In healthy young men (age, 20–22 years), we tested the role of prostanoids produced by the COX (cyclo-oxygenase) pathway in cutaneous vasodilatation evoked in the finger by ACh (acetylcholine). To this end, changes in cutaneous RCF (red cell flux), recorded by the laser Doppler technique, evoked by a series of iontophoretic pulses of ACh were tested before and after oral aspirin (600 mg). Increases in RCF produced by successive pulses of ACh up to a mean change of 125.5±11.8 PU (perfusion units) were potentiated 30 min after aspirin (160.0±12.4 PU; P<0.05). By contrast, aspirin had no effect on increases in RCF evoked by iontophoretic application of the NO (nitric oxide) donor and endothelium-independent dilator sodium nitroprusside (mean increases in RCF were 73.8±9.8 PU before and 79.1±12.2 PU after aspirin). The ACh-evoked increases in RCF were also potentiated 3 h after oral administration of the antioxidant vitamin C (1000 mg; 139.1±15.4 PU before and 170.5±13.5 PU after vitamin C; P<0.05). We propose that, in healthy young men, cutaneous vasodilatation evoked in the finger by the endothelium-dependent dilator ACh is limited by constrictor products of the COX pathway, including PGH2 (prostaglandin H2), TXA2 (thromboxane A2) and/or superoxide anions. This effect of the COX products may be an early marker of the increased risk of cardiovascular disease in men compared with women.
It is well established that vasodilator responses evoked by ACh (acetylcholine) may involve NO (nitric oxide), prostanoids and/or EDHF (endothelium-derived hyperpolarizing factor). However, it is not clear how these various factors contribute to ACh-induced dilatation in the cutaneous circulation of human subjects.
Morris and Shore , who used the laser Doppler technique to record cutaneous RCF (red cell flux) and applied ACh by iontophoresis, reported that aspirin, given at a dose sufficient to block COX (cyclo-oxygenase) activity, had no significant effect on the cutaneous vasodilatation evoked in the forearm of male subjects by ACh. Thus they concluded that prostanoids do not contribute to this response and that the dilatation is mediated by NO and/or EDHF. Abou-Elenin et al.  obtained similar results. By contrast, Khan et al.  and Noon et al. , who also used laser Doppler fluximetry and iontophoresis, showed that aspirin reduced ACh-induced dilatation in the forearm, but that inhibition of NO synthesis had no effect. They [3,4] therefore implicated vasodilator prostanoids in the ACh response.
On the other hand, there is evidence that ACh can release vasoconstrictor products of the COX pathway, at least under some circumstances . Thus resting tone and the transient contraction evoked by ACh in canine cerebral arteries were reduced by a COX inhibitor, a TXA2 (thromboxane A2) synthetase inhibitor and a TXA2/PGH2 (prostaglandin H2) receptor antagonist : PGH2 (endoperoxide) is an intermediate of the COX pathway, from which TXA2 and PGI2 are generated . Furthermore, contraction evoked by ACh in the aortae of SHRs (spontaneously hypertensive rats) was reduced by PGH synthetase inhibitors and by TXA2/PGH2 receptor antagonists, although not by a specific TXA2 synthetase inhibitor [8–10]. Moreover, ACh evoked a release of PGH2 from the aortae of SHRs, whereas PGH2 evoked contraction with greater potency in the aortae of SHRs than of normotensive rats [9,10]. In addition, O2− (superoxide anions), which can be generated by PGH synthetase , were shown to generate TXA2 in the smooth muscle of SHR aortae and to cause contraction that was inhibited by a TXA2 synthetase inhibitor and a TXA2/PGH2 receptor antagonist [8,12]. Thus PGH2 and TXA2 generated directly by the COX pathway and indirectly, via the action of O2−, have been implicated in ACh-evoked vascular responses and may play a particular role in hypertension.
