Abstract
Sildenafil, a selective inhibitor of the cyclic guanosine monophosphate (cGMP) degrading phosphodiestrase 5 (PDE5), induced migraine without aura in 10 of 12 migraine patients and in healthy subjects it induced significantly more headache than placebo. The aim of the present study was to determine whether the pain-inducing effects of sildenafil would be reflected in plasma levels of important signalling molecules in migraine: cGMP, cyclic adenosine monophosphate (cAMP) and calcitonin gene-related peptide (CGRP). Ten healthy subjects (four women, six men) and 12 patients (12 women) suffering from migraine without aura were included in two separate double-blind, placebo-controlled, cross-over studies in which placebo or sildenafil 100 mg was administered orally. Plasma levels of CGRP, cAMP and cGMP were determined in blood from the antecubital vein. Despite the ability of sildenafil to induce headache and migraine, no significant differences in plasma levels of CGRP, cGMP and cAMP were detected after sildenafil compared with placebo. In conclusion, plasma levels of CGRP, cGMP and cAMP remain normal during sildenafil-induced headache or migraine. However, since previous studies indicate an important role of these signalling molecules, the present study questions whether cAMP and cGMP in peripheral blood can be used for monitoring pathophysiological events in headache and migraine mechanisms.
Introduction
The intercellular signalling molecule nitric oxide (NO) and the neuropeptide calcitonin gene-related peptide (CGRP) are believed to play a major role in the pathophysiology of migraine (1, 2).
NO is able to increase intracellular cyclic guanosine monophosphate (cGMP) by stimulation of soluble guanylate cyclase in smooth muscle cells or neuronal cells (3, 4). CGRP, on the other hand, stimulates receptors on the endothelium or smooth muscle cells, thus activating adenylate cyclase and increasing intracellular cyclic adenosine monophosphate (cAMP) (5). How the different signalling molecules are involved in the pain-generating mechanisms is not fully understood.
We recently found that the selective inhibitor of the cGMP degrading phosphodiestrase 5 (PDE5), sildenafil, induced migraine in 10 of 12 patients suffering from migraine without aura (6) and that sildenafil induced significantly more headache than placebo in healthy subjects (7). The mechanism of action of sildenafil could be elevation of cGMP levels in neuronal cells, thus increasing neuronal excitability. However, whether this possible mechanism is reflected in the blood is not known.
Increased release of CGRP to extracerebral venous blood has been reported after stimulation of the trigeminal ganglion and during spontaneous attacks of migraine with and without aura (2, 8).
Since sildenafil, just as NO donors, is able to induce headache and migraine, the aim of the present study was to determine possible changes induced by sildenafil in plasma levels of cyclic nucleotides in particular, but also of CGRP in blood from the antecubital vein of healthy subjects and migraine patients. There are no previous reports on plasma levels of cAMP and cGMP or CGRP after sildenafil-induced headache. We also chose to compare baseline differences between healthy subjects and migraine patients. The blood samples were taken from both healthy subjects and migraine patients during the in-hospital observation period of the above-mentioned studies (6, 7) when the initial pain-generating processes were believed to take place.
Methods
Two separate studies including either healthy subjects or patients suffering from migraine without aura were performed. Results of headache and migraine induction as well as cerebral haemodynamics have been reported previously (6, 7). Blood samples were drawn from the antecubital vein.
For 8 h prior to the study day, the healthy subjects and migraine patients were not allowed to eat or smoke. However, the intake of water was permitted. No intake of any medication except contraceptive pills was allowed 72 h before the study day.
Study 1
Ten healthy subjects (four women, six men) were included in a double-blind, placebo-controlled, cross-over study in which placebo or sildenafil 100 mg was administered orally inserted in small capsules on two separate days at least 1 week apart. The subjects reported having no first-degree relatives (parents, siblings, children) suffering from migraine. The mean age of the subjects was 24 ± 1 years and mean body weight was 77.3 ± 1.3 kg.
The subjects were headache-free on the study day. After arrival at 08.30 h, a cannula was inserted in the left antecubital vein for retrieval of blood samples. Blood samples were taken at baseline, 60 min and 120 min after administration of the capsules.
A pilot study was performed, including a further two healthy subjects who received only 100 mg sildenafil and no placebo. Blood samples were also taken at baseline, 60 min and 120 min.
Study 2
Twelve patients (12 women) suffering from migraine without aura were included in a double-blind, placebo-controlled, cross-over study similar to the above, but with an observation period of 180 min instead of 120 min. The subjects were healthy except for their migraine without aura. Their mean age was 37 ± 3 years and mean body weight 69 ± 2.8 kg. Inclusion criteria were migraine frequencies of a maximum of two per month and a minimum one every 6 weeks, with a maximum 6 days of tension-type headache a month.
