Identification of X-Linked Genes in Migraine: Fine Mapping and Candidate Gene Studies
Author Maher, Bridget Helen
Published 2012
Thesis Type Thesis (PhD Doctorate)
School School of Medical Science
DOI https://doi.org/10.25904/1912/2744
Copyright Statement The author owns the copyright in this thesis, unless stated otherwise.
Downloaded from http://hdl.handle.net/10072/367770
Griffith Research Online https://research-repository.griffith.edu.au
Identification of X-Linked Genes in Migraine: Fine Mapping and Candidate Gene Studies
Bridget Helen Maher BBiotechInnov(Hons), MAppLaw(IPLaw)
School of Medical Science Griffith Health Griffith University
Submitted in fulfilment of the requirements of the degree of Doctor of Philosophy
September 2011 ii Abstract
Migraine is a painful temporarily incapacitating disorder that affects an estimated 12% of the general population including 18% of adult women and 6% of adult men. The disorder involves two main subtypes termed migraine with or without aura (MA and MO respectively). Migraine can present with a variety of symptoms that vary between individuals and between episodes experienced by a single individual. This disorder causes significant social and economic burden and alarmingly is often poorly treated. A direct cause of this is a lack of understanding of the underlying pathology of migraine.
Migraine is believed to be a neurogenic disorder that involves temporary disruption of pathways that receive and respond to sensory signals. While numerous environmental triggers may have been identified the exact mechanisms that cause the disruption are still largely unknown. However, familial aggregation of migraine suggests significant genetic contributors.
A large proportion of migraine research has therefore focused on identification of the genes involved in this disorder. There have been successes with the rare migraine subtype Familial Hemiplegic Migraine - that exhibits autosomal dominant inheritance. However, these contrast with the numerous studies in pedigrees affected by common migraine that have show multifactorial inheritance. Furthermore, the variable phenotypic presentation of migraine and the results of many genetic studies to date suggest that the disorder is polygenic. Currently, it is unknown how many genes and variants are potentially interacting and how these are influenced by environmental factors. Consequently many genes have been implicated through genetic studies, yet few causative variants have been conclusively confirmed.
An interesting aspect to the common migraine subtypes is the observed female preponderance. Often this is ascribed to hormonal influences due to the epidemiological data that demonstrates a strong correlation of migraine prevalence with reproductive milestones. An alternative explanation is that genes on the X chromosome may also be playing a role. In particular, variants in genes coded on the
iii X chromosome that exhibit a dominant effect may provide an additional mechanism through which a female preponderance could be caused. Due to the polygenic nature of migraine it is likely that factors coded on the X chromosome and regulation of hormonal factors at a gene level are both contributing to this preponderance.
Therefore the aim of the research described in this thesis was to identify candidate genes on the X chromosome that influence migraine susceptibility. Prior to the work described here, a susceptibility locus at Xq24-Xq28 had been identified by our laboratory in 2 large multigenerational Australian pedigrees. Analysis suggested locus heterogeneity within the region at Xq24, Xq27 and Xq28. This work therefore proceeded to investigate these regions in further detail.
The Xq27 and Xq28 regions were analysed in an additional 6 migraine families that showed no inheritance of migraine from father to son, suggesting an X-linked component. The results of this study identified 2 additional X-linked migraine families designated MF47 and MF879. These families showed evidence of excess allele sharing to different loci at Xq28 and Xq27 respectively, without cross-over of linkage regions. Importantly, this provided evidence to suggest that the loci are distinct and potentially harbour independent candidate genes. Combined, 4 migraine pedigrees have now been identified as showing linkage to either the Xq27 or Xq28 regions. A global analysis of these four families gave maximum LOD scores of 2.47 (P=0.0003) and 1.9 (P=0.001) at Xq27 and Xq28 respectively. Through haplotype analysis of MF879 we were able to refine the implicated Xq27 region to an area between markers DXS8043 and DXS297 that equates to an approx 2Mb region. Similar analysis of MF47 did not refine the Xq28 region from that previously identified. This locus spans from DXS8061 to Xqter translating to an approximate 5Mb region.
This preliminary evidence suggested that the Xq27 and Xq28 susceptibility loci were not specific to rare migraine families. We therefore tested a key marker from each locus in a migraine association population consisting of 500 migraineurs age, sex and ethnicity (Caucasian origin) matched to 500 controls. The results of this analysis showed evidence of association in the tested female populations (P=0.009 and P=0.07) and particularly with the MO subtype at the Xq28 marker (P=0.05). This
iv result provided the impetus to begin candidate gene studies within these loci in the association population.
Eleven candidate genes were selected, 2 at Xq27 and 9 at Xq28 for analysis. These genes were selected using evidence from the literature that suggested functions in neurological, hormonal or vascular systems that were related to migraine. Alternatively some genes were also chosen based on expression studies in brain. In total 32 polymorphisms - selected through a tag SNP approach, were genotyped across the eleven genes; however none showed evidence of association. However, further investigation of the SNPs that surrounded the initial microsatellite genotyped at the Xq28 region did identify a 3 marker risk haplotype for MO in the NSDHL gene. This gene was investigated due to its role in cholesterol synthesis and the high co- morbidity of migraine with cardiovascular disease.
The Xq24 locus identified previously in the 2 Australian migraine families was excluded from this fine-mapping and SNP analysis. This is because a clear candidate gene, GRIA3 is coded in this region. GRIA3 codes for a subunit of the glutamate receptors and numerous studies have previously implicated a role for glutamate in migraine. Furthermore, association has been identified at this gene in an Italian migraine cohort. Therefore analysis of this region was limited to replicating the association previously observed with this gene. The results of this study confirmed association at the previously identified rs3761555 polymorphism that is located within the promoter region of the gene and provides further proof that this GRIA3 promoter variant may play a role in migraine.
Finally a secondary aim of the research described herein was to undertake an analysis of the X chromosome in the genetic isolate of Norfolk Island. Migraine prevalence in the Norfolk Island pedigree is double the observed prevalence of outbred populations and family histories can be traced back 11 generations to the original founders. An X chromosome scan utilising 14,124 genotyped SNPs was conducted. Analysis of this data was undertaken using a two-step approach that firstly analysed the entire chromosome by logistic regression accounting for age, sex and the relatedness of the cohort - using a predetermined kinship coefficient, to maintain computational
v efficiency. A secondary analysis then analysed the top prioritised SNPs using a pedigree based regression that accounts for the relatedness of the cohort exactly.
SNP prioritisation of the Norfolk data revealed clustering of SNPs at a novel Xq12 locus with 10 of the top 25 SNPs, prioritised by P value, localising to this region. The strongest association was observed at rs599958 (P = 8.92x10-4). Haplotype analysis at this region identified two haplotype blocks with risk haplotypes identified within each block (P = 1.1x10-4 and P = 1.6x10-4). The V-set and Immunoglobin domain containing 4 gene (VSIG4), potentially involved in inflammation is coded within the first haplotype block. The 5’ end of the Hephaestin (HEPH) gene is within the second block. The HEPH gene is involved in iron homeostasis in the brain - interestingly elevated iron levels in the brains of migraineurs has been previously reported.
The Norfolk Island X chromosome scan also revealed clustering of 11 of the top 25 ranked SNPs at the previously identified Xq27 susceptibility locus. Haplotype analysis at this region similarly identified 2 risk haplotypes at the Xq27 locus (P = 4.6 x10-4 and P = 1.6x10-3). However few genes are coded within this region and no annotated genes map to these haplotype blocks. Therefore identifying potential causative factors at this locus will require novel, likely bioinformatic approaches.
Norfolk follow-up replication studies were also undertaken in 2 independent cohorts. The first cohort was the Women’s Genome Health Study that showed association to the Xq12 region, however showed no association to the Xq27 loci. In contrast the Australian migraine association population showed association at Xq27 but not Xq12.
The conflicting results of this study exemplify a common occurrence in migraine genetic studies. These results may indicate that either of these loci are false positives, however the replication present for both loci argues for, and supports, involvement of these loci. Alternatively conflicting results may be attributed to differences in the migraine cohorts used and the heterogeneity of the disorder. Overall, this may signify a need for better migraine classification in our cohorts, particularly as they become larger to provide power for new approaches such as genome-wide association studies.
vi In summary, this research used a variety of genetic techniques and analysis methods to fine-map susceptibility regions and identify candidate genes. In particular, four genes were implicated from this research including HEPH and VSIG4 at Xq12, GRIA3 at Xq24 and NSDHL at Xq28. In addition, further evidence for involvement of the Xq27 region in migraine was provided. Finally, this is also the first study that has undertaken an X chromosomal GWAS and the first that has used the Norfolk Island genetic isolate to localise X chromosomal regions implicated in migraine. Further detailed analysis of the candidate genes implicated from this research is required to elucidate the mechanisms through which they may influence the disorder. Overall, the results of this study provide new avenues for migraine investigation and ultimately contribute to the understanding of the genetics of this common debilitating disease.
vii viii Statement of Originality
This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.
______Bridget H. Maher
ix x Table of Contents
Abstract ...... iii Statement of Originality ...... ix Table of Contents ...... xi List of Tables ...... xvii List of Figures ...... xix List of Abbreviations ...... xxi Gene Abbreviations ...... xxiii Acknowledgements ...... xxv Publications Arising from this Thesis ...... xxvii Introduction to this Thesis ...... 1 Significance...... 2 Aims ...... 3 Structure of this Thesis ...... 4 CHAPTER ONE: Introduction to Migraine ...... 5 1.1 Migraine Classification ...... 5 1.2 Migraine Epidemiology ...... 7 1.2.1 General Population...... 7 1.2.2 Migraine Families ...... 10 1.2.3 Norfolk Island Isolated Population ...... 11 1.3 Social and Economic Impact of Migraine ...... 11 1.4 Migraine Treatments ...... 13 1.4.1 Prophylactic Therapies...... 13 1.4.2 Acute Therapies ...... 15 1.5 Migraine Co-morbidity ...... 17 1.5.1 Migraine and Epilepsy ...... 18 1.5.2 Migraine and Psychiatric Disorders ...... 20 1.5.3 Migraine and Cardiovascular Conditions ...... 22 1.5.3.1 Patent Foramen Ovale and Migraine ...... 22 1.5.3.2 Stroke and Migraine ...... 24 1.5.3.3 Migraine and Hypertension ...... 26 1.5.4 Migraine and Other Medical Disorders ...... 28
xi 1.6 Migraine Pathophysiology ...... 28 1.6.1 Migraine Triggers ...... 29 1.6.2 Central Processes in Migraine Pain ...... 30 CHAPTER TWO: Genetic Analysis in Complex Disease ...... 33 2.1 Heritability ...... 33 2.2 Twin Studies ...... 34 2.3 Identifying Genetic contributors in Complex disease ...... 34 2.3.1 Genetic Markers ...... 34 2.3.2 Polymerase Chain Reaction ...... 35 2.3.3 Genotyping methods ...... 37 2.3.3.1 RFLP ...... 37 2.3.3.2 HRM ...... 38 2.3.3.3 Capillary Electrophoresis ...... 39 2.3.3.4 Sequencing ...... 40 2.3.3.5 Arrays ...... 41 2.3.3.6 Mass Spectrometry ...... 42 2.3.4 Analysis Strategies ...... 42 2.3.4.1 Linkage Disequilibrium ...... 43 2.3.4.2 GWAS ...... 44 2.3.4.3 Linkage Studies ...... 46 2.3.4.4 Candidate Gene Studies ...... 47 2.3 Implications of Phenotypic Diversity in Complex disease ...... 49 CHAPTER THREE: Molecular Genetics of Migraine ...... 51 3.1 Molecular Genetics of Severe Migraine Subtypes ...... 51 3.2 Common Migraine Subtypes (MA and MO) ...... 54 3.2.1 Linkage Analysis ...... 54 3.2.1.1 Candidate Genes and 19p13 (MGR5, MIM ID: 607508) ...... 58 3.2.1.2 Candidate Genes and 15q11-q13 (MGR5, MIM ID: 609179) ...... 59 3.2.1.3 Candidate Genes and 10q22-23 (MGR12 MIM ID: 611706) ...... 59 3.2.1.4 Tresk and 10q25 (MG13 MIM ID: #613656) ...... 60 3.2.2 GWAS ...... 61 3.2.3 Candidate Genes ...... 62 3.2.3.1 Neurological Candidate Genes ...... 62 3.2.3.2 Hormone Candidate genes ...... 65
