Effects of nitro-butoxyl- and butyl-esters of non-steroidal anti-inflammatory drugs compared with parent compounds on the contractility of digital arterial smooth muscle from the fallow deer (Dama dama)

Background Non-steroidal anti-inflammatory drugs (NSAIDs) are a major cause of upper gastro-intestinal (GI) ulceration and bleeding as well as cardiovascular (CV) diseases (e.g., myocardial infarction and stroke). A feature common to both these adverse events is a variety of vascular reactions. One approach to overcome these side effects has been the development of nitric-oxide (NO)-donating NSAIDs. The NO is considered to overcome some of these vascular reactions caused by NSAIDs. Unfortunately, the NO-NSAIDs developed so far have not had the expected benefits compared with NSAIDs alone. Objectives Using in vitro preparations it is hoped to gain insight into the vascular and smooth muscle reactions induced by NO-NSAIDs compared with NSAIDs as a basis for improving the protective responses attributed to the NO-donating properties of these drugs. Methods A range of NO-NSAIDs was synthesized based on the esterification of NSAIDs with the nitro-butoxylate as a prototype of an NO-donor. These compounds, as well as NO-donor agents and NSAIDS, were examined for their possible effects on isolated segments of digital arteries of fallow deer, which provide a robust model for determining the effects of vasodilator and vasoconstrictor activities, in comparison with those of standard pharmacological agents. Results The NO-NSAIDs were found to antagonise the smooth muscle contractions produced by 5-hydroxytryptamine (serotonin, 5-HT). However, while almost all their parent NSAIDs had little or no effect, with the exception of the R-(−)-isomers of both ibuprofen and flurbiprofen, which caused vasodilatation, all the NO-NSAIDs tested antagonised the increase in tension produced by 5-HT. Conclusions R-(−)-ibuprofen and R-(−)-flurbiprofen, along with the nitro-butoxyl esters of the NSAIDs examined, produce relaxation of segments of deer digital artery smooth muscle in vitro. The evidence presented suggests that their mechanism involves the release of NO or its products.


