RESEARCH ARTICLE Genetic Susceptibility Loci in Genomewide Association Study of Cluster Headache Aster V.E. Harder, MD ,1,2† Bendik S. Winsvold, MD, PhD ,3,4,5† Raymond Noordam, PhD ,6† Lisanne S. Vijfhuizen, BSc ,2 Sigrid Børte, MD, PhD,3,4,7 Lisette J.A. Kogelman, PhD ,8 Irene de Boer, MD ,1 Erling Tronvik, MD, PhD,9,10 Frits R. Rosendaal, MD, PhD,11 Ko Willems van Dijk, PhD,2,12,13 Emer O’Connor, MD,14,15,16 Carmen Fourier, MSc ,17 Laurent F. Thomas, PhD ,4,18,19,20 Espen S. Kristoffersen, MD, PhD ,3,21,22 Cluster Headache Genetics Working Group, Rolf Fronczek, MD, PhD ,1 Patricia Pozo-Rosich, MD,23,24 Rigmor H. Jensen, MD, DrMedSci, PhD,8 Michel D. Ferrari, MD, PhD,1 Thomas F. Hansen, PhD,8 John-Anker Zwart, MD, PhD,3,4,7# Gisela M. Terwindt, MD, PhD ,1# and Arn M.J.M van den Maagdenberg, PhD 1,2# View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.26146 Received Sep 25, 2020, and in revised form Jun 17, 2021. Accepted for publication Jun 22, 2021. Address correspondence to Prof. Arn M.J.M. van den Maagdenberg, Department of Human Genetics; Leiden University Medical Center; PO Box 9600 2300 RC Leiden, The Netherlands. E-mail: a.m.j.m.van_den_maagdenberg@lumc.nl † A.V.E.H., B.S.W., and R.N. contributed equally to this work and should be considered co-first authors. # J.A.Z., G.M.T., and A.M.J.M.vdM. contributed equally to this work and should be considered co-last authors. From the 1Department of Neurology, Leiden University Medical Center, Leiden, The Netherlands; 2Department of Human Genetics, Leiden University Medical Center, Leiden, The Netherlands; 3Department of Research, Innovation and Education, Division of Clinical Neuroscience, Oslo University Hospital, Oslo, Norway; 4K.G. Jebsen Center for Genetic Epidemiology, Department of Public Health and Nursing, Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology, Trondheim, Norway; 5Department of Neurology, Oslo University Hospital, Oslo, Norway; 6Department of Internal Medicine, Section of Gerontology and Geriatrics, Leiden University Medical Center, Leiden, The Netherlands; 7Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway; 8Department of Neurology, Danish Headache Center, Rigshospitalet, Glostrup, Denmark; 9 Department of Neuromedicine and Movement Science, Norwegian University of Science and Technology, Trondheim, Norway; 10National Advisory Unit on Headaches, Department of Neurology and Clinical Neurophysiology, St. Olav’s Hospital, Trondheim, Norway; 11Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, The Netherlands; 12Einthoven Laboratory for Experimental Vascular Medicine, Leiden University Medical Center, Leiden, The Netherlands; 13Department of Internal Medicine, Division of Endocrinology, Leiden University Medical Center, Leiden, The Netherlands; 14 Department of Neuromuscular Diseases, Institute of Neurology, University College London, London, UK; 15Neurogenetics Laboratory, Institute of Neurology, University College London, London, UK; 16Headache and Facial Pain Group, The National Hospital for Neurology and Neurosurgery, London, UK; 17Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden; 18Department of Clinical and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway; 19BioCore - Bioinformatics Core Facility, Norwegian University of Science and Technology, Trondheim, Norway; 20Clinic of Laboratory Medicine, St. Olavs Hospital, Trondheim, Norway; 21Department of General Practice, HELSAM, University of Oslo, Oslo, Norway; 22Department of Neurology, Akershus University Hospital, Lørenskog, Norway; 23Headache Research Group, Vall d’Hebron Institute of Research (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain; and 24España Unidad de Cefalea, Servicio de Neurología, Hospital Universitari Vall d’Hebron, Barcelona, Spain © 2021 The Authors. Annals of Neurology published by Wiley Periodicals LLC on behalf of American Neurological Association. 203 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. ANNALS of Neurology Objective: Identifying common genetic variants that confer genetic risk for cluster headache. Methods: We conducted a case–control study in the Dutch Leiden University Cluster headache neuro-Analysis program (LUCA) study population (n = 840) and unselected controls from the Netherlands Epidemiology of Obesity Study (NEO; n = 1,457). Replication was performed in a Norwegian sample of 144 cases from the Trondheim Cluster head- ache sample and 1,800 controls from the Nord-Trøndelag Health Survey (HUNT). Gene set and tissue enrichment ana- lyses, blood cell-derived RNA-sequencing of genes around the risk loci and linkage disequilibrium score regression were part of the downstream analyses. Results: An association was found with cluster headache for 4 independent loci (r2 < 0.1) with genomewide significance (p < 5  108), rs11579212 (odds ratio [OR] = 1.51, 95% confidence interval [CI] = 1.33–1.72 near RP11-815 M8.1), rs6541998 (OR = 1.53, 95% CI = 1.37–1.74 near MERTK), rs10184573 (OR = 1.43, 95% CI = 1.26–1.61 near AC093590.1), and rs2499799 (OR = 0.62, 95% CI = 0.54–0.73 near UFL1/FHL5), collectively explaining 7.2% of the var- iance of cluster headache. SNPs rs11579212, rs10184573, and rs976357, as proxy SNP for rs2499799 (r2 = 1.0), repli- cated in the Norwegian sample (p < 0.05). Gene-based mapping yielded ASZ1 as possible fifth locus. RNA-sequencing indicated differential expression of POLR1B and TMEM87B in cluster headache patients. Interpretation: This genomewide association study (GWAS) identified and replicated genetic risk loci for cluster head- ache with effect sizes larger than those typically seen in complex genetic disorders. ANN NEUROL 2021;90:203–216 variants in ADCYAP1R1 and MME,13 but these findings C luster headache (CH) is a primary headache disorder characterized by attacks of intense unilateral orbital, supraorbital and/or temporal pain that last for 15 to were not replicated in a larger Swedish sample.14 To detect genetic variants for CH, we conducted a 180 minutes and are associated with ipsilateral facial auto- GWAS in a Dutch sample of 840 patients with CH and nomic symptoms and/or restlessness. The majority of 1,457 controls from the same geographical region. Results patients have episodic CH, with periods of attacks of were replicated in a Norwegian sample. Downstream analyses weeks to months, alternating with attack-free periods of at further assessed genes and mechanisms contributing to the least 3 months. In 10 to 15% of patients, cluster periods pathogenesis of CH. never remit for longer than 3 months for at least 1 year, classifying them as chronic CH. The male-to-female ratio Methods is 2:1.1 Smoking and psychiatric comorbidities are preva- Patient Recruitment and Sample Collection lent.2 Current treatment strategies include aborting acute The Dutch cluster headache study included 862 Dutch attacks and aim to reduce attack frequency with preventive patients with CH from the clinic-based Leiden University