New directions in migraine
© Weir and Cader; licensee BioMed Central Ltd. 2011
Received: 10 June 2011
Accepted: 25 October 2011
Published: 25 October 2011
Skip to main content
© Weir and Cader; licensee BioMed Central Ltd. 2011
Received: 10 June 2011
Accepted: 25 October 2011
Published: 25 October 2011
Migraine is a highly prevalent neurological disorder imparting a major burden on health care around the world. The primary pathology may be a state of hyperresponsiveness of the nervous system, but the molecular mechanisms are yet to be fully elucidated. We could now be at a watershed moment in this respect, as the genetic loci associated with typical forms of migraine are being revealed. The genetic discoveries are the latest step in the evolution of our understanding of migraine, which was initially considered a cerebrovascular condition, then a neuroinflammatory process and now primarily a neurogenic disorder. Indeed, the genetic findings, which have revealed ion channels and transporter mutations as causative of migraine, are a powerful argument for the neurogenic basis of migraine. Modulations of ion channels leading to amelioration of the migraine 'hyperresponsive' brain represent attractive targets for drug discovery. There lies ahead an exciting and rapidly progressing phase of migraine translational research, and in this review we highlight recent genetic findings and consider how these may affect the future of migraine neurobiology and therapy.
Migraine is a common, episodic neurological disorder characterised by severe headaches, and in about one-third of cases it is preceded by a focal, transient neurological phenomenon termed the 'aura'. Whether migraine with aura (MA) and migraine without aura (MO) are distinct or overlapping entities is somewhat contested, but it is not uncommon for patients to experience both forms of migraine. Migraine affects up to 18% of women and 6% of men, with a peak incidence between 25 and 55 years of age . The disease has long fascinated clinicians and scientists, with descriptions of migraine-like symptoms appearing in antiquity; however, only recently have the pathogenic mechanisms started to be unravelled. This is particularly true for MA, which is perhaps a more homogeneous and tractable condition for study than MO, because the phenotype of MA is well defined and genetics likely play a greater role .
The migraine aura is likely caused by 'cortical spreading depression' (CSD), a wave of intense neuronal and glial depolarization followed by a period of inactivity, slowly progressing over the cortex [3, 4]. The trigeminal system (TGVS), consisting of the meningeal and superficial cortical blood vessels that are innervated by the trigeminal nerve, is strongly implicated in the initiation of the headache pain . The TGVS projects to the trigeminal nucleus caudalis (TNC) in the brainstem, which in turn projects to higher-order pain centres. Animal models provide evidence that CSD could be linked to activation of the TGVS, thus providing a hypothetical triggering mechanism for the migraine headache . However, such a link in the human brain is still controversial , and although silent CSD has been speculated to be a cause, the relationship of CSD and MO is uncertain.
In this review, we consider the emerging genetic discoveries about migraine, beginning with familial hemiplegic migraine, where the search for causative genes has been particularly rewarding. The more typical forms of migraine have been considerably more challenging, but we describe the exciting discoveries of the past year. The discovery of a role for the TWIK-related spinal cord potassium channel (TRESK) in migraine, in particular resurrects the debate over the relative importance of peripheral and central mechanisms in migraine pain, and the origin of the headache is therefore considered. We next highlight the growing importance of glial cells in migraine pathogenesis, as evidenced by the expression pattern of migraine-causative genes, and finally we describe potential directions for migraine treatment.
Genetic factors are clearly involved in the aetiology of migraine, as illustrated by twin studies . Complex segregation analysis has illustrated a non-Mendelian inheritance pattern , revealing the multigenic nature of migraine. Often a profitable approach to studying such complex genetic diseases is to interrogate an autosomal dominant subtype of the condition. Familial hemiplegic migraine (FHM) is a rare, monogenic subtype of MA . Clinically, FHM patients experience a headache phase similar to that experienced by common migraineurs, but FHM is distinguished by additional prolonged hemiparesis. To date, three genes, when disrupted, have been identified as causing FHM: CACNA1A (FHM1) , encoding the α1 subunit of neuronal Cav2.1 Ca2+ channels; ATP1A2 (FHM2) , encoding the α2 subunit of Na+/K+ ATPase pumps; and SCNA1 (FHM3) , encoding the pore-forming α1 subunit of neuronal Nav1.1 Na+ channels. More recently, a mutation in SLC4A4  encoding the Na+-HCO3 - cotransporter NBCe1 has been identified in sisters with FHM, and a mutation in SLC1A3  encoding the glial glutamate transporter EAAT1 has been identified in one patient with pure hemiplegic migraine.