Consistent with these findings, Taddei et al.  proposed that a COX-dependent constrictor substance contributes to the impaired vasodilatation induced by ACh infusion in the forearm of middle-aged (31–60 years of age) patients with essential hypertension, for a normal response was restored in these patients by infusion of the COX inhibitor indomethacin. These COX-derived vasoconstrictor substances were thought to be O2−, for the dilator response to ACh was similarly restored by infusion of the antioxidant vitamin C . Moreover, given that either indomethacin or vitamin C infusion restored the attenuating effect of NO synthesis inhibition on the ACh-induced dilatation, it was proposed that, in essential hypertension, O2− produced by the COX pathway limits the contribution of NO to ACh-induced dilatation (see [13–15]). It has also been shown  that the forearm vasodilator response to ACh is gradually impaired with age, from approx. 30 years of age in men and following the menopause in women. In men, the early impairment seems to reflect a deficit in the L-arginine/NO pathway that can be reversed by L-arginine infusion, whereas, at approx. 60 years of age, there seems to be an additional impairment caused by COX-dependent vasoconstrictor substances that indomethacin can attenuate [13,15].
In view of these findings, the main aim of the present study was to elucidate the role of prostanoids in cutaneous vasodilatation evoked by ACh in the finger of healthy young men aged approx. 20 years of age. To this end, cutaneous RCF in the finger was recorded using the laser Doppler technique and ACh was applied by iontophoresis before and at intervals after oral administration of aspirin. To investigate the selectivity of the effects observed, vasodilator responses evoked by the endothelium-independent dilator SNP (sodium nitroprusside) and vasoconstrictor responses evoked by NA (noradrenaline) were tested before and after aspirin. Finally, dilator responses evoked by ACh were examined before and after a large oral dose of vitamin C.
Experiments were performed on healthy male volunteers who had no evidence of cardiovascular or other medical disorder and were not on vasoactive medication. They were asked not to smoke or drink alcohol within 24 h of the experiment, or take vigorous exercise, eat a heavy meal or consume caffeine within 2 h of the experiment. None was taking vitamin supplements. All subjects gave their fully informed consent to the study, which was approved by the South Birmingham Health Authority Local Research Ethics Committee.
The experiments were carried out in a quiet temperature-controlled room at 21–23 °C. The subjects sat in a comfortable armchair with arms supported at heart level. The equipment was arranged during an equilibration period of at least 30 min. The cuff of a semi-automatic sphygmomanometer was placed on the subject's right arm to allow ABP (arterial blood pressure) and HR (heart rate) to be recorded at appropriate times during the experiment (see below). The dorsal skin of the middle phalanx of the third or fourth finger on the left hand was gently cleansed with an alcohol swab, followed by de-ionized water. The method of iontophoresis was then applied to this finger in a similar manner to that described previously . A Perspex ring-shaped iontophoresis electrode chamber (30 mm total diameter and 7 mm height with a 8 mm diameter inner ‘drug’ chamber; Moor Instruments, Axminster, Devon, U.K.) was attached to the cleansed skin by means of a double-sided adhesive ring. A velcro strap soaked in saline was strapped round the left wrist to act as the indifferent electrode and to complete the circuit between the drug chamber and strap. The drug chamber was then filled with approx. 0.5 ml of the drug (see below). A laser Doppler probe (DPIT-V2; Moor Instruments) was then inserted into the drug chamber to allow a continuous recording to be made of cutaneous RCF. The electrodes were connected to a battery-powered iontophoresis controller (MIC 1; Moor Instruments) that provided the direct current for the iontophoresis. The Doppler probe was connected to a laser Doppler perfusion and temperature monitor (DRT 4; Moor Instruments). Both the Doppler probe and iontophoresis controller were connected to an Apple Macintosh computer (Power Mac 7100/66) using MacLab hardware, so that RCF and the iontophoresis current could be recorded; data were recorded at a sampling frequency of 40 Hz. The polarity of the electrodes was set according to the charge associated with the drug applied. Thus, as ACh and NA are positively charged, the electrode in the drug chamber was set to be the anode and, since SNP is negatively charged, the drug chamber electrode was set to be the cathode.
At the end of the equilibration period, ABP and HR were measured and baseline RCF was recorded for approx. 5 min. The iontophoresis protocol was then applied (see below) and, at its end, ABP and HR were measured again, and RCF was recorded for a further 5 min or until it returned to a baseline level.
