|Home | About | Journals | Submit | Contact Us | Français|
In a population-based genome-wide analysis including 5122 migraineurs and 18,108 non-migraineurs, rs2651899 (PRDM16), rs10166942 (TRMP8), and rs11172113 (LRP1) were among the top associations (p<5×10−6) with migraine. All three SNPs were significant in meta-analysis among replication cohorts and met genome-wide significance (p<4.3×10−9) in meta-analysis combining discovery and replication cohorts. Rs2651899 and rs10166942 associated with migraine compared to non-migraine headache; none of the three SNPs specifically associated with migraine subtypes or features.
Migraine is a common and often debilitating disorder affecting up to 20% of the population, women 3–4 times more often than men1. Clinically, migraine manifests with recurrent attacks of headache associated with gastrointestinal and autonomic nervous system symptoms2. Up to one third of patients may also experience transient focal neurological symptoms known as aura. Current concepts view migraine primarily as a multi-factorial brain disorder with heritability estimates as high as 50%3. Yet progress in genetic analysis has been largely restricted to rare monogenic subtypes of migraine, and very little is known about underlying genetic variants for common forms of migraine, including migraine with and without aura4. A recent genome-wide association study (GWAS) identified a genetic variant on chromosome 8q22.1 associated with migraine in a large clinic-based sample of European migraine patients5.
In order to identify common genetic variants for migraine at the population level, we performed a GWAS among 23,230 women with complete genotype and migraine information and verified European ancestry from the Women’s Genome Health Study (WGHS)6, a large population-based cohort (see Supplementary Material). Migraine was reported by 5122 women compared with 18,108 not reporting migraine. The clinical profile between migraineurs and non-migraineurs differed most strongly for age, postmenopausal hormone use, physical activity, and alcohol consumption (Supplementary Table S1).
In the discovery stage genome-wide scan among genotyped SNPs, no SNPs reached the conventional threshold association p-value <5×10−8 in age-adjusted logistic models assuming an additive relationship between the minor allele dose and log-odds of migraine. Nevertheless, the top SNPs were more significant than expected for the 95% confidence interval for the ordered test statistic under the null hypothesis and, except for one SNP, remained more significant after controlling for modest inflation of the test statistic (λGC=1.03, Supplementary Figure S1). Seven independent loci had at least one SNP with p<5×10−6. We selected the most significant SNP from each locus for further analysis (Table 1). The associations of all seven SNPs were specific for migraine, since none was significant (all p>0.05) for non-migraine headache (N=3,001 vs. 14,959 controls). The significance and magnitude of the associations of the seven SNPs were essentially unchanged in sensitivity analyses using 1) an allele frequency test, 2) logistic regression without adjustment, or 3) logistic regression adjusted for clinical characteristics, eigenvector parameters for sub-European population structure, or both (Supplementary Table S2). There was no evidence of non-additive modes of association with migraine for any of the top SNPs and no SNPs in the entire genome-wide scan reached genome-wide significance in non-additive models (data not shown). Finally, analysis performed with genotypes imputed for approximately 2.6 million SNPs in the HapMap (release 22)7 revealed the same top seven loci found with the genotyped SNPs.
We evaluated the seven SNPs in two additional, population-based cohorts, the Dutch Genetic Epidemiology of Migraine study (GEM; 774 migraineurs, 942 non-migraineurs) and the German Study of Health in Pomerania (SHIP; 306 migraineurs, 2260 non-migraineurs) as well as in the previously reported clinic-based case-control samples from the International Headache Genetics Consortium (IHGC; 2748 migraineurs, 10,747 population-based controls)5, all with European ancestry (see Supplementary methods). Genotyping failed for rs17172526 in both GEM and SHIP. The effect estimates for rs2651899, rs10166942, and rs11172113 were concordant in direction and magnitude to effects in the WGHS (Table 1; p=0.03 for consistency of direction, see Supplementary methods). Moreover, all seven SNPs were concordant between the WGHS and the IHGC replication (p=0.02 [=0.57×3 replication cohorts]). In IHGC, there were nominally significant associations for rs2078371 and the three SNPs that had concordant effects in all three replication cohorts. Rs10166942 in GEM and both rs2651899 and rs11172113 in SHIP were also nominally significant (Table 1).