xii 3.2.3.3 Vascular Candidate genes ...... 68 3.3 Migraine and the X Chromosome ...... 70 3.4 Conclusion ...... 73 CHAPTER FOUR: Materials and Methodology ...... 75 4.1 Sample Ascertainment, Structure and Epidemiology ...... 75 4.1.1 Norfolk Island Population ...... 75 4.1.2 Migraine Pedigrees ...... 77 4.1.3 Migraine Association Populations ...... 77 4.2 DNA Extractions ...... 85 4.2.1 Blood Extractions...... 85 4.2.2 Saliva Extractions ...... 86 4.2.3 Ethanol Precipitation ...... 87 4.2.4 DNA Quantification ...... 88 4.3 Selection of Polymorphisms ...... 88 4.3.1 Norfolk Island Study ...... 88 4.3.2 Pedigrees ...... 88 4.3.3 Candidate Gene Studies ...... 89 4.3.3.1 Selection of Candidate Genes ...... 89 4.3.3.2 Selection of SNPs ...... 90 4.4 Genotyping ...... 95 4.4.1 In-house Methods...... 95 4.4.1.1 PCR ...... 95 4.4.1.2 RFLP ...... 96 4.4.1.3 Microsatellite ...... 98 4.4.2 Outsourced Genotyping Methods ...... 102 4.4.2.1 Illumina ...... 102 4.4.2.2 MALDI-TOF MS ...... 102 4.5 Analysis...... 103 4.5.1 Input Files ...... 103 4.5.2 Power ...... 104 4.5.3 Chi Square ...... 105 4.5.3.1 HWE ...... 106 4.5.3.2 Significance level ...... 107 4.5.4 CLUMP ...... 107
xiii 4.5.5 Linkage Analysis ...... 108 4.5.5.1 Error Checking ...... 108 4.5.5.2 GENEHUNTER-PLUS and ASM ...... 110 4.5.6 Pedigree-based Association Incorporating Logistic Regression...... 113 4.5.6.1 SOLAR...... 115 4.5.7 Haplotype Analysis ...... 116 4.6 Methods Conclusions ...... 117 CHAPTER FIVE: Investigation of the Xq27 and Xq28 Susceptibility Regions 119 5.1 Introduction ...... 120 5.2 Materials and Methods ...... 122 5.2.1 Pedigree Study ...... 122 5.2.1.1 Pedigree Collection and Phenotyping ...... 122 5.2.1.2 Pedigree Genotyping ...... 122 5.2.1.3 Statistical Analysis of Pedigrees ...... 123 5.2.2 Association study ...... 124 5.2.2.1 Population Collection and Phenotyping ...... 124 5.2.2.2 Microsatellite Markers...... 124 5.2.2.3 Candidate Gene Studies ...... 124 5.2.2.4 Statistical Analysis of Case-control Cohort ...... 125 5.3 Results ...... 126 5.3.1 Pedigree Analysis...... 126 5.3.2 Investigation of Key Microsatellite Markers in a Case-control Cohort .... 131 5.3.3 Xq27 Candidate Gene Study ...... 131 5.3.4 Xq28 Candidate Gene Study ...... 132 5.4 Discussion ...... 134 5.5 Conclusion ...... 136 CHAPTER SIX: Investigation of the Xq24 candidate gene GRIA3 ...... 139 6.1 Introduction ...... 139 6.2 Methods...... 141 6.2.1 Population ...... 141 6.2.2 Genotyping ...... 142 6.2.2.1 GRIA1 ...... 142 6.2.2.2 GRIA3 ...... 142 6.2.3 Statistical Analysis ...... 143
xiv 6.3 Results ...... 143 6.3.1 GRIA1 ...... 143 6.3.2 GRIA3 ...... 145 6.4 Discussion ...... 146 6.5 Conclusion ...... 148 CHAPTER SEVEN: Investigation of the X chromosome in the Norfolk Island Isolated Population ...... 149 7.1 Introduction ...... 150 7.2 Methods...... 151 7.2.1 Norfolk Island Study ...... 151 7.2.1.1 Study Population ...... 151 7.2.1.2 Genotyping ...... 152 7.2.1.3 Statistical Analysis ...... 152 7.2.2 Replication Populations ...... 153 7.2.2.1 Genotyping ...... 154 7.2.2.2 Statistical Analysis ...... 154 7.3 Results ...... 154 7.3.1 Single Marker Analysis...... 155 7.3.2 Haplotype Analysis ...... 156 7.3.3 Replication Analysis ...... 160 7.4 Discussion ...... 164 7.4.1 Replication Studies ...... 164 7.4.2 The Novel Xq12 Locus ...... 166 7.5 Conclusion ...... 169 CHAPTER EIGHT: Discussion and Future Directions ...... 171 8.1 Research Overview ...... 171 8.1.1 Xq12 ...... 172 8.1.2 Xq24 ...... 175 8.1.3 Xq27 ...... 176 8.1.4 Xq28 ...... 178 8.2 Research Summary ...... 180 8.3 Future Directions ...... 181 8.3.1 Phenotypic Heterogeneity of Migraine ...... 181 8.3.2 Sequencing ...... 182
xv 8.3.3 Bioinformatics...... 185 8.4 Conclusion ...... 186 APPENDICES ...... 189 Appendix A ICHD-II Part One: The Primary Headaches 1. Migraine 1.1-1.2 ..... 190 Appendix B NCBI Build 37.2 Gene Map DXS1123 - Xqtel ...... 194 Appendix C NCBI Build 37.2 Gene Map DXS8043 - DXS297 ...... 197 Appendix D Xq27 and Xq28 Candidate Genes ...... 198 Appendix E PED files ...... 199 Appendix F Individual Analysis of Migraine Families ...... 207 Appendix G Global Analysis of Migraine Families ...... 215 REFERENCES ...... 219
xvi List of Tables
Table 1-1 Part I Migraine prevalence in different ethnicities ...... 8 Part II Migraine prevalence in different ethnicities ...... 9 Table 1-2 Relative risk of migraine to first degree relatives compared to the general population ...... 10 Table 1-3 Prophylactic treatments ...... 14 Table 1-4 Migraine co-morbidities...... 18 Table 3-1 Summary of linkage studies (Part I) ...... 56 Summary of linkage studies (Part II) ...... 57 Table 3-2 Positive migraine association Studies: neurological genes ...... 63 Table 3-3 Positive migraine association Studies: hormone related genes...... 67 Table 3-4 Positive migraine association studies: vascular related genes ...... 67 Table 4-1 Existing genotyping in migraine families ...... 88 Table 4-2 GABRA3 polymorphisms ...... 93 Table 4-3 Primer sequences ...... 97 Table 4-4 PCR protocol ...... 97 Table 4-5 Microsatellite multiplexes ...... 99 Table 4-6 PED file ...... 103 Table 4-7 Power parameters ...... 105 Table 5-1 Microsatellite analysis in case-control cohort ...... 131 Table 5-2 Xq27 Candidate gene study ...... 132 Table 5-3 Xq28 Candidate gene study ...... 133 Table 5-4 Xq28 MO Risk haplotype rs5970389, rs6653488, rs2071256 ...... 133 Table 6-1 GRIA1 rs548294 analysis ...... 144 Table 6-2 GRIA1 rs2195450 analysis ...... 145 Table 6-3 GRIA3 rs3761555 analysis ...... 146 Table 7-1 Single marker analysis results: Top 25 SNPs after prioritization ...... 156 Table 7-2 Xq12 Haplotype analysis ...... 158 Table 7-3 Xq27 Haplotype analysis ...... 160 Table 7-4 WGHS Replication analysis (Part I) ...... 161 WGHS Replication analysis (Part II) ...... 162 Table 7-5 Australian migraine population replication analysis...... 163
xvii xviii List of Figures Figure 1-1 Phases of the migraine and epilepsy episode ...... 19 Figure 2-1 Polymerase chain reaction ...... 36 Figure 2-2 Gel electrophoresis of PCR product ...... 37 Figure 2-3 Restriction digest gel...... 38 Figure 2-4 Capillary electrophoresis electrophoretogram ...... 39 Figure 2-5 Sequencing reaction and analysis ...... 41 Figure 2-6 A simple example of linkage between a trait and a marker ...... 46 Figure 2-7 Association study design...... 48 Figure 3-1 Xq24-q28 analysis ...... 71 Figure 4-1 MF6 ...... 78 Figure 4-2 MF7 ...... 79 Figure 4-3 MF14 ...... 80 Figure 4-4 MF47 ...... 81 Figure 4-5 MF55 ...... 81 Figure 4-6 MF541 ...... 82 Figure 4-7 MF878 ...... 83 Figure 4-8 MF879 ...... 84 Figure 4-9 CNGA2 gene structure and polymorphisms ...... 91 Figure 4-10 ABCD1 gene structure and polymorphisms ...... 92 Figure 4-11 GABRA3 gene structure and polymorphisms ...... 92 Figure 4-12 NSDHL gene structure and polymorphisms ...... 93 Figure 4-13 ATP2B3 gene structure and polymorphisms ...... 94 Figure 4-14 FLNA gene structure and polymorphisms ...... 94 Figure 4-15 CLIC2 gene structure and polymorphisms ...... 95 Figure 4-16 Electrophoretogram raw data male ...... 100 Figure 4-17 Electrophoretogram raw data female ...... 100 Figure 4-18 DXS8043 GENEMAPPER genotyping electrophoretograms ...... 101 Figure 4-19 DXS1108 GENEMAPPER genotyping electrophoretograms ...... 101 Figure 4-20 CLUMP DXS8043 files ...... 109 Figure 4-21 CLUMP DXS8061 files ...... 109 Figure 4-22 GENEHUNTER-PLUS DAT file ...... 112 Figure 5-1 Independent analysis of migraine families ...... 127
xix Figure 5-2 Global analysis of migraine families ...... 128 Figure 5-3 MF879 haplotypes ...... 129 Figure 5-4 MF47 haplotypes ...... 130 Figure 7-1 Manhattan plot of PLINK logistic regression analysis ...... 155 Figure 7-2 LD Plot Xq12in Norfolk Population ...... 157 Figure 7-3 LD Plot Xq27 in Norfolk Population ...... 159 Figure 8-1 Xq12 Summary ...... 174 Figure 8-2 Xq27 Summary ...... 177 Figure 8-3 Xq28 Summary ...... 179
xx List of Abbreviations
5-HT Serotonin AMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic Acid Receptor ASM Allele Sharing Model CADASIL Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy CE Capillary Electrophoresis CGRP Cacitonin-Gene Related Peptide CSD Cortical Spreading Depression CVD Cardiovascular Disease ddNTP Di-Deoxynucleotide-Tri Phosphate dNTP Deoxynucleotide-Tri Phosphate EAAT Excitory Amino Acid Transporter FHM Familial Hemiplegic Migraine GABA Gamma-Aminobutyric Acid GWAS Genome-Wide Association Study HRM High Resolution Melt IBD Identity By Descent ICHD-II International Classification of Headache Disorders 2nd edition IHS International Headache Society LCA Latent Class Analysis MA Migraine with Aura MALDI-TOF Matrix Assisted Laser Desorption-Ionisation Time of Flight MF Migraine Family/Families MIRACLES MIgraine and Hypertension RelAtionship: Comorbidity and risk of CerebrovascuLar EventS MO Migraine without Aura MS Mass Spectrometry NI Norfolk Island NMDA N-Methyl-D-aspartic acid NO Nitric Oxide NRM Nucleus Raphe Magnus NSAIDs Non-Steroidal Anti-Inflammatory drugs
xxi PAG Periaqueductal Grey Matter PCR Polymerase Chain Reaction PFO Patent Foramen Ovale RFLP Restriction Fragment Length Polymorphism SHM Sporadic Hemiplegic Migraine SOLAR Sequential Oligogenic Linkage Analysis Routines TCA Trait Component Analysis TE Tris-EDTA TGVS Trigeminovascular System TNC Trigeminal Nucleus Caudalis TNS Trigeminal Nucleus System TRESK TWIK-Related Spinal cord Potassium Channel WGHS Women’s Genome Health Study WHS Women's Health Study
xxii Gene Abbreviations
5-HT1B (HTR1B) 5-hydroxytryptamine (serotonin) receptor 1B 5-HT1D (HTR1D) 5-hydroxytryptamine (serotonin) receptor 1D 5-HT1E (HTR1E) 5-hydroxytryptamine (serotonin) receptor 1E 5-HT1F (HTR1F) 5-hydroxytryptamine (serotonin) receptor 1F 5-HT2A (HTR2A) 5-hydroxytryptamine (serotonin) receptor 2A 5-HT2C (HTR2C) 5-hydroxytryptamine (serotonin) receptor 2C 5-HT3A (HTR3A) 5-hydroxytryptamine (serotonin) receptor 3A 5-HT3B (HTR3B) 5-hydroxytryptamine (serotonin) receptor 3B 5-HT4 ('HTR4) 5-hydroxytryptamine (serotonin) receptor 4 5-HT7 (HTR7) 5-hydroxytryptamine (serotonin) receptor 7 ACE Angiotensin I converting enzyme (peptidyl-dipeptidase A) 1 ATP1A2 ATPase, Na+/K+ transporting, alpha 2 polypeptide ATP2B3 ATPase, Ca++ transporting, plasma membrane 3 CACNA1A Calcium channel, voltage-dependent, P/Q type, alpha 1A subunit CLIC2 Chloride intracellular channel 2 CNGA2 Cyclic nucleotide gated channel alpha 2 CXorf1 Chromosome X open reading frame 1 DBH Dopamine beta-hydroxylase (dopamine beta-monooxygenase) DDC Dopa Decarboxylase (aromatic L-amino acid decarboxylase) DRD2 Dopamine receptor D2 DRD3 Dopamine receptor D3 EDNRA Endothelin receptor type A ESR1 Estrogen receptor 1 [ FSHR Follicle stimulating hormone receptor GABRA3 Gamma-aminobutyric acid (GABA) A receptor, alpha 3 GABRA5 Gamma-aminobutyric acid (GABA) A receptor, alpha 5 GABRB3 Gamma-aminobutyric acid (GABA) A receptor, beta 3 GABRE Gamma-aminobutyric acid (GABA) A receptor, epsilon GABRQ Gamma-aminobutyric acid (GABA) receptor, theta GDI1 GDP dissociation inhibitor 1 GPR50 G-Protein Coupled Receptor 50
xxiii GRIA1 Glutamate receptor, ionotropic, AMPA 1 GRIA3 Glutamate receptor, ionotropic, AMPA 3 GRID1 Glutamate receptor, ionotropic, delta 1 HEPH Hephaestin INSR Insulin receptor KCNK18 Potassium channel, subfamily K, member 18 KCNMA1 Potassium large conductance calcium-activated channel, subfamily M, alpha member 1 LDLR Low density lipoprotein receptor MAOA Monoamine oxidase A MTDH Metadherin MTHFR Methylenetetrahydrofolate reductase (NAD(P)H) NOS Nitric oxide synthase NRG Neuregulin PGCP Plasma glutamate carboxypeptidase SLC1A3 Solute carrier family 1 (glial high affinity glutamate transporter), member 3 SLC4A4 Solute carrier family 4, sodium bicarbonate cotransporter, member 4 SLC6A3 Solute carrier family 6 (neurotransmitter transporter, dopamine), member 3 SLC6A4 Solute carrier family 6 (neurotransmitter transporter, serotonin), member 4 SLITRK2 SLIT and NTRK-like family, member 2 TPH2 Tryptophan hydroxylase 2 VSIG4 V-set Immunoglobin domain containing 4
xxiv Acknowledgements
It is with pleasure that I thank and acknowledge the people that made this thesis possible and those that provided ongoing support and encouragement. First and foremost my sincere thanks go to my supervisor Professor Lyn Griffiths, who took a chance on me - a rather inexperienced research student. The support and opportunities provided allowed me to develop skills and knowledge in the field of molecular genetics and I am sure that without such a great supervisor the journey would have been far more arduous and a lot less enjoyable.
