Sumatriptan can inhibit trigeminal afferents by an exclusively neural mechanism

Transcription

1 Brain (1996), 119, Sumatriptan can inhibit trigeminal afferents by an exclusively neural mechanism Karen L. Hoskin, 1 Holger Kaube 2 and Peter J. Goadsby 1 1 Institute of Neurology, The National Hospital for Neurology and Neurosurgery, London, UK and 2 Neurologische Klinik und Poliklinik University of Essen, Essen, Germany Summmary Mechanical distortion of the human cranial venous sinuses is painful as is cranial venous sinus distension during migraine. Sumatriptan, the serotonin (5HT) ib/d -like receptor agonist, is highly effective in relieving migraine headache and part of its action may be due to constriction of cranial dural blood vessels. Using immunohistochemical detection of the immediate early gene Fos, we have mapped the spatial pattern of neural activation in the caudal medulla and the upper cervical spinal cord (CI, C2 and C3) in cats following either electrical or mechanical stimulation of the superior sagittal sinus. Fourteen cats were anaesthetized with a- chloralose and prepared for physiological monitoring of blood pressure, heart rate, rectal temperature and expired COi- Electrical stimulation evoked significant increases in the (median) numbers of Fos-positive cells in laminae I and Ho of the superficial dorsal horn of CI, C2 and C3 cervical spinal cord (88, 92 and 18 cells, respectively) and of the Keywords: migraine; headache; vascular pain; serotonin; cerebrovascular Correspondence to: Dr Peter J. Goadsby, Institute of Neurology, Queen Square, London, WCIN 3BG, UK trigeminal nucleus caudalis (TNC) (81 cells). Mechanical stimulation revealed a similar pattern of neural activation but with reduced intensity in laminae I and Ho of the TNC (38 cells) and of CI and C2 (32 and 31 cells, respectively). The temporalis muscle was stimulated mechanically in the control group and the numbers and distribution of Fospositive cells were no different from those in non-stimulated controls. Treatment with sumatriptan reduced the numbers of Fos-positive cells found in laminae I and Ho of the TNC and C2 (6, 13 cells and 9 cells, respectively) after mechanical stimulation. These data suggest that the neural effect of sumatriptan alone is sufficient for significant attenuation of transmission in the trigeminal system. The fact that sumatriptan can inhibit trigeminal activation without its vascular effects suggests that drugs without a significant activity on blood vessels may be effective in the treatment of migraine. Abbreviations: DAB = diaminobenzidine; 5HT = serotonin (5-hydroxytryptamine); SSS = superior sagittal sinus; TNC = trigeminal nucleus caudalis Introduction Headache in general, and migraine in particular, form a common and often debilitating series of problems whose pathophysiology is poorly understood (Goadsby, 1994). Laboratory (Goadsby et ai, 1991a) and recent clinical evidence (Weiller et al., 1995) points to the importance of the CNS in migraine. Recent therapeutic advances suggesting a role for 5HT B/D -like agonists in the management of acute attacks of migraine (Goadsby et al., 199\b\ Subcutaneous Sumatriptan International Study Group, 1991) have stimulated interest in this class of compounds and in their mode of action. While the rationale for the development of sumatriptan was as a crania! vasoconstrictor (Saxena, 1991) this mode of action has recently been vigorously debated (Humphrey and Goadsby, 1994). There is little doubt, based on clinical observation, that the pain in migraine in some way involves trigeminal structures (Lance, 1993). Moreover, the long-recognized vascular component of the syndrome has focused attention on the trigeminal innervation of the pain-sensitive intracranial structures, such as the dura mater and large vessels (Feindel et al., 1960; McNaughton, 1966). It has been shown during neurosurgical procedures that mechanical or electrical stimulation of the dura mater or blood vessels of conscious humans leads to pain referred to the ophthalmic (first) division Oxford University Press 1996

2 1420 K. L. Hoskin et al. of the trigeminal nerve. Because responses seen in cats are the same as those seen in humans (Goadsby et al., 1988), and more specific to the cranial circulation, we have turned to intracranial vessel stimulation to model trigeminovascular nociception. The structures chosen, the superior sagittal sinus and middle meningeal artery, have the advantage of being both pain-sensitive in humans (Wolff, 1963) and of being largely innervated by small unmyelinated C fibres (Keller et al., 1985). Stimulation of the superior sagittal sinus has been shown to alter CBF in the cat (Lambert et al., 1988) as well as lead to changes in neuropeptide levels (Zagami et al., 1990) similar to those seen in humans during migraine (Goadsby et al., 1990) and cluster headache (Goadsby and Edvinsson, 1994; Fanciullacci et al., 1995). Sumatriptan is clearly a potent and relatively specific cranial vasoconstrictor (Humphrey et al., 1991). Sumatriptan also potently blocks