Release of Calcitonin Gene-Related Peptide (CGRP) from Guinea Pig Dura Mater in vitro is Inhibited by Sumatriptan But Unaffected by Nitric Oxide

Abstract

Migraine attacks can be provoked by administration of nitroglycerin, suggesting a role for nitric oxide (NO). The fact that release of the neuropeptide CGRP from trigeminal sensory nerves occurs during the pain phase of migraine and that NO can augment transmitter release prompted us to study CGRP release from the in situ dura mater in guinea pig skulls. Release of CGRP by capsaicin or by high potassium concentration was concentration-dependent and counteracted in calcium-free medium. The anti-migraine compound, sumatriptan, inhibited CGRP release via the 5-HT₁-receptor. The NO donors, nitroglycerin, sodium nitroprusside and S-nitroso-N-acetylpenicillamine did not influence CGRP release, alone or together with the stimulants. We concluded that the skull preparation is well suited for scrutinizing CGRP release from dura mater. The fact that sumatriptan inhibits CGRP release as in migraine patients suggests a use for the present preparation in headache research.

Introduction

The neuropeptide CGRP is present in dura mater and cranial arteries together with other neuropeptides, e.g. substance P and neurokinin A in nociceptive sensory fibres, that have been implicated in migraine (1, 2). CGRP is released from the trigeminal sensory nerves to the venous outflow in migraine patients during the headache phase (3). The reason why the nerves that innervate dura mater and vessels are activated during the migraine attack is unknown but the release of neuropeptides may contribute to sensitization of the nociceptors and probably also to neurogenic inflammation. This type of aseptic inflammation has been implicated in the development of migraine and other vascular headaches (4) and is characterized by vasodi-latation, plasma protein extravasation and mast cell degranulation (4). It can be induced experimentally in the dura mater by antidromic electrical stimulation of the trigeminal ganglion or by chemical stimulation by i.v. injection of capsaicin (5–7).

The anti-migraine drug sumatriptan has been shown to counteract the headache as well as the elevated CGRP levels in cranial venous blood of migraine patients (8). Moreover sumatriptan attenuates the release of CGRP after trigeminal nerve stimulation in cats and rats (8, 9) and the stimulated CGRP release from cultured trigeminal neurones (10).

Whilst the pathogenesis of migraine remains unclear, it has been shown that nitroglycerin, i.e. glyceryltri-nitrate (GTN) and other nitric oxide (NO) donors are able to induced headache in normal persons and migraine attacks in migraineurs (11). It is thought that NO by its powerful vasodilatation of the cephalic arteries induces headache, but activation/sensitization of central and peripheral nociceptive pathways may also be involved. The fact that NO has been shown to augment transmitter release in central neurones (12–16) and the trigeminal cerebrovascular system (17) could also be of importance to understanding migraine pathophysiology.

In the present work we studied the release of CGRP from the in situ dura mater of isolated guinea pig skulls. The advantage of this preparation is that the trigeminal nerve and the dura mater remain intact and that exogenous compounds can readily stimulate the nerves. We chose the guinea pig because the 5-HTj-receptors are similar to the human in amino acid sequence and in the binding affinity of many serotonergic compounds (18, 19). The purpose of the present study was to characterize the release of CGRP from the in vitro dura preparation and to study whether the established anti-migraine compound sumatriptan was able to interfere with CGRP release and whether NO donors were able to augment the release. We show that in our model sumatriptan indeed is able to attenuate the release of CGRP, while NO donors do not enhance the release.

Materials and methods

Skull preparation

Male guinea pigs (200-350 g; obtained from Charles River, Sulzfeld, Germany) were stunned by a C0₂ atmosphere and decapitated using a guillotine. The lower jaw was removed. The skin of the skull was cut along the midline and retracted from the calvarium, which was then divided along the sagittal suture into two halves by a hand-held saw. After the brain halves were carefully removed without lesioning the dura mater or the trigeminal ganglion, the skulls were washed for 30 min in HEPES buffer (room temperature, pH 7.4) containing (in mM): 10.0 HEPES; 120.0 NaCl; 5.0 KC1; 0.62 MgS0₄, 7 H₂0; 1.0 CaCl₂, 2 H₂0 and 6.1 glucose. Then the skulls were mounted in a mould of modelling wax with the lateral side downwards and placed in a closed chamber situated on a water bath which maintained the chamber at 37C, relative humidity at 100%, and which was supplied with a steady supply of oxygen. The skull cavities were filled with HEPES buffer (pH 7.4; 37C) and left untouched for 10 min before the experiment started. The essence was to measure the CGRP, which was released from the dura when exposed to a HEPES buffer with or without high potassium concentration/capsaicin added. The skull, without touching the dura, was repeatedly and carefully emptied and filled with the same volume of various solutions as described below. Between two exposures to HEPES buffer with capsaicin/high potassium the skull was rinsed with HEPES buffer during two 5-min periods. The relevant samples were immediately transferred to test tubes placed on ice before they were stored at - 20C for later analysis of the CGRP contents.

