitive to 4-AP.Finally, it has also been demonstrated that the VIP receptors strongly contribute to the relaxation response of gentisic acid in the guinea pig trachea.

Figure 1 -
Figure 1 -Mean relaxant concentration-response curves for gentisic acid in the guinea pig isolated trachea, without (filled circles) or with (open circles) epithelium.Data are reported as means ± SEM of 10 experiments.P<0.05 for the difference between points (Student t-test).