trigeminal-induced plasma protein extravasation (Moskowitz, 1992) and calcitonin gene-related peptide release in the cranial circulation (Goadsby and Edvinsson, 1993). However, Moskowitz observed that exogenously administered substance P induced a plasma protein extravasation that was not blocked by sumatriptan (Buzzi and Moskowitz, 1990) and, similarly, in a model of chemical meningitis that both dihydroergotamine and sumatriptan blocked trigeminal nucleus c-fos expression (Nozaki et al., 1992). Furthermore, the conformationally restricted analogue of sumatriptan, CP122,288, can block plasma protein extravasation at a dose at which it has no vascular effects (Lee and Moskowitz, 1993). These data taken together suggest that sumatriptan may act sufficiently at a prejunctional 5HT D -like receptor to inhibit neural transmission without necessarily having a vascular effect. In these studies the superior sagittal sinus was mechanically dilated to determine whether sumatriptan was still effective in attenuating trigeminal neural traffic. This unique method offers the possibility to separate the vasoactive and neural properties of this class of compounds. Methods Cats were anaesthetized initially with halothane and then occhloralose (60 mg kg" 1, intraperitoneally) (Sigma, St Louis, USA) and prepared for physiological monitoring. The femoral artery and vein were cannulated in order to measure blood pressure and heart rate and provide access for drug and fluid administration, respectively. Cardiovascular parameters and pupillary reaction to noxious pinching of the forepaw were used to determine the need for supplementary anaesthesia. The animals were endotracheally intubated, ventilated with 40% oxygen and paralysed after the surgical procedures with repeated doses of gallamine triethiodide (6 mg kg" 1 intravenously) (May and Baker, UK). Body temperature and end-expiratory CO 2 were monitored and maintained within physiological limits. The animals were mounted in a stereotactic frame and a circular midline craniotomy (2 cm in diameter) was performed for access to the superior sagittal sinus. The adjacent dura and falx were dissected parallel to the sinus over mm. To prevent dehydration and for electrical insulation against the cortex, a paraffin bath was built with a dam of dental acrylic around the craniotomy and, additionally, a small polyethylene sheet inserted under the vessel. Fluid (4% glucose with 0.18% saline or normal saline) was administered intravenously at a rate of 35 ml kg" 1 h" 1, while gallamine (6 mg kg" 1 intravenously) and a- chloralose (15 mg kg" 1 intravenously) were administered every 2 h. Prior to readministration of gallamine, adequacy of anaesthesia was monitored by observing cardiovascular and pupillary changes, and by checking the withdrawal reflex, after pinching the forepaw. Blood pressure and heart rate were stable and within physiological ranges for all animals throughout the whole experiment. Arterial blood gas parameters were monitored intermittently as a guide to the end-expiratory CO 2 output. Study design and stimulation Electrical stimulation Following completion of surgery the animal was maintained essentially undisturbed for the following 24 h. Animals were then randomized to have electrical stimulation or mechanical stimulation, the latter with or without treatment with sumatriptan (85 ig kg" 1, intravenously) or vehicle administered in a manner blind to the investigator performing the stimulation. After this resting phase the superior sagittal sinus (SSS) was stimulated. For electrical stimulation, the SSS was suspended over a pair of stainless steel hook electrodes. We have previously shown that this would, of itself, not provoke significant Fos activation (Kaube et al., 1993b). The SSS was stimulated with a Grass S88 stimulator driving a stimulus isolation unit (SIU5A, Grass Instruments, Quincy, Mass., USA; 150 V, 250 us duration) at a rate of 0.3 s" 1 for 1 h. Mechanical stimulation For mechanical stimulation a device was custom made and some of its characteristics have been described (Kaube et al., 1992). It consisted of a stainless steel stimulator with a vertical hollow shaft, held fixed to a small brass plate, with an inner thinner steel rod connected to a small solenoid. During stimulation the inner rod moved in relationship to the fixed outer piece creating a vertical displacement that distended the sinus. Both pieces were bent to 90, mounted on a stereotaxic carrier (Kopf Instruments, Tunjunga, Calif., USA) and inserted into the sinus under microscopic control after a small cut was made for entry. With practice, insertion was achieved with very minimal blood loss and certainly without effect upon resting cardiovascular parameters. The solenoid was driven by a function generator to allow modulation of the excursions of the device so that a clear expansion of the sinus could be seen macroscopically. The sinus was stretched with a sinusoidal pattern at a frequency