While FHM and migraine overlap significantly, phenotypic differences do exist . Despite this, FHM serves as a highly valuable model of migraine. The usefulness of FHM genetic studies extends beyond simply identifying genes responsible for rare migraine subtypes. More importantly, they have provided mechanistic insights into the pathobiology of migraine in general, which seems likely to be a disease related to ionic disturbances with a resultant altered excitability of certain areas of the brain. Adopting this hypothesis may help prioritize findings among the large volume of data arising from genome-wide association studies and genome sequencing by examining ionic homeostasis pathways.
Although multiple genes have been implicated in FHM, evidence for genetic alterations associated with common forms of migraine had been lacking. There have been numerous linkage studies, and despite some replication, the genetic abnormalities at these loci remain unknown [26, 27]. Very many association studies of candidate genes have also been undertaken but have been hampered by the relatively small sample sizes. Since 2010, several genetic loci have been identified in association with 'typical' migraine. The first was found in a large, genome-wide association study comprising 2,731 MA patients at three separate European headache clinics. The minor allele of rs1835740 on chromosome 8q22.1 was shown to be linked to migraine with and without aura . Two genes, MTDH and PGCP, flank the region. More functional work is needed to determine the relevance of such polymorphisms to MTDH and PGCP gene expression; however, it is interesting to note that both genes have a role in regulating levels of extracellular glutamate in the brain, a neurotransmitter heavily implicated in migraine pathogenesis . A second study published in 2011 identified three additional susceptibility loci for common migraine . The three SNPs mapped within or near transcribed regions of known genes: LRP1, PRDM16 and TRPM8. LRP1 is highly expressed by neurons and is found in association with NMDA receptors in dendritic synapses  linking the SNP with disrupted central glutamate signalling. While the plausible link between PRDM16 and migraine is not obvious, the association of TRPM8 is intriguing. The gene encodes for a sensor of cold and cold-induced burning pain, which is primarily expressed in sensory neurons such as the trigeminal ganglion and the dorsal root ganglion . TRPM8 is considered to have a role in animal models of neuropathic pain , a disorder that shares some similarities with migraine .
The genome-wide association studies in migraine have yielded plausible susceptibility loci, but further work is still required to confirm that the identified SNPs are causally related and the adjacent genes are relevant to migraine. An alternative approach that can yield more immediate insights into disease pathogenesis is family-based genetic studies. A candidate gene approach followed up by linkage and functional analysis revealed a mutation in KCNK18, the gene encoding TRESK, that segregated with MA in a large pedigree. TRESK is a member of the two-pore potassium channel family showing neuronal expression . KCNK18 was chosen for screening, as the channel was known to play a prominent role in neuronal excitability and therefore was postulated to participate in pain pathways . Furthermore, volatile anaesthetics such as halothane have been shown to activate TRESK and are known to inhibit CSD . One migraine proband was identified as carrying a two-bp deletion in the KCNK18 gene, which leads to premature truncation of the protein. The frameshift mutation F139WfsX24 was found to cosegregate perfectly with the migraine phenotype in relatives of the proband. Patch-clamp recordings from Xenopus oocytes revealed a loss of protein function and a dominant-negative action of mutant subunits on wild-type TRESK channels. Further work confirmed that TRESK is expressed in migraine-salient areas such as the trigeminal ganglion, the cortex and the dorsal root ganglion . In a mouse knockout model, loss of TRESK results in a 20% increase in sensitivity to thermal pain . It is expected that a loss of K+ leak current would lead to the reduced rheobase current required for neuronal spiking, as shown in dorsal root ganglion neurons of TRESK-/- mice . The TRESK mutation was identified in only one large family, and therefore further studies are required to confirm any broader role in migraine pathogenesis. Nevertheless, given the strong expression and physiological function of TRESK in trigeminal ganglia and the central role of the trigeminal ganglia in migraine, modulation of TRESK is a viable target for migraine therapy. Furthermore, the protein has no close paralogues, and known compounds have already been shown to influence channel activity .