treatment.3 CH shows some phenotypic overlap with Cluster headache neuro-Analysis program (LUCA) that other trigeminal neuralgias, but also with migraine, for were recruited between 2010 and 2015 via the project’s example, in that some patients with migraine may also website. Patients with CH aged 18 years or older were report autonomic features. Certain similar pathophysiolog- included. Participants fulfilling the screening criteria ical pathways are hypothesized to be involved in both CH were asked to complete an extended questionnaire that and migraine.4 Although these disorders share prominent focused on signs and symptoms of CH as outlined in the features, they are clinically well distinguishable.5 International Classification of Headache Disorders The pathophysiology of CH is poorly understood, (ICHD-II or ICHD-III) criteria for CH.15,16 Individual although vasomotor changes, inflammation, hypothalamic diagnoses were made upon visiting the outpatient clinic or dysfunction, and dysregulation of the autonomic nervous sys- using a validated algorithm (positive predictive value: tem have been implicated as potential disease mechanisms.6  92%) based on ICHD criteria.17 CH cases were diag- Twin and family studies have highlighted the involvement of nosed in specialized headache centers to minimize mis- genetic factors in CH.7 Thus far, most genetic studies used a classification. Controls (n = 1,671) were obtained from hypothesis-driven approach and have examined a limited the Netherlands Epidemiology of Obesity Study (NEO) number of variants in genes linked to presumed pathways in study,18 a population-based sample that includes individ- CH. Most studied are variants in HCRTR2, which encodes uals aged 45 to 65 years living in a nearby municipality the hypocretin (orexin) type 2 receptor that binds neuropep- (Leiderdorp, The Netherlands) recruited between 2008 tides hypocretin-1 and -2 in the central nervous system. Still, and 2012. Most cases and all controls originated from the initially positive genetic findings for HCRTR2 associations8–10 same geographical region, in the Western part of the were not replicated in better-powered studies.11,12 Finally, Netherlands. All participants were unrelated and of the first, although very small hypothesis-free, Italian geno- European ancestry. The local ethics committees approved mewide association study (GWAS) investigating 99 patients the study. Written informed consent was obtained from with CH reported suggestive associations with genetic all participants. 204 Volume 90, No. 2 Harder et al: Genetic Loci in Cluster Headache Genotyping, Quality Control, and Imputation in apart than 500 kb, with association p < 5  108. Posi- the Discovery Stage tional gene mapping and fine mapping of significant loci Genomic DNA was extracted from peripheral blood leu- was performed using Functional Mapping and Annota- kocytes according to standard protocols and genotyping of tion version 1.3.6 (FUMA), Probabilistic Identification both cases and controls was performed using the Illumina of Causal SNPs (PICS), and Locuszoom.23–25 The pro- Infinium CoreExome-24 version 1.1 array according to portion of variance explained by a given SNP was calcu- the protocol from the manufacturer. Cases were gen- lated using Nagelkerke pseudo R2. otyped at the Genomics-Core Facility at the Norwegian University of Science and Technology (Trondheim, Patient Recruitment and Data Generation in the Norway) and controls at the Centre National de Gén- Replication Stage otypage (Paris, France). For the cases, variant calling was Cases were recruited at the Norwegian Advisory Unit on performed with Genome Studio 2.0 following a standard Headaches, St Olav’s Hospital, Trondheim (Norway) quality protocol,19 and the CHARGE best practice calling of between 2005 and 2016, with the inclusion criterion being the HumanExome Bead chip.20 For the controls, calling was the definite diagnosis of CH according to ICHD-II or performed using the GenCall algorithm using standard set- ICHD-III,15,16 made by a neurologist with special compe- tings as provided by Illumina. Quality control was performed tence in headache disorders to minimize misclassification. As according to standard procedures.21 Markers with high mis- controls, we used a random subset of 1,800 adult partici- singness rates (≥ 2%), monomorphic variants and those fail- pants from the Nord-Trøndelag Health Study (HUNT) ing the Hardy–Weinberg equilibrium were excluded. who did not have CH defined by the International Classifi- Individuals were excluded if they had a high proportion of cation of Disease (ICD)-10 diagnosis G44.0 (Cluster head- missing genotype data (≥ 2%), inconsistent sex information, ache syndrome) or the ICD-9 diagnosis 346.2 (Migraine were related (PI-HAT ≥ 0.2), or were heterozygosity outliers. variants, including cluster headache).26 Principal component analysis (PCA) was performed on the A sample of 159 CH cases were genotyped with the pruned data set (with a 50-kb sliding window, r2 > 0.2) Illumina Infinium CoreExome-24 version 1.1. Calling was using PLINK and population outliers were excluded. No performed with Genome Studio 2.0, using the cluster file overt population substructure between cases and controls was from the largest batch of 58,996 HUNT All-in controls (see observed (data not shown). After combining the genotyped below). The analysis followed the Genome Studio quality single nucleotide polymorphism (SNP) information from protocol,19 and the CHARGE best practice calling of the LUCA and NEO, imputation was performed on the Michi- HumanExome Bead chip.20 The HUNT control samples gan Imputation Server (https://imputationserver.sph.umich. were genotyped on 3 different Illumina HumanCoreExome edu/) using Haplotype Reference Consortium (HRC version arrays (HumanCoreExome12 version 1.0, HumanCore- 1.1 2016) as a reference panel after phasing by Eagle (version Exome12 version 1.1, and UM HUNT Biobank version 2.3),22 using the default parameters. 