In ten male subjects (aged, 22.2±1.0 years; value is the mean±S.E.M.), ACh (1% in mannitol and water, see below) was iontophoresed by applying seven pulses of 0.1 mA for 20 s each, followed by one pulse of 0.2 mA for 20 s, with 60 s between successive pulses. This is the same iontrophoresis protocol used by Morris and Shore . When baseline RCF had been achieved again, the subject was given 600 mg of soluble aspirin dissolved in 200 ml of diluted orange squash to disguise the taste and appearance of aspirin. This dose of aspirin produced 86% inhibition of bradykinin-induced production of PGI2 and 99% inhibition of TXA2 production by platelets at 30 min, and 70% inhibition of PGI2 production at 90 min; PGI2 production recovered after approx. 6 h, whereas platelet production of TXA2 was still inhibited by 99% at this time . The subject rested for 30–40 min and then the ACh iontophoresis protocol was re-applied. The protocol was applied again at 3–4 h after aspirin and again at 24 h. Each time the iontophoresis was carried out at the same site on the same finger, because there is variation between different sites as to the baseline level of RCF and the change in RCF evoked by iontophoretic application of agonists (for example, see ). This procedure therefore helped to minimize the inherent variability in the results obtained.
In six male subjects (age, 21.1±0.3 years), ACh was applied on four occasions exactly as described in Protocol 1, but these subjects did not receive aspirin. Instead they drank 200 ml of diluted orange squash after the first iontophoresis session. Individual subjects who took part in Protocol 1 or 2 were not told whether they were taking aspirin or orange squash alone. On the basis of the information provided by the manufacturer, we estimate that 200 ml of diluted orange squash contained <2 mg of vitamin C.
In ten male subjects (age, 21.3±0.3 years), SNP (0.01% in water) and NA (0.5 mM in water) were iontophoresed in random order before and after aspirin. SNP was iontophoresed by applying five pulses of 0.1 mA for 20 s, followed by one pulse of 0.2 mA for 20 s, with 120 s between successive pulses. This was a slight modification of the protocol used by Morris and Shore . The dilator response to SNP takes longer to develop than that to ACh and, when using this concentration and current in this range, the responses are less pronounced. In pilot experiments, we found that our SNP protocol produced a maximum change in RCF that was more comparable with that achieved with ACh than that induced by the protocol described by Morris and Shore .
NA was iontophoresed by applying seven pulses of 0.1 mA for 30 s, followed by one pulse of 0.2 mA for 30 s with 60 s intervals between successive pulses. This protocol was based on that described by Drummond : the current and pulse duration were adjusted in pilot experiments so as to produce a substantial decrease in RCF with each pulse (see the Results section).
Aspirin was given after the SNP and NA iontophoresis sessions and the whole protocol was repeated at 30 min as in Protocol 1. As SNP and NA were applied within the same protocol to each individual, the finger used for SNP and NA was kept constant within an experiment, but the choice of whether the third or fourth finger was used was randomized between experiments.
In ten male subjects (age, 20.7±0.3 years), ACh was iontophoresed as described in Protocol 1. The subject was then given 1000 mg of vitamin C dissolved in 200 ml of water. The protocol was repeated after 3 h, as this time period has been shown to allow maximal absorption of vitamin C .
ACh (1% in a vehicle of 3% mannitol in water for injection) was obtained as Miochol (CIBA Vision Ophthalmics, Southampton, U.K.) or was prepared by using ACh (Sigma, Poole, Dorset, U.K.), D-mannitol (Sigma) and water for injection (Norton, Steri-amp Water for injection; BP, Cheshire, U.K.). SNP (0.01%; Faulding Pharmaceuticals, Leamington Spa, Warwickshire, U.K.) and NA (0.5 mM; Abbott Laboratories, Queenborough, Kent, U.K.) were dissolved in water for injection (as above). Stock solutions of the drugs were prepared and stored in 3 ml vials at −60 °C. As SNP and NA are light-sensitive, the vials were wrapped in foil and kept in the dark. Soluble aspirin and effervescent vitamin C were obtained from The Boots Company, Nottingham, U.K.