We then performed meta-analysis of the results for the primary SNPs using a fixed-effects model with inverse-variance weighting. Combining only the replication cohorts (Table 2), the meta-analysis supported association (p<0.008[=0.05/6]) for three SNPs from WGHS: rs2651899 (OR [95%CI]=1.08 [1.03–1.14], p=4.2×10−3), rs10166942 (OR [95%CI]=0.84 [0.79–0.91], p=5.0×10−6), and rs11172113 (OR [95%CI]=0.90 [0.85–0.95], p=3.0×10−4). There was no strong evidence of heterogeneity among the studies (all I2<22%, phet>0.28). Furthermore, all three SNPs met genome-wide standards of significance in meta-analysis combining the discovery cohort with the replication cohorts (Table 2, all p<4.3×10−9). None of the remaining SNPs reached either nominal significance in meta-analysis of the replication cohorts or genome-wide significance in meta-analysis combining all cohorts.
All three replicating SNPs map within or near transcribed regions of known genes. Rs2651899 is within the first intron of PRDM16 in a block of moderate LD extending about 22 kb in the 5’ direction and about 25 kb in the 3’ direction (Supplementary Figure S2A). The transcript for the second closest gene, ACTRT2, terminates 144kb from rs2651899. Rs10166942 is 950bp 5’ to the transcription start site for TRPM8 in a block of moderate LD spanning approximately 168 kb. This LD block begins about 117 kb 5’ from the SNP and extends through TRPM8 (Supplementary Figure 2B), a gene previously identified as a potential candidate for association with migraine in the IHGC cohort on the basis of sub-genome-wide significance5. The second closest gene, HJURP, is situated 61.9 kb from rs10166942. Rs11172113 maps to the first intron of LRP1 in a gene rich region that also includes the second closest gene STAT6 22.1 kb away (Supplementary Figure S2C). At each of three loci, only the primary SNP was retained in stepwise model selection for association with migraine suggesting the absence of additional, non-redundant or conditional associations (see Supplementary methods).
Since migraine is more prevalent in women, we examined the effects of the three replicating SNPs on migraine in the replication cohorts, which include both women and men in contrast to the WGHS. In meta-analysis of the six sex-specific strata among the three replication cohorts, effect estimates were comparable to the main analysis that was not stratified by sex (compare Table S3, left and Table 2). Notably, the heterogeneity estimate (I2) for rs10166942 increased from 0% to 45.8%, although neither the heterogeneity p-value nor a potential sex effect at this SNP in meta-regression was significant (p=0.10 and 0.23, respectively, Table S3, right). However, a potential sex interaction for rs10166942 may have been confounded by differences in study design and migraine ascertainment between IHGC (clinic-based) and GEM and SHIP (both population-based). Meta-analysis by sex strata restricted to the population-based studies revealed even greater heterogeneity for rs10166942 (I2:67.2%, heterogeneity p-value=0.03; Table S4, left). In meta-regression of this SNP among these two studies alone, the association with migraine suggested a significant sex interaction (p=0.004) and essentially no evidence of association in men (OR [95% CI]=1.08 [0.85–1.38], p=0.52; Table S4, right). The potential differential association according to sex did not appear to be related to the estrogen receptor 1, since there was no interaction between rs10166942 and ESR1 SNPs in the WGHS, considering SNPs associated with other clinical traits (Supplementary methods, data not shown). Similarly, none of the other primary SNPs showed an interaction with ESR1 SNPs (data not shown).