1.0), and called as described elsewhere.27 Markers with high In total, 345,064 SNPs from 2,297 individuals missingness rates (≥ 2%), monomorphic variants and those (840 cases and 1,457 controls) were available for impu- failing the Hardy–Weinberg equilibrium were excluded. tation. Prior to analyses, variants with minor allele fre- Individuals with high missingness rates (≥ 2%) or whose quency (MAF) ≤ 0.01 or imputation INFO score ≤ 0.6 inferred sex contradicted with reported sex were excluded. A were excluded, resulting in 7,578,399 SNPs. second round of quality control was performed after merging cases and all HUNT controls, excluding variants that were Statistical Analysis in the Discovery Stage monomorphic, deviated from Hardy–Weinberg equilibrium, Case–control SNP association analysis was performed or had different genotype rate between cases and controls. using a logistic regression model implemented in Individuals were excluded if they had missingness ≥ 2%, out- SNPTEST (version 2.5.2) for autosomal variants, with lying heterozygosity rate or were duplicates. Population out- case–control status as outcome and assuming additive liers and non-European samples were excluded. No overt allelic effects. The model was adjusted for sex. In addi- population substructure between cases and controls was tion, the model was adjusted for the first 4 principal observed (data not shown). A total of 69,440 individuals pas- components to minimize effects of confounding and sed quality control, including 144 cases. A dataset including population stratification. A Manhattan and a quantile- the 144 cases and 1,800 randomly selected controls was quantile (QQ) plot for test statistics were generated using imputed using Minimac3 (version 2.0.1) and the Hapmap R version 3.6.1 (R Foundation for Statistical Computing, r22 CEU panel. Variants with minor allele count < 3 or with Vienna, Austria). We determined lead SNPs that were imputation quality r2 < 0.3 were excluded, resulting in independent from each other at r2 < 0.1 and further 2,363,678 well-imputed variants for 144 cases (38 women August 2021 205 ANNALS of Neurology and 106 men) and 1,800 controls (952 women and assumes a single causal SNP for each trait. For CH, we used 848 men). The study was approved by the local ethics com- the summary statistics from the discovery cohort and for mittees. Written informed consent was obtained from all migraine we used the summary statistics from Gormley participants. et al35 without 23andMe (30,465 migraine cases and 143,147 controls); both populations are of European ances- Replication Analysis try. Colocalization was tested for the region between the Association analysis was performed using a mixed logistic 2 nearest recombination hotspots. regression model implemented in SAIGE (version 0.35.8.3), where CH was modeled as the dependent variable, and the Genetic Correlation genetic variants as the independent variable. Sex and the first LDSC was also used to calculate genetic correlation between 8 principal components were included as covariates. From CH and migraine.33 For migraine, we used summary statistics each independent significant locus (p < 5  108) in the dis- from Gormley et al35 without 23andMe (30,465 migraine covery sample, the lead SNP, or a proxy SNP, was selected cases and 143,147 controls), excluding variants with for replication. To correct for multiple testing, Bonferroni MAF ≤ 0.01, INFO score ≤ 0.6, large-effect variants or vari- correction was applied for the number of loci tested (n = 4). ants in an HLA region. In addition, the 38 genomewide signif- icant migraine loci were tested for association with CH.35 Sex Stratified Analysis Using the cor.test function in R, the correlation of the effect Analyses stratified for men and women were performed in size (beta) between migraine and CH (current study) was SNPTEST to examine possible sex-specific genetic effects. calculated. Both models were adjusted for the first 4 principal components. Gene-Based Analysis We performed the MAGMA gene-based association analy- Previously Reported Cluster Headache Loci sis implemented in FUMA, using default settings to iden- The 9 SNPs previously significantly associated with tify genes associated with CH.24 This calculates a gene CH8–10,13,28–32 were tested for association in our dis- test-statistic (p value) based on all SNPs located within covery analysis. The p values were adjusted for multi- genes. SNPs were assigned to the genes obtained from ple-testing using Bonferroni correction. Ensembl build 85 (only protein-coding genes). Univariate Linkage Disequilibrium Score Tissue Specificity Analyses Regression To further test the relationship between tissue-specific Linkage Disequilibrium Score Regression (LDSC version expression and genetic associations to CH, we examined all 1.0.1) was used to estimate the proportion of a true poly- SNPs and their respective effect on the expression of genes genic signal versus confounding factors, such as population up to 1 Mb away (cis-expression quantitative trait locus stratification, and to calculate SNP-based heritability.31 Vari- [eQTL]), using FUMA quantitative trait locus (eQTL) map- ants present in the HapMap 3 reference set were used, after ping (https://fuma.gtlab.nl/tutorial#eQTLs); all SNPs were excluding variants (1) with large-effect, explaining > 1% of mapped based on each of the tissues in the Genotype Tissue phenotype variation, or variants in liquid disequilibrium Expression (GTEx) version 8 dataset using default setting.24 (LD) with such; (2) with MAF ≤ 0.01 or imputation INFO Additionally, we performed tissue expression analysis based score ≤ 0.9; and (3) in the HLA region. Heritability esti- on the MAGMA gene property in FUMA.24 This analysis mates were converted to the liability scale assuming a popula- tests for positive relationships between tissue-specific gene tion prevalence of CH of 0.1%.6 expression in 30 general tissue types and 54 specific tissue types in the GTEx version 8 RNA-seq data and gene-based Colocalization Analysis p values from the gene-based analysis described above. To test whether the association signals for CH and migraine, on chromosome 6 near UFL1/FHL5, are consistent with a RNA-Sequencing of Patients With CH and shared causal variant, we used a Bayesian colocalization pro- Controls cedure using the R package “coloc” with default settings.34 The genes identified by eQTL mapping with FUMA (see This test generates posterior probabilities for each locus above) were further interrogated using existing RNA weighting the evidence for 5 competing hypotheses regarding sequencing (RNA-seq) data generated from peripheral the sharing of causal variants, namely H0 (no causal variant venous blood samples from 39 patients with CH and for either trait); H1 or H2 (a causal variant only for trait 20 controls matched for age, sex, and smoking habits. 