Introduction
Non-steroidal anti-inflammatory drugs (NSAIDs) are amongst the most widely used drugs for prescription and non-prescription ('over-the-counter' or OTC) medications for the treatment of musculo-skeletal and various acute and chronic painful and inflammatory conditions (Rainsford 2007). Their use is associated with the development of serious adverse drug reactions (ADRs) especially in the gastro-intestinal (GI) tract of elderly patients with compromised health status (Rainsford et al. 2008;Lanas 2010;Lanas et al. 2010;Rahme and Bernatsky 2010) or those 1 3 with compromised cytochrome CYP2C9 metabolism (Carbonell et al. 2010); Süleyman et al. 2007). Over recent decades there has been increasing concern about the risks of NSAIDs, especially the cyclo-oxygenase (COX)-selective agents or coxibs, being associated with cardiovascular (CV) and cerebrovascular reactions including increased risk of myocardial infarction (Antman et al. 2007;McGettigan and Henry 2011;Olsen et al. 2012;Shau et al. 2012;Sudano et al. 2012;Caughey et al. 2011) and stroke (Barthélémy et al. 2011;Caughey et al. 2011;Varas-Lorenzo et al. 2011). These reactions are primarily related to hypertension that is exacerbated by NSAIDs (Barthélémy et al. 2011;Varas-Lorenzo et al. 2011) as well as T-cell associated plaqueinstability in atherosclerosis (Padol and Hunt 2010;Rainsford 2010). The atherogenic promoting effects of NSAIDs may also be related to their propensity to divert arachidonic acid through the 5-lipoxygenase pathway (Yu et al. 2012).
Current concerns regarding the safe use of NSAIDs have centred on the combined GI and CV risks of these drugs Scheiman and Hindley 2010;Salvo et al. 2011). A general feature that is common to both these adverse reactions is the effects of the NSAIDs on vascular reactions. Thus, in addition to the abovementioned vascular effects in CV disease, NSAIDs also cause microvascular injury in the early stages of the development of gastric mucosal damage (Rainsford 1983(Rainsford , 1992(Rainsford , 1993a1999;Gyömber et al. 1996a, b;Pasa et al. 2009;Tarnawski et al. 2012). The NSAID-induced impairment of platelet aggregation contributes to the extravasation of blood from the damaged microvasculature into the interstitial space, ischaemia and subsequent bleeding that accompanies the pathological injury to the gastric mucosa (Rainsford 1986(Rainsford , 1992Gyömber et al. 1996a, b;Tarnawski et al. 2012). The initiation of vascular constriction by NSAIDs is considered to be related to excess production of vasoconstrictor peptido-leukotrienes which occurs from the diversion of arachidonic acid through the 5-lipoxygenase pathway as a result of NSAIDs inhibiting the cyclo-oxygenases (Rainsford 1986(Rainsford , 1993a(Rainsford , b, 1999Rainsford et al. 1995;Gyömber et al. 1996a, b). This is accompanied by accumulation, endothelial interactions and activation of polymorphonuclear (neutrophil) and other leucocytes that contribute to mucosal damage (Rainsford et al. 1995;2012;McCafferty et al. 1995;Appleyard et al. 1996;Wallace 1997;Wallace et al. 1999;Muscará, et al. 2000). Aside from arachidonate metabolites (prostanoids, leukotrienes, lipoxins), nitric oxide (NO) is known to have a central role in the control of vascular smooth muscle contraction, blood flow and plateletendothelial interactions (Brzozowski et al. 2008;Palileo and Kaunitz 2011;Tarnawski et al. 2012). However, it has also been shown that NO has actions that could be seen to be protective in the GI tract, by reducing vascular injury, enhancing production of protective mucus, reducing the effects of acid-pepsin and promotion of anti-thrombotic effects (Wallace et al. 1993(Wallace et al. , 1999Fiorucci and Distrutti 2011).
To this end, a range of NSAIDs coupled to an NOreleasing moiety (NO-NSAIDs) has been developed, in the hope that such compounds, by releasing NO in the mucosa, would be less damaging to the GI tract. Some of these NO-NSAIDs have been shown experimentally to cause less gastric injury than the parent NSAIDs (Wallace et al. , 1999Fiorucci and Distrutti 2011;Gund et al. 2014). The actions of these drugs in preventing GI injury are considered to result from the hydrolysis of an NO-ester link. Despite an immense amount of research, the outcomes from the longterm studies with candidate NO-NSAIDs (e.g. NO-naproxen or naproxcinod) have been disappointing (Milton et al. 1999;Lowry 2010), although NO-aspirin may have potential as an anti-thrombotic agent (Wallace et al. 1999;Callingham et al. 2012). In the present study, the actions of some NO-NSAIDs were compared with established NSAIDs and NO-donating analogues on the contractility of isolated segments common digital arteries of fallow deer (Dama dama; Callingham et al. 2012).
The propionic acids, ibuprofen and flurbiprofen, are referred to as their racemic mixtures (rac). The R-(−)-and S-( +)-isomers of these drugs were gifts from Boots Healthcare International, Nottingham, UK.

Chemistry
The NO-NSAIDs (3a-i), were synthesized by a modification of the literature method Wallace et al. 1995) that is shown in Figs. 1 and 2.

General methods
Melting points are uncorrected and were determined on Stuart Scientific SMP3 apparatus. Infrared spectra were recorded with an ATI Mattson Genesis series FTIR spectrophotometer. 1 H NMR and 13 C NMR spectra were recorded in CDCl 3 using a Brucker AC 250 spectrometer operating at 250 and 62.9 MHz, respectively. Chemical shifts (δ) are recorded in ppm downfield from Me 4 Si as internal standard and J values are given in Hz. Mass spectra were recorded with EI-VG 7070E mass spectrometer. Accurate masses were determined on VG Autospec EI mass spectrometer with magnetic sector instrument. Optical rotations were measured at 23 °C with a Bellingham and Stanley ADP 440 polarimeter using dichloromethane as the solvent. All solvents were dried and distilled by standard techniques.