3 Mechanical stimulation of the sagittal sinus 1421 Table 1 Physiological data prior to stimulation (mean±sd) Weight (kg) Arterial blood pressure (mmhg) PH PCO 2 (mm Hg) po 2 (mm Hg) Electrical stimulation (n = 5) Mechanical stimulation (n = 6) 2.4± ± ±8 101±7 7.32± ± ±5 36±5 193±29 221±23 of 3 s" 1, also for a period of 1 h. After completion of the period of stimulation 1 h was allowed to elapse prior to perfusion. In control animals, mechanical stimulation was carried out on the skin overlying the anterior temporalis muscle to assess any possible vibration effect of the device. Perfusion Cats were perfused transcardially with of 0.9% saline after a bolus injection of 1000 IU of heparin and 0.5 ml of 1% sodium nitrite. This was followed by 2 1 of 4% paraformaldehyde in phosphate buffer (ph 7.4) and finally by ml of 30% sucrose solution in phosphate buffer. The brain and cervical spinal cord were removed and stored in 50% sucrose with azide. Coronal sections (40 u.m) of the caudal medulla and upper cervical spinal cord were cut on a freezing microtome and every fifth section was collected for processing. Sections were cut from a block beginning at the level of the obex and ending at the C3 segment of the cervical cord. The c-fos procedure Free-floating sections were incubated at 4 C for 3-7 days in a commercially available rabbit, polyclonal anti-body to Fos protein ('Ab-2\ Oncogene Science Ltd, Uniondale, NY, USA) in a 1:1000 dilution with 1% phosphate buffered horse serum, containing 0.1% bovine serum albumin and 0.2% Triton-XlOO. Fas-like immunoreactivity (hereafter simply called Fos) was visualized using standard avidin-biotin peroxidase immunohistochemical techniques. Following the primary incubation, sections were washed in 0.1 M phosphate buffer (ph 7.4) for 30 min and then incubated in a biotinylated goat anti-rabbit IgG (1:200 dilution) (Vector Labs, Peterborough, USA) for a minimum of 2 h at room temperature on a rotating table. Following the second incubation, the sections were washed again in 0. IM phosphate buffer (ph 7.4) for 30 min. The sections were incubated for 2.5 h in a 1:1000 dilution of ExtrAvidin-peroxidase (Sigma, London, USA) and then washed again in 0.1M phosphate buffer (ph 7.4) for 30 min. The sections were then incubated in 20 ml of 0.1 M phosphate buffer containing 0.05% diaminobenzidine (DAB) (Sigma), 0.005% of 4% ammonium chloride, 0.005% of 20% D-glucose and 0.02% of a 1% solution of nickel ammonium sulphate for 20 min (DAB reaction). The sections were then placed in a fresh identical 20 ml solution and 20 jxl of glucose oxidase (Sigma) was added to initiate the chromogenic reaction. The reaction was allowed to proceed until Fcs-positive nuclei could be clearly seen under the microscope. The DAB reaction product was visible as a black precipitate due to the presence of the nickel ammonium sulphate. Following this reaction the sections were washed two or three times in 0.1 M phosphate buffer (ph 7.4) to terminate the reaction and then mounted. The Fas-positive cells were distinguished from the background by their black nickel enhanced nuclei. Using the procedure adopted by Hammond et al. (1992), cells were only considered positive if the black precipitate of the DAB reaction within the cell nucleus was distinguishable from the background throughout a range of magnifications between X20 and X4. Fas-positive cells were plotted onto schematic drawings of the caudal medulla and upper cervical spinal cord modified from the atlases of Berman (1968) and Rexed (1954). Control incubations in the presence of the antigen were not carried out in this series of experiments. However, the omission of the primary antibody in related experiments performed in this laboratory did not produce positive staining. Furthermore, preabsorption of the antibody with Fos protein has demonstrated its specificity in other studies. Plotting and statistics Distributions of cells were quantified for each individual animal by taking 10 sections at random, from each of the levels (TNC, Cl, C2 and C3) and plotting the label from a single side on to one of the schematic sections described above. The plotting was performed by one person who, although they had knowledge of the experimental design, was not aware of the experimental group to which each animal belonged. The data are reported as a median with interquartile (25%, 75%) ranges and have been compared using a Mann-Whitney U test in view of the fact that the c-fos method as applied here can only yield non-continuous rather than interval data (Siegel, 1956). Results All animals included in this study were successfully maintained within normal physiological limits for the anaesthetized cat. There was no difference in physiological data between the group which received mechanical stimulation and the one which received electrical stimulation