1 or 2); H3 (distinct causal variants, for each trait); and H4 Data generation and quality control is described in detail (a single causal variant common to both traits). The analysis elsewhere.36 In short, RNA was extracted, using the 206 Volume 90, No. 2 Harder et al: Genetic Loci in Cluster Headache PAXgene Blood miRNA kit, and sequenced using lead SNPs had either a call rate or imputation metric close Illumina Hiseq4000. RNA-seq reads were aligned and to 100%. Three of the 4 lead SNPs were present in the processed using the in-house transcriptome analysis pipe- replication sample (rs11579212, rs6541998, and line Gentrap (version 0.3.1). Within this pipeline, rs10184573), whereas for the SNP on chromosome sequencing reads were aligned to the human genome refer- 6 (rs2499799) we selected a proxy SNP (rs976357, ence GRCh38 using TopHat (version 2.0.13) and coun- r2 = 1.0, D0 = 1). Lead SNPs of loci rs11579212 on chro- ted per gene using Htseq (version 0.6.1p1). The data were mosome 1, (OR = 1.58, 95% CI = 1.16–2.15) normalized for between-sample variation and for within- rs10184573 on chromosome 2 (OR = 1.74, 95% sample variation, using the Limma voom transformation. CI = 1.29–2.34) and rs976357 on chromosome 6 Differential expression analysis was performed in Limma, (OR = 0.44, 95% CI = 0.30–0.64) replicated after fitting a linear model correcting for age, gender, current Bonferroni correction (Table 3). smoking status, and leukocyte counts. The p values were The genomic inflation factor (λ) was 1.069 in the dis- adjusted for multiple-testing using Bonferroni correction. covery analysis, whereas the LD score regression intercept To determine the specificity of differential expression was 1.044 (SE 0.0077), indicating moderate inflation due to results obtained for CH, we examined the (nominally) signif- factors other than polygenic architecture. We estimated the icant genes from the CH RNA-seq analysis in RNA-seq data SNP-based heritability (h2) of CH at 30.3% (SE = 19.4%) obtained from 26 patients with migraine and 20 age- and on the observed scale. Assuming a population prevalence of sex-matched controls. Data generation and quality control 0.1% for CH this corresponds to a h2 of 11.5% have been previously described.37 In short, peripheral venous (SE = 7.4%) on the liability scale. blood samples were drawn when the patients with migraine Fine mapping with PICS identified two variants with were migraine-free for at least 5 days and headache-free for causal probability larger than 0.2, at rs11579212 (PICS 24 hours. RNA was extracted using PAXgene Blood RNA probability = 0.40) and rs10184573 (PICS probabil- kit, sequenced (using Illumina Novaseq) and aligned. RNA- ity = 1.0), respectively. seq reads were, after quality control, aligned to the human reference transcriptome, using kallisto (version 0.42.5). Resulting count matrices were corrected for library size and Genetic Correlation of Cluster Headache with gene length, and normalized using the R package DESeq2. Migraine Differential expression was performed using the R package The observed h2 for migraine was 17.1% (SE = 1.56%). DESeq2 by fitting a generalized linear model, correcting The genetic correlation between CH and migraine was 0.33 for age. (SE = 0.021, p = 0.12). Next, we examined the 38 migraine-associated loci reported in Gormley et al.35 Of Results the 37 migraine loci that were represented in our data set Study Participants directly or by variants in high LD, one, located on chromo- The clinical characteristics of cases and controls of the some 6, was associated with CH rs2971606, a proxy discovery sample are summarized in Table 1. There was (r2 = 1.0) for the migraine index variant rs67338227, in a higher proportion of men (69% vs 44%) and smokers FHL5, p = 1.39  108 (Bonferroni corrected (52% vs 14%) among the cases compared to controls. pcorr = 0.5  106). The association had the same effect Most patients had episodic CH (69%). A total of 13% direction for migraine and CH. There was also moderate LD of cases had migraine. between the lead SNP for migraine and CH (r2 = 0.64 in data from 1000 Genomes Project Phase 3 CEU). Still, Association Analysis colocalization analysis revealed a 67.5% posterior probability Overall association results are shown in the Manhattan for the hypothesis (H3) that the causal variants for CH and plot (Fig 1A) and the QQ plot (Fig 2). In total, 4 indepen- migraine are distinct, higher than the 32.5% posterior proba- dent loci showed genomewide significant (p < 5  108) bility for hypothesis (H4) that CH and migraine share a associations with CH (Fig 1B–E) with a combined causal variant in this region. explained variance of 7.2%. More specifically, we identi- The other 36 migraine loci were not associated with fied rs11579212 (odds ratio ]OR = 1.51, 95% confidence CH (data not shown), with the second strongest associa- interval [CI] = 1.33–1.72 near RP11-815 M8.1), tion seen for rs10786156 in PLCE1 (p = 2.82  103, rs6541998 (OR = 1.53, 95% CI = 1.37–1.74 near pcorr = 0.10). The migraine locus near MED14 on chro- MERTK), rs10184573 (OR = 1.43, 95% CI = 1.26– mosome X (rs12845494) was not represented in our 1.61 near AC093590.1), and rs2499799 (OR = 0.62, dataset. The effect sizes for the 37 loci combined corre- 95% CI = 0.54–0.73 near UFL1/FHL5; Table 2). These lated with those of CH, Pearson’s r(35) = 0.59 August 2021 207 ANNALS of Neurology TABLE 1. Clinical Characteristics of the Study Samplea Discovery sample Characteristics Patients with CH (n = 840) Controls (n = 1,457) p valueb Men 579 (68.9) 636 (43.7) < 0.001 Current daily smoking 440 (52.4) 202 (13.9) < 0.001 Episodic cluster headache 577 (68.7) - - Chronic cluster headache 233 (27.7) - - Migraine comorbidity 106 (12.6) - - Data are expressed as numbers (percentages) unless otherwise stated. a Numbers and proportions may not add up to total of 100 due to rounding or missing values. b The p values of chi-square test for categorical variables. CH = cluster headache. (p = 1.36  104), even disregarding the overlapping genes derived from FUMA. Eleven genes were expressed in FHL5 locus (Pearson’s r(34) = 0.58, p = 2.18  104). the samples, of which one was differentially lower expressed (POLR1B, p = 7.50  105, pcorr = 8.3  104) and one Sex-Stratified Analyses was nominally differentially lower expressed (TMEM87B, The low number of female cases gave limited power for p = 0.014, pcorr = 0.15) in CH cases than in controls, both the women-only analysis. Rs6541998 was genomewide genes representing the rs6541998 locus (Table 5). The significant in men using sex-stratified analyses; all other 2 genes were not differentially expressed in RNA-seq data loci were nominally significant (p < 1  103) for both when comparing 26 patients with migraine with 20 controls men and women with effects in the same direction. Using (p = 0.50 and 0.45, respectively; see Table 5). the method suggested by Clogg et al,38 we found no sig- nificant differences for the regression coefficients between men and women at the 4 lead SNPs (p value 0.54 for Discussion rs11579212; p value 0.62 for rs6541998; p value 0.57 for rs10184573; and p value 0.59 for rs2499799). We performed a GWAS in CH and identified 4 independent genetic risk loci, of which 3 replicated in an independent Previously Reported Cluster Headache Loci sample. The association effect sizes, with ORs around 1.5, Of the 9 different SNPs previously associated with CH, are high compared to those usually observed in GWAS one replicated, rs1800759 in ADH4 (p = 0.00039, (https://www.ebi.ac.uk/gwas/).39 Whereas this may indicate Bonferroni corrected pcorr = 0.0035; Table 4). In contrast, that the risk for CH is driven by a limited number of loci none of the previously reported associations in HCRTR2, with strong associations with CH, it is likely to be expected ADCYAP1R1, CLOCK, CHRNA3-CHRNA5, and MME that follow-up studies with larger sample sizes also will iden- were replicated in our sample (see Table 4). tify loci with smaller effect sizes. Except for the MERTK locus (rs6541998), all loci replicated in our replication sam- Downstream Bioinformatic Analysis ple, suggesting that the signals are genuine. Gene-based map- Using FUMA gene-based eQTL mapping, 16 genes were ping additionally found that expression of the ASZ1 gene mapped to the 4 loci (see Table 2). Additionally, gene- may be influenced by one or more CH loci, providing a pos- mapping with MAGMA identified 5 genes whose expression sible additional locus. RNA-seq results show altered expres- was significantly influenced by the CH loci, TMEM87B, sion in patients with CH of POLR1B and TMEM87B, MERTK, FHL5, UFL1, and ASZ1 (Fig 3). Finally, we per- suggesting their involvement in CH. Although there seems formed a MAGMA tissue expression analysis, which did not to be a considerable SNP-based heritability for CH, a robust render any significant results (data not shown). estimation of SNP-based heritability is not possible given the small sample size, hence heritability estimates should be RNA-Seq Analyses interpreted with caution. Using RNA-seq data from white blood cells of 39 patients The main limitations of our study are that, with CH and 20 controls, we assessed the 16 eQTL-mapped (1) although we identified and replicated genomewide 208 Volume 90, No. 2 Harder et al: Genetic Loci in Cluster Headache FIGURE 1: Manhattan plot and reginal plots for the discovery analysis. (A) Manhattan plot showing the -log10 p value for each SNP. Each marker was tested for association using an additive genetic model by logistic regression. The horizontal axis shows the chromosomal position and the vertical axis shows the significance of tested markers from logistic regression. The threshold for genome wide significance (p < 5  108) is indicated by a red dotted line. Markers that reach genomewide significance are shown in blue. (B–E) Regional Manhattan plots of the 4 genomewide significant cluster headache loci, with +/ 600 kb-window. Each dot represents an SNP, the horizontal axis gives the genomic coordinate and the vertical axis the significance level (log10 p value). The index SNP for each locus is marked with a purple diamond and annotated with its corresponding location number (CRCh37/hg19). SNPs are colored based on their correlation (r2) with the labeled lead SNP according to the legend. The solid blue line shows the recombination rate from 1000 Genomes (EUR) data (right vertical axis). Gencode genes are shown. Figures were obtained from LocusZoom.24 (B) Locus: rs11579212 and 1:222072819. (C) Locus: rs6541998 and 2:112785237. (D) Locus: rs10184573 and 2:200448253. (E) Locus: rs2499799 and 6:96851676. SNP = single nucleotide polymorphisms. August 2021 209 ANNALS of Neurology other than European ancestry; and (3) although cases and controls were genotyped using the same platform, genotyping was performed in different laboratories possi- bly introducing batch effects. Therefore, we made signifi- cant efforts to circumvent possible problems arising from our design by rigorous quality control. Overall, our case sample was representative of the general CH population with an  2:1 male–female ratio, chronic CH of  30% and without any familial confounder, as familial cases were removed in the quality control steps.1 The difference in the percentage of men and women for cases and controls was corrected for in the statistical analysis. Among previously suggested loci to be involved in CH, we found evidence for significant association to the alcohol dehydrogenase 4 gene (ADH4), although the effect identified is opposite to what was previously reported and FIGURE 2: Quantile-quantile (QQ) plot for association with at the genome wide level it was not significant.32 In previ- cluster headache (CH). The horizontal axis shows -log10 p ous studies, ADH4 was investigated mainly because alco- values expected under the null distribution. The vertical axis hol is considered both a trigger and possible risk factor for shows observed -log10 p values. Red = common SNPs (MAF transformation from episodic to chronic CH.31,32 Of ≥0.05), blue = low frequency SNPs (MAF = 0.005–0.05). Genomic inflation factor (λ) = 1.069. MAF = minor allele note, we did not find evidence for an association of frequency; SNPs = single nucleotide polymorphisms. HCRTR2, as reported previously,8–10 nor for any of the other previously reported loci in CH. significant loci, the relatively small number of cases in the A remarkable finding in our study was that one of discovery sample will leave loci with smaller effect sizes or the leading loci, represented by rs2499799, which covers lower allele frequencies hidden; (2) it is unclear to what both FHL5 and UFL1, has previously been identified as a extent the present results can be extrapolated to ancestries migraine risk locus.35 FHL5 encodes a transcription factor TABLE 2. SNPs at the Four Loci Associated With Cluster Headache Discovery Sample Chr:Pos SNP EA:NEAa EAF OR [95% CI] p valueb Nearest genec eQTL mapped genesd rs11579212 1:222072819 0.34 1.51 [1.33–1.72] 4.78  1010 RP11-815 M8.1 DUSP10 C:A rs6541998 2:112785237 0.63 1.53 [1.37–1.74] 1.91  1010 MERTK TTL POLR1B FBLN7 C:T ZC3H8 MERTK TMEM87B RGPD8 ZC3H6 rs10184573 2:200448253 T: 0.44 1.43 [1.26–1.61] 2.20  108 AC093590.1 SATB2 FTCDNL1 G rs2499799 6:96851676 C: 0.81 0.62 [0.54–0.73] 1.29  109 UFL1/FHL5 UFL1 FHL5 T GPR63 MMS22L FUT9 a Chromosomal positions in GRCh37/hg19 coordinates. b Significant result (p < 5  108). c The nearest gene is based on ANNOVAR annotations with Ensembl build version 85. d The eQTL mapping was done in FUMA based on GTEx version 8. Chr = chromosome; CI = confidence interval; EA = effect allele; EAF = effect allele frequency; NEA = non effect allele; OR = odds ratio; Pos = position; SNP = single nucleotide polymorphism. 210 Volume 90, No. 2 Harder et al: Genetic Loci in Cluster Headache TABLE 3. Replication of the Significant Loci in an Independent Sample L Chr SNP discovery sample SNP replication sample Posa OR [95% CI] p valueb Directionc 1 1 rs11579212 rs11579212 222072819 1.58 [1.16–2.15] 3.50  103 + 2 2 rs6541998 rs6541998 112785237 1.04 [0.78–1.40] 0.78 + 3 2 rs10184573 rs10184573 200448253 1.74 [1.29–2.34] 2.78  104 + 5 4 6 rs2499799 rs976357 (r = 1.0) 2 96849679 0.44 [0.30–0.64] 2.76  10 + a Chromosomal positions in GRCh37/hg19 coordinates for the replication SNP. b Significant result (p < 0.05/4). c Direction; Same (+) or opposite () direction of association for discovery and replication analyses. Chr = chromosome; CI = confidence interval; L = locus number; OR = odds ratio; Pos = chromosomal position; SNP = single nucleotide polymorphism. that regulates cAMP-responsive elements CREB6 and ubiquitin protease system (UPS) has been associated as a CREM, which play a role in synaptic plasticity and mem- pathway in neuropsychiatric and neurodegenerative disor- ory formation.40 UFL1 codes for the ubiquitin-fold modi- ders.42 