With the exception of 5-HT₃ receptors, all known biogenic amine receptors belong to the superfamily of rhodopsinlike G protein-coupled receptors (GPCRs) (56). Based on crystal structure data, hydropathy profile analyses, and phylogenetic comparisons, these receptors share the common motif of seven transmembrane (TM) domains. The membrane-spanning regions are linked by three extracellular loops that alternate with three intracellular loops. Activation of the receptors occurs by binding of specific biogenic amines to a binding pocket formed by the TM regions. Specific residues in different TM segments interact with functional groups of the biogenic amines. In particular, an aspartic acid residue in TM3, serine residues in TM5, and a phenylalanine residue in TM6 determine the ligand-binding properties of biogenic amine receptors. Ligand binding to its receptor induces a conformational change, initiating the activation of second messenger systems via specific G proteins. Residues in close vicinity to the plasma membrane in intracellular loops 2, 3, and 4 determine the specificity and efficacy of G-protein activation. Receptor-mediated signaling is turned off by phosphorylation of serine and threonine residues in the C-terminus and third intracellular loops, as it is known from rhodopsin (22).

The initial step in the synthesis of 5-HT is the facilitated transport of L-tryptophan from blood into brain. Because other neutral amino acids, such as phenylalanine, leucine, and methionine, are transported by the same carrier, tryptophan entry into brain is related not only to its blood concentration but also to the concentrations of other neutral amino acids. Tryptophan hydroxylase catalyzes the tetrahydrobiopterin-dependent hydroxylation of tryptophan to 5-hydroxytryptophan, the first and rate-limiting step in the biosynthesis of 5-HT. The next step involves the decarboxylation of 5-hydroxytryptophan by L-aromatic amino acid decarboxylase (this enzyme is also present in catecholaminergic neurons, where it
converts 3,4-dihydroxyphenylalanine to dopamine, see below).

The highest 5-HT concentrations are found in blood platelets and in the gastrointestinal tract (enterochromaffin cells of the mucosa contain more than 90% of the body’s 5-HT). Lesser amounts are found in the brain and retina. Most serotonergic soma are located in morphologically diverse neurons clustered along the midline of the brainstem and reticular formation. Dahlstrom and Fuxe (16) have described nine groups of 5-HT-containing cell bodies, designated B1 through B9, which correspond for the most part with the raphe nuclei. Although there are only about 20,000 serotonergic neurons in the rat brain (around 300,000 in humans), the extensive axonal projection system arising from these neurons densely innervates nearly all regions of the CNS.

As with other biogenic amine transmitters, 5-HT is stored in vesicles and released by a Ca²⁺-dependent exocytotic mechanism. The extent of 5-HT release depends on the firing rate of serotonergic neurons. Drugs that decrease the firing rate of serotonergic soma (e.g., 5-HT_1A-receptor agonists) decrease the release of 5-HT in the projection areas. Administration of 5-HT_1B-receptor agonists into areas receiving serotonergic innervation activates serotonergic autoreceptors in terminal fields and decreases 5-HT synthesis and release. However, in contrast to the activation of 5-HT_1A somatodendritic autoreceptors, these effects are not caused by decreases in the firing rate of serotonergic neurons.

Most of the released 5-HT is recaptured by an active reuptake mechanism via the serotonin transporter (SERT) located on serotonergic neurons. SERT exhibits about 50% homology with the transporters for norepinephrine and dopamine. The selectivity of monoamine reuptake inhibitors for these transporters varies greatly. For instance, tricyclic antidepressants such as desipramine are 25- to 150-fold more potent at inhibiting transport of norepinephrine than 5-HT. In contrast, selective serotonin reuptake inhibitors (SSRIs), such as fluoxetine, citalopram, sertraline, and paroxetine, are 15 to 75 times more potent at inhibiting the uptake of 5-HT than the uptake of norepinephrine. Various heterocyclic antidepressants (amitriptyline, imipramine, nortriptyline, clomipramine, doxepin, trazodone) are used in migraine prophylaxis, but the efficacy of tricyclic antidepressants in migraine does not seem to depend on their antidepressant action, and the strongest evidence of efficacy is for amitriptyline (2,62). Venlafaxine, a structurally novel bicyclic antidepressant that inhibits the reuptake of 5-HT and norepinephrine, has been shown to have a favorable efficacy and side effect profile when compared to amitriptyline (11). The use of the SSRIs (e.g., fluoxetine, sertraline, and paroxetine) has also been advocated in migraine prophylaxis, but data supporting the efficacy of these drugs are scant (15).