To Dr Rod Lea and Dr Larisa Haupt, many thanks go to you both for your mentoring over the course of my PhD and assistance in designing the experimental and analysis work required for the research in this thesis. I am indebted to you both for your time and patience as well as for providing ongoing support when it felt like progress was two steps forward and one step back.
Many members of the GRC also deserve my thanks. In particular to Rachel Okolicsanyi - for many, many coffees; helpful discussions and for taking my mind off things when needed. Likewise to Larisa, the advice has been greatly appreciated, made the journey worthwhile and contributed to many memorable moments. Also my sincere thanks goes to Christine for your mentoring in the years prior to commencing my PhD, for putting me in contact with Lyn, and for all your ongoing and still continuing support.
And most importantly my appreciation goes to my family. To my parents - Chris and Tony; and sisters – Rachel and Rebecca; for supporting me and encouraging me to give this a go in the first place, and especially for your patience when I didn’t want to talk about it and willingness to listen when I did. And finally to my husband Simon, I cannot express my gratitude enough. Thank-you so much for being there through this journey. Your endless optimism and ability to convince me that everything will be fine has been absolutely essential and has often provided a source of motivation when needed.
xxv xxvi Publications Arising from this Thesis Journal Articles:
Maher BH, Griffiths LR (2011) Identification of Molecular Genetic Factors that Influence Migraine. Molecular Genetics and Genomics 285(6):433-46
Maher BH, Kerr M, Cox HC, MacMillan JC, Brimage PJ, Esposito T, Gianfrancesco F, Haupt LM, Nyholt DR, Lea RA, Griffiths LR (2011) Confirmation that Xq27 and Xq28 are susceptibility loci for migraine in independent pedigrees and a case-control cohort Neurogenetics 13(1):97-101
Maher BH, Lea R, Follet J, Cox HC, Fernandez F, Esposito T, Gianfrancesco F, Haupt LM, Griffiths LR (2011) Association of a GRIA3 gene polymorphism and migraine susceptibility in an Australian case-control cohort (Under review - Cephalalgia)
Maher BH, Lea RA, Benton M, Cox HC, Bellis C, Carless M, Dyer TD, Curran J, Charlesworth JC, Buring JE, Kurth T, Chasman DI, Ridker PM, Schürks M, Blangero J, Griffiths LR (2011) An X chromosomal scan in a large pedigree from the Norfolk Island genetic isolate provides evidence for a novel migraine susceptibility locus at Xq12 (Manuscript under review - PlosOne)
Conferences:
Maher BH, Cox HC, Haupt L, Bellis C, Blangero J, Curran J, Lea RA, Griffiths LR (2011) Evidence for a Novel Migraine Susceptibility Locus on Xq12 detected in a large pedigree from the genetic isolate of Norfolk Island Human Genetics Society of Australasia, July 31 - August 3, Gold Coast Qld, Australia (Oral presentation – Awarded best student presentation)
xxvii Griffiths LR, Maher BH, Cox HC, Haupt L, Bellis C, Blangero J, Curran J, Lea RA (2011) Evidence for a Novel Migraine Susceptibility Locus on Xq12 detected in a large pedigree from the genetic isolate of Norfolk Island 15th International Headache Congress June 23-26, Berlin, Germany.
Maher BH, Kerr M, Nyholt DR, Cox HC, Haupt LM, Lea RA, Griffiths LR. (2010) Identifying Migraine Susceptibility Genes at the Xq28 Locus Gold Coast Health and Medical Research Conference, December 2-3, Gold Coast, QLD, Australia (Poster presentation)
Maher BH, Kerr M, Nyholt DR, Cox HC, Lea RA, Griffiths LR. (2009) Suggestive Linkage to Xq28 Locus in 3 Independent Migraine Families Australian Society for Medical Research National Scientific Conference, November 15-17, Hobart, Tasmania, Australia (Poster presentation)
Maher BH, Kerr M, Nyholt D, Esposito T, Cox H, Gianfranceso F, Lea RA, Griffiths LR (2009) Fine Mapping and Candidate Gene Studies of the Xq28 Locus in Migraine Genemappers, April 14-17, Katoomba, NSW, Australia (Poster presentation)
Maher BH, Kerr M, Nyholt D, Lea RA, Griffiths LR (2008) Identification of an X- linked gene in Migraine Gold Coast Health and Medical Research Conference, December 4-5, Gold Coast, QLD, Australia (Poster presentation)
xxviii Introduction to this Thesis
Headaches are an ailment that have been recognised throughout history. Their existence in early times is evident in documentation from Mesopotamian and Egyptian societies. These historical documents outline varying methods of curing the sufferer. Headache relief in early civilisations predominantly focused on the supernatural and religious beliefs of the culture and therefore consisted of prayer and rituals to customary deities (Rapoport and Edmeads 2000). Since early civilisation the understanding and treatment of headaches - and particularly of migraine, has significantly evolved.
Hippocrates, around 400BCE, was the first to describe and recognise the symptoms of what is now known as migraine. Migraine is often described as an episodic, neurological disorder that presents with variable clinical phenotypes. The common forms of migraine – Migraine with (MA) or without Aura (MO), are diagnosed by the presence of recurrent headache that lasts 4-72 hours and is generally accompanied by nausea, photophobia, phonophobia, aggravation by physical activity and possible neurological symptoms, as outlined by the International Classification of Headache Disorders 2nd Edition (ICHD-II) (IHS 2004). Numerous theories regarding the causes and underlying mechanisms that result in migraine symptoms have been suggested. However to date, no clear pathophysiology has been determined. Presently migraine is largely accepted as a neurogenic disorder caused by a dysfunction in the activation of the trigeminovascular system and/or an abnormal processing of normal sensory signals (Borsook et al. 2006; Lambert and Zagami 2008; Moskowitz 2008).
While the aetiology of migraine is far from being completely understood, it has been observed that migraine has a strong tendency to run in families and a substantial genetic predisposition to migraine has been demonstrated in numerous studies. In addition epidemiological data shows that migraine has a strong female preponderance, of 3:1 and in some sufferers clearly shows an alignment of migraine frequency with reproductive milestones (Maggioni et al. 1997; Lipton et al. 2001a; Stovner et al. 2006; Lipton et al. 2007; MacGregor 2009). Combined, this evidence is indicative of a hormonal influence on migraine perhaps directed at a genetic level. An alternative
1 theory is that the X chromosome may also harbour susceptibility genes that exert an effect on migraine suscpetibility.
The aim of the work described in this thesis was to contribute to the growing body of knowledge of the genetic origins of migraine. In particular this work focused on the X chromosome with some emphasis on investigating a previously identified susceptibility locus (Nyholt et al. 1998a; Nyholt et al. 2000) to identify the gene or genes located in this region that increase migraine susceptibility.
Significance
Migraine affects approximately 12% of the population and has significant personal, social and economic burdens (Lipton et al. 2007). The high prevalence of this disorder and the temporary impairment of normal function results in significant economic burden to both families and employers. Migraine prevalence varies with age however it is highest at the peak economic years of 30-39 (Jensen and Stovner 2008) and numerous studies have indicated that the economic impact of migraine is significant, largely resulting from indirect costs such as reduced productivity at work (Berg 2004; Goetzel et al. 2004). Migraine has consequently been classified by the World Health Organisation as the 19th leading disease caused disability (WHO 2001).
Despite the clear burden to society the underlying cause is unknown and treatment options are limited and can be ineffectual. This may be partially attributable to the difficulty in accurately diagnosing the common forms of migraine. Currently no lab based diagnostic exists for the common forms of migraine therefore clinicians rely on the ICHD-II that was developed by the International Headache Society (IHS). However a key limitation of this approach is the reliance on the observations and accounts of the sufferer. Migraine also has a number of co-morbidities and symptomatic overlap with other neurological disorders. Combined these factors can lead to misdiagnosis and poor treatment plans. Thus there is strong impetus for a lab based diagnostic for common migraine types.
The genetic influence on migraine has been clearly demonstrated and heritability and familial aggregation studies suggest a genetic component of 34-57% (Mulder et al.
2 2003). Studies have also shown that migraine is a complex polygenic disorder and phenotypic outcome may be conferred by interactions of many genes of small effect as well as interacting environmental factors. The identification of migraine susceptibility genes that exist in this interplay of environmental and genetic factors is therefore of great importance - firstly to contribute to the understanding of migraine pathophysiology and in the long-term the development of a lab based diagnostic. Consequently this may then also aid in the development of effective treatments, reducing the burden of this highly prevalent disorder.
Aims
The overall aim of this research was to identify migraine susceptibility genes in the Xq24-28 susceptibility region that has been previously identified in 2 large multigenerational Australian pedigrees. To achieve this aim a number of approaches were employed. Firstly we sought to determine if this susceptibility locus was specific to migraine in the previously identified pedigrees or could be observed in familial migraine more generally. Six additional migraine families were therefore analysed for excess allele sharing at the Xq27 and Xq28 regions.
Subsequent to this, we sought to determine if the susceptibility locus was contributing to migraine in the general population or was limited to migraine pedigrees. Key markers identifying the locus were therefore genotyped in a migraine association population consisting of 500 migraineurs diagnosed according to ICHD-II criteria and 500 age and sex matched controls.
Finally, upon confirmation that this susceptibility locus was contributing to migraine generally, we directed the research to identify the gene(s) contributing to migraine using a candidate gene approach in the migraine association population. A secondary aim of the project was to expand the X chromosome analysis to determine if other susceptibility loci could also be detected. This was investigated using a chromosome-wide scan in the isolated population of Norfolk Island where migraine is highly prevalent.
3 Thus specific aims of this research were to:
i) Test 6 migraine pedigrees for potential linkage to the Xq27 and Xq28 regions ii) Undertake association analysis of key SNPs in implicated X chromosomal regions iii) Undertake molecular analysis of the Xq24 candidate gene GRIA3 iv) Perform a full X chromosome scan in the Norfolk Island genetic isolate
Structure of this Thesis This thesis will provide a background to migraine as a disorder as well as a background to analysis of a complex disease - such as migraine, presented as the first two chapters of this thesis. The current state of molecular genetics research into migraine will then be presented to provide the landscape in which the research investigated in this thesis is couched. This background has recently been published as a review in Molecular Genetics and Genomics however; some additional information has also been included. In addition the discussion regarding the implications of phenotypic diversity that was included in the review article has been removed and included in a more relevant section in chapter two of this thesis.
A general methodology section will follow outlining the key lab-based and statistical techniques used in the research. The results of this PhD have then been divided into 3 sections. Firstly, an analysis of the Xq27 and Xq28 migraine susceptibility loci will be presented consisting of a linkage analysis in migraine pedigrees and candidate gene studies in a migraine association population. An analysis of the candidate gene GRIA3 which is located in the Xq24 region, will also then be explored. These two chapters represent independent scientific articles that have been included in a format based on the submitted manuscripts. The final results chapter presents the expansion of the X chromosome investigation through a chromosome-wide scan in the Norfolk Island population with follow-up studies in two independent populations. Finally, the overall impact of the combined results will be discussed with implications to migraine research and future directions that lead from this work.
4 CHAPTER ONE: Introduction to Migraine
Migraine is a common debilitating neurological disorder that causes temporary, reversible, incapacitation of the sufferer. The disorder presents with variable clinical phenotypes including nausea, photophobia, phonophobia, aggravation by physical activity and possible neurological symptoms, as outlined by the ICHD-II (IHS 2004). This disorder is highly prevalent in society with significant social and economic burdens. However it is also largely undermanaged due to incomplete understanding of the pathophysiology of disease, poor diagnosis and limited effective treatments. This chapter will present migraine as a disorder in terms of current diagnostic criteria, as well as considering epidemiology, social and economic impacts, available treatments, known co-morbidities and current understanding of the pathology of the disease.
1.1 Migraine Classification
Migraine is an episodic, neurological disorder that presents with variable clinical phenotypes. The migrainous episode is commonly divided into 4 phases: the prodrome, the aura, the headache and finally the postdrome. However, not all phases are experienced in every attack or by all migraine sufferers.
The prodrome consists of premonitory symptoms that some sufferers view as a warning signal for an oncoming episode. While symptoms and duration can be varied the most common symptoms reportedly consist of tiredness, general malaise and fatigue, as well as mood changes and gastrointestinal disturbances (Kelman 2004). Not all migraine sufferers experience the prodrome phase with frequency varying significantly across a number of studies.
The second phase is the Aura that is defined by the IHS as the “complex of neurological symptoms that occur just prior to or at the onset of the migraine headache.” Commonly auras consist of a temporary visual, sensory and/or speech disturbances. The occurrence of the Aura phase is the distinction between the common migraine types MA and MO. The MA subtype is further divided into 6 subcategories: Typical Aura with Migraine Headache; Typical Aura with Non-
5 Migraine Headache; Typical Aura without Headache; Familial Hemiplegic Migraine (FHM); Sporadic Hemiplegic Migraine (SHM); and Basilar-Type Migraine. These subcategories vary depending on the symptoms of the aura and/or the head pain that follows (See Appendix A for full diagnostic criteria).
The third phase consists of the headache. During this stage migraineurs experience a moderate to severe intensity attack that lasts 4-72 hours. The pain is often of a throbbing nature and is unilateral. Physical activity exacerbates the intensity and this is often accompanied by nausea and vomiting as well as possible photophobia and phonophobia.