4 1422 K. L. Hoskin et al. Table 2 Fos expression following electrical stimulation of the SSS and for mechanical controls* Table 3 Effect <Df sumatriptan on Fos expression after mechanical stimulation of the superior sagittal1 sinus Region Electrica1 stimulation (n = 5) Mechanical control (n = :I) Region Mechanica1 stimulation. (71 = 4) Mechanical stimulation with sumatriptan (TI = 3) Median Ranged Median Ranged Median Range 1 ' Median Range 1 Caudal medulla cnts* Cl C2 C *In mechanical controls the skin and muscle of the temporoparietal region was stimulated mechanically; ^interquartile range; *cnts = medial nucleus of the solitary tract. (Table 1). Some of the cats included here have provided data for a previous study (Kaube et al., 1993b) as this general group of studies is ongoing. In laminae I/IIo of the trigeminal nucleus caudalis (TNC) and Cl and C2 cervical spinal cord, electrical stimulation produced a median of 81 (range ), 88 (range ) and 92 (range 70-97) fas-positive cells, respectively (Table 2), compared with unstimulated median control levels of seven, five and two cells positive, respectively. Mechanical stimulation of the SSS (n = 4) evoked a marked and reproducible Fos expression that was very similar to that seen with electrical stimulation, although less dense; in the superficial laminae of the caudal TNC the median number of cells was 38 (range 33^43) and in the dorsal horn of Cl, 32 cells (range 25^0) and of C2, 31 cells (range 19-42). This mechanical stimulation of the SSS produced no more than control levels of Fos expression at the level of C3. The positive control stimulus (consisting of mechanical stimulation the skin and musculature of the temporo-parietal region; n = 2) produced no significant levels of Fos expression (Fig. 1). Effect of sumatriptan Treatment with sumatriptan (85 jag kg" 1, intravenously; n = 3) essentially blocked the expression of Fos induced by mechanical stimulation of the superior sagittal sinus in all the above-mentioned areas. In the TNC, C and C2 (laminae I/IIo) the median number of cells detected fell to 6 (range 5-6; P < 0.05), 13 (range 11-19; P < 0.05) and 9 (range 8-9; P < 0.05), respectively (see Table 3 and Figs 2 and 3). Discussion These data demonstrate that activation of the neuronal 5HT ]D - like receptor can inhibit trigeminally mediated neural traffic. Caudal medulla cnts* Cl C2 C ^ ^ ^1 3-6 *P < 0.05; interquartile range; *cnts solitary tract. 6* 11 13* 7 9* = medial nucleus of the The model employed activates vascular nociceptive afferents in the trigeminal nucleus and this activity is monitored by c-fos immunohistochemistry. Both electrical and mechanical stimulation result in robust expression of c-fos which is well above control levels. Mechanical stimulation of the superior sagittal sinus allows activation of the trigeminovascular pathway through a more physiological transduction mechanism. The fact that the 5HTi B/D -like receptor agonist sumatriptan inhibits activation of trigeminal afferents normally evoked by mechanical distension suggests that the vascular action of this class of compounds is not required for their trigeminal neuronal inhibitory effects. The interpretation of these data is relatively straightforward, although several elements are required to describe the model system and its relevance to migraine. The Fos procedure is not specific for painful stimuli and the relevance of the data results from the model system employed rather than the method of analysis. It is crucial that the model employed has relevance to human vascular headaches as it does in these studies. The superior sagittal sinus has been used to activate the trigeminovascular system for several reasons. It is known that the sinus is pain-sensitive in humans. Electrical or mechanical displacement of the sinus and surrounding dura mater is reported as painful in patients (Feindel et al., 1960; McNaughton and Feindel, 1977). Moreover, its innervation is by small fibres (Penfield and McNaughton, 1940; Kimmel, 1961; Keller et al., 1985) as is much of the dura mater in most species studied (Keller and Marfurt, 1991). The trigeminal innervation of the sinus and dura mater is in turn taken by the ophthalmic division of the trigeminal nerve (Steiger et al., 1982) via the trigeminal ganglion to the trigeminal nucleus. Similarly, the large cranial vessels are pain-sensitive (Ray and Wolff, 1940; Wolff, 1963; Martins et al., 1993) and their innervation is also predominantly through the ophthalmic division of the trigeminal nerve (Liu-Chen et al., 1983). This pattern of

5 Active Control Mechanical stimulation of the sagittal sinus 1423 Legend TNC Cl C2 C3 Fig. 1 Individual Fes-positive cells in the control animals in which the temporalis muscle was stimulated mechanically. These Fos-positive cells (.) were from 40 a.m sections of the caudal medulla (TNC) and Cl, C2 and C3 cervical spinal cord segments. The modest number of Fos-positive cells here is similar to the number and distribution seen in unstimulated animals. Some anatomically relevant areas are indicated on the right: I/IIo = laminae I and IIo of the dorsal horn; III/IV = laminae III and IV; X = lamina X; Cu = nucleus cuneatus; cnts = medial nucleus of the solitary tract; Gr = nucleus gracilis; LCN = lateral cervical nucleus; NRA = nucleus retroambigualis; INT = intermediate grey matter; VH = ventral horn; NRA = nucleus retroambigualis.