Deer common digital artery contractility studies
The methods employed were those previously described (Callingham et al. 2012 (Callingham et al. 2012). On arrival at the laboratory, the vessels were dissected free of extraneous tissues and stored, until required, in fresh aerated PSS at 4 °C. With changes of PSS daily, the vessels remained viable for up to 10 days.
Segments (approximately 3 mm in length), were mounted, in 10 ml, water-jacketed organ baths at 37 °C and attached to Harvard isometric transducers (0-50 g sensitivity), connected, via Harvard amplifiers and A/D converters (Pow-erLab® 8/35, ADInstruments Ltd, Bishops Mews, Transport Way, Oxford, OX4 6HD, UK) for computer recording of developed tension. Resting tension was adjusted to 3 g, which was maintained during a 45 min period of acclimatisation and beyond. The integrity of the vascular endothelium was tested by measuring the relaxation produced by addition of 10 -6 M histamine to segments pre-contracted with either 10 -6 M 5-HT or 10 -6 M PHE; since acetylcholine, the agent normally employed to detect functional endothelium is without effect in this preparation. In each experiment, the vessel rings were contracted, either with single concentrations or graded concentrations of (5-HT).
Cumulative changes in tension to applied agents were plotted as percentages of maximum responses against log concentrations of the relevant agent and fitted to the Hill equation by use of the non-linear regression facility in Kaleidagraph® (Synergy Software, 2457 Perkiomen Ave., Reading PA, USA 19,606) with n-values referring to the number of animals used. Tests for statistical significance were performed using the unpaired t-test.
In Figs. 3, 4, 5, 6, 7 inclusive, parameters for 5-HT (EC 50 ± s.e.m. and maximum tension ± s.e.m.) were 1 3 obtained from the mean regressions. In Figs. 8,9,10,11 inclusive, while the regressions were derived as above, tests for statistical significance were applied to individual mean data points and identified by asterisks as appropriate.

Drugs and reagents
Stock solutions of the NSAIDs were made by first dissolving a few milligrams of the compound in 0.25 ml of DMSO (dimethyl sulphoxide) and made up to 10 -2 M with an appropriate volume of deionised water. These solutions, together with any dilutions, were kept on ice until used.
Stock solutions of 10 −2 M 5-HT), phenylephrine (PHE) and histamine were prepared and kept at 4 °C and diluted with deionised water for use on the day of the experiment and kept on ice. Solutions of methylene blue for use as an inhibitor of nitric oxide synthase (Mayer et al. 1993) were made up on the day they were required.

Experimental protocol
Rings of 2-3 mm length were cut from the digital arteries using scissors and mounted in the organ bath by sliding the two hooks into the lumen of the artery. Each water bath was filled with PSS (buffered salt solution) and continuously aerated with 95% O 2 5% CO 2 . The jackets surrounding the water baths had water heated to 37 °C continuously pumped through them to maintain physiological temperature in the water baths. The day's stock solution flask of aerated PSS was also kept submerged in the water bath so that it was at the correct temperature when it was added to the organ baths. The tension pulled by the rings was adjusted to 3 g before each experiment was begun.
On the morning of each day of experiments, the artery segments were pre-contracted with 10 −5 M 5-HT as this concentration was sufficient to achieve the maximum contractile response; previous studies had shown to induce the rings to respond well to subsequent drug additions. When the vessels had reached maximum contraction, 10 −6 M histamine was added to the organ baths to test for the presence of a functional endothelium. The organ baths were then washed out and filled with fresh PSS. The rings were left to relax for an hour, with the tension returned to 3 g at intervals and the experiment proper was begun.