At the beginning of each experiment the skull was filled with HEPES buffer, which after 5 min was removed and the skull was refilled with the same volume of HEPES buffer. After an additional 5 min the second sample, the basal level, was collected. During the next periods the skull was repeatedly stimulated with HEPES buffers containing various compounds as indicated.

Compounds

The following drugs were used: capsaicin (Sigma Chemical Co., St Louis, MO), sumatriptan succinate (Glaxo Wellcome, Brondby, Denmark), GTN (The Dispensary at Glostrup Hospital), sodium nitroprusside (SNP) (Sigma), S-nitroso-N-acetylpenicillamine (SNAP) (Calbiochem, La Jolla, CA) and methiothepin (Sigma). Capsaicin (10 mM) was dissolved in ethanol (48%) and diluted with HEPES buffer. Potassium solutions were made by replacing an equal concentration of NaCl with KC1 in the HEPES buffer. The required amount of NO donor was weighted in test tubes and dissolved in the HEPES buffer immediately before use. In solutions of SNP>30mM, SNP replaced an equal concentration of NaCl in the HEPES buffer. Sumatriptan and methiothepin were dissolved in HEPES buffer.

Assay of CGRP

The samples from individual skulls were thawed at room temperature and then used directly to determine the CGRP content using a commercial Rat CGRP Enzyme Immunometric Assay (EIA) Kit (SPIbio Company, Paris, France) on a 96-well plate. This assay is based on a double-antibody sandwich technique. The wells of the microtitre plate are coated with a monoclonal antibody specific for CGRP. An acetylcholinesterase (AchE)-Fab’ conjugate, which binds selectively to a different epitope on the CGRP molecule, is also added to the wells. The two antibodies will form a sandwich by binding on different parts of the rat CGRP molecule. AchE catalyses the formation of a yellow dye from the reagents, which is measured photometrically on a microplate reader. Standard curves were made from CGRP dissolved in the HEPES buffer and included on each 96-well plate. The minimum detection limit of the assay is approximately 1 pg/ml. None of the tested drugs interfered with the present assay.

Statistical analysis

All data are expressed as means+ sem. Differences among the groups were evaluated by non-parametric statistics since a Gaussian distribution could not be verified in all cases. The Kruskal-Wallis test (anova) with Dunn's post test was used when more than two groups were involved; otherwise Mann-Whitney IHest, intra-individual comparisons or the Wilcoxon matched pairs test were used (significance level P<0.05).

Results

Stimulated release of CGRP from the dura mater by capsaicin and potassium

The release of CGRP was assessed by the amount of CGRP that had accumulated in the skull reservoir during a 5-min period. Stimulation with capsaicin caused a dose-dependent release of CGRP from dura mater (Fig. la). The skull preparation was stimulated four consecutive times for 5 min with increasing concentrations of capsaicin. Before each stimulation period the skull was exposed to control HEPES buffer for 2X5 min. There was a notable increase with 0.1 and 1 um capsaicin (2.3- and 6.6-fold, respectively) of which only the latter was significant (P < 0.001, anova). Exposure of the skulls to increasing concentrations (25, 50 and 100 mM) of potassium for 5 min during three consecutive episodes, each preceded by two 5-min exposures to control HEPES buffer, caused a 3.2-, 5.0- and 6.8-fold increase, respectively (NS, P<0.05, P<0.01, anova) as shown in Fig. lb.

Figure 1

Release of CGRP from the in situ dura mater of guinea pig skull during consecutive 5-min periods with increasing concentrations of either capsaicin (a) (n = 10) or potassium (b) (n = 4) in the HEPES buffer. The skull was exposed twice (2X5 min) to control HEPES buffer between stimulations. (Mean + sem.) ∗P<0.05; ∗∗P<0.01; ∗∗∗P< 0.001 compared with the basal level (anova).

