Serotonin Receptor Ligands: Treatments of Acute Migraine and Cluster Headache P. J. Goadsby Institute of Neurology, Queen Square, London WC1N 3BG, UK
Abstract Fuelled by the development of the serotonin 5-HT1B/1D receptor agonists, the triptans, the last 15 years has seen an explosion of interest in the treatment of acute migraine and cluster headache. Sumatriptan was the ﬁrst of these agonists, and it launched a wave of therapeutic advances. These medicines are effective and safe. Triptans were developed as cranial vasoconstrictors to mimic the desirable effects of serotonin, while avoiding its side- effects. It has subsequently been shown that the triptans’ major action is neuronal, with both peripheral and central trigeminal inhibitory effects, as well as actions in the thalamus and at central modulatory sites, such as the periaqueductal grey matter. Further reﬁnements may be possible as the 5-HT1D and 5-HT1F receptor agonists are explored. Serotonin receptor pharmacology has contributed much to the better management of patients with primary headache disorders.
Keywords Triptans · Trigeminovascular · Migraine · Cluster headache
The last 15 years has seen an explosion of interest in the treatment of acute migraine. The description of serotonin 5-HT1B/1D receptor agonists, triptans (Goadsby 2000), as acute anti-migraine compounds (Doenicke et al. 1987), and the subsequent validation of sumatriptan (Doenicke et al. 1988) launched a wave of therapeutic advances. Many countries now have a triptan available and some have up to seven. These medicines are safe (Dodick et al. 2004) and effective (Ferrari et al. 2001); they have done much good both for patients and for the headache ﬁeld more broadly.
Triptans were developed as cranial vasoconstrictors to mimic the desirable effects of serotonin (Kimball et al. 1960; Lance et al. 1967), while avoiding its side-effects (Humphrey et al. 1990). The developmental rationale (Humphrey et al. 1990) was based on clinical data, i.e. that of serotonin and methysergide (Lance et al. 1963), and a pre-clinical rationale of cranial vasoconstriction, speciﬁcally selective constriction of arteriovenous anastomoses (Johnston and Saxena 1978). However, as the 1990s unfolded, views on migraine have changed. It is now generally accepted that migraine is a disorder of brain (Goadsby et al.
2002), with neurovascular changes being secondary to brain activity. I will review here the models used for triptan development, explore the speciﬁc pharmacology and experimental data regarding triptans, using a largely neu- roanatomical approach (Table 1), and then review the possibility for even more speciﬁc 5-HT ligands in migraine.
The Trigeminovascular System
The brain itself is largely insensate and, with some exceptions (Raskin et al.
1987; Veloso et al. 1998), does not seem to produce headache when stimulated (Wolff 1948). The pain-producing innervation of the intracranial contents is largely supplied by branches of the ophthalmic (ﬁrst) division of the trigeminal nerve (Penﬁeld 1932a, b, 1934; Penﬁeld and McNaughton 1940; Feindel et al.
1960; McNaughton 1966). The crucial structures that produce pain seem to be the dura mater and the large intracranial blood vessels (Wolff 1948). Electrical
stimulation of large venous sinuses, such as the superior sagittal sinus (SSS), will produce pain (Ray and Wolff 1940) although, remarkably, mechanical stimulation is much less likely to do so (Wolff 1948). Because these structures—
dura mater and dural/large intracranial vessels—are pain-producing, they have been used to model trigeminovascular nociception (Edvinsson 1999), thus helping us make predictions concerning new anti-migraine compounds.
Modelling the Trigeminovascular System
Several approaches are used to model the trigeminovascular system. One can look at pre- and post-junctional targets peripherally, i.e. the nerve-vessel inter-
Serotonin Receptor Ligands: Treatments of Acute Migraine and Cluster Headache
a Central nervous system sites where serotonin 5-HT1B/1D receptor agonists modulate trigeminovascular nociceptive processing
face, or pre- and post-synaptic targets centrally, i.e. the second order neuron interface (De Vries et al. 1999; Goadsby and Kaube 2000). The ﬁrst model for post-junctional craniovascular targets was the arteriovenous anastomoses (Saxena and De Boer 1991), which at the time were considered as a target inde- pendent of the trigeminal nerve as such. Interestingly, a pre-junctional action for ergotamine is not a new concept (Saxena and Cairo-Rawlins 1979). Sub- sequently, the plasma-protein extravasation model was developed (Markowitz et al. 1987) and used to explore anti-migraine compounds (Markowitz et al.