Data recording
The transducers were calibrated by use of the PowerLab® calibration facility and tested for linearity of response by attaching weights from 1 to 20 g. All data were processed by use of LabChart® (ADInstruments) on the recording computer.
The cumulative changes in tension to applied agents were plotted as percentages of maximum responses against log concentrations of the relevant agent and fitted to the Hill equation by non-linear regression, with n-values referring to the number of animals used. Only rings from left feet were used after ensuring, having previously that there were no differences in responses between rings taken from either foot, to ensure that the n-values truly represented individual animals. Tests for statistical significance were performed using the unpaired t-test.
When the maximum tension that could be developed by the segments, in response to applied 5-HT, was examined, the relaxation in tension produced by NO-aspirin alone was reduced from 50 to 30% in the presence of methylene blue (Fig. 4). Of the other NSAIDs and their nitroxy-derivatives, examined, indomethacin and naproxen, produced similar results (Figs. 5, 6, 7).
However, when the effects produced by racemic (rac)ibuprofen and nitroxybutyl-ibuprofen (NO-ibuprofen) were compared, on 5-HT pre-contracted arterial segments, both were effective at reducing the responses to electrical stimulation, with no significant difference (p > 0.05) in effect between them (Fig. 7). Another phenyl-propioinic acid, racflurbiprofen and its nitroxybutyl derivative (NO-flurbiprofen) produced similar results (data not shown). It was also found that rac-ibuprofen produced a reversible relaxation of vessel segments, when they had been pre-contracted with 3 × 10 -6 M phenylephrine (PHE), to a maximum tension of 16.5 ± 15% of control, with an EC50 of 2.97 × 10 -4 ± 10 -5 M (n = 7: p < 0.01). In view of this unexpected relaxation produced by ibuprofen and flurbiprofen, further experiments were done to attempt to discover their mode of action.
Removal of the vascular endothelium (a source of NO) reduced (p < 0.001) but did not eliminate the vasodilator actions of R-(−)-ibuprofen (Fig. 10.), suggesting a role for NO in the relaxation produced.