The migraine episode may then conclude with the postdrome phase which is possibly the least studied of the phases. During this time the sufferer may feel weak or tired while recovering from the migrainous episode.
While the phases of the migraine episode have been characterised there can be significant variability in the phenotypic features of the migraine both within an individual from attack to attack, as well as between individual sufferers causing difficulties in accurate diagnosis. Migraine is currently diagnosed according to the ICHD-II (IHS 2004). The ICHD-II defines the common forms of migraine by the presence of recurrent headache that lasts 4-72 hours and is generally accompanied by nausea, photophobia, phonophobia, aggravation by physical activity and possible neurological symptoms (See Appendix A for full diagnostic criteria).
The ICHD-II is used as the standard reference for migraine classification in the absence of lab based diagnostics - as is the current situation for MA and MO. The criteria was initially released in 1988 and subsequently updated in 2004. While the classification system is widely accepted; it is limited by its basis on the subject’s recall of the migraine attack and the characteristics and symptoms that accompany it.
6 1.2 Migraine Epidemiology
1.2.1 General Population Migraine prevalence is estimated to be approximately 12% (Lipton et al. 2007) but estimates vary widely. In early studies some of this variance was attributable to differing migraine classification methods, however since the introduction of the ICHD most epidemiological studies have adhered to these criteria. Recent studies have therefore highlighted the ethnic variations of migraine prevalence. In particular migraine appears to occur most frequently in Caucasian populations followed by African and Asian populations (Stovner et al. 2007). Migraine also has a well documented preponderance in females over males at a ratio of approximately 3:1 and approximate prevalence of 18% and 6% respectively (Stovner et al. 2006; Lipton et al. 2007). The reported prevalence of migraine in males and females has also varied significantly across studies in different ethnic backgrounds (Table 1-1).
Migraine prevalence has also been observed to vary over age increasing until the highest occurrence in the peak economic years of 30-39, and then declining thereafter (Stewart et al. 1992; Jensen and Stovner 2008). In children and adolescents migraine prevalence increases with age with 3% of 3-7 year olds affected, 11% of 7-11 year olds and up to 23% of adolescents above the age of 11 experiencing migraine (Stewart et al. 1992; Lewis et al. 2004; Eiland et al. 2007). In children prevalence is observed equally in males and females; however the female preponderance can be observed in adolescence (Wang et al. 2000; Shivpuri et al. 2003; Laurell et al. 2004; Zencir et al. 2004; Zwart et al. 2004; Eiland et al. 2007).
The prevalence of migraine has been compared to that of other common diseases with migraine prevalence in America estimated as equalling that of asthma and diabetes combined (Unger 2006). The relatively high occurence of migraine in the general population has allowed for development of large case-control populations in which to screen for genetic contributors to this disease. In this study a number of different case- control populations are used for this purpose.
7 Table 1-1 Part I Migraine prevalence in different ethnicities Country Sample Method Age Prevalence Reference Males Females Africa Ethiopia 15,500 Questionnaire ≥20 1.7 4.2 (Tekle Haimanot et al. 1995) Tanzania 3,351 Interview ≥11 2.5 7.0 (Dent et al. 2004) Asia Hong Kong 7,356 Telephone interview ≥15 0.6 1.5 (Wong et al. 1995) Hong Kong 1,436 Telephone interview ≥15 3.0 6.2 (Cheung 2000) Japan 4,795 Questionnaire ≥15 2.3 9.1 (Takeshima et al. 2004) Japan 4,029 questionnaire ≥15 3.6 12.9 (Sakai and Igarashi 1997) Malaysia 595 Door to door ≥6 6.7 11.3 (Alders et al. 1996) Taiwan 3,377 Questionnaire ≥15 4.5 14.4 (Wang et al. 2000) Europe Austria 997 Interview ≥15 6.1 13.8 (Lampl et al. 2003) Croatia 3,794 Door to door 15-65 12.3 18.0 (Zivadinov et al. 2001) Denmark 740 Interview 21-30 5.9 15.3 (Rasmussen et al. 1991) Denmark 207 Interview 25-36 5.4 23.5 (Lyngberg et al. 2005) France 10,585 Interview ≥15 4.0 11.2 (Henry et al. 2002) Hungary 813 Mailed questionnaire 15-80 4.3 10.7 (Bank and Marton 2000) Netherlands 1,680 Questionnaire ≥12 5.0 12.0 (van Roijen et al. 1995) Netherlands 6,491 Questionnaire 20-65 7.5 25.0 (Launer et al. 1999) Spain 5,668 Telephone interview 18-65 8.0 17.2 (Matias-Guiu et al. 2010) Sweden 728 Interview 40-74 - 18.0 (Mattsson et al. 2000) Sweden 1,668 Telephone interview 18-74 9.5 16.7 (Dahlof and Linde 2001) Turkey 1,320 Interview ≥7 7.9 17.1 (Kececi and Dener 2002) (Adapted from (Manzoni and Torelli 2003; Stovner et al. 2007))
8
Table 1-1 Part II Migraine prevalence in different ethnicities Country Sample Method Age Prevalence Reference Males Females North America Canada 2,922 Telephone interview ≥18 7.4 21.9 (O'Brien et al. 1994) USA 20,334 Mailed questionnaire 12-85 5.7 17.6 (Stewart et al. 1992) USA (Caucasian) Telephone interview 18-65 8.6 20.4 (Stewart et al. 1996) USA (African) Telephone interview 18-65 7.2 16.2 (Stewart et al. 1996) USA (Asian) Telephone interview 18-65 4.8 9.2 (Stewart et al. 1996) USA 29,727 Mailed questionnaire ≥12 6.5 18.2 (Lipton et al. 2001b) USA 4,376 Interview 18-65 6.0 17.2 (Lipton et al. 2002) South America Colombia Questionnaire ≥15 4.8 13.8 (Morillo et al. 2005) Ecuador Questionnaire ≥15 2.9 13.5 (Morillo et al. 2005) Mexico Questionnaire ≥15 3.9 12.1 (Morillo et al. 2005) Peru 3,246 Interview ≥15 2.3 7.8 (Jaillard et al. 1997) Venezuala Questionnaire ≥15 4.7 12.2 (Morillo et al. 2005) (Adapted from (Manzoni and Torelli 2003; Stovner et al. 2007))
9 1.2.2 Migraine Families Migraine displays significant familial aggregation. Family and twin studies have clearly demonstrated that both the rare and common forms of migraine have a significant genetic basis (Gervil et al. 1999; Mulder et al. 2003; Stewart et al. 2006). Heritability is a measurement of the genetic component of complex disease. Twin studies across 6 countries estimate migraine heritability to range from 0.34 – 0.57 depending on the population (Mulder et al. 2003). Twin studies have also demonstrated that monozygotic twins have approximately twice the concordance rate of migraine compared to dizygotic twins (Gervil et al. 1999), indicating that migraine evidently requires both genetic and environmental influences to manifest in a predisposed individual.
Family studies have also been used to assess the genetic component in migraine with estimates that first degree relatives of migraineurs are 1.88 times more likely to suffer migraine then first degree relatives of non-migraineurs (Stewart et al. 2006). Table 1- 2 shows a breakdown of migraine risk to first degree relatives according to MA or MO diagnosis of the proband.
Table 1-2 Relative risk of migraine to first degree relatives compared to the general population Diagnosis RR to first degree relatives proband MA MO MA 3.8 - MO 1.4 1.9 (adapted from (Russell and Olesen 1995))
Despite the strong evidence for migraine heritability, no clear mode of inheritance has been identified for the common migraine subtypes. This has been attributed to the fact that migraine is a result of the contribution of genetic variants from several chromosomal loci combined with the interplay of environmental triggers that are considerably varied amongst sufferers (Wessman et al. 2004).
Migraine families represent another population in which to test for genetic contributors to disease. Migraine families are also used in this study to analyse inheritance of migraine and to localise regions within which to look for susceptibility genes.
10 1.2.3 Norfolk Island Isolated Population Genetic isolates represent a novel population in which to study complex disease due to the potential reduction in genetic heterogeneity and increase in clinical homogeneity. The Norfolk Island population is one example of a genetic isolate that has been used to analyse a number of complex traits including cardiovascular disease (Bellis et al. 2005; Bellis et al. 2008a; Cox et al. 2009) and migraine (Cox et al. 2011). Heritability of migraine in this pedigree is 0.53 (Cox et al. 2011) similar to family studies reported previously. However migraine prevalence is 25.5%, approximately twice as high as is generally observed in an outbred population.
The Norfolk Island pedigree is descendent from 9 British ‘Bounty’ mutineers and 6 Tahitian women and 2 further Caucasian males that joined the colony early in the nineteenth century. Genealogical information has been used to recreate the Norfolk Island pedigree spanning 11 generations and ~6000 individuals back to the original 17 founders. Pedigree analysis has identified a core 377-member pedigree with associated phenotype data that can be used for gene mapping studies (Cox et al. 2011). Furthermore, the unique nature of the geographical and cultural isolation has led to shared lifestyles and customs consequently reducing the effect of non-genetic influences on complex traits.
The Norfolk Island pedigree is also used for migraine studies in this thesis. This population represents a third category of samples – an isolated population, that is analysed alongside pedigrees and case-control cohorts for identification of genetic contributors to disease.
1.3 Social and Economic Impact of Migraine
The prevalence and severity of migraine, particularly in the most economically productive age groups dictates that migraine is a significant health problem with substantial social and economic impact. However due to the episodic nature of migraine this burden is generally overlooked. The economic impact of this disorder has been evaluated in the United States and Europe in terms of both direct and indirect costs. Direct costs often consist of medication, physician services, emergency
11 department visits, diagnostic services and management of treatment side effects. Indirect costs relate to absent work days and reduced productivity while at work due to migraine.
In Western Europe the cost of a migraine patient per year has been estimated at €461, largely consisting of indirect costs caused by reduced productivity at work (Berg 2004). Costs in the USA have been determined as a national burden of illness, estimated at $11.07 billion in direct costs (Hawkins et al. 2008) and $12 billion in indirect costs largely consisting of absenteeism in the workplace (Hawkins et al. 2007).
Studies have suggested that reduce productivity at work is the biggest contributor to total indirect costs (42-89%) and therefore Hawkins et al (2008) may have significantly underestimated the burden of migraine in the USA (Berg 2004; Goetzel et al. 2004). In addition, a number of studies have also shown that most sufferers do not consult physicians and therefore do not receive correct diagnosis or treatment of migraine. Consequently the estimates of indirect costs in most studies are thought to be largely underestimated (Hu et al. 1999; Lipton et al. 2001a; Berg 2004)
The American Migraine Study reported that 91% of migraineurs have functional impairment associated with the headache which clearly has significant consequences to employers and families (Hawkins et al. 2007). The economic studies have attempted to quantify this impact and, although the total cost estimates vary, all studies clearly demonstrate that the burden to society is substantial and the greatest proportion of that burden results from the indirect costs of this disorder.
Another aspect to the economic burden of migraine is the impact on the family of migraineurs. Families where a child suffers migraine have been reported with an average increase of $600 in total healthcare costs, while families with a parent and child sufferer incur $2500 more in total healthcare costs compared to non-migraine families (Stang et al. 2004). In addition it has been found that first-degree relatives of a sufferer have 25% higher total healthcare costs compared to matched non-migraine controls (Stang et al. 2004). It has been suggested that this may be in part due to the stress caused from dealing with the temporary disability that migraine can cause. The
12 burden of migraine on the family unit has been shown to be substantial, not only in the economic terms discussed here but also in social terms. In a study by Lipton it was found that 32% of sufferers and 12% of their partners also avoided making social plans because of the chance of a migraine episode. Furthermore, 43% of sufferers believed they would make better parents if they did not suffer migraines (Lipton et al. 2003a).
The social effects of this disorder are substantial. In 2001 the World Health Organisation classified migraine as the 19th leading disease caused disability (WHO 2001). A number of studies have been carried out assessing the health related quality of life of migraine patients with all indicating an inverse correlation between migraine disability and quality of life (Lipton et al. 2003a; Lipton et al. 2003b; Leonardi et al. 2010).
1.4 Migraine Treatments
Migraine is generally poorly managed, which contributes significantly to the burden of this disease. The American Migraine Study II showed that only 48% of migraineurs reported physician diagnosed migraine and 57% of IHS diagnosed migraineurs rely solely on over the counter products to treat their attacks (Lipton et al. 2001a). A recent survey of migraine management in Australian GP’s reported that only 8.3% of patients took prophylactic medication, while 15% of patients had previously used prophylactic medication but 74% of these patients had discontinued use due to lack of efficacy or side-effects. In contrast 79.3% of patients reported using acute medication (Ducros et al. 2011). Consequently many patients are not receiving adequate migraine treatment. Approaches to migraine treatment can be broken down into prophylactic and acute therapies.
1.4.1 Prophylactic Therapies Prophylactic therapies are generally used in cases where acute therapies are ineffective and migraine caused disability is a common occurrence. The primary goal of prophylactic medication in migraine is to reduce attack frequency by ≥50% (Barbanti et al. 2011). Prophylactic treatment options generally include
13 Table 1-3 Prophylactic treatments Class Drug Mode of Action Side-effects Antiepileptic Sodium Enhancement of GABA-mediated neurotransmission Nausea Volproate Attenuation of low threshold T-type Ca2+ channels Fatigue Blockage of voltage dependant Na+ channels Weight gain Reduction of plasma extravasation Dizziness Topiramate Ca+ and NA+ channel blocker Weight loss Anorexia Taste disturbances Memory problems Nausea Fatigue Gabapentin Enhances GABA turnover and release Somnolence Decrease noradrenaline, dopamine and serotonin release mediated through Dizziness Ca2+ channels β-blockers examples Blockade of β1-mediated effects and consequent inhibition of NA+ release Drowsiness include: and tyrosine hydroxylase activity Propanolol, Timolol, Metoprolol Anti- Tricyclics e.g. Potentiate 5-HT transmission Dry mouth depressants Aimtriptyline block activation of the trigeminovascular system Drowsiness Anti- Pizotifen Blocks 5-HT2 and 5-HT1C receptors Weight gain serotonergic Sedation agents Methysergide Blocks 5-HT2 and 5-HT1 receptors Fibrosis
Calcium Flunarizine Regulates neuronal excitability and attenuates dural vasodilation Weight gain antagonists
14 avoidance of known triggers, β-blockers, tricyclic antidepressants, antiepileptics and/or calcium-channel blockers (Pringsheim et al.; Barbanti et al. 2011). Examples of prophylactic treatment classes are provided in Table 1-3.