6 1424 K. L. Hoskin et al. Mechanical Stimulation Mechanical Stimulation and Sumatriptan TNC Cl C2 C3 Fig. 2 The effect of sumatriptan on Fos activation by mechanical stimulation. Plots of Fas-positive cells (.) at the level of the (TNC and Cl, C2 and C3 cervical spinal cord. Mechanical stimulation evokes a pattern of Fos activation very similar to that seen with electrical stimulation (Kaube el al., 19936). This activation pattern is blocked after administration of sumatriptan. intracranial vascular innervation is thought to underlie the expression of migraine as an often throbbing or pulsating pain. The superior sagittal sinus has thus been chosen as a structure for stimulation because of its known pain-sensitivity and because its size and position makes it accessible for study. Stimulation of the superior sagittal sinus results in activation of neurons predominantly in the most caudal part of the trigeminal nucleus caudalis and in the dorsal horn of the Cl and C2 cervical spinal segments. This has been demonstrated by monitoring cellular activity (Lambert et al.,

7 Mechanical stimulation of the sagittal sinus 1425 Mechanical Stimulation S o VI VI 1 VI a. o s B B \ i i r Mechanical Stimulation 30 min post Sumatriptan TNC Cl C2 Fig. 3 Box plots (median with quartile ranges) demonstrating the effect of mechanical stimulation of the SSS on various regions of the caudal medulla: TNC, Cl and C2 cervical spinal cord segments studied using the Fos method. The median number of cells found in each area for the entire cohort of animals studied is plotted. The effect of mechanical stimulation is blocked by sumatriptan. I/IIo = laminae I and IIo of the dorsal horn; III/IV = laminae III and IV; cnts = medial nucleus of the solitary tract; NRA = nucleus retroambigualis; RF = reticular formation; VH = ventral horn; X = lamina X. 1988; Goadsby and Hoskin, 1996; Hoskin et al., 1996), 2-deoxyglucose metabolism (Goadsby and Zagami, 1991) and c-fos immunohistochemistry (Kaube et al., 1993/?; Hoskin et al., 1996). Activation of these cells produces changes in cerebral blood flow (Lambert et al., 1988) just as does trigeminal stimulation in cat (Goadsby and Duckworth, 1987) and humans (Tran-Dinh et al., 1992). Moreover, trigeminal ganglion stimulation (Goadsby et al., 1988; Buzzi et al., 1991) and superior sagittal sinus stimulation lead to release of sensory neuropeptides into the cranial circulation (Zagami et al., 1990) in a similar manner to that seen in migraine (Goadsby et al., 1990; Gallai et al., 1995) and cluster headache (Goadsby and Edvinsson, 1994; Fanciullacci et al., 1995). Superior sagittal sinus stimulation, therefore, results in a pattern of activation of the trigeminal system that approximates that seen in migraine. The introduction into therapeutic use of the novel antimigraine agent sumatriptan and the demonstration of its utility in treating migraine attacks (Goadsby et al., 1991b; Subcutaneous Sumatriptan International Study Group, 1991) has resulted in considerable discussion as to its mode of action. The compound was developed as a 5HT agonist to constrict the dog saphenous vein and close cranial arteriovenous anastomoses (Humphrey et al., 1991; Saxena, 1991) on the basis that migraine was a primarily vascular disorder (Wolff, 1963). Due to this design, there is a very small, yet well-recognized risk of cardiovascular problems associated with the use of drugs from this class of compounds. It has emerged that the 5HT m receptor (Hoyer et al., 1994; Hartig et al., 1996) at which sumatriptan is active may be present on trigeminal nerves. It has been shown that while sumatriptan inhibits protein plasma extravasation in the dura mater elicited by trigeminal ganglion stimulation, substance P-elicited plasma protein extravasation is not affected (Buzzi et al., 1991). Similarly, sumatriptan blocks release of a peptide marker of trigeminal activation, calcitonin generelated peptide, in rat (Buzzi et al., 1991) and cat (Goadsby and Edvinsson, 1993). Moreover, sumatriptan blocks calcitonin gene-related peptide release in migraine sufferers if their headache is aborted by the drug (Goadsby and

8 1426 K. L. Hoskin et al. Edvinsson, 1993). Finally, the development of conformationally restricted analogues of sumatriptan that can block plasma protein extravasation at doses at which they have no vascular effects (Lee and Moskowitz, 1993) raises the possibility that the two effects could be dissociated in the clinical context too. Since it is known that sumatriptan acts at both the 5HT 1B and 5HT, D receptors (Weinshank et al, 1992; Hoyer et al., 1994) it may be possible to design drugs that affect only one receptor and selectively target the trigeminal nerve (Rebeck et al., 1994), although recent data suggest that both receptor subtypes may be present in both cranial vessels and nerves in different amounts (E. Hamel, personal communication). Pharmacological approaches to dissecting the 5HT 1B/D -like