In the latest migraine GWAS meta-analysis, the fier 1 (UFM1)-specific ligase 1, an ubiquitin-like protein UFL1/FHL5 locus had an OR of 1.09 (1.08–1.11) based that allows UFL1 to conjugate to its substrates.41 The on the primary signal (rs67338227).35 The direction of TABLE 4. The Association of Previously Reported Cluster Headache Loci in Our Discovery Cluster Headache Sample Association with CH discovery Previously reported CH loci sample Nearest coding gene Index SNP EA OR [95%CI] p value Ref OR [95%CI] pcorr-valuea ADCYAP1R1 rs12668955 G 0.48 [0.34–0.07] 9.1  106 13 1.05 [0.91–1.21] 1 ADH4 rs1126671 A 2.33 [1.25–4.37] 0.006 b 29 0.87 [0.76–0.99] 0.36 ADH4 rs1126671 A - 0.03 30 0.87 [0.76–0.99] 0.36 ADH4 rs1800759 A - 0.03 30 0.80 [0.70–0.90] 0.0035 CLOCK rs12649507 A 1.29 [1.08–1.54] 0.02 28 0.92 [0.80–1.05] 1 CHRNA3-CHRNA5 rs578776 A - 0.038 26 0.95 [0.83–1.09] 1 HCRTR2 rs2653349 G 6.79 [2.25–22.99] < 0.0002 8 1.08 [0.93–1.26] 1 HCRTR2 rs2653349 G 1.97 [1.32–2.92] 0.0007 10 1.08 [0.93–1.26] 1 HCRTR2 rs3122156 G 0.82 [0.68–0.99] 0.0421 (Pcorr 9 0.92 [0.80–1.06] 1 0.126) HCRTR2 rs10498801 G 0.69 [0.49–0.97] 0.030 27 1.03 [0.88–1.21] 1 MME rs147564881 C - 0.019 13 0.24 [0.42–41.51] 1 The SNPs previously reported to be associated with cluster headache and the corresponding OR (based on the same EA) and p values for these SNPs in the discovery sample. a The p values were Bonferroni corrected for 9 tests. b The p value is based on the carriers with homozygous AA genotype compared with GG/GA genotypes. CH = cluster headache; CI = confidence interval; EA = effect allele; OR = odds ratio; SNP = single nucleotide polymorphism. August 2021 211 ANNALS of Neurology FIGURE 3: Gene-based Manhattan plot. Input SNPs were mapped to 18,795 protein coding genes. The horizontal axis shows the chromosomal position and the vertical axis shows the significance of tested markers. The threshold for genome wide significance (p = 0.05/18,795 = 2.66  106) is indicated by a dotted line. SNP = single nucleotide polymorphism. [Color figure can be viewed at www.annalsofneurology.org] the effect in the UFL1/FHL5 locus in our dataset was the colocalization analysis suggested that CH and migraine are same in both migraine and CH and, the lead SNPs for more likely caused by distinct variants at this locus. migraine and CH were in LD (r2 = 0.64). Our Admittedly, this finding could also be a result of different TABLE 5. RNA-seq Expression Data Gene Expression in blood Locus number CH p value CH pcorr valuea M p value POLR1B Yes 2 7.50  105 8.3  104 0.50 TMEM87B Yes 2 0.014 0.15 0.45 ZC3H8 Yes 2 0.084 0.92 - DUSP10 Yes 1 0.725 1 - MERTK Yes 2 0.527 1 - TTL Yes 2 0.465 1 - MMS22L Yes 4 0.424 1 - FBLN7 Yes 2 0.361 1 - ZC3H6 Yes 2 0.285 1 - FTCDNL1 Yes 3 0.123 1 - UFL1 Yes 4 0.123 1 - FHL5 No 4 - - - FUT9 No 4 - - - GPR63 No 4 - - - RGPD8 No 2 - - - SATB2 No 3 - - - 22 Genes were selected based in the eQTL mapping in FUMA. a The p values were Bonferroni corrected for 11 tests. CH = cluster headache; eQTL = expression quantitative trait locus; FUMA = Functional Mapping and Annotation; M = migraine. 212 Volume 90, No. 2 Harder et al: Genetic Loci in Cluster Headache LD patterns in the samples that were compared in the POST-SCRIPT PARAGRAPH colocalization analysis. The other 36 independent loci Two parallel manuscripts (Harder et al and O’Connor implicated in migraine showed no association with et al), submitted to the journal, report the first replicated CH. Our results suggest though that the UFL1/FHL5 genomic loci associated with CH. Whereas Harder et al locus is “specific” for CH and that the association is not due investigated Dutch CH cases (n = 840) and controls to the mere presence of comorbid migraine among patients (n = 1,457) and Norwegian CH cases (n = 144) and with CH. This is further supported by the similar prevalence controls (n = 1,800), O’Connor et al investigated UK of migraine among cases in our discovery sample (13%) and cases (n = 852) and controls (n = 5,614) as well as Swed- the expected population prevalence (10–17%), although the ish cases (n = 591) and controls n = 1,134). The 4 loci number of migraine cases in controls was not collected.43 reported by Harder et al correspond to 4 loci reported by While no other migraine locus reached significance in our O’Connor et al, with the index variants reported in the study individually, there was a moderate correlation between 2 studies being in linkage disequilibrium with each other association effect sizes of CH and migraine for the 37 exam- (D0 = 0.86 and r2 = 0.36 for rs11579212 and ined migraine loci. This may reflect a shared genetic architec- rs12121134; D0 = 0.98 and r2 = 0.95 for rs6541998 and ture underlying both disorders, which is not surprising given rs4519530; D0 = 0.95 and r2 = 0.34 for rs10184573 and that they share pathophysiological features, such as the rs113658130; and D0 = 0.93 and r2 = 0.38 for involvement of the trigeminovascular system and efficacy of rs2499799 and rs11153082, in the 1000 Genomes data calcitonin gene-related peptide (CGRP) monoclonal anti- for European populations). The independent discovery of bodies and triptans.4 It is possible that future studies with the 4 loci in the 2 studies provides additional support that larger sample sizes may identify the involvement of more they represent genuine risk loci for cluster headache. migraine loci in CH. Next, we combined the summary statistics from the With respect to the other replicated loci, four studies (Dutch, Norwegian, United Kingdom, and rs11579212 and rs10184573, which mapped to Swedish) using inverse-variance weighted meta-analysis as RP11-815 M8.1 and AC093590.1, respectively, they have implemented in METAL (with the “STDERR” option), not previously been related to disease, and their role in after harmonizing the datasets using EasyQC.47,48 In total, CH pathogenesis remains unclear. Although rs6541998 8,039,373 variants were analyzed. The association to CH did not replicate in the small replication sample, 2 genes remained significant for all 8 index variants (in the 4 loci) (POLR1B and TMEM87B) in the locus showed differen- reported in the 2 papers: rs11579212 (effect allele, EA: C), tial expression in CH compared to controls in RNA-seq OR = 1.31 (95% CI = 1.21–1.41), p value 8.98  1013; analyses, whereas no such effect was seen in migraine data. rs12121134 (EA: T), OR = 1.40 (95% CI = 1.29–1.53), p POLR1B, encoding DNA-directed RNA polymerase I value 9.18  1015; rs6541998 (EA: C), OR = 1.40 (95% subunit RPA2, has been associated with Treacher Collins CI = 1.30–1.51), p value 2.37  1019; rs4519530 (EA: and TMEM87B, encoding transmembrane protein 87B C), OR = 1.41 (95% CI = 1.31–1.52), p value may be involved in restrictive cardiomyopathy.44,45 4.18  1029; rs10184573 (EA: T), OR = 1.38 (95% MERTK, the nearest gene, encodes a receptor tyrosine CI = 1.28–1.50), p value 3.35  1016; rs113658130 (EA: kinase of the TAM (Tyro3, Axl, and