How these varied drug types contribute to preventing or reducing the severity of migraine is not completely understood. Many have been recognised as treatment options for migraine due to treatment of co-morbidities such as hypertension (β- blockers) and depression (antidepressants). It is hypothesised that these drugs act through mechanisms involved in central hyperexcitability and/or anti-nociceptive pathways (Mueller 2007).
1.4.2 Acute Therapies Acute therapies are those taken upon onset of migraine attack during the early stages of migraine development. Treatment options can be broken down into non-specific and specific. Non-specific treatments include aspirin, non-steroidal anti-inflammatory drugs (NSAIDS) and analgesics. Specific drugs include ergot derivatives and the triptans class (Goadsby et al. 2002). Currently the triptans are the most commonly used acute medication (Humphrey 2008).
The mode of action for triptans is debatable. The triptans are generally considered to target the 5-HT1b, d and f receptors. The 5-HT1b receptors have been found on the smooth muscle cells of the coronary vessels, although are more predominant in the cerebral arteries. Numerous studies have demonstrated that triptans induce vasoconstriction mediated through this receptor. The 5-HT1d receptor is found predominantly on neuronal tissue and activation of this receptor is known to inhibit trigeminovascular neurons in the Trigeminal Nucleus Caudalis (TNC) through inhibition of neurotransmitter release as well as being thought to reduce neurogenic inflammation through inhibition of neuropeptide release (e.g. CGRP) (Neeb et al. 2010). It was hypothesised that triptans generally work through both inhibiting vasoconstriction of the cerebral vessels through the 5-HT1b receptor as well as inhibiting neurotransmitter release through the 5-HT1d receptor.
15 Despite the efficacy of the available triptan treatment options numerous migraine patients still do not respond to this therapy or are contraindicated due to uncontrolled hypertension, ischemic heart disease and various other cardiovascular conditions (Johnston and Rapoport 2010). Therefore research has focused on eliminating the vasoconstrictive effects of triptans mediated through the 5-HT1b receptor. Selective agents for the 5-HT1d receptor have been developed to reduce the cardiovascular risk profile. Unfortunately, the selective agonist in clinical trials was found to be ineffective against migraine suggesting that the effects of triptans in migraine function through other receptor subtypes (Gomez-Mancilla et al. 2001). Recently work has turned toward the 5-HT1f receptor. This receptor is located in parts of the trigeminal nucleus system (TNS) and, similar to the 1d receptor subtype, activation of 5-HT1f does not lead to vasoconstriction of the cerebral vessels (Razzaque et al. 1999;
Bouchelet et al. 2000). Recent work has focused on the 5-HT-1F agonist lasmiditan which lacks the vascular effects of the known triptans yet appears to have comparable efficacy (Ferrari et al. 2010).
Clinical trials are continuing to assess the efficacy of selective inhibitors of the 5-HT1f receptor with promising results suggesting the ability to alleviate acute migraine attacks (Ferrari et al. 2010). These results indicate that vasoconstriction of the cerebral vessels through the 5-HT1b receptor subtype may not be required for therapeutic treatment of a migraine attack (Neeb et al. 2010). Due to the efficacy of the selective
5-HT1f agonists and the lack of vasoconstrictive effects it has been theorised that the triptans may instead function through inhibition of glutamate release. Glutamate is a neurotransmitter that co-localises with the 5-HT receptors and is believed to be required to activate the TNS and central sensitization. Thus inhibition of glutamate release culminates in an anti-nociceptive effect (Neeb et al. 2010).
There are currently 7 triptans available on the market (sumatriptan, zolmitriptan, rizatriptan, naratriptan, almotriptan, eletriptan and frovatriptan). While these are pharmacologically similar the efficacy and tolerability differs due to varying methods of administration as well as pharmacokinetics (Johnston and Rapoport 2010). However, this allows migraine sufferers a number of options when determining a suitable treatment method.
16 Other emerging therapies include a new class of CGRP antagonists the Gepants e.g. telcagepant, which appear to have a good tolerance and safety profile (Ho et al. 2008) and has been shown in phase III trials to be effective at providing pain relief (Connor et al. 2009). It is hypothesised that several mechanisms contribute to the efficacy of CGRP receptor antagonists including reduction of cerebral vasodilation, neurotransmission and release of inflammatory mediators such as nitric oxide (NO) and Substance P. Other targets for new acute migraine treatments include ion channels, nitric oxide synthase (NOS) and glutamate receptors (Magis and Schoenen 2011; Olesen and Ashina 2011). Identifying molecular mechanisms contributing to migraine and understanding how triggers cause migraine onset may provide novel targets for this purpose. In addition identifying susceptibility variants and examining co-morbidities also provide new insights into the disease that may assist in tailoring effective treatments.
1.5 Migraine Co-morbidity
Co-morbidity generally refers to the occurrence of two conditions in an individual that occur at a greater then coincidental level of association (Lipton and Bigal 2007). Co- morbidities exist for a number of potential reasons; firstly they may simply result from uncertainty in the diagnosis due to shared symptomatic profiles. Alternatively, conditions with similar symptomatic profiles could arise in the same individual because they share similar genetic or environmental susceptibility factors. Finally, one condition may also cause the manifestation of the other.
Understanding co-morbidities of complex diseases is important to dissect the molecular mechanisms underlying one or both disease states. Co-morbidities may also assist to recognise environmental risk factors common to both disease types. Ultimately these studies aid in the development of diagnostic criteria to distinguish disorders that manifest in similar symptoms. Furthermore the characterization of co- morbidities assist to develop treatment regimes that accommodate the complexities of the co-morbidity particularly where a standard treatment exacerbates the effects of one of the disorders (Elston Lafata et al. 2004).
17 The relatively high prevalence of migraine has consequently led to the identification of a number of other disorders that could be classified as migraine co-morbidities. Table 1-4 lists those that have been previously identified, some of which are discussed in further detail below.
Table 1-4 Migraine co-morbidities Class Disorder Neurological disorders: Epilepsy
Psychiatric disorders: Depression Anxiety Panic attacks
Heart/vascular conditions: Patent Foramen Ovale Stroke Hypertension
Other medical disorders: Asthma Allergies Endometriosis Diabetes
1.5.1 Migraine and Epilepsy Migraine and epilepsy are both heterogeneous families of neurological disorders that share various risk factors, clinical features and treatment options (Bigal et al. 2003). Both are chronic disorders defined by a characteristic temporary neurological dysfunction that share similar stages of progression during an episode (Bigal et al. 2003) (Figure 1-1).
In addition, migraine and epilepsy share a number of common symptoms often leading to one disorder being mistakenly diagnosed as the other. Symptoms shared by these disorders include: post-event lethargy, impaired or loss of consciousness, visual disturbances, visual and hormonal triggering factors, vertigo, paraesthesias, hemiparesis and aphasia (De Simone et al. 2007).
The overlap in symptoms has consequently also led to an overlap in treatment options. Anti-epileptic drugs, e.g. topiramate and valproate, which act by modulating neuronal function are often used as prophylactic migraine treatments and have demonstrated efficacy in double-blind trials (Bigal et al. 2003; De Simone et al. 2007; Haan et al. 2008).
18
Migraine Epilepsy Phase Shared Symptoms Symptoms Symptoms Premonitory Some examples include: - - fatigue, nausea, blurred visions
Visual disturbances and Slow onset lasting Rapid onset and Aura may involve up to 60 mins brief somatosensory or motor phenomena
Ictus - Head pain lasting Seizures including up to 72hrs partial or generalised that vary depending on epilepsy type
Various symptoms - Can be prolonged in Postdrome usually associated with the epilepsy patient fatigue
Figure 1-1 Phases of the migraine and epilepsy episode
There is significant debate therefore whether co-morbidity between migraine and epilepsy does exist or whether it is an artefact of poor diagnosis. Evidence for a common link between these disorders is demonstrated by the association of migraine aura with the triggering of seizures. The ICHD-II codes this as migraine triggered seizure where a seizure fulfilling criteria for any type of epileptic attack occurs within one hour after a migraine aura (IHS 2004). This occurrence is sometimes referred to as migalepsy. Post-ictal headache is also thought to occur in up to 50% of epilepsy sufferers (Leniger et al. 2001; De Simone et al. 2007). This close temporal occurrence of migraine and seizures contributes to the significant debate regarding the relationship of migraine and epilepsy and whether one may cause the other adding to the confusion associated with diagnosing either disorder (Davies and Panayiotopoulos 2011).
Epidemiology studies also suggest that a link beyond phenotypic similarities does exist. Lipton et al. (1994) determined that the incidence of migraine in epilepsy sufferers is 2.4 times greater then those without epilepsy. Stevenson (2006) has also reported that the prevalence of migraine in children with epilepsy was 14.7% which is higher than the prevalence found in non-epileptic children. Ludvigsson et al. (2006)
19 conducted a study breaking down MO and MA and found that children with MA had a 3.7 fold increased risk of developing epilepsy whereas there was no increased risk in MO sufferers.
Studies from a pathophysiological viewpoint also provide some evidence to suggest a commonality between these disorders. Piccinelli et al. (2006) and Leniger et al. (2003) both hypothesised that neuronal hyperexcitability was the basis of the relationship. Leniger found that the frequency of migraine with aura was significantly higher in co-morbid patients (41%) than migraine alone (25.8%, P= 0.019) (Leniger et al. 2003). This theory was supported by Piccinelli’s results (Piccinelli et al. 2006). In contrast however, studies by Karaali-Savran (2002) and Nryen (2006) found no association between epilepsy and migraine.
Ottman and Lipton (1996) also conducted a study to determine if there was a shared genetic susceptibility between these disorders suggesting a bidirectional co-morbidity model. They assessed the risk of migraine in relatives of probands with genetic versus idiopathic forms of epilepsy. However, overall the results of this study did not support the theory that migraine and epilepsy share a genetic predisposition.
Despite the results of this study it is known that the three genes that cause the rare migraine subtype Familial Hemiplegic Migraine (FHM) can also cause epilepsy (Gargus and Tournay 2007; Vanmolkot et al. 2007; Haan et al. 2008). However, the genetic basis for the complex migraine subtypes, and the co-morbidity between these subtypes and epilepsy remain unknown. It therefore remains a priority to uncover the similarities and differences of these disorders to gain insights into the pathophysiology for the development of targeted treatments and clear diagnostics.
1.5.2 Migraine and Psychiatric Disorders Migraine has been associated with a number of psychiatric conditions, in particular depression, anxiety and panic disorders (Hamelsky and Lipton 2006). The association between these disorders has been observed in several studies using both population and clinical settings and a variety of classification criteria (Zwart et al. 2003; Bag et al. 2005; Mongini et al. 2006).
20 The prevalence of depression in migraine sufferers is estimated at 28% (McWilliams et al. 2004; Patel et al. 2004). Several studies have observed a strong association between these conditions with reports that migraineurs are 2.2-4.0 times (Moschiano et al. 2011) more likely to suffer from depression then controls. Similarly migraine and anxiety (Lanteri-Minet et al. 2005; Tan et al. 2007) and panic disorders (Stewart et al. 1994) have also shown strong association with migraine. A study using data from the 2002 Canadian Community Health Survey of 36,984 subjects concluded that major depression and panic disorder were both twice as common in the migraine group compared to the controls (Jette et al. 2008). A recent meta-analysis by Antonaci and colleagues (Antonaci et al. 2011) considering 12 studies also confirmed this finding.
Studies over the past two decades have largely indicated positive associations between psychiatric disorders and migraine. Some evidence (Stam et al. 2010), particularly the work by Breslau (Breslau et al. 2000; Breslau et al. 2003), suggests bidirectional aetiology rather then the manifestation of one causing the other. Radat and Swendsen (2005) propose that each disorder is a risk factor for the other suggesting a common genetic or environmental susceptibility factor. However, despite the significant amount of literature concerning the co-morbidity of psychiatric disorders and migraine little is understood regarding the mechanisms that lead to this phenomenon. It has been suggested that a shared biology may lead to the manifestation of both disorders; key candidates being considered include dopamine and serotonin pathways as well as ovarian hormones (Antonaci et al. 2011). Further support for a shared biology lies in the treatment of these disorders where there is evidence suggesting that some anti-depressants are effective in the management of migraine (Smitherman et al. 2011).
Understanding the mechanisms that underly migraine psychiatric co-morbidity remains a priority for future research. Studies have shown that both conditions are associated with poor health-related outcomes and a decreased quality of life that is intensified in an individual suffering from both (Jette et al. 2008). Therefore investigation into understanding these co-morbidities may aid in the development of treatment options and significantly assist, from a public health perspective, to provide appropriate services and reduce the burden of mental health disorders in society.
21 1.5.3 Migraine and Cardiovascular Conditions The relationship between migraine and the cardiovascular system has also been the topic of extensive studies. Significant research has considered causative roles for the vasculature in migraine as well as the recognition of co-morbidities including Patent Foramen Ovale (PFO), stroke and hypertension.
1.5.3.1 Patent Foramen Ovale and Migraine The Foramen Ovale is a hole in the atrial wall of the foetal heart that shunts blood from the right to left atrium during foetal development bypassing the lungs (Bandolier 2005). Generally the hole closes after birth so that blood flows to the functioning lungs for oxygenation. However, in approximately 25% of the general population the hole remains open and is known as a PFO (Hagen et al. 1984). The PFO is similar to a flap valve that opens under certain pressure, such as a cough or sneeze. When this occurs blood may travel directly into the left atrium and circulate on through the arterial system bypassing the lungs.