The multigenic nature of migraine lends itself to the idea that multiple perturbations in homeostasis are required for disease pathology. Evidence derived from FHM mutations points to an important role of astrocytes in migraine. During late gestation and the early neonatal stage, ATP1A2 is expressed predominantly in neurons; however, by adulthood, the gene is expressed primarily in astrocytes. This expression switch may explain the clinical phenotype switch seen in patients. Young FHM2 patients often have infantile epilepsy (presumably as a consequence of neuronal hyperexcitability), whereas by adulthood the symptoms are more of the classical FHM2 type, that is, migraine attacks caused by impaired neurotransmitter uptake by astrocytes .
Astrocytes are critical in modulating the neuronal microenvironment, as they control extracellular ionic composition and prevent accumulation of neurotransmitters in the synaptic cleft . Spinal glial activation has been suggested to play a role in a variety of pain states . Intracellular Ca2+ oscillation in cultured astrocytes shows significant similarity to CSD  and has been suggested to contribute to the propagation of spreading depression . However, the role of glial calcium waves in vivo, where intracellular Ca2+ oscillation has yet to be robustly shown, is contested . The release of proinflammatory mediators by glia may also contribute to central sensitisation of migraine and to the development of symptoms such as allodynia [32, 50]. Chronic opioid treatment can result in hyperalgesia and allodynia, potentially by activating glial cells , whilst administration of naloxone, which blocks microglial activation, enhances the analgesic effect of morphine .
It is therefore becoming increasingly clear that glia have a role beyond simply myelination and maintaining neuronal homeostasis. They appear to directly and actively influence the excitability of neurons and as such may represent a novel means of modulating neuronal activity and responsiveness.
Therapies are broadly divided into acute symptomatic and prophylactic. To date, triptans are the gold standard in acute migraine therapy. They are potent 5-hydroxytryptamine (serotonin) subtypes 1B and 1D receptor agonists that, when delivered orally, are effective in 29% to 64% of patients, depending on the criteria used to define pain relief: (1) pain-free after two hours and (2) moderate or severe pain to mild or no pain after two hours, respectively [53, 54]. However, clinical use can be limited by their potent vasoconstriction . The most significant recent advance in acute treatments relates to calcitonin gene-related peptide (CGRP) receptor antagonists. Stimulation of TG neurons results in the release of CGRP ; high levels of CGRP are found in external jugular venous blood during migraine attacks ; and infusion of CGRP can cause MO . Telcagepant (the first orally administered CGRP antagonist) has been shown to possess efficacy similar to that of zolmitriptan as acute therapy . It also shows improved safety and tolerability , although it caused elevated levels of liver transaminases in a minority of patients taking the drug twice daily for three months . Insights gleaned from genetic studies offer novel drug targets and potential prophylactic agents through alterations in neuronal excitability. Indeed, current prophylactics such as valproate, verapamil, topiramate and lamotrigine are known to have effects on migraine-associated ion channels [62, 63]. Several drugs that block glial cell activation are also available and have approved clinical safety profiles. These drugs include naltrexone, naloxone, minocycline and ibudilast, and some have already been shown to be effective as prophylactics for migraine [24, 64, 65].
Migraine is a multigenic condition in which aberrant ion channels are becoming increasingly implicated in pathogenesis as inducing an altered state of peripheral and central neuronal excitability. It is realistic to anticipate that further genetic studies may reveal other tractable drug targets, and this will help unravel the pathogenic process in migraine, which is essential for the rational development of effective treatments. Although each protein may be genetically relevant to only a small subset of migraineurs, each one reveals important information about migraine mechanisms potentially applicable to the wider migraine population. It is likely that therapies must be targeted against general pathways involved in neuronal excitability, which will mean addressing both neuronal and glial activation. Furthermore, it may be crucial to understand the interplay of neurons and glia, given that at least two FHM genes are of key importance for glial function. Indeed, simply targeting the neuronal ion channels may be either partially or completely ineffective if glial adaptive changes circumvent any therapeutic intervention. The use of triptans (and soon likely CGRP antagonists) has made major leaps forward in patient care, but for a multigenic, complex condition such as migraine, it may be that only therapies targeting multiple pathways and cell types will lead to drugs with broader and more sustained efficacy. Equally true is that a better understanding of the genetic architecture of migraine with the future availability of rapid whole-genome sequencing in an individual could lead to the ever sought after personalized medicine.
GAW and MZC are funded by the Medical Research Council.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.