receptors have been disappointing. This reflects, in part, species variations which have been addressed in a recent reclassification (Hartig et al., 1996) and the possibility that there is as yet an undiscovered receptor. The paradox remains that, whereas the vasoconstrictor effect of sumatriptan is blocked by methiothepin (Hamel and Bouchard, 1991), a potent but not specific 5HT 1B/D antagonist (Peroutka and McCarthy, 1989; Schoeffter and Hoyer, 1989), the response is not antagonized by metergoline either in vivo (Perren et al., 1991; Villalon et al., 1992) or in vitro (Bax et al., 1992), although metergoline is more potent than methiothepin at the 5HT 1D receptor (Schoeffter et al, 1988; Waeber et al, 1988). These data may simply be an amalgam of species variations and varying antagonist potencies of metergoline and methiothepin, although the intriguing possibility remains that a further 5HTl-like receptor is yet to be discovered. The development of the potent 5HT 1D antagonist GR (Clitherow et al, 1994) does not hold an immediate key to this issue since it has some agonist activity itself (Pauwels and Colpaert, 1995). Moreover, the specificity of GR relates to what is known and if there is a further, as yet undefined, receptor in the 5HT1 class, GR may also be active at that site. The experiments reported here were undertaken since it has been impossible to disentangle the vascular and neuronal effects of sumatriptan completely in vivo (Humphrey and Goadsby, 1994). In this study, the sinus was held apart by the stimulator so that no vasoconstrictor effect could have contributed to the inhibition of c-fos expression seen. The simplest reasonable conclusion is that sumatriptan inhibited trigeminal afferents by a direct neuronal mechanism at the peripheral terminal. An action at the trigeminal ganglion is unlikely since direct application of sumatriptan to trigeminal ganglion cells does not inhibit their activity (O'Shaughnessy et al, 1993). Moreover, since it has been shown that to affect central trigeminal neurons, sumatriptan must be allowed to access such sites by artificial disruption of the blood-brain barrier (Kaube et al, 1993a; Shepheard et al, 1995), central inhibition could not account for these results. In summary, these studies demonstrate that mechanical distension of a pain-sensitive intracranial structure is capable of activating neurons in the caudal trigeminal nucleus caudalis and dorsal horns of the Cl and C2 cervical spinal cord. In the presence of a mechanical distension that would block any possible vasoconstrictor action of sumatriptan, inhibition of trigeminal neurons is seen. The only site available to mediate such an action would then be a neuronal site, probably on the trigeminal nerve terminals. These data suggest that a purely neuronally active inhibitor of trigeminal nerve responses would be capable of blocking trigeminal nucleus activation. Whether neuronally selective drugs would be effective in migraine and whether they would be superior to the current first generation 5HT 1B/D -like drugs remains a question that must be answered in the clinic with migraine patients. These data do not establish that nonvascular drugs would be superior but they do provide a rationale for their development and study in the clinic. The prospect of eliminating vascular effects from the management of migraine is sufficiently important alone to justify a clinical study of such compounds. Acknowledgments These studies arose from a valuable discussion and challenge from Professor P. P. A. Humphrey (Humphrey and Goadsby, 1994). The authors acknowledge the valuable collaboration of Dr Kevin Keay in our helping us establish the methodology. These studies have been supported by grants from the Wellcome Trust and the Migraine Trust. P.J.G. is a Wellcome Senior Research Fellow. These data have been presented in preliminary form at the Vllth International Headache Congress, Toronto, September, 1995 (Hoskin et al, 1995). References Bax WA, Van Heuven-Nolsen D, Simoons ML, Saxena PR. 5- Hydroxytryptamine-induced contractions of the human isolated saphenous vein: involvement of 5-HT 2 and 5-HT 1D -like receptors, and a comparison with grafted veins. Naunyn Schmiedebergs Arch Pharmacol 1992; 345: Berman AN. The brain stem of the cat. A cytoarchitectonic atlas with stereotaxic co-ordinates. Madison: University of Wisconsin Press, Buzzi MG, Moskowitz MA. The antimigraine drug, sumatriptan (GR43175), selectively blocks neurogenic plasma extravasation from blood vessels in dura mater. Br J Pharmacol 1990; 99: Buzzi MG, Carter WB, Shimizu T, Heath H, Moskowitz MA. Dihydroergotamine and sumatriptan attenuate levels of calcitonin gene-related peptide in plasma in rat superior sagittal sinus during electrical stimulation of the trigeminal ganglion. Neuropharmacology 1991; 30: Clitherow JW, Scopes DI, Skingle M, Jordan CC, Feniuk W, Campbell IB, et al. Evolution of a novel series of [(N,Ndimethylamino) propyl]- and piperazinylbenzanilides as the first selective 5-HT 1D antagonists. J Med Chem 1994; 37: Fanciullacci M, Alessandri M, Figini M, Geppetti P, Michelacci