We investigated whether there was evidence for disproportionate association of the three genome-wide significant SNPs from the meta-analysis with migraine aura status and features recognized by International Headache Society (IHS) diagnostic criteria (unilateral pain location, pulsating pain quality, sensitivity to light or sound, attack duration, nausea/vomiting, aggravation by physical activity, inhibition of physical activity; Supplementary Methods). None of these associations was more significant than for overall migraine (Supplementary Table S5). Moreover, allele frequency differences of the three SNPs between migraineurs with or without each of the features were not significant enough to overcome correction for multiple hypotheses testing although some met nominal significance (Supplementary Table S6). In contrast, rs2651899 and rs10166942 (but not rs11172113) were significantly associated with migraine compared to non-migraine headache (rs2651899 OR [95% CI]=1.12 [1.05–1.20], p=5×10−4; rs10166942 OR [95% CI]=0.90 [0.82–0.98], p=0.01) with effects similar to the association of migraine compared to non-migraine controls, reinforcing their specificity for migraine (Supplementary Table S7, compare with Table 1).
TRPM8 encodes a sensor for cold and cold-induced burning pain8, primarily expressed in sensory neurons and dorsal root ganglion neurons9. Members of the mammalian TRP superfamily are channels activated by stimuli of chemo- and somatosensation. TRPM8 particularly, is a target in animal models of neuropathic pain10. As migraine shares some characteristics with neuropathic pain disorders11 TRPM8 could be a pathophysiological link between both pain syndromes.
LRP1 is expressed in many tissues including brain and vasculature, where it modulates synaptic transmission12. As a member of the lipoprotein receptor family, it serves as a sensor of the extracellular environment. LRP1 and glutamate (NMDA) receptors are co-localized on neurons and interact. This integrates well with findings from the previous GWAS in migraine reporting a genetic variant implicated in glutamate homeostasis5 as well as recent pharmacological approaches to migraine aiming at glutamate receptors13.
A potential role of PRDM16 protein in migraine is unclear. PRDM16, was originally identified near a chromosomal breakpoint associated with myelodysplastic syndrome and acute myeloid leukemia14, while subsequent research has focused on its transcriptional role in brown fat development15. Structurally, PRDM16 contains two arrays of C2H2 zinc-finger domain repeats, often linked to transcriptional activity; it also contains a putative SET domain, a conserved region among histone lysine methyltransferases.
Our findings may be compared with the recent GWAS reporting an association of rs1835740 at 8q22.1 for migraine, especially migraine with aura5. The nearby candidate genes, PGCP and MTDH, are involved in glutamate homeostasis, consistent with current concepts of migraine pathophysiology. However, in the WGHS, rs1835740 is neither associated with overall migraine (p=0.22) nor migraine with or without aura separately (data not shown). Similarly, across the entire region, no SNP reached locus-wide significant thresholds for association with migraine, and there was no evidence of stronger associations with migraine with aura than without aura (Supplementary Figure S3). The most significant SNP for association with migraine in the WGHS was rs1864729 (OR [95%CI]=1.15 [1.06–1.25], p=0.0001). The differential findings between the current and the previous study5 may partly relate to the differences in migraine ascertainment (see also analysis of sex interaction).
There are three major implications of our study. First, we have identified three SNPs with genome-wide association for common migraine at the population level. Two of the SNPs significantly distinguish migraine from non-migraine headache. In addition, the association of rs10166942 may be stronger among women, which may be related to but would not explain the higher prevalence of migraine in women. Second, while one novel locus (LRP1) supports prevailing (glutamatergic) concepts of neurotransmitter pathways in migraine4, we also identified a second novel locus (TRPM8) explicitly implicated in a pain related pathway. The functional significance of the third locus (PRDM16) pathway is still unknown. Third, while some studies have focused on differences between migraine with and without aura4, our results suggest shared pathophysiology among common types of migraine (Supplementary Tables S5 and S6).