Collection of venous blood
Blood samples were drawn from the left antecubital vein after at least 30 min of rest following the insertion of the cannula (Venflon). For cyclic nucleotide analysis blood was collected into two precooled tubes containing EDTA-sodium (5 mmol/l). For CGRP analysis the blood was collected into precooled tubes prepared with aprotinin (Trasylol, 350 µl, 10 000 kIU) and EDTA. Samples were kept cool, centrifuged at 4 °C and the plasma was stored at − 20 °C until analysis. Cyclic nucleotides were measured separately for healthy subjects and migraine patients, whereas CGRP measurements were performed simultaneously for the two groups. All plasma samples were analysed in duplicates and sample identity regarding treatment, and for CGRP also regarding subject status (healthy or migraine patient), was concealed during analysis.
Analysis of cyclic nucleotides
Plasma cGMP was measured after Sep-pak C-18 extraction as previously described (9). Plasma cAMP was measured by radioimmunoassay employing a monoclonal antibody (MCC-1002-01) from Peninsula Laboratories (Belmont, CA, USA) and in-house prepared iodinated tracer (10). Analyses were performed separately for healthy subjects and migraine patients using separate batches for reference values because of the time interval between analyses. The cGMP reference values were slightly lower for migraine patients and this accounts for a slight difference of factor 1.5 in cGMP values between migraine patients and control. The reference values were stabile during the analyses.
Analysis of CGRP
Plasma levels of CGRP were measured with a validated radioimmunoassay for human CGRP using antibody raised against the C-terminal of human α-CGRP. The detection limit of the assay is < 1 pmol/l (11). Tracer was prepared by the method of Iodo-General (Pierce, Rockford, IL, USA). Tyr(25–37)amide-α-CGRP (Multiple Peptide Systems, San Diego, CA, USA) was used as substrate. Standards of human α-CGRP (Peninsula Laboratory Europe, St Helens, UK) were used.
In contrast to the measurements of cyclic nucleotides, analyses of CGRP were performed for migraine patients and healthy subjects together, using the same batch.
Statistical analysis
All values are presented as means ± SEM.
The calculated area under the curve and the peak of plasma levels were chosen as summary measures when analysing difference in response between treatments (12). The areas and peaks were compared using a paired t-test. Changes over time for each variable were analysed with two-way analyses of variance with factors time and subjects and repeated for each single variable (Statgraphics 3.3). P < 0.05 was considered significant.
Results
Healthy subjects
There were no significant changes in the area under the curve for plasma levels of either cGMP, cAMP or CGRP after sildenafil administration compared with placebo. Neither was a significant difference in peak values found for cGMP (P = 0.12), cAMP (P = 0.61) or CGRP (P = 0.09). Absolute values, mean and SEM, are presented in Table 1.
Healthy subjects
Plasma values for healthy subjects (mean ± SEM) at baseline for cyclic guanosine monophosphate (cGMP), cyclic adenosine monophosphate (cAMP) and calcitonin gene-related peptide (CGRP) and the corresponding P-values between values after placebo and sildenafil (paired t-test of area under the curve) are shown above. AUC, Area under the curve.
There was an increase (29.8 ± 10.0%) in plasma cGMP over time after sildenafil from 5.9 ± 0.8 nmol/l at baseline to 7.5 ± 1.2 nmol/l at 120 min (P = 0.007). However, the increase was not consistent in all subjects and thus not significantly different from placebo.
As reported previously, nine of 10 subjects reported headache on the day of sildenafil; three of these fulfilled the International Headache Society criteria for migraine without aura (13).
Migraine patients
In migraine patients sildenafil 100 mg did not cause any significant changes in cGMP, cAMP or CGRP compared with placebo. Absolute values, mean and SEM, are given in Table 2.
Migraine patients
Plasma values for migraine patients (mean ± SEM) at baseline, 60, 120 and 180 min for cyclic guanosine monophosphate (cGMP), cyclic adenosine monophosphate (cAMP) and calcitonin gene-related peptide (CGRP) together with the corresponding P-values are shown. There was no significant increase in plasma CGRP in the patients who experienced migraine during the observation period (P = 0.41). AUC, Area under the curve.
The plasma levels of CGRP increased significantly over time compared with baseline on the day of sildenafil from 39.0 ± 5.6 to 42.8 ± 6.4, but since the change was very small (8.9 ± 3.4%), it was not significantly different from placebo.