MERTK) family, is C), OR = 1.54 (95% CI = 1.41–1.69), p value among other tissues expressed in oligodendrocytes, astro- 1.28  1021; rs2499799 (EA: C), OR = 0.77 (95% cytes, and microglia in the brain and has an effect on the CI = 0.70–0.84), p value 2.73  108; rs11153082 (EA: immune response.46 Unfortunately, the number of associ- G), OR = 1.33 (95% CI = 1.23–1.43), p value ated genes with CH is not large enough to perform mean- 2.98  1014. The 8 index variants in the overlapping loci ingful further downstream pathway analyses. Based on the showed a consistent effect direction across the 2 studies. regression coefficients, we found no evidence for a differ- Colocalization analysis, to determine whether the reported ent effect for the lead SNPs between men and women. loci of both papers represent the same causal variants, identi- In conclusion, this GWAS of CH reveals 4 genetic fied a high posterior probability for 3 loci (those on chromo- risk loci for CH with unusually high effect sizes for a com- somes 1 and 2) to likely represent the same causal variant.34 plex disorder, of which 3 replicated in an independent Rs12121134 and rs11579212 have a posterior probability sample. One of the loci has previously been identified as a that the causal variants are the same (H4) of 80.4%, for migraine risk locus. Our results suggest several genes to be rs4519530 and rs6541998 H4 is 87.4% and involved in the pathogenesis of CH and offer a starting for rs113658130 and rs10184573 H4 is 96.9%. For the point for future research to elucidate the molecular mech- locus on chromosome 6, the colocalization analysis shows a anisms of this severe disease. higher probability that the loci in the 2 studies represent August 2021 213 ANNALS of Neurology distinct causal variants (H3: 78.7%) rather than the same Genomics Core Facility (GCF), Norwegian University of causal variant (H4: 21.2%). Science and Technology (NTNU). The GCF is funded Finally, the meta-analysis resulted in three additional by the Faculty of Medicine and Health Sciences at loci becoming genomewide significant: (1) a locus on NTNU and Central Norway Regional Health Authority. chromosome 7 with 31 significant (p value <5  108) The NEO study is supported by the participating Depart- variants with index variant rs6966836 (chr7:117002998, ments, the Division and the Board of Directors of the Lei- EA: C), OR = 1.25 (95% CI = 1.16–1.35), p value den University Medical Centre, and by the Leiden 2.06  109; (2) a locus on chromosome 10 with 2 signif- University, Research Profile Area “Vascular and Regenera- icant variants with index variant rs10786156 tive Medicine.” The funding organizations had no role in (chr10:96014622, EA: C), OR = 1.24 (95% CI = 1.15– the design and conduct of the study; collection, manage- 1.33), p value 7.61  109; and (3) a locus on chromo- ment, analysis, and interpretation of the data; preparation, some 19 with 2 significant variants with index variant review, or approval of the manuscript; and decision to rs60690598 (chr19:55052198, EA: T), OR = 1.87 (95% submit the manuscript for publication. CI = 1.51–2.33), p value 1.70  108. Author Contributions A.V.E.H., B.S.W., R.N., E.T., T.F.H., R.H.J., M.D.F., Acknowledgments J.-A.Z., G.M.T., and A.M.J.M.v.d.M. contributed to the The authors thank the patients for their participation in conception and design of the study. A.V.E.H., B.S.W., this project. The Nord-Trøndelag Health Study (The R.N., L.S.V., S.B., L.J.A.K., I.d.B., F.R.R., K.W.v.D., HUNT Study) is a collaboration between HUNT E.O.C., C.F., E.S.K., L.F.T., R.F., and P.P.-R. contrib- Research Centre (Faculty of Medicine and Health Sci- uted to the acquisition and analysis of data. A.V.E.H., ences, NTNU, Norwegian University of Science and B.S.W., and R.N. contributed to drafting the text and Technology), Trøndelag County Council, Central Norway preparing the figures. Regional Health Authority, and the Norwegian Institute of Public Health. The genotyping was financed by the Members of “Cluster Headache Genetics National Institute of health (NIH), University of Michi- Working Group” gan, The Norwegian Research council, and Central Nor- All members of the “Cluster Headache Genetics Working way Regional Health Authority and the Faculty of Group” were actively involved in patient collection. The Medicine and Health Sciences, Norwegian University members of the Cluster Headache Genetics Working group of Science and Technology (NTNU). The genotype qual- include Roemer B. Brandt, MD, Ilse F. de Coo, MD, and ity control and imputation has been conducted by the Patty G.G. Doesborg, MD, all at the Leiden University K.G. Jebsen Center for Genetic Epidemiology, Depart- Medical Center (LUMC), Leiden, The Netherlands; Roser ment of Public Health and Nursing, Faculty of Medicine Corominas, PhD, from the University of Barcelona, Barce- and Health Sciences, Norwegian University of Science lona, Spain; Institut de Biomedicina de la Universitat de and Technology (NTNU). We express our gratitude to all Barcelona (IBUB), Barcelona, Spain; Instituto de Salud individuals who participated in the Netherlands Epidemi- Carlos III, Spain; Institut de Recerca Sant Joan de Déu ology in Obesity study. We are grateful to all participating (IR-SJD), Esplugues de Llobregat, Spain; Victor J. Gallardo, general practitioners for inviting eligible participants. We MSc, at the Headache Research Group, Vall d’Hebron Insti- also thank P. van Beelen and all research nurses for col- tute of Research (VHIR), Universitat Autònoma de Barce- lecting the data and P. Noordijk and her team for sample lona, Barcelona, Spain; Nunu Lund, MD, PhD, from the handling and storage and I. de Jonge, MSc for data man- Danish Headache Center, Rigshospitalet Glostrup, Glostrup, agement of the NEO study. This work was supported by Denmark; Paavo Häppölä, MSc, at University of Helsinki, the South-Eastern Norway Regional Health Authority Helsinki, Finland. (grant no. 2015089 to J.A.Z.), and grants of the Nether- lands Organization for Scientific Research, that is the Potential Conflicts of Interest Center of Medical System Biology established by The authors declared no conflict of interest. the Netherlands Genomics Initiative/Netherlands Organi- sation for Scientific Research (to A.M.J.M.v.d.M.), Spi- noza 2009 (to M.D.F.) and EU-funded FP7 “EUROHEADPAIN” (grant no. 6026337 to M.D.F., References 1. Lund N, Barloese M, Petersen A, et al. Chronobiology differs G.M.T, and A.M.J.M.v.d.M.). The genotyping of Dutch between men and women with cluster headache, clinical phenotype and Norwegian case samples was provided by the does not. Neurology 2017;88:1069–1076. 214 Volume 90, No. 2 Harder et al: Genetic Loci in Cluster Headache 2. Lund N, Petersen A, Snoer A, et al. Cluster headache is associated 24. Watanabe K, Taskesen E, van Bochoven A, et al. Functional mapping with unhealthy lifestyle and lifestyle-related comorbid diseases: and annotation of genetic associations with FUMA. Nat Commun results from the Danish cluster headache survey. Cephalalgia 2019; 2017;8:1826. 