Ongoing large GWAS will continue to identify additional genetic risk variants for migraine and further delineate the pathophysiological basis of migraine. Meanwhile, the three novel loci identified in the present work provide hypotheses for immediate further exploration.
This study is supported by a grant from the National Institute of Neurological Disorders and Stroke (NS-061836). The “Women’s Health Study” and “Women’s Genome Health Study” are supported by grants from the National Heart, Lung, and Blood Institute (HL-043851, HL-080467, HL-099355), and the National Cancer Institute (CA-47988). Part of the research for this work was supported by grants from the Donald W. Reynolds Foundation and the Leducq Foundation. Genome-wide genotyping and collaborative scientific support was provided by Amgen.
Genotyping in the “Genetic Epidemiology of Migraine Study” was supported by the Netherlands Organisation for Scientific Research (NWO) VICI (918.56.602) and Spinoza (2009) grants and the Center for Medical Systems Biology (CMSB) established by the Netherlands Genomics Initiative/Netherlands Organisation for Scientific Research (NGI/NWO), project no. 050-060-409. The GEM study was supported by the Ministry of Health, Welfare and Sport and the National Institute of Public Health and the Environment, the Netherlands.
SHIP is part of the Community Medicine Research net of the University of Greifswald, Germany, which is funded by the Federal Ministry of Education and Research (grants no. 01ZZ9603, 01ZZ0103, and 01ZZ0403), the Ministry of Cultural Affairs as well as the Social Ministry of the Federal State of Mecklenburg-West Pomerania. Genome-wide data have been supported by the Federal Ministry of Education and Research (grant no. 03ZIK012) and a joint grant from Siemens Healthcare, Erlangen, Germany and the Federal State of Mecklenburg-West Pomerania. The SHIP authors are grateful to the contribution of Alexander Teumer, Anja Hoffmann, and Astrid Petersmann in generating the SNP data. The University of Greifswald is a member of the ‘Center of Knowledge Interchange’ program of the Siemens AG.
The IHGC study was supported, among others, by the Academy of Finland (200923 to A.Palotie), the Wellcome Trust (grant number 089062), the European Community's Seventh Framework Programme [FP7/2007-2013] (through the SYNSYS Consortium [grant agreement no. 242167] and the ENGAGE Consortium [grant agreement no. 201413]), the Helsinki University Central Hospital (to M. Kallela), and the Finnish Culture Foundation (to V. Anttila). Funding by the German Federal Ministry of Education and Research (BMBF) within the National Genome Research Network (NGFNplus, EMINet-01GS08120 for C. Kubisch, 01GS08121 to M. Dichgans), the Deutsche Forschungsgemeinschaft (to C. Kubisch) and the Center for Molecular Medicine Cologne (to C.K.). For a full list, please see reference5.
Competing Financial Interests
None of the authors has a competing financial interest with regard to this manuscript.
Author contributions (alphabetical)Obtaining funding: J.E. Buring, M.D. Ferrari, W. Hoffmann, T. Kurth, A.M.J.M. van den Maagdenberg, A. Palotie, P.M Ridker, U. Schminke, H. Völzke, R.Y.L. Zee.
Overall study design: D.I. Chasman, T. Kurth, M. Schürks.
Cohort supervision and phenotyping: V. Anttila, J. Buring, K. Fendrich, M.D. Ferrari, T. Freilinger, W. Hoffmann, M. Kallela, C. Kubisch, T. Kurth, L.J. Launer, A. Palotie, P.M Ridker, U. Schminke, M. Schürks, G.M. Terwindt, H. Völzke.
Analysis and genotyping: V. Anttila, D.I. Chasman, K. Fendrich, L.R. Griffiths, W. Hoffmannm, A.M.J.M. van den Maagdenberg, P.M Ridker, U. Schminke, M. Schürks, U. Völker, B. de Vries, R.Y.L. Zee.
Manuscript writing: D.I. Chasman, M. Schürks.
All authors participated in critical review of the manuscript for intellectual content.