9 S. Increase in plasma calcitonin generelated peptide from the extracerebral circulation during nitroglycerin-induced cluster headache attack. Pain 1995; 60: Feindel W, Penfield W, McNaughton F. The tentorial nerves and localization of intracranial pain in man. Neurology 1960; 10: Gallai V, Sarchielli P, Floridi A, Franceschini M, Codini M, Glioti G, et al. Vasoactive peptide levels in the plasma of young migraine patients with and without aura assessed both interictally and ictally. Cephalalgia 1995; 15: Goadsby PJ. The challenge of headache for the 1990s. Curr Opin Neural 1994; 7: Goadsby PJ, Duckworth JW. Effect of stimulation of trigeminal ganglion on regional cerebral blood flow in cats. Am J Physiol 1987; 253: R Goadsby PJ, Edvinsson L. The trigeminovascular system and migraine: studies characterizing cerebrovascular and neuropeptide changes seen in humans and cats. Ann Neurol 1993; 33: Goadsby PJ, Edvinsson L. Human in vivo evidence for trigeminovascular activation in cluster headache. Brain 1994; 117: Goadsby PJ, Hoskin KL. Inhibition of trigeminal neurons by intravenous administration of the serotonin (5HT)-1-D receptor agonist zolmitriptan (311C90): are brain stem sites a therapeutic target in migraine? Pain In press. Goadsby PJ, Zagami AS. Stimulation of the superior sagittal sinus increases metabolic activity and blood flow in certain regions of the brainstem and upper cervical spinal cord of the cat. Brain 1991; 114: Goadsby PJ, Edvinsson L, Ekman R. Release of vasoactive peptides in the extracerebral circulation of humans and the cat during activation of the trigeminovascular system. Ann Neurol 1988; 23: Goadsby PJ, Edvinsson L, Ekman R. Vasoactive peptide release in the extracerebral circulation of humans during migraine headache. Ann Neurol 1990; 28: Goadsby PJ, Zagami AS, Lambert GA. Neural processing of craniovascular pain: a synthesis of the central structures involved in migraine. Headache 1991a; 31: Goadsby PJ, Zagami AS, Donnan GA, Symington G, Anthony M, Bladin PF, et al. Oral sumatriptan in acute migraine. Lancet 1991b; 338: Hamel E, Bouchard D. Contractile 5-HT, receptors in human isolated pial arterioles: correlation with 5HT D binding sites. Br J Pharmacol 1991; 102: Hammond DL, Presley R, Gogas KR, Basbaum Al. Morphine or U-50, 488 suppresses Fos protein-like immunoreactivity in the spinal cord and nucleus tractus solitarii evoked by a noxious visceral stimulus in the rat. J Comp Neurol 1992; 315: Hartig PR, Hoyer D, Humphrey PPA, Martin GR. Alignment of receptor nomenclature with the humam genome: classification of 5HT-1B and 5-HT1D receptor subtypes. Trends Pharmacol Sci 1996; 17: Hoskin KL, Kaube H, Goadsby PJ. Mechanical distension of the Mechanical stimulation of the sagittal sinus 1427 superior sagittal sinus evokes c-fos expression in trigeminal neuronstowards a better model of migraine [abstract]. Cephalalgia 1995; 15 Suppl 14: 104. Hoskin KL, Kaube H, Goadsby PJ. Central activation of the trigeminovascular pathway in the cat is inhibited by dihydroergotamine: a c-fos and electrophysiological study. Brain 1996; 119: Hoyer D, Clarke DE, Fozard JR, Hartig PR, Martin GR, Mylecharane EJ, et al. International Union of Pharmacology classification of receptors for 5-hydroxytryptamine (Serotonin). [Review]. Pharmacol Rev 1994; 46: Humphrey PPA, Goadsby PJ. The mode of action of sumatriptan is vascular? A debate. [Review]. Cephalalgia 1994; 14: Humphrey PPA, Feniuk W, Marriott AS, Tanner RJN, Jackson MR, Tucker ML. Preclinical studies on the anti-migraine drug, sumatriptan. [Review]. Eur Neurol 1991; 31: Kaube H, Hoskin KL, Goadsby PJ. Activation of the trigeminovascular system by mechanical distension of the superior sagittal sinus in the cat [see comments]. Cephalalgia 1992; 12: Comment in: Cephalalgia 1992; 12: 127. Kaube H, Hoskin KL, Goadsby PJ. Inhibition by sumatriptan of central trigeminal neurones only after blood-brain barrier disruption. Br J Pharmacol 1993a; 109: Kaube H, Keay KA, Hoskin KL, Bandler R, Goadsby PJ. Expression of c-fos-like immunoreactivity in the caudal medulla and upper cervical spinal cord following stimulation of the superior sagittal sinus in the cat. Brain Res 1993b; 629: Keller JT, Marfurt CF. Peptidergic and serotoninergic innervation of the rat dura mater. J Comp Neurol 1991; 309: Keller JT, Saunders MC, Beduk A, Jollis JG. Innervation of the posterior fossa dura of the cat. Brain Res Bull 1985; 14: Kimmel DL. The nerves of the cranial dura mater and their significance in dural headache and referred pain. Clin Med School Q 1961; 22: Lambert GA, Goadsby PJ, Zagami AS, Duckworth JW. Comparative effects of stimulation of the trigeminal ganglion and the superior sagittal sinus on cerebral blood flow and evoked potentials in the cat. Brain Res 1988; 453: Lance JW. Mechanism and management of headache. 5th ed. London: Butterworth-Heinemann, Lee WS, Moskowitz MA. Conformationally restricted sumatriptan analogues, CP-122,288 and CP-122,638 exhibit enhanced potency against neurogenic inflammation in dura mater. Brain Res 1993; 626: Liu-Chen LY, Mayberg MR, Moskowitz MA. Immunohistochemical evidence for a substance P-containing trigeminovascular pathway to pial arteries in cats. Brain Res 1983; 268: Martins IP, Baeta E, Paiva T, Campos J, Gomes L. Headaches during intracranial endovascular procedures: a possible model of vascular headache. Headache 1993; 33: McNaughton FL. The innervation of the intracranial blood vessels and the dural sinuses. In: Cobb S, Frantz AM, Penfield W, Riley