Five patients fulfilled the criteria for migraine when one or more blood samples were taken. None of these showed significant elevations in CGRP (P = 0.41). Likewise, there were no changes in cGMP or cAMP at the onset of migraine. One patient suffered from a migraine attack after intake of placebo and had to receive rescue medication (sumatriptan 6 mg, s.c.) at 90 min and was excluded from the analysis after this time point. Results are presented in Fig. 1a,b and Table 2.
Absolute values (nmol/l) of plasma cyclic guanosine monophosphate (cGMP) in migraine patients are shown over time after placebo (a) and after sildenafil (b). Individual values are shown in thin lines and mean values are shown by a thick line with a triangle (▴) representing placebo and a square (▪) representing sildenafil. Dotted lines indicate patients who fulfilled the criteria for migraine without aura during the observation period.
Migraine patients and healthy subjects showed no significant differences between plasma CGRP (P = 0.88) or cAMP (P = 0.17) at baseline. There was a significant difference between baseline values of cGMP (P = 0.02) with lower values in migraine patients; however, this does not seem to represent a true difference. It was caused by the method of analysis where different reference values were used, because the analyses of the blood from healthy subjects and migraine patients were performed separately and with a time interval.
Discussion
The present data show that sildenafil 100 mg, which induces headache in healthy subjects and migraine in migraine patients with no changes in cerebral circulation (6, 7), did not change plasma levels of cAMP, cGMP or CGRP in healthy subjects or migraine patients in blood drawn from the antecubital vein during the first hours postdose compared with placebo. The headache and migraine-inducing processes of sildenafil are thus not reflected in the plasma levels of these signalling molecules in peripheral blood.
A full understanding of the role of various signalling molecules in migraine pathology has not yet been reached, but one way of monitoring and examining the underlying pathology of migraine and headache has been to measure signalling molecules released from the brain to venous blood during experimental and spontaneous headache and migraine. These results are used to show involvement of the various signalling systems in the pain processes (2, 14–17).
Plasma levels of cAMP and cGMP in migraine
Only few studies have investigated the plasma levels of either cAMP or cGMP or both in migraine without aura. Some studies have concentrated on the cyclic nucleotide levels in platelets, but with variable results (18–21). A significant increase in plasma cAMP, believed to represent the actions of catecholamines, was found during migraine attacks (22). Plasma cGMP was increased during migraine attacks in 37 patients suffering from migraine with and without aura compared with 40 normal subjects (23) and was reported to decrease in response to sumatriptan. However, cGMP levels were not correlated to pain intensities or clinical symptoms and large interindividual variations in cGMP levels were observed.
Plasma levels of cyclic nucleotides in blood taken from the internal jugular vein during migraine attacks in seven patients showed an increase in cGMP concomitant with an increase in nitrite levels, with peak levels 1 h after initiation of the attack and a rapid return to normal values at cessation of the attack (15). Plasma cAMP showed a steady increase over time, reaching its peak levels at 4 h, and the increase was of longer duration than seen for cGMP (15).
In a recent study using a rat model of cortical spreading depression, however, plasma levels of cGMP from the jugular vein showed a significant decrease after 3 h which returned to baseline after 3 days, while a significant increase in cGMP was found in the brain stem and cortex. Thus the neuronal level of cGMP seems not to be reflected in the plasma (24). The most likely explanation is the diffusion barrier represented by the cellular wall and the blood–brain barrier.
Plasma levels of CGRP in migraine
The sensory neuropeptide CGRP is released to the blood stream during exercise and hypoxia (25) as well as in complex regional pain syndrome (26). Increased plasma levels of CGRP are believed to be responsible for hot flushes in postmenopausal women (27).
The presence of CGRP is described in cerebrovascular nerve fibres and the trigeminal ganglion, both believed to be important structures in migraine pathophysiology (28, 29). Several studies, both animal and human, show evidence of a role of CGRP, released from the perivascular nerves, in the pain mechanisms of primary headaches (2, 17, 30). In one human study, CGRP was increased in external jugular venous blood during migraine attack, but not in the antecubital vein (2). Increased plasma CGRP in the antecubital vein during migraine attacks was, however, found in young migraine patients (14). Results from internal jugular venous blood have been varied, since one study found significantly increased plasma CGRP during migraine attacks (15) and one did not (31). It has recently been suggested that CGRP outside of an attack may be increased in migraine patients in general compared with controls (16).
In the present study, we found no significant difference in antecubital venous blood between placebo and sildenafil 100 mg. Neither was a difference found between healthy subjects and migraine patients. However, the latter finding should be interpreted with caution, since the study was not designed to detect a difference between healthy subjects and migraine patients and the two populations differed in sex and age composition.