Several studies have demonstrated an association between PFO and migraine. Schwedt estimated in a meta-analysis that PFO prevalence ranged from 40.9-72% in MA and 16.2-33.7% in MO (Schwedt et al. 2008). The particular association between MA and PFO has been reported in a number of studies (Del Sette et al. 1998; Anzola et al. 1999; Dalla Volta et al. 2005; Schwerzmann et al. 2005; Carod-Artal et al. 2006; Domitrz et al. 2007; Tatlidede et al. 2007). One case-control study was conducted assessing PFO size in MA. It was determined that 38% of the MA patients with PFO had a moderate to large shunt (as measured by the number of bubbles crossing the septum in 4 cardiac cycles) while only 8% of controls had a PFO of similar size. The authors concluded that the existence of more than a moderate sized shunt increased the risk of co-morbid MA by 7.78-fold (p ≤ 0.001).
Despite the studies indicating an association between these disorders it is still debated whether the association is causal or merely coincidental. This is largely due to the high prevalence of both conditions in the general population (PFO ~25% and migraine 12%). However, the association studies provide enough evidence to warrant
22 further investigation into a causal link, particularly in MA populations. A number of theories have been put forward as to how these conditions may be related:
1. The PFO allows bypass of the pulmonary filtering allowing passage of vasoactive substances into the arterial system at higher concentrations then normally found and this contributes to the precipitation of a migraine attack (Schwedt et al. 2008); 2. A common genetic factor might result in both an atrial septal defect and migraine, as PFO and migraine have been found to be inherited and transmitted simultaneously (Carod-Artal et al. 2006); 3. Embolisms of thrombi (consisting of fibrin or platelets) cross through the PFO into the arterial system and provide a trigger for cortical spreading depression upon reaching the occipital cortex resulting in aura (Tepper et al. 2007).
The search for a causal link has been motivated by a number of observational studies regarding the effect of PFO closure on migraine. Retrospective studies by Schwerzmann and Reismann suggest that >50% of migraine patients have complete resolution of migraine symptoms post PFO closure (Schwerzmann et al. 2004; Reisman et al. 2005). A randomized, double-blind, sham-controlled trial to investigate the effect of PFO closure on migraine as a treatment option, called the Migraine Intervention with Starflex technology (MIST), has also been conducted (Dowson et al. 2008). However, this study failed to reach the set primary endpoint of complete migraine resolution or the secondary end-point of change in incidence, severity, characteristics of the migraine when compared to implant sham groups.
The methodologies of these studies have been the subject of significant criticism. Consequently the results have been questioned halting any progression of this surgical intervention as a potential migraine solution. The key concerns of these studies include:
1. Most are retrospective relying on subjective recall of attack frequency instead of objective means such as a headache diary; this allows the study to be potentially affected by recall bias of the subjects.
23 2. Migraine studies have shown the disorder to have a high placebo effect, which can occur in 20-40% of subjects in therapeutic studies. 3. In most studies no control groups were used therefore the placebo effect to migraine improvement post closure cannot be measured. 4. Subjects received anti-platelet medication after PFO closure as a necessary measure post surgery; however this may have contributed to the improvement in migraine symptoms. 5. The completeness of PFO closure was found not to affect migraine relief therefore bringing into question whether a causal link exists at all and highlighting the risk of migraine improvement being a placebo effect. 6. Studies have not always distinguished between migraine subtypes therefore it is not clear if PFO is associated with a particular migraine type. Given the invasive nature of this solution, if proven, it is critical to identify patients who are most likely to benefit due to the risky potential side-effects of surgery. (Tsimikas 2005; Carod-Artal et al. 2006; Schwedt and Dodick 2006; Carroll 2008).
The results of these studies have yet to prove that closure of PFO has any real effect on migraine and that a causal relationship does in fact exist. However, a number of studies are underway which aim to address these issues so that the risk to reward ratio of this surgical solution to migraine can be properly assessed.
1.5.3.2 Stroke and Migraine Stroke is believed to occur in approximately 2 per 1000 people per year affecting a mean age of 70 with a male preponderance of 2:1 (Bousser and Welch 2005). Stroke and migraine share a number of risk factors despite the difference in epidemiological data for these disorders. PFO, hypertension and familial occurrence (MacClellan et al. 2007) as well as shared clinical features such as platelet aggregation, focal neurological and ophthalmologic signs as well as a vasospastic component are some factors recognised in both disorders (Stang et al. 2005).
The ICHD-II recognises a particular relationship between migraine and stroke, known as the migrainous infarction. This is defined as “one or more migrainous aura
24 symptoms associated with an ischaemic brain lesion in appropriate territory demonstrated by neuroimaging.’ For diagnosis the ICHD-II stipulates:
a. The present attack in a patient with 1.2 Migraine with Aura is typical of previous aura attacks except that one or more aura symptoms persists for >60 minutes. b. Neuroimaging demonstrates ischaemic infarction in a relevant area. c. Not attributed to another disorder (IHS 2004).
However migrainous infarct represents only a small portion of ischaemic stroke patients and furthermore can only occur in MA sufferers according to the ICHD-II guidelines (Lampl and Marecek 2006). Therefore studies have focused on the occurrence of any stroke type in migraineurs. In population based studies Becker found that cases with a history of migraine were 2.7 times more likely to suffer stroke then controls (Becker et al. 2007). Similarly MacClellan (2007) found that migraine was a significant risk factor for stroke in women without a history of hypertension, diabetes or myocardial infarction (Odds ratios of 1.7, 1.5 and 1.6 respectively). Furthermore women with probable migraine with visual aura who smoked and used oral contraceptives were 7 and 10 times more likely to suffer stroke compared to migraine with visual aura sufferers and non-migraineurs who did not smoke or take oral contraceptives, respectively (MacClellan et al. 2007). Stang et al. (2005) also reported a significant association between ischaemic stroke and MA which did not change after controlling for covariates such as smoking, family history or hypertension. A meta-analysis by Etminan et al. (2005) also found the risk for ischaemic stroke in migraine sufferers was 2.16 and increased to 2.27 for MA independently
Recent reports have also considered large prospective cohort studies including data from the Women’s Health Study (WHS) (Kurth et al. 2005) and the Arthrosclerosis Risk in Communities Study (Stang et al. 2005). Both studies identified an increased risk for ischemic stroke in migraineurs. In particular the WHS results demonstrated a 1.7-fold risk in women with MA. The Stroke Prevention in Young Women Study (MacClellan et al. 2007) has also analysed this association identifying a 1.5 fold increased risk in women with probable migraine with visual aura.
25 The association between migraine and stroke is well documented particularly in young women and the MA subtype. However, the reason for this relationship is yet to be definitively identified. A number of review papers have considered this issue presenting a variety of hypotheses. Firstly it has been suggested that a shared genetic component exists whereby shared mutations lead to either occurrences or one mediates the event of the other through complex interplay of genetic and environmental factors. Alternatively migraine pathophysiology accentuates existing risk factors for stroke and both jointly increase the risk of stroke outside a migraine episode. Finally it has also been proposed that shared risk factors lead to both migraine and stroke. However studies controlling for these factors (such as smoking and contraceptive use) have shown strong associations indicating that the link may be independent of these identified factors (Pierangeli et al. 2004; Bousser and Welch 2005; Diener and Kurth 2005; Katsarava et al. 2008; Narbone et al. 2008).
Further research is required to define the link between migraine and stroke. Both are complex disorders attributing attacks to both genetic and environmental factors. The identification of common factors is crucial to aiding in the diagnosis and prevention of both attacks in order to improve and/or prevent the further deterioration of quality of life due to severe stroke or migraine events, and, if so identified, to halt the progression of one event to the other.
1.5.3.3 Migraine and Hypertension Several studies have investigated the association between migraine and hypertension; however the resulting data has been conflicting. Both disorders are known to have elevated prevalence in the general population (Fanciullacci 2004), suggesting that any relationship between migraine and hypertension is merely coincidental. However, studies have shown significant associations between the disorders (Cirillo et al. 1999; Grebe et al. 2001), indicating that there may be shared vascular contributors.
A study by Williams et al. (2004) used twin studies to examine whether hypertension, migraine, Reynauds phenomenon and coronary artery disease were linked by common environmental factors or an underlying genetic propensity. They determined through variance components analysis that these conditions were linked through a single common genetic factor (Williams et al. 2004). Over time a number of different
26 genetic factors have been put forward as a potential link between hypertension and migraine. These include mutated or misregulated calcitonin gene-regulated peptide (CGRP), serotonin (5-HT) (Gudmundsson et al. 2006), Angiotensin-converting enzyme (ACE) (Fanciullacci 2004) and methylenetetrahydrofolate reductase (MTHFR), all of which require further investigation to conclusively identify any as a potential link.
In contrast to these studies, a number of recent investigations have failed to show statistically significant associations. Hagen et al. conducted a study of 22,685 adults examining the risk of developing headache through an 11 year prospective follow-up study. The study found no significant association between headache and blood pressure, however headache status at the initial consultation was based only on analgesic use, rather then direct questioning (Hagen et al. 2002). Wiehe et al. (2002) and colleagues also conducted an investigation using 1,171 cases from Brazil and showed that normotensive individuals were more likely to suffer migraine then hypertensive patients. Similarly two studies of patients in healthcare settings failed to provide strong evidence to support a causal relationship, although both did find prevalence of hypertension in MA and MO to be higher then in the general population, and one determined that hypertension was a contributing factor to increased severity and frequency of migraine episodes (Pietrini et al. 2005; Prudenzano et al. 2005).
A final study worth noting is the MIRACLES study (Migraine and hypertension relationship: Comorbidity and risk of cerebrovascular EventS). This study was a cross-sectional survey carried out in Italy involving 2973 patients with a GP diagnosis of either hypertension or migraine. This study identified 17% of enrolled patients suffered both migraine and hypertension and significantly this group showed a 1.7 higher probability of stroke compared to the non-migraine hypertensive patients (Mancia et al. 2011). This study not only recognises the co-morbidity of hypertension and migraine, but highlights the risk that migraine may contribute to cardiovascular events such as stroke. Therefore treatment of these disorders should consider the co- morbidity and address any increased risk that may not be ordinarily addressed in a patient presenting with one of the disorders alone.
27 1.5.4 Migraine and Other Medical Disorders Migraine in general is a condition that has a high occurrence of co-morbidities. In addition to the disorders mentioned, studies have also indicated that asthma, allergies and endometriosis are also common co-morbidities (Davey et al. 2002; Ku et al. 2006; Aamodt et al. 2007; Tietjen et al. 2007; Mehle 2008). In the majority of cases it is still unknown whether these are causal or merely coincidental due to the high prevalence of migraine in the general population. Nonetheless, the study of co-morbidities is an important area of migraine research as known pathophysiologies of co-morbid conditions may transpire to contribute to migraine as well. This knowledge may provide a platform to leverage further research into how the common cause may interplay with migraine particularly in aiding to identify candidate genes for analysis.
1.6 Migraine Pathophysiology
The understanding of migraine pathophysiology has evolved significantly over time. Initially thought to be of vascular origin, theories have moved through neurological to neurovascular hypothesis and most recently to a neurogenic theory involving a disturbance in the sensory modulatory systems involving a variety of brain structures.
Experiments into blood flow have long disproven the theory that migraine is purely of vascular origin. This theory was hypothesised due to the vasoconstrictive effects of treatments such as triptans and 5-HT receptor blockers as well as vasodilation being initially thought to be the cause for the pulsating quality of migraine. However studies into blood flow determined that vascular changes did not necessarily occur during an attack and where changes were observed these were not necessarily correlated with any particular phase of the migraine episode (Olesen et al. 1990). These studies therefore relegated vascular change, when it occurs in a migraine episode, to an epiphenomenon status. Furthermore, new treatments have shown efficacy with no vasoconstrictive effects (Goadsby 2009).
In genetic studies, understanding the pathophysiology of disease is important to identify candidate genes for further investigation and analysis. Evidence shows that migraine symptoms result from a combination of vascular, neurological and sensory
28 events. Some of these events include Cortical Spreading Depression (CSD), possible neurological inflammation, sensitization, protein extravasation, changes in the meningeal vasculature and ultimate activation of the trigeminal system and downstream nociceptive pathways. However, the exact order for these pathological events and their interactions is still widely debated (Goadsby 2009; Levy 2010). Recently, research has focused on central neuronal hyperexcitability, the trigeminovascular system and abnormal processing of normal sensory signals (Borsook et al. 2006; Lambert and Zagami 2008; Moskowitz 2008).
1.6.1 Migraine Triggers Migraine sufferers have reported numerous and varied triggers for their migraine attacks. These include external factors such as light, strong smells, food, wine, environmental changes and stress (Martin and MacLeod 2009). Due to the complex nature of migraine and the inherent interplay of genetic and environmental factors, it has been proposed that migraine onset may involve a threshold (Hargreaves and Shepheard 1999) which is lowered due to certain undefined genetic factors and breached in the presence of certain environmental triggers such as those listed previously. Identifying the specific factors that therefore predispose an individual to the onset of a migraine episode is consequently an important factor in current migraine management strategies. In addition to external environmental triggers some of the internal factors that may also contribute include hormones or CSD.
The exact mechanism through which hormones may play a role in triggering a migraine episode is still unclear. Migraine prevalence in women increases considerably post puberty, peaking at 35-45 years of age and declining post menopause (Stewart et al. 1992; Jensen and Stovner 2008), indicating an involvement of hormones or hormonal variation in migraine manifestation. Furthermore, 7-14% of migraineurs experience the ICHD-II classified, Pure Menstrual Migraine, where MO occurs only on day 1±2 days of menstruation in every 2 of 3 cycles (Russell 2010). While it is not known specifically how hormones play a role, evidence has shown that hormones, particularly estrogen, interact with a number of mediators that have been indicated in migraine e.g. CGRP and 5-HT and their receptors (Gupta et al. 2007).