CGRP release is calcium-dependent

Figure 2 shows the CGRP concentration of the skull fluid when the skull preparation was stimulated twice, during periods of 5 min, with 50 mM potassium and finally with 1 um capsaicin (open bars), each stimulation preceded by control HEPES buffer for 2X5 min. The fact that the two potassium stimulations gave a similar release of CGRP suggests that this procedure can be used to evaluate the influence of compounds on CGRP release. Hence in another trial (solid bars) the second potassium as well as the final capsaicin challenge was made with calcium-free solutions. The lack of calcium significantly diminished the CGRP release provoked by both types of stimulation (Fig. 2).

Figure 2

Release of CGRP from the in situ dura mater during consecutive 5-min stimulations with HEPES buffer containing 50 mM potassium (K⁺) (twice) followed by HEPES buffer with 1 |iM capsaicin added. The skull was exposed twice (2X5 min) to control HEPES buffer between stimulations. ?, The control trial (mean + sem, n = 8); ?, the trial in which the second K⁺ stimulation and the capsaicin stimulation were devoid of calcium (mean + sem, n = 4). ∗P<0.05; ∗∗P<0.01 (Mann-Whitney test).

Effect of sumatriptan on CGRP release

The ability of sumatriptan to interfere with CGRP release was tested by adding 1, 10 or 50 um sumatriptan to the buffer solutions. Sumatriptan at these concentrations had no effect on basal CGRP release (results not shown). We used the double potassium stimulation to show that sumatriptan at 10 um and 50 um significantly attenuated the release of CGRP by 31% and 32%, respectively (P<0.05), while 1 um was without effect (Fig. 3a). In order to assess the mechanism of inhibition we used the 5-HTx-receptor antagonist methiothepin (20). The compound at 100 um had no effect on basal (result not shown) or stimulated CGRP release but was able to block the inhibitory effect of sumatriptan on potassium-provoked CGRP release from dura mater (Fig. 3a).

Figure 3

Effect of the 5-HTj-receptor agonist sumatriptan (suma) and the 5-HTj-receptor antagonist methiothepin (met) on potassium- or capsaicin-stimulated CGRP release from the in situ dura mater. The skull was exposed twice (2X5 min) to control HEPES buffer between stimulations, (a) The skull was stimulated twice with 50 mil K⁺, of which the second stimulation contained sumatriptan, methiothepin or both. The columns show the result from the second K⁺ stimulation (K⁺ ₂) expressed in percent of the first K stimulation (K⁺j) (mean + sem; 50K, n = 8; 50K + 1 mm suma, n = 8; 50K + 10 mm suma, h = 6; 50K + 50 hm suma, n = 8; 50K + 50 mm suma + met, n = 4; 50K + met, n = 2; ∗P<0.05 compared with first K stimulation, Wilcoxon test), (b) The skull was exposed to capsaicin during four consecutive 5-min periods with HEPES buffer containing increasing concentrations of capsaicin (0.001-1 mm). The figure shows the results using 1 mm capsaicin (mean sem, capsaicin, n = 10; capsaicin + suma, n = 6; capsaicin + met, n = 2; capsaicin + suma + met, n = 4; ∗P<0.05 compared with capsaicin alone, anova).

The effect of sumatriptan on capsaicin-evoked release was tested during four consecutive stimulations with increasing concentrations of capsaicin (as shown in Fig. la) with sumatriptan added. Before each stimulation period the skull was exposed to control HEPES buffer for 2X5 min, of which the latter contained sumatriptan. Five trials were performed and Fig. 3b show the results from using the highest concentration of capsaicin (1 lim). Sumatriptan 10 lim had no effect on the capsaicin-mediated release, whereas 50 lim attenuated the release, albeit non-significantly. The presence of 100 lim methiothepin counteracted the effect of sumatriptan (P< 0.05, anova).

Nitric oxide donors

The three different NO donors GTN, SNP and SNAP were used to evaluate the effect of NO. An augmentation of CGRP release from dura mater had been anticipated, but this was not the case. The donors were added to the control buffer in increasing concentrations during three consecutive exposures (Table 1). GTN from 0.01 to 1 lim (n=2) and 10 to 1000 lim (n = 2), SNP (1, 10 and 30 dim) (m=4) and SNAP 1 mM three times (n = 2). Likewise, when two consecutive stimulations were performed with 50 mM potassium while SNP 30 mM or SNAP 1 mM were added to the second stimulation, there was no detectable increase of the release of CGRP (n = 2, data not shown). Lastly when GTN 1 Lim was added to the capsaicin buffer there was no significant augmentation of the resulting CGRP release (Fig. 4).

