5-HT, originally discovered as a serum vasoconstrictor agent, was isolated and characterized more than 50 years ago (46), and later identified as an important brain neurotransmitter (54). 5-HT is actively taken up and stored in blood platelets, is synthesized in mast cells and nerve terminals, including those that innervate the cranial and cerebral vasculature (8,9,33,40,54). Shortly after its discovery, 5-HT was implicated in the pathophysiology of migraine based on the original finding by Sicuteri et al. (52) of an increased urinary excretion of its primary metabolite, 5-hydroxyindoleacetic acid (5-HIAA), during migraine attacks. This observation, however, has proven not to be as reproducible as expected (for references, see Ferrari and Saxena [18]), possibly owing to individual variations in metabolism between various populations of migraine sufferers (25). In contrast, consistent and indisputable increases in plasma 5-HT content concomitant with decreases in 5-HIAA levels have been found during migraine attacks, whereas the reverse is true between attacks, and this irrespective of migraine with and without aura (for references see Ferrari and Saxena [18], Humphrey [25], and Saxena [47]). In keeping with these observations, platelet 5-HT content has been shown to decrease during migraine attacks (18,25), a change that may be absent or much more subtle in migraine without aura (17). The changes in the levels of 5-HT and its metabolite seen in plasma and urine are thought to reflect dysfunctions occurring not only at the level of blood platelets, but primarily so in the brain (25). This is supported by the reported increase in cerebrospinal fluid 5-HIAA levels in migraine patients (18), a finding that further suggests that migraine is accompanied by an enhanced turnover of central 5-HT. Together, these observations lead to the hypothesis that a chronically low serotonin disposition may form the biochemical basis of migraine etiology (17), and that migraine attacks are triggered by a sudden raise in 5-HT release (18,25,29). However, arguments that do not substantiate this view have also been considered (for details, see Panconesi A, Sicuteri F [41]).

The synthesis of brain serotonin depends on the activity of its rate-limiting enzyme tryptophan hydroxylase and the availability of its precursor, the essential amino acid tryptophan that originates exclusively from dietary intake. Similar to migraine, depression is considered as a disorder of low brain serotonergic activity and, as such, the mechanism of action of most effective antidepressant drugs is to enhance and stabilize 5-HT neurotransmission (2). Epidemiologic studies have shown the comorbidity of migraine with depression and anxiety (38), observations that support a common locus of neurochemical disturbances. Interestingly, pharmacologically controlled depressed patients submitted to a rapid tryptophan-depleting diet experienced a depressive relapse together with some symptoms reminiscent of migraine, such as increased nausea or vomiting, drowsiness, and, in some cases, headache (15). Moreover, tricyclic antidepressants are effective in migraine prophylaxis, albeit at a fraction of the dose used in depression, possibly reflecting different mechanisms of action in their clinical efficacy in theses two disorders (for references, see Pringsheim [44]). Recently, chronic administration of low-dose amitriptyline in rats, compatible with that used in the treatment of migraine (1), selectively decreased 5-HT synthetic rate in brainstem raphé nuclei while maintaining a preserved synthetic activity in projection areas. These findings would be compatible with the ability of amitriptyline to ensure an adequate regulation of the pain pathways implicated in migraine headache, an effect possibly related to its antimigraine efficacy (44).

Convincing evidence in favor of migraine being a low serotonergic syndrome came from physiologic studies on the evaluation of auditory evoked potentials, the amplitude of which being inversely related to central serotonergic neurotransmission (24). In migraine patients, a marked increase in amplitude was observed between attacks, in support of a low 5-HT transmission and abnormal cortical processing of sensory information (56). The implication of serotonergic pathways in this response has further been confirmed by the observation that 5-HT_1B/1D receptor agonists, which can penetrate the brain and activate cortical inhibitory prejunctional 5-HT_1B/1D autoreceptors, are able to increase the amplitude of auditory evoked potentials in normal subjects as well as migraine sufferers (45). Together with previous findings of increased amplitude of visual evoked potentials in migraine sufferers (10), these observations all converge to support that migraine is associated with a cortical hypersensitivity to stimulus in the headache-free interval which might be caused, at least in part, by low 5-HT transmission (24,56). Such a scenario of a chronically low availability of 5-HT should result in 5-HT hypersensitivity in migraine sufferers with sensitization of all or specific populations of 5-HT receptors.

Arguments in favor of the biochemical-triggering event being an increase in 5-HT availability include the observations that several drugs which release 5-HT from neurons and blood platelets (fenfluramine and reserpine) and some 5-HT reuptake inhibitors (zimeldine and femoxetine) are all able to provoke migraine attacks and more frequently so in migraine subjects than in controls (18,25,41). However, fenfluramine and reserpine used for an extended period can confer resistance to migraine headache, and 5-HT given during an attack induces headache relief. It seems therefore that the mobilization of 5-HT from intracellular stores at an early stage triggers migraine attacks (20,25,29). Also, the receptors inducing the migraine attack (possibly the 5-HT_2B) appear to be different from those (5-HT₁ family) that relieve migraine headache (see Chapters 51 and 55). The ability of the 5-HT_2B/2C receptor agonist m-chlorophenylpiperazine (m-CPP) to induce attacks, indistinguishable from spontaneous ones, in migraine sufferers (5), together with the 5-HT_2B affinity of prophylactic antimigraine 5-HT₂ receptor antagonists, support the involvement of 5-HT_2B receptor supersensitivity in migraine attacks (20,29,51). On the other hand, one must point out that m-CPP is not selective for 5-HT_2B receptor and that several 5-HT_2B/2C receptor antagonists, including mianserin and cyproheptadine are not very effective antimigraine agents (see Saxena and Den Boer [49], Tfelt-Hansen [53], Villalón et al. [55]). In the event that 5-HT_2B receptor is sensitized in migraine sufferers, it would appear interesting to verify if genetic variations in these receptors might explain the susceptibility to migraine. This does not appear to be the case for the 5-HT_2A and 5-HT_2C receptors, and they have been excluded as candidate genes for migraine susceptibility (6). Similarly, the genetic variants of 5-HT_1B (35) and 5-HT_1F (34) receptors do not seem to influence the clinical response to sumatriptan in migraine patients. It has to be conceded that these genetic variations were “silent” because the nucleotide changes did not affect the peptide composition of either the 5-HT_1B or 5-HT_1F receptor.