1988). The intravital model of Williamson and Hargreaves (Williamson et al.
1997) studies both pre- and post-junctional effects of compounds and has been very useful in evaluating new receptor targets. At the other end of the trigemi- novascular neuron one can study pre-synaptic and post-synaptic events in the trigeminocervical complex to begin to infer possible brain effects of anti- migraine compounds (Kaube et al. 1992). These approaches have all been exploited to characterise the effects of various transmitters on the trigemino- vascular system, particularly the pre-junctional/pre-synaptic effects that may be a crucial part of the action of these compounds.
Triptans, 5-HT1B/1D Receptor Agonists: Peripheral Pharmacology
Plasma Protein Extravasation
Moskowitz (1990) has provided a series of experiments to suggest that the pain of migraine may be a form of sterile neurogenic inﬂammation. Although many aspects of this view seem clinically implausible, the model system has been helpful in understanding some aspects of trigeminovascular pharmacology. Neurogenic plasma extravasation can be seen during electrical stimulation of the trigeminal ganglion in the rat (Markowitz et al. 1987). Plasma extravasation can be blocked by ergot alkaloids, indomethacin, acetylsalicylic acid, and the serotonin-5-HT1B/1D agonist, sumatriptan (Moskowitz and Cutrer 1993). The pharmacology of abortive anti-migraine drugs has been reviewed in detail (Cutrer et al. 1997). In addition, there are structural changes in the dura mater that are observed after trigeminal ganglion stimulation. These include mast cell degranulation and changes in post-capillary venules including platelet aggregation (Dimitriadou et al. 1991, 1992). While it is generally accepted that such changes, and particularly the initiation of a sterile inﬂammatory response, would cause pain (Strassman et al. 1996; Burstein et al. 1998), it is not clear whether this is sufﬁcient of itself or requires other stimulators or promoters. Preclinical studies suggest that cortical spreading depression may be a sufﬁcient stimulus to activate trigeminal neurons (Bolay et al. 2002), although this has been a controversial area (Moskowitz et al. 1993; Ingvardsen et al. 1997, 1998; Ebersberger et al. 2001; Goadsby 2001).
Although plasma extravasation in the retina, which is blocked by sumatrip- tan, can be seen after trigeminal ganglion stimulation in experimental animals, no changes are seen with retinal angiography during acute attacks of migraine or cluster headache (May et al. 1998b). A limitation of this study was the prob- able sampling of both retina and choroid elements in rat, given that choroidal vessels have fenestrated capillaries (Steuer et al. 2004). Clearly, however, block- ade of neurogenic plasma protein extravasation is not completely predictive of anti-migraine efﬁcacy in humans as evidenced by the failure in clinical trials of substance P, neurokinin-1 antagonists (Goldstein et al. 1997; Connor et al.
1998; Norman et al. 1998; Diener and The RPR100893 Study Group 2003), spe- ciﬁc blockers of dural plasma protein extravasation (PPE) [CP122,288 (Roon et al. 1997) and 4991w93 (Earl et al. 1999)], an endothelin antagonist (May et al.
1996) and a neurosteroid (Data et al. 1998).
Sensitisation and Migraine
While it is highly doubtful that there is a signiﬁcant sterile inﬂammatory response in the dura mater during migraine, it is clear that some form of
sensitisation takes place during migraine, since allodynia is common. About two-thirds of patients complain of pain from non-noxious stimuli (Selby and Lance 1960; Burstein et al. 2000a,b). Sensitisation in migraine may be periph- eral with local release of inﬂammatory markers, which would certainly activate trigeminal nociceptors (Strassman et al. 1996). More likely in migraine there is a form of central sensitisation—which may be classical central sensitisation (Burstein et al. 1998)—or a form of disinhibitory sensitisation with dysfunc- tion of descending modulatory pathways (Knight et al. 2002). Just as dihy- droergotamine (DHE) can block trigeminovascular nociceptive transmission (Hoskin et al. 1996), probably at least by a local effect in the trigeminocervical complex (Lambert et al. 1992; Storer and Goadsby 1997), DHE can block cen- tral sensitisation associated with dural stimulation by an inﬂammatory soup (Pozo-Rosich and Oshinsky 2005).