29 CSD is another mechanism historically considered to play a role in triggering the migraine episode. CSD is an intense, self-propagating wave of depolarisation of neurons across the cortex. It is associated with a disruption of ionic gradients resulting in release of intracellular potassium, an influx of sodium and release of neurotransmitters (Moskowitz 2007). CSD is widely accepted to be the cause of visual auras. Studies of the rare MA subtype FHM suggest that predisposition to hyperexcitability is caused by mutations in ion translocation genes resulting in the abnormal clearance or release of neurotransmitters (van den Maagdenberg et al. 2004; de Vries et al. 2009).
There is significant debate whether CSD is simply one of many potential triggers, or is integral to the migraine phenotype. It is a well established phenomenon in MA sufferers. However, there is some argument that a ‘silent’ aura may be occurring in many MO patients and the manifestation of the aura is dependent on the point of origin of the CSD (Moskowitz 2007). Furthermore, there is evidence to suggest that CSD can activate the trigeminovascular system in genetically susceptible individuals. This is thought to be through release of ions and neurotransmitters in sufficient quantities to activate perivascular nerve afferents (Bolay et al. 2002). In contrast however, it is also known that CSD may cause visual aura which are not always followed by migraine supporting the argument that CSD is not a key mechanism in migraine occurrence. The ICHD-II defines this as Typical Aura without headache.
Despite the plethora of triggers for some patients no common specific factor precipitating migraine onset is apparent. Furthermore, research has yet to decipher the exact mechanisms that occur from interactions with known triggers to actual migraine onset.
1.6.2 Central Processes in Migraine Pain Allodynia is common in migraine sufferers suggesting that some form of sensitisation is taking place. Primary sensitization is thought to occur when the meningeal afferents respond excessively to stimuli causing the throbbing phenotype and the intensification of pain with movement that is commonly associated with migraine. Upon continued stimulation the neurons in the TNC (second order neurons) may also begin to
30 discharge spontaneously and excessively. This event is known as secondary sensitisation and is commonly associated with allodynia - a pain response to non- painful stimuli (Borsook et al. 2006; Moskowitz 2007; Moskowitz 2008).
The trigeminovascular system is believed to be integral to the onset and transmission of pain migraine attacks. The system includes nerve connections between the meningeal tissues and the brain stem. Trigeminal fibres that innervate the cerebral vessels arise from neurons in the trigeminal ganglion. One theory of migraine suggests that activation of the meningeal afferents results in the release of neurotransmitter peptides and Calcitonin Gene related peptide (CGRP). These peptides trigger vasodilation and possibly protein extravasation of the meningeal vasculature as well as initiation of signal impulses to the TNC. Second order neurons within the TNC transmit signals to the brain stem to process arriving signals and possibly register pain (Moskowitz 2008).
Lambert and Zagami (2008) have recently also presented an alternative hypothesis that the brainstem not the trigeminovascular system is the originator of pain in migraine. They present evidence suggesting that excitation of cortical neurons can cause withdrawal of sensory inhibition in the brainstem. They argue that continuous discharge from neurons in the Periaqueductal Gray Matter (PAG) and the Nucleus Raphe Magnus (NRM) of the brainstem inhibit sensory input, particularly from cranial structures, by maintaining the discharge of the second-order trigeminovascular neurons at a low rate. Upon cortical activation by various triggers the neuronal discharge of the PAG and NRM are inhibited therefore facilitating excessive discharge in the trigeminovascular system causing sensitisation (Lambert and Zagami 2008).
The understanding of migraine pathophysiology has progressed from early hypothesis of vascular disorders to current theories of neurovascular origin and/or the abnormal processing of normal sensory signals. However, it still remains to be established where the origin of migraine begins, be that peripherally or centrally and which pain modulatory circuits are involved in triggering, maintaining and intensifying pain in migraine. Further studies using emerging imaging and screening techniques as well as animal models and candidate gene studies that target vascular or central nervous
31 systems will help to provide clarity to the mechanisms that cause and sustain a migraine episode.
32 CHAPTER TWO: Genetic Analysis in Complex Disease
Complex disorders pose a number of obstacles to unravelling the genetic basis of disease. By definition complex disorders are attributable to both genetic and environmental factors. Furthermore complex disease may be polygenic with many genes involved in causing the phenotypic expression and any number of these genes may interact with one or more environmental factors (Mayeux 2005). Other difficulties that arise when considering complex disease include incomplete penetrance, phenocopies and the potential for both allelic (different mutation in the same gene) or locus (different genes) heterogeneity of the disease (Strauch et al. 2003).
In the case of migraine, heritability studies have shown that both genetic and environmental factors play a role in manifesting the migraine phenotype. Furthermore, genetic studies have implicated an array of chromosomal loci that would suggest that common migraine subtypes are polygenic. There are a variety of measures and approaches used to decipher the genetic factors of complex disease. Some of the approaches that have relevance to the techniques used in this study are explained in detail in this chapter.
2.1 Heritability
Heritability is a measure of the proportion of phenotypic variation that is attributable to variation in genetic factors (Visscher et al. 2008). If a heritability factor of 1 is determined for a specific trait, this implies that 100% of the trait is due to the genetic makeup of the individual. Measuring the heritability of complex diseases such as migraine is important in understanding the extent to which the manifestation of disease can be affected by environmental factors. In the case of migraine, heritability has been estimated to be between 34-57% (Mulder et al. 2003) with variations observed in different populations.
33 2.2 Twin Studies
Studies of both mono- and dizygotic twins are helpful to assess the environmental impact on complex disease. Monozygotic twins share the same genetic makeup therefore determining concordance between identical twins can assist in measuring the environmental influence. Twin studies in migraine have determined that monozygotic twins have approximately twice the concordance rate of migraine compared to dizygotic twins (Gervil et al. 1999) as would be expected on the assumption that dizygotic twins only share half their genetic makeup.
2.3 Identifying Genetic contributors in Complex disease
2.3.1 Genetic Markers A genetic marker is a variation at a specific location within the genome that can be used to identify a specific gene or trait. Commonly these markers are unique sites not repeated elsewhere in the genome and can be subject to significant variation between individuals. These variations define the differences that occur in the genetic makeup of individuals in a population and contribute significantly to the heritability of disease. Genetic markers are key to studying the genetic basis of complex disease.
Some examples of genetic markers include Single Nucleotide Polymorphisms (SNPs) and Microsatellites. SNPs are the most common contributor to variation within the human genome. They are variations of single nucleotides within the genetic sequence that occur at a low mutation rate. The heterozygosity rate for SNPs is estimated at ~1 in every 1000bp (Altshuler et al. 2008). These variations can occur anywhere within the genome and be causal or silent. Causal SNPs generally occur within coding or regulatory regions and cause changes to amino acid sequence thus affecting protein function or affect protein availability and therefore possibly contributing to disease. Alternatively SNPs may exist within introns or intergenic regions having no effect but nevertheless acting as useful markers for genes within the region (Gray et al. 2000).
Microsatellites are repeats of small nucleotide sequences. These are highly variable within the population and most people will be heterozygous at any given locus
34 (Burton et al. 2005). Due to their polymorphic nature microsatellites are considered more informative than SNPs and are often used in linkage studies. Other variations that exist include deletions of individual or multiple nucleotides and insertions or duplications of extra nucleotides in a sequence.
Determining the variations of these genetic markers in a population is critical to studying complex disease. In particular identifying that a specific marker is co- inherited with a phenotype or trait, can suggest a causal link. Alternatively in unrelated individuals the occurrence of a particular variation that is more common in an affected population may also be indicative of a causal link.
2.3.2 Polymerase Chain Reaction Polymerase Chain Reaction (PCR) is a technique used in molecular biology to specifically amplify a region of interest within the genome for further studies. The basic principle of PCR is to amplify a section of the genome using two primers that are complementary to opposite strands of the DNA sequence of interest. A series of heat changes allows for i) the denaturing of the double stranded DNA, ii) annealing of the primers to their complementary strands and iii) extension of the complementary sequence in the 5’-3’ direction using dNTPs and a Taq polymerase. The cycle is repeated ~30-40 times with each cycle theoretically doubling the amount of the sequence of interest until the availability of dNTPs are exhausted (Lynch and Brown 1990; Baumforth et al. 1999)
The PCR primers are critical to the success of the PCR reaction (Abd-Elsalam 2003). PCR primers are designed to flank the region of interest. They are short specific sequences generally of 17-25 nucleotides in length. These sequences preferably only bind to one specific sequence in the genome of interest to ensure clean amplification of only one fragment of the genome. PCR primers should have 50-60% Guanine/Cytosine content, should avoid long runs of single nucleotide repeats and should not form secondary structures. The primer pairs should also have similar melting temperatures around 55-65°c, and complementary sequences should be avoided to prevent ‘primer dimer’ where the primers bind to each other instead of the genomic sequence (Wu et al. 1991; Roux 1995; Baumforth et al. 1999).
35
Figure 2-1 Polymerase chain reaction
Once the region of interest has been amplified the DNA segment can be visualised on agarose gel with ethidium bromide to check for successful amplification. This method involves creating a charge in the gel which forces the DNA to migrate due to the slight negative charge on the DNA backbone, smaller fragments will move faster through the gel matrix separating into size bands. Ethidium bromide is an intercalating agent that binds to the DNA and can be added to the gel. The DNA can then be visualised under UV light (Voytas 2001). In addition to running the PCR product on the gel to check for successful amplification a DNA size ladder can also be run. This is a mixture of DNA amplicons of specified sizes (e.g. 100bp, 200bp, 300bp, 400bp 500bp 1000bp etc). The size of the PCR product can then be checked against this banding pattern to ensure the amplicon is the expected size and therefore the primers are amplifying the correct sequence (see Figure 2-2) (Baumforth et al. 1999).
Once a region of interest has been successfully amplified the sequence of that amplicon can be genotyped for variations that may be associated with disease.
36 2.3.3 Genotyping methods There are a variety of methods to determine the genetic variant that an individual may have at a specific polymorphic site. Some of the common methods used from low to high throughput analysis include Restriction Fragment Length Polymorphism (RFLP), High Resolution Melt Analysis (HRM), Capillary electrophoresis (CE), sequencing, arrays and mass spectrometry (MS).
Figure 2-2 Gel electrophoresis of PCR product
2.3.3.1 RFLP RFLP is a genotyping method that employs the unique site recognition and digestion characteristics of restriction endonucleases. These enzymes recognise a specific DNA sequence and cut double stranded or single stranded DNA into fragments at the site. In genetic analysis these enzymes are used when a polymorphism is introduced into the sequence that either creates or abolishes a sequence recognition site (Chuang et al. 2008).
Using this genotyping method a region of interest is amplified by PCR and then incubated with the relevant restriction enzyme that recognises a short sequence that includes the polymorphic site. Following incubation the digested PCR product is run on an agarose gel to consider the band patterns. In the simplest example (Figure 2-3)
37 an enzyme will only cut once in the amplicon therefore creating 2 fragments smaller then the initial PCR amplicon. Therefore a homozygous individual carrying the recognition site will show 2 bands on a gel. A heterozygous individual will also show the same 2 bands as well as a 3rd band which is the uncut PCR product. Similarly an individual homozygous for the alternate allele will only show the one band at the same size as the initial PCR product as the cut site is not present and therefore the restriction enzyme does not function. From this analysis the genotype of the individual can be deduced (Todd et al. 2001).
Figure 2-3 Restriction digest gel
2.3.3.2 HRM HRM is another PCR based genotyping technique. This method involves the inclusion of an additional fluorescent reagent, such as SYTO®9 into the PCR. This reagent is an intercalating dye that binds to double stranded DNA. HRM is run on specialised equipment such as the Corbett Rotor-gene that perform the thermal changes required for the PCR as well as measuring the fluorescence emitted by the SYTO®9 as the amount of dsDNA in the PCR reaction increases (Vossen et al. 2009). The cycle then concludes with a melt phase of increasing temperature in which the PCR product denatures and therefore the level of fluorescent accordingly decreases. The key to genotyping by this method is in the variation in the melt curve observed as a
38 polymorphism is introduced. Due to the different nucleotide make-up of the amplified sequence the temperature at which the amplicon dissociates shifts as the chemical bond between the strands are strengthened or weakened compared to the wildtype (Gundry et al. 2003).
2.3.3.3 Capillary Electrophoresis Microsatellites are particularly useful in genetic studies due to their high variability. One polymorphic site may have numerous alleles based on the number of repeats present. However, this advantage is also problematic when using standard visualisation methods such as agarose or polyacrylamide gels to genotype the PCR product as the resolution is not sufficient to determine an allele difference of 1 or 2 nucleotides. This problem has been overcome with the use of capillary electrophoresis (CE) and fluorescent tags (Ziegle et al. 1992). CE offers numerous advantages over standard slab gel methods. Particularly it offers significantly improved resolution and sensitivity as well as utilising automated systems increasing throughput ability (Kan et al. 2004).
Figure 2-4 Capillary electrophoresis electrophoretogram
39 Using this method either the forward or reverse primer is tagged with a fluorescent probe and the region of interest amplified by standard PCR. The PCR product then migrates through a polymer filled capillary passing a laser that excites the fluorescent tag on the primer sequence. A sensor detects the intensity, wavelength and time at which the amplicon passes the sensor which are then displayed as peaks on an electrophoretogram. The size of the fragment is gauged to 1bp resolution by reference to a known size standard which has been tagged with a different fluorescent probe. Due to this high resolution a number of polymorphic sites can be genotyped in a single multiplexed sample if i) the approximate length range of each amplified fragement is known and ii) if each polymorphic site produce PCR amplicons of sufficiently different lengths ≥10bp. In addition automated systems can detect a number of different wavelengths in the one sample allowing for multiplexing of PCR products of similar fragment length using different fluorescent tags (Mansfield et al. 1998; Moretti et al. 2001).