10 1428 K. L. Hoskin et al. HA, editors. The circulation of the brain and spinal cord. New York: Hafner Publishing, 1966: McNaughton FL, Feindel WH. Innervation of intracranial structures: a reappraisal. In: Rose FC, editor. Physiological aspects of clinical neurology. Oxford: Blackwell Scientific, 1977: Moskowitz MA. Neurogenic versus vascular mechanisms of sumatriptan and ergot alkaloids in migraine. [Review]. Trends Pharmacol Sci 1992; 13: Nozaki K, Moskowitz MA, Boccalini P. CP-93,129, sumatriptan, dihydroergotamine block c-fos expression within rat trigeminal nucleus caudalis caused by chemical stimulation of the meninges. Br J Pharmacol 1992; 106: O'Shaughnessy CT, Connor HE, Feniuk W. Extracellular recordings of membrane potential from guinea-pig isolated trigeminal ganglion: lack of effect of sumatriptan. Cephalalgia 1993; 13: Pauwels PJ, Colpaert FC. The 5-HT m receptor antagonist GR 127,935 is an agonist at cloned human 5-HT 1Da receptor sites. Neuropharmacology 1995; 34: Penfield W, McNaughton FL. Dural headache and innervation of the dura mater. Arch Neurol Psychiatry 1940; 44: Peroutka SJ, McCarthy BG. Sumatriptan (GR 43175) interacts selectively with 5-HTJB and 5-HT 1D binding sites. Eur J Pharmacol 1989; 163: Perren MJ, Feniuk W, Humphrey PPA. Vascular 5-HT r like receptors that mediate contraction of the dog isolated saphenous vein and carotid arterial vasoconstriction in anaesthetized dogs are not of the 5-HT, A or 5-HT 1D subtype. Br J Pharmacol 1991; 102: Ray BS, Wolff HG. Experimental studies on headache. Pain sensitive structures of the head and their significance in headache. Arch Surg 1940; 41: Rebeck GW, Maynard KI, Hyman BT, Moskowitz MA. Selective 5-HT1D alpha serotonin receptor gene expression in trigeminal ganglia: implications for antimigraine drug development. Proc Natl Acad Sci USA 1994; 91: Rexed B. A cytoarchitectonic atlas of the spinal cord in the cat. J Comp Neurol 1954; 96: Saxena PR. 5-HT in migraine - an introduction.[review]. J Neurol 1991; 238 Suppl 1: S36-7. Schoeffter P, Hoyer D. How selective is GR 43175? Interactions with functional 5-HT1A, 5-HT1B, 5-HT1C and 5-HT1D receptors. Naunyn Schmiedebergs Arch Pharmacol 1989; 340: Schoeffter P, Waeber C, Palacios JM, Hoyer D. The 5-hydroxytryptamine 5-HTi D receptor subtype is negatively coupled to adenylate cyclase in calf substantia nigra. Naunyn Schmiedebergs Arch Pharmacol 1988; 337: Shepheard SL, Williamson DJ, Williams J, Hill RG, Hargreaves RJ. Comparison of the effects of sumatriptan and the NK1 antagonist CP-99,994 on plasma extravasation in the dura mater and c-fos mrna expression in the trigeminal nucleus caudalis of rats. Neuropharmacology 1995; 34: Siegel S. Non-parametric statistics for the behavioral sciences. New York: McGraw-Hill, Steiger HJ, Tew JM, Keller JT. The sensory representation of the dura mater in the trigeminal ganglion of the cat. Neurosci Lett 1982; 31: Subcutaneous Sumatriptan International Study Group. Treatment of migraine attacks with sumatriptan. N Engl J Med 1991; 325: Tran Dinh YR, Thurel C, Cunin G, Seme A, Seylaz J. Cerebral vasodilation after the thermocoagulation of the trigeminal ganglion in humans. Neurosurgery 1992; 31: Villalon CM, Bom AH, Den Boer MO, Heiligers JP, Saxena PR. Effects of S9977 and dihydroergotamine in an animal experimental model for migraine. Pharmacol Res 1992; 25: Waeber C, Schoeffter P, Palacios JM, Hoyer D. Molecular pharmacology of 5-HT 1D recognition sites: radioligand binding studies in human, pig and calf brain membranes. Naunyn Schmiedebergs Arch Pharmacol 1988; 337: Weiller C, May A, Limmroth V, Juptner M, Kaube H, Schayck RV, Coenen HH, Diener HC. Brain stem activation in spontaneous human migraine attacks. Nat Med 1995; 1: Weinshank RL, Zgombick JM, Macchi MJ, Branchek TA, Hartig PR. Human serotonin 1D receptor is encoded by a subfamily of two distinct genes: 5-HT 1Da and 5-HT 1D p. Proc Natl Acad Sci USA 1992; 89: 3630-^. Wolff HG. Headache and other head pain. 2nd ed. New York: Oxford University Press, Zagami AS, Goadsby PJ, Edvinsson L. Stimulation of the superior sagittal sinus in the cat causes release of vasoactive peptides. Neuropeptides 1990; 16: Received February 9, Revised May 1, Accepted May 24, 1996