Triptans, 5-HT1B/1D Receptor Agonists: Central Pharmacology
The Trigeminocervical Complex
Fos immunohistochemistry is a method for looking at activated cells by plot- ting the expression of Fos protein (Morgan and Curran 1991). After meningeal irritation with blood, Fos expression is noted in the trigeminal nucleus cau- dalis (Nozaki et al. 1992), while after stimulation of the superior sagittal sinus, Fos-like immunoreactivity is seen in the trigeminal nucleus caudalis and in the dorsal horn at the C1 and C2 levels in the cat (Kaube et al. 1993c) and monkey (Goadsby and Hoskin 1997; Hoskin et al. 1999). These latter ﬁndings are in ac- cordance with similar data using 2-deoxyglucose measurements with superior sagittal sinus stimulation (Goadsby and Zagami 1991). Similarly, stimulation of a branch of C2 , the greater occipital nerve, increases metabolic activity in the same regions, i.e. trigeminal nucleus caudalis and C1/2 dorsal horn (Goadsby et al. 1997). In experimental animals one can record directly from trigeminal neurons with both supratentorial trigeminal input and input from the greater occipital nerve, a branch of the C2 dorsal root (Bartsch and Goadsby 2002). Stimulation of the greater occipital nerve for 5 min results in substantial in- creases in responses to supratentorial dural stimulation, which can last for over an hour (Bartsch and Goadsby 2002). Conversely, stimulation of the middle meningeal artery dura mater with the C-ﬁbre irritant mustard oil sensitises responses to occipital muscle stimulation (Bartsch and Goadsby 2003). Taken together these data suggest convergence of cervical and ophthalmic inputs at the level of the second order neuron. Moreover, stimulation of a lateralised structure, the middle meningeal artery, produces Fos expression bilaterally in both cat and monkey brain (Hoskin et al. 1999). This group of neurons from
the superﬁcial laminae of trigeminal nucleus caudalis and C1/2 dorsal horns should be regarded functionally as the trigeminocervical complex.
These data demonstrate that trigeminovascular nociceptive information
comes by way of the most caudal cells. This concept provides an anatomi- cal explanation for the referral of pain to the back of the head in migraine. Moreover, experimental pharmacological evidence suggests that some abortive anti-migraine drugs, such as, ergot derivatives (Lambert et al. 1992; Hoskin et al. 1996), acetylsalicylic acid (Kaube et al. 1993b), sumatriptan (Kaube et al.
1993a; Levy et al. 2004), eletriptan (Goadsby and Hoskin 1999; Lambert et al.
2002), naratriptan (Goadsby and Knight 1997; Cumberbatch et al. 1998), riza- triptan (Cumberbatch et al. 1997) and zolmitriptan (Goadsby and Hoskin
1996) can have actions at these second order neurons that reduce cell activity and suggest a further possible site for therapeutic intervention in migraine.
This action can be dissected out to involve each of the 5-HT1B, 5-HT1D and
5-HT1F receptor subtypes (Goadsby and Classey 2003), and are consistent with
the localisation of these receptors on peptidergic nociceptors (Potrebic et al.
2003). Interestingly, triptans also inﬂuence the calcitonin gene-related peptide (CGRP) promoter (Durham et al. 1997) and regulate CGRP secretion from neu- rons in culture (Durham and Russo 1999). Furthermore, the demonstration that some part of this action is post-synaptic with either 5-HT1B or 5-HT1D re- ceptors located non-presynaptically (Goadsby et al. 2001; Maneesi et al. 2004) offers a prospect of highly anatomically localised treatment options.