Table 1

Effect of nitric oxide donors on CGRP release from dura mater

Control		1
GTN	0.01 μM	0.97±0.06
	0.1 μM	0.99±0.05
	1 μM	1.05±0.01
GTN	10 μM	0.90±0.22
	100 μM	0.61±0.09
	1000 μM	0.49±0.12
SNP	1 mM	0.61±0.09
	10 mM	0.57±0.18
	30 mM	0.58±0.24
SNAP	1 mM (1st stimulation)	0.96±0.96
	(2nd stimulation)	1.1±0.32
	(3rd stimulation)	0.55±0.33

The concentration of CGRP was measured in control HEPES buffer from the skull with and without nitric oxide (NO) donors added following three consecutive exposures to increasing concentrations of nitroglycerin (GTN) (0.01-1 mm or 10-1000 mm, h = 2), sodium nitroprusside (SNP) (1, 10 and 30 mM, n = 4) or S-nitroso-N-acetylpenicillamin (SNAP) (1 mMX3, n = 2). The skull was exposed twice (2X5 min) to control HEPES buffer between stimulations. (Mean ± sem.) The values are expressed in relation to the control value. The HEPES buffers containing SNP were not significantly different from the control value (anova).

Figure 4

Effect of the nitric oxide donor nitroglycerine (GTN) on the capsaicin-stimulated release of CGRP from dura mater. The skull was stimulated four consecutive times with increasing concentrations of capsaicin with (?. n = 6) or without (?, n = 10) GTN (1 mm) added. The skull was exposed twice (2X5 min) to control HEPES buffer between stimulations. (Mean + SEM; NS, not significant; Mann-Whitney test.)

Discussion

The present study demonstrates that the intact dura mater encephali maintained in a skull preparation in vitro is able to release the neuropeptide CGRP by direct chemical stimulations in a calcium-mediated manner. Our study was inspired by Ebersberger et al. (21), who showed that release of the neuropeptides CGRP and substance P from the rat skull preparation could be provoked by electrical stimulation of the trigeminal ganglion. We used depolarization by high potassium concentration or by the pungent compound, capsaicin, that selectively activates A5- and C-fibre sensory neurones via the vanilloid receptors (22). As such the preparation is well suited to study mechanisms of neuropeptide release from the trigeminal nerve that is the main sensory nerve in the head region. The skull CGRP release should give a direct measure of trigeminal CGRP release and is (present study and (21)) quantitatively larger than the release that can be inferred from determination of the plasma CGRP content during trigeminal stimulation (7, 8) and migraine attacks (3). Several factors may be important for the difference. The in vivo measurements are dampened by diffusion limitation of CGRP from nerve to blood, by degradation of the peptide, and by dilution. Moreover, the fact that the stimulus used here in the skull preparation is direct and involves most of the dural nerve fibres is bound to result in quantitatively more CGRP being released. The question, however, remains whether the preparation is relevant to events taking place during headache, e.g. migraine. Three observations suggest that it may be used in this connection: CGRP is released from excited nociceptive fibres in the dura, suggesting that pain sensations would have occurred in the intact nerve system. Moreover, CGRP is indeed released from intracranial structures during a migraine attack (3), and finally, the present investigation shows that the anti-migraine drug sumatriptan was able to interfere with CGRP release as it does in migraine patients (8)

Sumatriptan attenuates dural CGRP release

Sumatriptan was able to attenuate both potassium- and capsaicin-stimulated CGRP release in a dose-dependent manner. The findings are in agreement with those of Durham & Russo (10) who, in cultured trigeminal neurones, showed that sumatriptan at the same concentrations as used here inhibited the release of CGRP caused by potassium or a mixture of inflammatory mediators at low pH. The mechanism by which sumatriptan acts is not fully understood, but it involves 5-HTj-receptor subtypes. Sumatriptan, originally developed as a vasoconstrictor selectively acting as an agonist at 5-HT_1B/1D-receptors, also inhibits CGRP release from trigeminal sensory nerves in vivo (8, 23). This effect is mediated via 5HT_1D-receptor, since the 5HT_1D-receptor antagonist GR-127.935 prevented the inhibitory effects of sumatriptan on neurogenic plasma extravasation (24). Likewise, the present work shows that the sumatriptan effect is receptor-specific, since the 5-HTx antagonist methiothepin removed the inhibitory effect of sumatriptan on CGRP release, in agreement with Durham & Russo (10). Activation of the 5-HTx-receptors is thought to mediate a decrease in intracellular cAMP levels through G_rproteins, but recently it was shown that sumatriptan did not change cAMP level in the cultured trigeminal cells but rather caused a prolonged increase of the intracellular calcium level, and perhaps thereby activated calcium-dependent phosphatases which interfere with the exocytotic process (10). It is highly likely that this mechanism also applies to the trigeminal neurones in situ.