Our finding of an unchanged CGRP plasma level could be caused by the blood sampling in the early phase of the headache and migraine, or by the peripheral site of blood sampling where the possible small amount of CGRP released may have been too diluted. In animal studies it seems necessary to disrupt the blood–brain barrier in order for human α-CGRP to reach the smooth muscle cell receptors on the arteries and arterioles (32), so CGRP, like cyclic nucleotides, may only enter the blood stream in very small amounts when an intact blood–brain barrier is present and thus escape detection when sampling far from the release site.
Sildenafil and migraine mechanisms
Sildenafil (Viagra©) is a highly selective inhibitor of intracellular cGMP degradation and used for treatment of erectile dysfunction. The drug has a C max of 514–746 ng/ml and T max of ∼ 1 h, it is easily absorbed from the gut, and has a bioavailability of 40–50% (33, 34). It is a very lipophilic molecule and is reported to have central effects, thus indicating that it passes the blood–brain barrier (35).
The most likely mechanism of action for sildenafil seems to be intracellular cGMP elevation. Sildenafil has not been found to affect other signalling systems and the target cells seem to be either smooth muscle or neuronal cells (36). Administration of sildenafil in migraine patients causes migraine attacks similar to the patients’ usual migraine attacks, as seen after infusion of NO donors. This may indicate a similar hyper-reactivity towards the NO–cGMP signalling cascade compared with healthy subjects in the pain processing pathway (6, 37). The target tissue in migraine induction may be either the vascular smooth muscle cells or neuronal cells, probably the latter because of the lack of haemodynamic effects, suggesting a central mechanism in migraine pathology (6).
The present study is the first to measure plasma levels of cAMP, cGMP and CGRP after sildenafil administration in migraine patients. In healthy subjects only one study has been published and it was reported that sildenafil (100–200 mg) increased plasma cGMP in blood from the antecubital vein but only when investigating the area under the curve for cGMP measurements over 12 h and with a very high intersubject variability (33). From in vitro studies, it is known that only minor efflux of cGMP from neuronal cells is possible (38) and the efflux of cGMP from smooth muscle cells after administration of sildenafil is < 1% (39). These data are in line with the unchanged plasma levels of cGMP after sildenafil administration seen in the present study. Thus circulating levels of cGMP may not play a role in the pain mechanisms, but a local increase in plasma cGMP of the cerebral circulation can not be excluded, since only peripheral blood samples were taken. However, in light of the limited efflux from the cells, this seems unlikely.
The difference in baseline cGMP levels between healthy subjects and migraine patients can not be used for drawing conclusions on the pathophysiology of migraine, since it does not represent a true difference, but is probably caused by the methods of analysis.
Peripheral CGRP and cAMP could have been increased as a result of sildenafil administration, since CGRP may be released during induced migraine, and plasma levels of cAMP may be increased by either CGRP or indirectly by the increased levels of cGMP. Increase in cGMP can inhibit the mainly cAMP-degrading enzyme, PDE3. However, no increase in peripheral CGRP compared with placebo was seen in either healthy subjects, migraine patients or the subgroup of migraine patients who fulfilled the migraine criteria in hospital. Since cAMP levels were unaffected by sildenafil, it could furthermore indicate that PDE3 is not involved in the pain process, or that cAMP, like cGMP, enters the peripheral blood stream only in small amounts.
In conclusion, sildenafil 100 mg did not alter cyclic nucleotides or CGRP concentrations in antecubital venous blood, either in healthy subjects or in patients suffering from migraine without aura. The presence of an intracellular change in the investigated signalling molecules can not be excluded by the present study. It seems likely that the processes initiating headache and migraine take place within cells, probably neuronal cells, since no change in cerebral haemodynamics was found previously (6).
Our study, together with experimental studies showing little or no efflux of intracellular second messengers and little efflux of transmitters across the blood–brain barrier, indicates that measurement of messenger molecules in the peripheral blood is not likely to contribute to the further understanding of migraine mechanisms. Evidence of no increase in peripheral plasma levels of these molecules may not exclude that cAMP, cGMP or CGRP play a vital role in migraine pathophysiology within the cells, in the perivascular space or local brain areas, as suggested by previous studies.
Footnotes
Acknowledgements
We thank Kirsten Enghave, Lene Elkjær, Kirsten Bruunsgaard, Alice Rudbøl and Birgit Vingaard for the assistance in blood sampling and analysis. For financial support we thank P.C. Petersen Foundation and University of Copenhagen.