5-HT is primarily degraded by the enzyme monoamine oxidase (MAO), which occurs as two isoforms (MAO-A and MAO-B). MAO converts 5-HT to 5-hydroxyindole acetaldehyde, which in turn is metabolized by aldehyde dehydrogenase to produce 5-hydroxyindole acetic acid, the major excreted 5-HT metabolite. Both 5-HT and norepinephrine are metabolized preferentially by MAO-A. Inhibition of MAO-A activity has been linked to the antidepressant properties of a number of subtype selective (e.g., moclobemide) and nonselective (e.g., phenelzine) MAO inhibitors in clinical use. Interestingly, there is more MAO-A than MAO-B throughout rat brain, whereas human brain contains more MAO-B. Although serotonergic cell bodies contain predominantly MAO-B (for which 5-HT is not a preferred substrate), treatment of rats with clorgyline, a selective MAO-A inhibitor, raises the brain content of 5-HT and reduces the conversion of 5-HT to 5-hydroxyindole acetic acid. Thus, 5-HT may well be oxidized preferentially by MAO-A in vivo, as it is in vitro, even though serotonergic neurons do not contain much of this form of the enzyme.

With the exception of 5-HT₃ receptors, 5-HT receptors belong to the GPCR superfamily. With at least 14 distinct members, they are one of the most complex families of neurotransmitter receptors. Adding to this complexity, splice variants of various subtypes (5-HT₄, 5-HT₇) or RNA-edited isoforms (5-HT_2C) have been described (27). The 5-HT₁ receptor class is composed of five receptor subtypes (5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1E and 5-HT_1F), which, in humans, share 40 to 63% overall sequence identity and couple preferentially to Gi/o proteins, with resulting inhibition of cyclic adenosine monophosphate (cAMP) formation. The 5-HT₂ receptor class is composed of the 5-HT_2A, 5-HT_2B and 5-HT_2C receptors, which exhibit 46 to 50% overall sequence identity and couple preferentially to Gq/11 to increase the hydrolysis of inositol phosphates and elevate cytosolic [Ca²⁺].

Based on their overall electrophysiologic features and sequence, 5-HT₃ receptors have been placed within the intrinsic ligand-gated ion channel receptor superfamily, together with nicotinic acetylcholine or GABA_A receptors. 5-HT₃ receptors are found on both central and peripheral neurons, where they trigger rapid depolarization because of a transient inward current, subsequent to the opening of nonselective cation channels. A cDNA clone encoding a single subunit of the 5-HT_3A receptor was initially isolated from a neuron-derived cell line. Two splice variants were subsequently described in neuroblastoma-glioma cells and rat native tissues. A second subunit, 5-HT_3B, has been cloned. It appears that the heteromeric combination of 5-HT_3A and 5-HT_3B subunits is necessary to provide the full functional features of the 5-HT₃ receptor (61).

Although 5-HT₄, 5-HT₆, and 5-HT₇ receptors all couple preferentially to Gs and promote cAMP formation,
they are classified as distinct receptor classes because of their limited (<35%) overall sequence identities (9,24). There have been no published reports to date concerning a physiologic functional response or specific binding to a 5-HT₅ recognition site. Two subtypes of the 5-HT₅ receptor (5-HT_5A and 5-HT_5B), sharing 70% overall sequence identity, have been found in rodents. The 5-HT_5A subtype has also been found in humans, but the human 5-HT_5B receptor gene does not encode a functional protein; there are stop codons in its coding sequence. In the rat, the recombinant 5-HT_5A receptor may be negatively coupled to adenylate cyclase activity (24).