Analysis of results
All results are expressed as means±S.E.M. Responses evoked in the cutaneous circulation by iontophoresis of drug were expressed as change from baseline RCF in PU (perfusion units), where 100 PU represents 1 V recorded by the laser Doppler meter. For responses to ACh, the change was calculated from the average RCF in the final 20 s of the interval between successive iontophoretic pulses and at 40–60 s after the final pulse, when the response to each pulse had reached its maximum. Similarly, for SNP, average RCF was calculated over the final 30 s of the interval between successive pulses and at 90–120 s after the final pulse. The responses evoked by each pulse of NA were much shorter-lasting, so we took the maximum fall in RCF evoked by each pulse. These procedures allowed the mean change in RCF evoked by each iontophoretic pulse in each protocol to be calculated, as described by Morris and Shore . In addition, a compacted mean was calculated for each subject's response to each session of iontophoresis by combining the values to each pulse (see above), so as to obtain a single value that represented the response to the whole iontophoresis session. Differences between sessions within Protocols were assessed by using ANOVA for repeated measures with Fisher's post-hoc test when appropriate. Statistical significance was taken as P<0.05. Differences between baselines within groups were analysed by paired Student's t test.
The baseline values of cutaneous RCF, mean ABP and HR are shown for each Protocol in Table 1. There were no differences between Protocols.
|RCF (PU)||Mean ABP (mmHg)||HR (beats/min)|
|RCF (PU)||Mean ABP (mmHg)||HR (beats/min)|
Successive pulses of ACh evoked graded increases in RCF that reached a maximum after the fourth or fifth pulse (Figure 1), as described by Morris and Shore  who used a similar protocol on forearm skin. ABP did not change significantly between the beginning and end of this or any other iontophoresis session (results not shown).
Effects of aspirin (600 mg) on cutaneous vascular responses evoked by iontophoresis of ACh in finger of young male subjects
At 30–40 min, 3–4 h and 24 h after aspirin, baseline RCF was not significantly different from the original baseline (36.8±4.9, 37.8±2.6, 38.6±4.3 and 46.7±7.9 PU respectively). However, at 30–40 min after aspirin, the increases in RCF evoked by ACh were potentiated relative to the control responses, as can be seen from the original data and the compacted mean values of RCF (Figure 1A). At 3–4 h and 24 h, the ACh-evoked responses were not significantly different from the control responses (P=0.22 and 0.86 respectively).
This Protocol provided the time control for Protocol 1. Baseline RCF did not change between sessions, neither were there any differences between the ACh-evoked responses at 30 min, 3–4 h and 24 h after the original session (Figure 1B).
Successive pulses of SNP evoked graded increases in RCF that did not reach an obvious maximum (Figure 2A). At 30–40 min after aspirin, baseline RCF was not different from the original baseline (29.6±4.5 and 33.1±4.4 PU respectively), and the responses evoked by SNP were not different from the control response.
Lack of effect of aspirin (600 mg) on cutaneous vascular responses evoked by iontophoresis of SNP (A) and NA (B) in finger of young male subjects
The magnitude of the decreases in RCF evoked by NA varied substantially between individuals. Nevertheless, successive pulses of NA evoked progressive reductions in RCF (Figure 2B). At 30–40 min after aspirin, baseline RCF was not different from the original baseline (35.05±3.7 and 34.33±7.6 PU respectively). Moreover, aspirin did not affect the changes in RCF evoked by NA (Figure 2B). As it seemed likely that the magnitude of the decrease in RCF evoked by NA might be dependent on the absolute value of baseline RCF, we also calculated the changes evoked by NA as a percentage change from baseline. This made no difference to the conclusion that aspirin had no effect on the NA-evoked responses (P=0.34).
The 1000 mg dose of vitamin C had no effect on baseline RCF recorded at 3 h (38.6±4.1 PU in control compared with 36.7±6.4 PU after vitamin C). However, at 3 h after vitamin C, responses evoked by ACh were potentiated relative to the control responses (Figure 3).