Discussion
These results demonstrate that the NO-donating analogues of aspirin, indomethacin, etc., significantly reduced the contractile responses of vascular smooth muscle to electrical stimulation and to applied 5-HT and PHE (results not shown), while, with the exception of ibuprofen and flurbiprofen, the parent NSAIDs were without effect. It was also shown that methylene blue (an inhibitor of NO action) significantly reduced the effect of NO-aspirin (Fig. 4), as well as other NO-NSAIDs (data not shown). In addition, the presence of haemoglobin had the same effect on NO-aspirin. This suggests that, in the presence of blood, in particular, the actions of NO-NSAIDs could be limited (data not shown).
The fact that R-(−)-ibuprofen produced a relaxation of a similar magnitude to racemic NO-ibuprofen suggests that either R-(−)-ibuprofen released NO on a similar scale to NO-ibuprofen, or that it caused relaxation by some other means. There are several other means possible, including the induction of iNOS or direct activation of soluble guanylate cyclase. Some previous work has been done on the possible involvement of ibuprofen with iNOS. One study suggests that the concentration of NO in cells can be raised by the presence of ibuprofen, through the induction of iNOS (Menzel and Kolarz 1997). This showed that, at therapeutically attainable concentrations (1-30 μM), iNOS was induced similarly by both stereoisomers of ibuprofen, although only slightly more by the R-(-)-enantiomer. In another study, ibuprofen significantly increased the spontaneous production of NO, which was unaffected by an iNOS inhibitor, suggesting instead that eNOS was involved (Miyamoto et al. 2007). This is relevant to the present study due to the observations that while S-( +)-ibuprofen was shown to have relatively little effect, this was not significantly different from the vehicle control and R-(−)-ibuprofen caused appreciable relaxation. However, contrary to this, there is evidence to suggest that ibuprofen, in fact, reduces NO produced in stressful situations, for example in the presence of bacterial endotoxin, where increased NO production leads to a fall in mean arterial blood pressure. Ibuprofen blunts this effect, and the data suggests that ibuprofen down-regulates NO production in human subjects (Vandivier et al. 1999).
The reduction in the relaxation caused by rac-ibuprofen was blocked by ODQ (Feelisch et al. 1999) (Fig. 11), strongly suggests that the relaxation is mediated through cGMP. Removing the endothelium of the vessels, which , showed that aspirin had no significant effect on the responses of the arterial rings to 5-HT (P > 0.05), while NO-aspirin, significantly reduced the maximum tension produced (P < 0.001) together with a significant increase in the EC 50 of applied 5-HT (P < 0.001), when compared with responses of control rings and rings in the presence of aspirin. Control: n = 19, EC 50 = 4.29 × 10 -7 ± 1.63 × 10 -8 M, max. developed tension (percent) = 100.9 ± 0.81. Vehicle: n = 8, EC 50 = 3.53 × 10 -7 ± 2.61 × 10 -8 M, max. developed tension (percent) = 91.24 ± 1.12. Aspirin: n = 8, EC 50 = 4.74 × 10 -7 ± 1.88 × 10 -8 M, max. developed tension (percent) = 93.43 ± 1.40. NO-Aspirin: n = 12, EC 50 = 2.05 × 10 -6 ± 1.02 × 10 -7 M, max. developed tension (percent) = 59.19 ± 0.83. There was no significant difference between the responses of control vessels and those to which vehicle had been added volumes appropriate to the concentrations of applied drugs (P > 0.05) should prevent any action of NOS, had no significant effect on the relaxation. Attempts to employ L-nitro arginine (L-NAME) to block endogenous NO production have been complicated by its action (after potentiating contraction as expected due to the reduction in local NO) to cause a reduction in tension on its own. By comparison, another diastereoisomeric propionic acid, rac-flurbiprofen had similar properties to the ibuprofen isomers, with the R-(-) enantiomer causing significantly greater relaxation than the S-( +)-enantiomer; the magnitude of the relaxation produced being less than with the same concentrations of ibuprofen enantiomers. The other difference is that the NO-flurbiprofen compounds appear to have a more potent vasorelaxant effect than the parent compound. This might be due to an increased ability to release NO. There is little difference between the magnitude of reduction in response by the two enantiomers of the NO-flurbiprofen, suggesting that they can release NO while not directly activating sGC. If activating sGC were important in their action, it would be expected that the R-enantiomer would have a greater effect than the S-enantiomer. As this is not the case, it seems likely that they are producing relaxation via NO.
The results overall suggest that R-(-)-ibuprofen directly activates sGC. They also suggest that NO-ibuprofen does not work in the same fashion. If it did then it would be likely to produce greater relaxation given its coupling with a nitric oxide-releasing moiety. The combination of the release of NO and direct activation of sGC by ibuprofen should produce a greater relaxation than just the activation of sGC alone but it does not, suggesting that the change in the chemical composition by esterification causes sufficient change in structure to prevent the compound working in the In the presence of 10 -4 M methylene blue, a direct nitric oxide synthase and guanylyl cyclase inhibitor (Mayer et al. 1993), the contractile responses of the arterial rings to 5-HT were enhanced, with a significant decrease (P < 0.001) of the EC 50 value, without effect on the maximum tension developed. In the absence of methylene blue, NO-aspirin produced a significant reduction in the maximum response to 5-HT (P < 0.01) and a significant increase in EC 50 value of 5-HT (P < 0.001). The presence of methylene blue, had no significant effect on the NO-aspirin induced increased EC 50 value but appeared to reduce its reduction of the maximum effect of 5-HT. Control: n = 3, EC 50 = 8.83 × 10 -7 ± 6.83 × 10 -8 M, max.

Fig. 10
Effect of increasing concentrations of R-(−)-ibuprofen on the tension produced in fallow deer isolated arterial rings by a constant concentration of 3 × 10 -6 M 5-HT, in the presence and absence of vascular endothelium. While there appears to be no significant difference (P > 0.05) between the maximum relaxation produced by R-(−)ibuprofen in the presence (n = 6) and absence (n = 5) of endothelium, at lower concentrations the difference is significant. At 10 -7 M R-(−)ibuprofen the relaxation in the presence of endothelium is significant (*P < 0.05), while in its absence it is not; a difference even more marked at 3 × 10 -7 M. (control tension: n = 6, the asterisks denote levels of significance between drug treated and control) Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.