Recent technical advances in genetics made large-scale genome-wide association studies (GWAS) in migraine feasible and have identified over 40 common DNA sequence variants that affect risk for migraine types. Most of the variants, which are all single nucleotide polymorphisms (SNPs), show robust association with migraine as evidenced by the fact that the vast majority replicate in subsequent independent studies. However, despite thorough bioinformatic efforts aimed at linking the migraine risk SNPs with genes and their molecular pathways, there remains quite some discussion as to how successful this endeavour has been, and their current practical use for the diagnosis and treatment of migraine patients. Although existing genetic information seems to favour involvement of vascular mechanisms, but also neuronal and other mechanisms such as metal ion homeostasis and neuronal migration, the complexity of the underlying genetic pathophysiology presents challenges to advancing genetic knowledge to clinical use. A major issue is to what extent one can rely on bioinformatics to pinpoint the actual disease genes, and from this the linked pathways. In this Commentary, we will provide an overview of findings from GWAS in migraine, current hypotheses of the disease pathways that emerged from these findings, and some of the major drawbacks of the approaches used to identify the genes and pathways. We argue that more functional research is urgently needed to turn the hypotheses that emerge from GWAS in migraine to clinically useful information.

Cortical spreading depolarization (CSD) is the electrophysiological substrate of migraine aura, and a putative trigger of trigeminovascular activation and migraine headache. Many migraineurs report stress or relief after a stress triggers an attack. We tested whether various stress conditions might modulate CSD susceptibility and whether this is dependent on genetic factors. Male and female wild type and familial hemiplegic migraine type1 (FHM1) knock-in mice heterozygous for the S218L missense mutation were subjected to acute or chronic stress, or chronic stress followed by relief (36 h). Acute stress was induced by restraint and exposure to bright light and white noise (3 h). Chronic stress was induced for 28 days by two cycles of repeated exposure of mice to a rat (7 days), physical restraint (3 days), and forced swimming (3 days). Electrical CSD threshold and KCl-induced (300 mM) CSD frequency were determined in occipital cortex in vivo at the end of each protocol. Relief after chronic stress reduced the electrical CSD threshold and increased the frequency of KCl-induced CSDs in FHM1 mutants only. Acute or chronic stress without relief did not affect CSD susceptibility in either strain. Stress status did not affect CSD propagation speed, duration or amplitude. In summary, relief after chronic stress, but not acute or chronic stress alone, augments CSD in genetically susceptible mice. Therefore, enhanced CSD susceptibility may explain why, in certain patients, migraine attacks typically occur during a period of stress relief such as weekends or holidays.

Migraine affects over a billion individuals worldwide but its genetic underpinning remains largely unknown. Here, we performed a genome-wide association study of 102,084 migraine cases and 771,257 controls and identified 123 loci, of which 86 are previously unknown. These loci provide an opportunity to evaluate shared and distinct genetic components in the two main migraine subtypes: migraine with aura and migraine without aura. Stratification of the risk loci using 29,679 cases with subtype information indicated three risk variants that seem specific for migraine with aura (in HMOX2, CACNA1A and MPPED2), two that seem specific for migraine without aura (near SPINK2 and near FECH) and nine that increase susceptibility for migraine regardless of subtype. The new risk loci include genes encoding recent migraine-specific drug targets, namely calcitonin gene-related peptide (CALCA/CALCB) and serotonin 1F receptor (HTR1F). Overall, genomic annotations among migraine-associated variants were enriched in both vascular and central nervous system tissue/cell types, supporting unequivocally that neurovascular mechanisms underlie migraine pathophysiology.

Hemiplegic migraine (HM) is a rare subtype of migraine with aura. Given that causal missense mutations in the voltage-gated calcium channel α1A subunit gene CACNA1A have been identified in a subset of HM patients, we investigated whether HM patients without a mutation have an increased burden of such variants in the "CACNA1x gene family". Whole exome sequencing data of an Australian cohort of unrelated HM patients (n = 184), along with public data from gnomAD, as controls, was used to assess the burden of missense variants in CACNA1x genes. We performed both a variant and a subject burden test. We found a significant burden for the number of variants in CACNA1E (p = 1.3 × 10-4), CACNA1H (p < 2.2 × 10-16) and CACNA1I (p < 2.2 × 10-16). There was also a significant burden of subjects with missense variants in CACNA1E (p = 6.2 × 10-3), CACNA1H (p < 2.2 × 10-16) and CACNA1I (p < 2.2 × 10-16). Both the number of variants and number of subjects were replicated for CACNA1H (p = 3.5 × 10-8; p = 0.012) and CACNA1I (p = 0.019, p = 0.044), respectively, in a Dutch clinical HM cohort (n = 32), albeit that CACNA1I did not remain significant after multiple testing correction. Our data suggest that HM, in the absence of a single causal mutation, is a complex trait, in which an increased burden of missense variants in CACNA1H and CACNA1I may contribute to the risk of disease.