39:254–263. 25. Pruim RJ, Welch RP, Sanna S, et al. LocusZoom: regional visualiza- 3. Brandt RB, Doesborg PGG, Haan J, et al. Pharmacotherapy for clus- tion of genome-wide association scan results. Bioinformatics 2010; ter headache. CNS Drugs 2020;34:171–184. 26:2336–2337. 4. Vollesen AL, Benemei S, Cortese F, et al. Migraine and cluster head- 26. Krokstad S, Langhammer A, Hveem K, et al. Cohort profile: the ache - the common link. J Headache Pain 2018;19:89. HUNT study, Norway. Int J Epidemiol 2013;42:968–977. 5. Headache classification Committee of the International Headache 27. Ferreira MA, Vonk JM, Baurecht H, et al. Shared genetic origin of Society (IHS) the international classification of headache disorders, asthma, hay fever and eczema elucidates allergic disease biology. 3rd edition. Cephalalgia 2018;38:1–211. Nat Genet 2017;49:1752–1757. 6. May A, Schwedt TJ, Magis D, et al. Cluster headache. Nat Rev Dis 28. Cainazzo MM, Tiraferri I, Ciccarese M, et al. O015. Evaluation of the Primers 2018;4:18006. genetic polymorphism of the α3 (CHRNA3) and α5 (CHRNA5) nico- tinic receptor subunits, in patients with cluster headache. 7. Sjaastad O, Shen JM, Stovner LJ, et al. Cluster headache in identical J Headache Pain 2015;16:A88. twins. Headache 1993;33:214–217. 29. Fan Z, Hou L, Wan D, et al. Genetic association of HCRTR2, ADH4 8. Rainero I, Gallone S, Valfrè W, et al. A polymorphism of the and CLOCK genes with cluster headache: a Chinese population- hypocretin receptor 2 gene is associated with cluster headache. Neu- based case-control study. J Headache Pain 2018;19:1. rology 2004;63:1286–1288. 30. Fourier C, Ran C, Zinnegger M, et al. A genetic CLOCK variant asso- 9. Fourier C, Ran C, Steinberg A, et al. Analysis of HCRTR2 gene vari- ciated with cluster headache causing increased mRNA levels. ants and cluster headache in Sweden. Headache 2019;59:410–417. Cephalalgia 2018;38:496–502. 10. Schürks M, Kurth T, Geissler I, et al. Cluster headache is associated 31. Rainero I, Rubino E, Gallone S, et al. Cluster headache is associated with the G1246A polymorphism in the hypocretin receptor 2 gene. with the alcohol dehydrogenase 4 (ADH4) gene. Headache 2010;50: Neurology 2006;66:1917–1919. 92–98. 32. Zarrilli F, Tomaiuolo R, Ceglia C, et al. Molecular analysis of cluster 11. Weller CM, Wilbrink LA, Houwing-Duistermaat JJ, et al. Cluster headache. Clin J Pain 2015;31:52–57. headache and the hypocretin receptor 2 reconsidered: a genetic association study and meta-analysis. Cephalalgia 2015;35:741–747. 33. Bulik-Sullivan BK, Loh PR, Finucane HK, et al. LD score regression distinguishes confounding from polygenicity in genome-wide associ- 12. Gibson KF, Santos AD, Lund N, et al. Genetics of cluster headache. ation studies. Nat Genet 2015;47:291–295. Cephalalgia 2019;39:1298–1312. 34. Giambartolomei C, Vukcevic D, Schadt EE, et al. Bayesian test for 13. Bacchelli E, Cainazzo MM, Cameli C, et al. A genome-wide analysis colocalisation between pairs of genetic association studies using in cluster headache points to neprilysin and PACAP receptor gene summary statistics. PLoS Genet 2014;10:e1004383. variants. J Headache Pain 2016;17:114. 35. Gormley P, Anttila V, Winsvold BS, et al. Meta-analysis of 375,000 14. Ran C, Fourier C, Michalska JM, et al. Screening of genetic variants individuals identifies 38 susceptibility loci for migraine. Nat Genet in ADCYAP1R1, MME and 14q21 in a Swedish cluster headache 2016;48:856–866. cohort. J Headache Pain 2017;18:88. 36. Eising E, Pelzer N, Vijfhuizen LS, et al. Identifying a gene expression 15. Pelzer N, Haan J, Stam AH, et al. Clinical spectrum of hemiplegic signature of cluster headache in blood. Sci Rep 2017;7:40218. migraine and chances of finding a pathogenic mutation. Neurology 2018;90:e575–e582. 37. Kogelman LJ, Falkenberg K, Halldorsson GH, et al. Comparing migraine with and without aura to healthy controls using RNA 16. Headache Classification Subcommittee of the International sequencing. Cephalalgia 2019;39:1435–1444. Headache Society. The International Classification of Headache Dis- orders: 2nd edition. Cephalalgia. 2004;24 Suppl 1:9–160. 38. Clogg CC, Petkova E, Haritou A. Statistical methods for comparing regression coefficients between models. Am J Sociol 1995;100: 17. Wilbrink LA, Weller CM, Cheung C, et al. Stepwise web-based ques- 1261–1293. tionnaires for diagnosing cluster headache: LUCA and QATCH. 39. Visscher PM, Wray NR, Zhang Q, et al. 10 years of GWAS discovery: Cephalalgia 2013;33:924–931. biology, function, and translation. Am J Hum Genet 2017;101:5–22. 18. de Mutsert R, den Heijer M, Rabelink TJ, et al. The Netherlands epi- 40. Anttila V, Winsvold BS, Gormley P, et al. Genome-wide meta-analysis demiology of obesity (NEO) study: study design and data collection. identifies new susceptibility loci for migraine. Nat Genet 2013;45: Eur J Epidemiol 2013;28:513–523. 912–917. 19. Guo Y, He J, Zhao S, et al. Illumina human exome genotyping array 41. Komatsu M, Chiba T, Tatsumi K, et al. A novel protein-conjugating sys- clustering and quality control. Nat Protoc 2014;9:2643–2662. tem for Ufm1, a ubiquitin-fold modifier. EMBO J 2004;23:1977–1986. 20. Grove ML, Yu B, Cochran BJ, et al. Best practices and joint calling of 42. Cheon S, Dean M, Chahrour M. The ubiquitin proteasome pathway in the HumanExome BeadChip: the CHARGE consortium. PLoS One neuropsychiatric disorders. Neurobiol Learn Mem 2019;165:106791. 2013;8:e68095. 43. Lipton RB, Stewart WF. Prevalence and impact of migraine. Neurol 21. Anderson CA, Pettersson FH, Clarke GM, et al. Data quality control Clin 1997;15:1–13. in genetic case-control association studies. Nat Protoc 2010;5:1564– 1573. 44. Sanchez E, Laplace-Builhe B, Mau-Them FT, et al. POLR1B and neu- ral crest cell anomalies in Treacher Collins syndrome type 4. Genet 22. Loh PR, Danecek P, Palamara PF, et al. Reference-based phasing Med 2020;22:547–556. using the haplotype reference consortium panel. Nat Genet 2016; 48:1443–1448. 45. Yu HC, Coughlin CR, Geiger EA, et al. Discovery of a potentially del- eterious variant in TMEM87B in a patient with a hemizygous 2q13 23. Farh KK, Marson A, Zhu J, et al. Genetic and epigenetic fine map- microdeletion suggests a recessive condition characterized by con- ping of causal autoimmune disease variants. Nature 2015;518: genital heart disease and restrictive cardiomyopathy. Cold Spring 337–343. Harb Mol Case Stud 2016;2:a000844. August 2021 215 ANNALS of Neurology 46. Tondo G, Perani D, Comi C. TAM receptor pathways at the cross- 48. Winkler TW, Day FR, Croteau-Chonka DC, et al. Quality control and roads of Neuroinflammation and neurodegeneration. Dis Markers conduct of genome-wide association meta-analyses. Nat Protoc 2019;2019:2387614. 2014;9:1192–1212. 47. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics 2010;26:2190–2191. 216 Volume 90, No. 2