Effect of vitamin C (1000 mg) on cutaneous vascular responses evoked by iontophoresis of ACh in finger of young male subjects
The present study showed that increases in RCF evoked in the cutaneous circulation of the finger of healthy young men were potentiated by orally administered aspirin and vitamin C. By contrast, aspirin had no effect on responses evoked by the NO donor SNP or the catecholamine NA. We accept it is a limitation of the present study that the experimenters were not blinded to whether aspirin or placebo was administered in Protocols 1 and 2. It is also a limitation that the experiments in Protocols 3 and 4 were not repeated with placebo rather than aspirin or vitamin C. However, the experimenters had no prior expectation of whether, or in which direction, cutaneous vascular responses evoked by the various agonists would be affected by aspirin. Moreover, it may be noted that the results obtained in Protocols 1 and 2 at time 0 and 3–4 h after aspirin or placebo can serve as controls for the results obtained in Protocol 4 at time 0 and at 3–4 h after vitamin C. These comparisons would not change the conclusion drawn from the results of Protocol 4, in that responses evoked by ACh at 3–4 h after aspirin or placebo were not different from those evoked at time 0, whereas those evoked at 3–4 h after vitamin C were potentiated relative to those evoked at time 0. Thus it seems very unlikely that the limitations of our present study design influenced the outcome of the results obtained.
The increases in cutaneous RCF evoked by ACh were similar to those evoked by the same protocol in the cutaneous circulation of the forearm by Morris and Shore . Since ABP did not change between the beginning and end of the iontophoresis protocol in the present study, we can assume the increases in RCF reflected cutaneous vasodilatation in the finger. At 30–40 min after 600 mg of aspirin, at a time when the COX pathway is maximally inhibited , there was no effect on baseline RCF. Thus neither vasodilator nor vasoconstrictor products of the COX pathway seem to have a net tonic effect on cutaneous circulation. Similar conclusions have been reached in previous studies on cutaneous circulation (for example, see [1–4]). Thus our finding that aspirin potentiated the increases in RCF evoked by ACh allows the proposal that constrictor factor(s) produced by the COX pathway as a consequence of the action of ACh on cholinergic receptors limited the evoked cutaneous vasodilatation. We accept that aspirin may have effects additional to those it has on the COX pathway. However, as we have no evidence of what these might be from the measurements of systemic and local cardiovascular variables, we obtained, we have interpreted our results in terms of effects on the COX pathway as have others in the field (for example, see [1–4] and below).
Our findings contrast both with previous reports that aspirin had no effect on cutaneous vasodilatation evoked in the forearm by ACh [1,2] and with reports that aspirin reduced such dilator responses [3,4]. The difference between the outcomes of these studies are difficult to reconcile and they cannot be directly compared with those of the present study. For, although the iontophoresis protocol used by Morris and Shore  was comparable with that of the present study, those used in the other studies were different [2–4]. Thus in one of the studies that suggested aspirin had no effect on forearm cutaneous vasodilatation, responses were tested 30 min after a maximal dose of aspirin  as in the present study, whereas in two studies that differed in their outcome, 500 or 600 mg of aspirin/day was taken for 3 days [2,3], and in the remaining study, aspirin was given intravenously. Furthermore, the subjects who took part in the previous studies were generally older (22–45 years of age [1–4]) than those of the present study; in three studies they were all male [1,3,4], whereas in the remaining study they were a mixed group of two females and seven males . It may be there is a real difference between the role of COX products in ACh-evoked vasodilatation in finger and forearm skin: they may have a predominant dilator role in forearm skin , but the ability to show this may be critically dependent on the details of the experimental protocol. However, the security of this proposal is complicated by the fact that, in studies performed on whole forearm vasculature by plethysmography, the vasodilator response to ACh was compromised by vasoconstrictor COX products with increasing age in men, in men compared with women and in hypertensives compared with normotensives (see the Introduction and [13–16]). None of these factors was taken into consideration in the studies on the cutaneous circulation of the forearm [1–4].