The reason why 50 lim sumatriptan was needed to attenuate the neuropeptide release caused by capsaicin, which takes place via the VR_X receptor, while 10 lim was sufficient when a general membrane depolarization was elicited by potassium, is unclear. Compared with the estimated plasma concentration of sumatriptan in migraine patients (0.2 lim (25)) and the concentration needed to cause contraction of vessels (26–28), the applied concentration appears high. Moreover, the CGRP release to the cranial circulation in experimental animals and during a migraine attack in man could also be counteracted by clinically relevant doses of sumatriptan (8, 9). We have no single explanation for the discrepancy, except that our way of provoking CGRP release probably represents a much stronger stimulation than that used in the other investigations and therefore requires a higher concentration of sumatriptan.

NO does not augment dural CGRP release

It is well established that NO by itself or together with other stimuli augments the release of neurotransmitters from brain slices, e.g. aspartate (12) and dopamine from striatum (13, 29, 30) and noradrenaline and glutamate from hippocampus (14, 31, 32). NO is believed to influence transmitter release via activation of guanylate cyclase, which leads to the formation of cGMP. The ensuing activation of cGMP-dependent protein kinase phosphorylates proteins that are involved in the neurotransmitter release process or in the activation, e.g. calcium channels. In the current study it was shown that none of the nitric oxide donors GTN, SNP or SNAP caused a release of CGRP from dura mater. In fact, there was a tendency towards a lowering of the release (Table 1), but the effect was not significant for SNP and for GTN and SNAP significance was not tested due to the limited number of experiments. In contrast to these observations, another study (33) indicated that SNP can stimulate CGRP release from dorsal horn slices from rat spinal cord. They, however, added SNP (30 mM) directly to the buffer, and it cannot be excluded that other mechanisms, e.g. hyperosmotic effects, could play a role (34). Also Dymshitz & Vasko (35) found in cultured DRG neurones that SNP was able to increase the release of CGRP provoked by capsaicin. However, only SNP, but not two other NO-donating compounds, SNAP and 3-morpholinosydnonimine, showed an effect, which again raises the possibility that NO is not responsible for the effect of SNP.

An explanation of why NO can augment the release of small transmitter substances while not affecting the release of CGRP could be the different storage conditions. Neuropeptides are presynaptically stored in large dense-cored vesicles, while the small neurotransmitters reside in clear small vesicles. While the release from both types of vesicles occurs by calcium-mediated exocytose, there are differences in the calcium requirement (36) and probably also in the cGMP-dependent processes.

There are several reasons why NO was expected to alter the release of CGRP. First, infusion of GTN, en exogenous source of NO, can induce migraine attacks, which are similar to spontaneous attacks (37). Second, GTN and SNP cause vasodilatation in the cerebral circulation in cats (17) and isolated rings of rat aorta (38). These effects are blocked by the CGRP antagonist CGRP₈_₃₇, suggesting that CGRP is responsible for at least a part of the vasodilatory effect of NO. In the study by Thomsen et al. (37) the concentrations of CGRP were not measured and it is therefore not known whether these migraine patients without aura experience an increased release of CGRP after GTN infusion. Increased levels of CGRP after GTN infusion have been shown in cluster headache patients (39), but only when the patient experienced a peak headache after the GTN infusion. Thus, it seems that GTN does not directly cause release of CGRP but triggers a cascade of events that subsequently leads to the release of CGRP. In contrast, in the two other studies (17, 38) performed in situ in cat and in vitro on rat aorta, NO seemed to cause an immediate release of CGRP. Why these results are different from that found in the present study is hard to explain, but could be due to a difference in species and vascular regions examined.

In conclusion, we have found that potassium depolarization and capsaicin, but not nitric oxide donors, cause the release of CGRP in the dura mater of isolated guinea pig skulls and that this release can be attenuated by the 5-HT_1B/id agonist sumatriptan.