Although most of the evidence for a direct role of 5-HT in the pathophysiology of migraine is circumstantial, there is a significant amount of literature linking this amine to migraine (see Chapter 29). For example, platelet 5-HT levels are reduced by 30% during attacks, and plasma concentrations are 60% lower; the biogenic amine-depleting drug reserpine causes a “typical headache” in migraineurs, probably by inducing 5-HT release from intracellular stores. Similarly, m-chlorophenylpiperazine (m-CPP), a major metabolite of the antidepressant trazodone, has been reported to cause migraine-like headaches in humans by activating 5-HT_2B or 5-HT_2C receptors. Perhaps the strongest evidence for a role of 5-HT in migraine is provided by the fact that some acute antimigraine drugs (ergot alkaloids, triptans) activate 5-HT₁ receptors (most probably the 5-HT_1B, 5-HT_1F, and/or 5-HT_1F subtypes), whereas various prophylactic agents (methysergide, pizotifen, cyproheptadine) are 5-HT₂ receptor antagonists (see Chapters 21, 50, 51, and 55).

Dopamine is a catecholamine neurotransmitter found predominately in the CNS. It is synthesized from tyrosine, which is converted to L-dihydroxyphenylalanine (L-DOPA) by the enzyme tyrosine hydroxylase. Dihydroxyphenylalanine is converted to dopamine by the cytoplasmic enzyme DOPA decarboxylase (or aromatic amino acid decarboxylase).

Five subtypes of dopamine receptor have been cloned. The D₁-like receptors (D₁ and D₅) are closely related. They couple to the G protein subunit G_sα and stimulate adenylyl cyclase activity. In contrast, D₂, D₃, and D₄ receptors (D₂-like) couple to G_iα and inhibit the formation of cAMP. The human D₂ gene consists of eight exons separated by seven introns (25). It undergoes alternative splicing to generate two molecularly distinct isoforms, D₂S and D₂L, that have distinct functions in vivo (39). D₂ receptor polymorphism has been hypothetically connected to a large number of neuropsychiatric disorders (65), but no conclusive link or association has been found. In a recent study, a large group of migraine patients with different D₂ genotypes showed similar clinical and psychological features (49). These results are at odds with a previous study showing that D₂ NcoI alleles significantly modify the clinical susceptibility to migraine with aura (47). In addition, NcoI C allele frequency has been reported to be significantly higher in individuals with migraine with aura, anxiety, or major depression than in individuals who have none of these disorders (46). This observation is unexpected, because the NcoI polymorphism involves a silent change at amino acid His³¹³ and the expression of different alleles results in identical receptor molecules. Subsequent studies have in fact been unable to confirm the association between migraine with aura and D₂ NcoI alleles (20,33). In contrast to NcoI polymorphism, the -141C Ins/Del polymorphism in the D₂ receptor gene is functional and the -141C Ins allele frequency is significantly higher in schizophrenic patients than in control subjects (41). However, no association has been found between this polymorphism and migraine (36). Focusing on an isolated genetic population, Del Zompo et al. (19) found a disequilibrium in D₂ genotypes, but only in a subgroup of patients experiencing both nausea and yawning immediately before or during the pain phase of migraine (so-called dopaminergic migraineurs). A recent study assessed the association between polymorphism of other dopamine-related genes and susceptibility to migraine (38). No significant differences were found between control and migraine groups for the frequency of a 40-bp tandem repeat in the dopamine transporter gene and that of a dinucleotide repeat in the dopamine β-hydroxylase gene, but migraine without aura, and not migraine with aura, showed significant genetic association with D₄ receptor polymorphism. The precise role of D₂-like receptor gene polymorphism in migraine is therefore presently unclear. There is no evidence for allelic association between D₁-like (D₁, D₃, and D₅) dopamine receptor gene polymorphism and migraine with or without aura (55).