By contrast, we can make direct comparisons between the results of the present study on young men with our own observations on women of a similar age . Thus in experiments in which we compared female patients with primary Raynaud's disease with control women, aspirin given orally and at the same dose as in the present study had no effect on increases in RCF evoked in the finger of the control women when ACh was delivered by the same protocol as in the present study. Thus it seems that the limitation of the cutaneous vasodilator response evoked in the finger by ACh by vasoconstrictor products of COX is a feature of young men, but not young women. Clearly, we have to also conclude that the factors responsible for the ACh-evoked vasodilatation in the finger of young men and young women must be mainly NO and/or EDHF. Others have already demonstrated that the cutaneous and muscle vasodilatation evoked in the forearm of normotensive subjects is mainly NO-dependent [14,21].
Since the potentiating effect of aspirin on the ACh-evoked dilator responses had faded by 3–4 h after aspirin and had disappeared at 24 h, it seems likely the vasoconstrictor factors that limited the response were produced by endothelium or vascular smooth muscle, rather than platelets. For example, PGI2 production in response to bradykinin, which can be attributed to COX metabolism by endothelium, recovered by 6 h after this dose of aspirin , whereas the effects aspirin has on TXA2 production by platelets persists for the lifetime of the platelet, approx. 10 days [17,22].
On the basis of previous studies (see the Introduction), the constrictor factors that limited ACh-evoked dilatation could be PGH2 or TXA2 produced directly by the action of ACh on the COX pathway in the endothelium or, if O2− were released by the increased COX activity (see below), O2− could have acted on the vascular smooth muscle to produce PGH2 or TXA2 [6,8–10,12]. The proposal that newly synthesized COX products were involved is consistent with the other findings of the present study. For, if any of these substances had been released tonically, rather than as a consequence of cholinergic receptor stimulation, dilator responses evoked by the NO donor SNP would have been likely to be potentiated by aspirin, particularly as O2− inactivate NO . This was not the case. Moreover, in previous studies, tonically released vasodilator prostanoids or dilator prostanoids released during vasoconstriction evoked by NA were shown to limit NA-evoked vasoconstriction: COX blockade potentiated the vasoconstriction (for example, see ). In the present study, vasoconstriction evoked by NA was not affected by aspirin. This suggests that vasodilator prostanoids are not released during NA-evoked constriction in the fingers of young men, and/or that the effect of any such dilator prostanoids was balanced by the production of vasoconstrictor COX products, such that aspirin had no effect on the NA response.
Finally, our finding that a large dose of vitamin C, which is an antioxidant that scavenges aqueous free radicals, such as O2− , had no effect on baseline RCF, but potentiated the ACh-evoked increases in RCF, suggests that any free radicals present at rest had little vasoactive effect. Rather, it is more likely that free radicals that were released in the finger during ACh iontophoresis limited the ACh-evoked dilatation. Given our findings on the effects of aspirin (see above), the most obvious interpretation is that O2− produced by the COX pathway somehow limited the ACh-evoked dilatation, although we have to accept that O2− or other free radicals may have been produced by other actions of ACh, possibly via the NO synthesis pathway . We also have to accept that vitamin C may have effects additional to its antioxidant properties . If our favoured interpretation is correct, then it seems likely that O2− generated by the COX pathway in the endothelium led to production of vasoconstrictor PGH2/TXA2 by the vascular smooth muscle via the COX pathway [8,12]. There remains the possibility that PGH2 and TXA2 generated in the endothelium also limited the ACh-evoked dilatation. In the absence of selective PGH2 and TXA receptor antagonists and synthetase inhibitors that can be used in human subjects, it is difficult to differentiate between these possibilities.
Irrespective of the precise mechanisms, the present study has demonstrated that the dilatation evoked by ACh in the finger circulation of young men has characteristics similar to those identified for ACh-evoked dilatation in the forearm of patients with essential hypertension and that appear in the forearm of normotensive men from 60 years of age. From these studies it has been concluded that endothelium-dependent dilatation is limited by COX products and free radicals [13–15]. This raises the possibility that the characteristics of the finger vascular response to ACh in young men provides an early index of the characteristics that predispose men to an increased risk of cardiovascular disease, such as hypertension and its consequences . It may be that, in women from puberty until menopause, oestrogen, with its known antioxidant and dilator properties [26,27], offers protection against these hypertensive properties of COX-dependent vasoconstrictor products (for example, see [16,28]).