The automated system used in this work is the ABI 3130 Genetic Analyser and related Genemapper software. This software produces an electrophoretogram of the sample from which sizes are determined and a genotype called. An example of this output is provided in Figure 2-4.
2.3.3.4 Sequencing Sequencing is another commonly used approach to identify new genetic variants contributing to disease. While historically sequencing was done on a gel, development of CE has allowed for vast improvements particularly in minimising time and amount of DNA required for analysis (Karger and Guttman 2009).
Sequencing also employs a standard PCR approach; however an additional PCR-like reaction is performed post amplification. This reaction takes the PCR product and adds a mixture of dNTPs and fluorescently labelled terminator dideoxyribonucleotides (ddNTPs) and uses a separate reaction for the forward and reverse primers. The cycle process is generally the same as a standard PCR. However when a ddNTP is included in the amplicon during the elongation step the DNA synthesis is terminated, creating a series of amplicons of varying lengths that are fluorescently labelled. Each ddNTP is labelled with a known probe of a different
40 wavelength so that this can be measured and represent a specific nucleotide. This sample is analysed using capillary electrophoresis so that the fragments are separated in size order and the wavelength of each amplicon is measured over time allowing the sequence of the amplicon to be read (See Figure 2-5) (Kan et al. 2004).
Figure 2-5 Sequencing reaction and analysis
2.3.3.5 Arrays Since the sequencing of the human genome there has been an explosion in the number of SNPs identified throughout the genome. Arrays provide a method for high- throughput genotyping of genomewide SNPs for genetic analysis (Grant and Hakonarson 2008). Numerous different array formats exist including chip and bead arrays, however most work off the same underlying principles. The arrays are designed around a solid plate format divided into known sections or beads each containing numerous copies of a single oligonucleotide. This oligonucleotide is specific to a sequence in the genome that contains a SNP. For each SNP recognised on the array generally a number of different oligonucleotides are included to reduce the chance of false positives or negatives. Furthermore oligonucleotide probes will be included for both the wildtype and the variation polymorphism. Therefore if DNA
41 binds to both probes the sample is heterozygous; alternatively binding at only one of the probes would indicate a homozygote sample (Bilitewski 2009).
The method for genotyping using an array varies depending on the array used. However, the process generally involves amplification of the sample DNA which is then incubated with the array to allow for binding to the probe sequences. The array is generally then washed to remove unbound DNA. A fluorescent marker is then used to indicate DNA that has bound to the array. In some methodologies this marker may be added as a wash step that attaches to a biotin attachment on the sample DNA. Alternatively some protocols incorporate a marker during the amplification stage (Dufva 2009). The array is then read and genotypes called based on the location of the bound DNA.
2.3.3.6 Mass Spectrometry A final medium throughput genotyping method is the use of MS. MS measures the mass to charge ratio of charged particles to determine molecular weight. The molecular weight of a sequence which contains a SNP will vary depending on the allele present. Using this technique short oligonucleotide strands are ionised (through various methods) and their molecular weights determined through mass spectrometry. The molecular weight of the ion is then compared to expected molecular weights based on the known sequence, therefore allowing for identification and genotyping of a sample (Ross et al. 1998; Kwok 2001).
Matrix assisted laser desorption-ionisation Time of Flight (MALDI-TOF) MS is commonly used in this genotyping approach. This method measures the time it takes for the ions to travel in a vacuum from the ion source to the detector in order to determine the molecular weight of the sample. This method is particularly favourable due to its speed and potential for automation (Jackson et al. 2000).
2.3.4 Analysis Strategies There are many different approaches to identifying which genetic variants are causative in complex disease. These approaches can typically be characterised as genome-wide or candidate gene. Genome-wide approaches attempt to identify
42 specific contributing loci in which candidate gene studies may then be carried out. Two main methods have been employed; linkage studies that analyse affected pedigrees and genome-wide association studies (GWAS) that use large case-control populations. Alternatively studies may take a candidate gene approach that focuses on a specific gene or family of genes that have functional relevance to the disorder.
2.3.4.1 Linkage Disequilibrium Genetic mapping studies attempt to identify a marker(s) in close proximity to a disease gene that will segregate with the disease phenotype if they are sufficiently close so as not to have undergone recombination (Rodriguez-Murillo and Greenberg 2008). This draws on the theory of Linkage disequilibrium (LD) which is the non- random pattern of alleles at different loci being found together more often then is expected. Therefore SNPs in close physical proximity are likely to be in LD so that the presence of a particular allele at one SNP is likely to indicate the presence of a particular allele at an adjacent or nearby SNP.
Regions of LD are significantly more pronounced in pedigree cohorts than in population based studies. This is due to recombination and the break down of common haplotypes over generations that consequently reduce the extent of shared regions amongst unrelated individuals (Williams et al. 2010). Therefore in pedigree based studies the number of SNPs required to predict the genetic variation within a given region will generally be less then the number required to predict the genetic variation across the same region in population based studies.
In genetic mapping studies this is especially important to determining the degree of coverage of genetic variation in the region of interest that can be obtained when genotyping a proposed set of SNPs. Measuring the LD between SNPs in a proposed set, can therefore determine the coverage provided by a SNP set, as well as identify any redundancy in the proposed subset where one SNP may be used to infer another (Goode et al. 2007; Painter et al. 2011).
The level of LD between adjacent or physically close markers can be calculated through determining Lewontins D’ coefficient where a D’ value of 1 indicates complete LD. This calculation can be carried out by a number of programs including
43 Haploview (Barrett 2009). Another measure of LD is the r2 statistic to calculate how reliably one can predict the allele at one polymorphism based on the known allele at another polymorphism. For the studies described here two SNPs with an r2 >0.8 in the HapMap CEU population indicated that one SNP could be suggested as ‘tagging’ the second SNP and therefore only one was chosen to for genotyping. Haploview can also be used to determine the r2 value of two SNPs (Barrett 2009).
2.3.4.2 GWAS The development of high throughput genotyping techniques, in combination with the identification of millions of genetic variants through projects such as HapMap has led to the development of the GWAS. This approach has revolutionised the understanding of many complex diseases including Crohn’s disease , prostate cancer, heart disease, rheumatoid arthritis and type 2 diabetes (Manolio et al. 2008).
GWAS use large case-control cohorts in which up to a million SNPs across the genome are genotyped and allele frequencies of the tested SNPs are compared between the groups and tested for association with disease (Pearson and Manolio 2008). The SNPs used in GWAS generally have a minor allele frequency (MAF) of >5% and are therefore ‘common’ in the tested population. Due to this it is expected that the SNPs used will ‘tag’ a block of genetic variations which they are in LD with, therefore significantly increasing the coverage of the genome. In this way hundreds of specific loci for common traits have been identified (Manolio et al. 2008).
Despite the successes already realised using this approach, there are several significant drawbacks that should also be recognised (Pearson and Manolio 2008; Gandhi and Wood 2011; Manolio 2011). Firstly, due to the number of SNPs analysed, the number of association tests are also significantly high. Therefore determining significant association requires very large population sizes and furthermore, requires replication in independent populations of similar characteristics.
Secondly it is unlikely that the common variants actually genotyped in these studies are causal. Therefore the GWAS approach generally identifies loci rather then causal variants per se. These loci therefore require further studies to identify causal variants and the affected genes. It has recently been noted also that a significant number of
44 identified loci do not contain genes that are obviously implicated through known pathologies or often contain annotated genes that have yet to have functions identified (Hirschhorn 2009). Furthermore it is clear that where a SNP is directly implicated many are not traditional ‘causal’ variants i.e. coding variants, suggesting a role in transcriptional regulation and gene expression.
Finally, GWAS are predominantly applicable to disorders where a common-disease common-variant hypothesis applies. This idea suggests that common disorders with a genetic component are attributable to a number of common variants in the population with low penetrance (Gandhi and Wood 2011). The nature of the design of the GWAS study means that these SNPs are likely to be ‘tagged’ by the SNPs tested and therefore may be identified. While many studies have identified common variants that confer low – modest increases in risk, these generally do not account for the entire genetic influence on disease (Hunter et al. 2008). This suggests that in some complex disease, rare variants contribute the greater risk or that increased power is required in these studies to identify new loci with small increases in risk associated with disease (Goldstein 2009; Kraft and Hunter 2009).
Addressing these issues is a challenge for GWAS studies. Firstly rare variants are poorly captured by the GWAS approach and secondly there is significant debate as to the value of identifying more variants that confer small effects. Some predict that identifying numerous loci with small effect sizes on a particular trait or disease will preclude the interpretation of meaningful results. An alternative view is that the value of employing a GWAS approach to identify loci of small effect sizes may not be in contributing to genetic risk profiles but may implicate new pathways in the molecular biology of a particular disease (Hirschhorn 2009).
Nonetheless surveying the genome through GWAS has presented significant insight into numerous complex diseases as well as directed research to new molecular pathways with unexpected involvement in disease. Furthermore, GWAS assists in directing targeted efforts to identify the functions of both annotated genes that have yet to be studied and variants that may be involved in the complexity of transcriptional regulation and gene expression.
45 2.3.4.3 Linkage Studies Linkage analysis is the traditional approach to identifying disease genes. This method uses families where a particular trait or disorder is inherited to map the causal location of the disease gene. Correlations are made between the inheritance patterns of phenotypes and multiple genotypes within an affected pedigree (Figure 2-6). The LOD score is used to measure the likelihood that the trait and marker (or multiple markers) are linked. Traditional methods require the mode of inheritance to be identified as well as penetrance and disease-gene frequencies. The LOD statistic is then calculated using programs such as Merlin (Abecasis et al. 2002) and GENEHUNTER-PLUS.
In complex disease it is not always easy to determine the mode of inheritance in a pedigree with confidence. Therefore non-parametric, i.e. ‘model-free’, linkage analysis methods have been developed (Shih and Whittemore 2001). These methods measure allele sharing among affected relatives and have been implemented for use in programs such as GENEHUNTER-PLUS.
Figure 2-6 A simple example of linkage between a trait and a marker
46 In linkage studies statistical significance is considered at a LOD score ≥ 3 (P=0.0001), while suggestive significance is generally at LOD ≥ 2 (P=0.001) (Nyholt 2000; Dawn Teare and Barrett 2005). However, on the X chromosome statistical significance is considered at a LOD ≥ 2 (P=0.001), while suggestive significance is generally LOD ≥ 1 (P=0.01).
While linkage analysis has been incredibly successful in identifying causal variants in monogenic diseases, there have been fewer gains made with complex disease. The inherent nature of complex diseases means that they are generally polygenic and have significant genetic heterogeneity (Kere 2010). The effects of this are twofold. Firstly, the polygenic nature generally indicates that a number of loci interact to cause the phenotype of interest. Therefore the mode of inheritance can often be unclear in the pedigree and results may not identify contributing factors when analysed in isolation. Secondly, the heterogeneity makes replication and confirmation of susceptibility loci increasingly difficult with numerous studies of the same disease type being conflicting in their findings. Furthermore individual families may carry rare variants contributing to disease making it unlikely that these results will translate to other pedigrees or to the disorder in the general population (Panoutsopoulou and Zeggini 2009).
Another limiting factor in linkage studies is the level of LD found within the pedigree. While LD in association studies may be useful to reducing the number of variants to test therefore keeping costs down it can also be limiting in pedigree studies. The regions identified as segregating with the disease can be substantially broad (Williams et al. 2010) and may code hundreds of potential candidate genes for further studies. Nonetheless linkage studies can be a powerful method for identifying susceptibility loci in affected families particularly when the heritability of the disorder is high.
2.3.4.4 Candidate Gene Studies Another approach to identifying genetic factors in complex disease is to directly test genes for association. Genes may be selected based on genomic location within a susceptibility region that has been implicated through the methods described previously. Alternatively genes may be considered as they are functionally relevant
47 due to known pathophysiology or because effective treatments target the pathway in which a gene interacts (Gardner 2006).
The candidate gene approach generally tests for variations in allele frequencies in case-control populations using a χ2 analysis to identify a significant difference. An increased frequency of a particular allele in a case population may indicate that the SNP is causal or is in LD with the variant that is contributing to the disease (Figure 2- 7) (Painter et al.). However an associated allele may not have a frequency of 100% in the case group due to genetic heterogeneity of the disorder or beacuse the SNP is not in LD with the causal variant in the entire case population (as would be expected in an outbred population). Nonetheless a significant increase in the presence of a particular allele in an affected population may indicate association and further investigation of the variant itself or nearby variants is warranted.
Figure 2-7 Association study design
The variations analysed ideally cause a change in the protein coded by the gene of interest (i.e. exonic SNPs). Alternatively the SNP may also affect the regulation and expression of the gene (i.e. 5’UTR, 3’UTR, and regulatory regions). However, depending on the genotyping method employed and the size of the gene, it may not
48 always be feasible to genotype potential causal variants or there may be too many. Therefore common SNPs within these regions as well as intronic and intergenic regions may also be used that potentially tag causal SNPs. This approach ultimately aims to narrow the number of potentially causal SNPs that may be required to be tested by more expensive means such as sequencing.
Recently it has also become clear that microRNAs and other non-protein coding sequences may be relevant to the progression of disease. Therefore exclusively focusing on coding or regulatory regions may be considered narrow. A tagging approach in candidate gene studies may be prudent to eliminate the possibility that the gene itself may not be contributing to disease but another factor coded within it. This is especially pertinent to candidate gene studies based on susceptibility loci where the connection of a gene to the disease biology is not necessarily clear.
2.3 Implications of Phenotypic Diversity in Complex disease
Genetic studies of complex disease can often be hindered by poor diagnosis and phenotyping. These factors lead to heterogeneity in case cohorts and consequently impede the replication of findings. For example, migraine can present with a variety of symptoms that may differ significantly between sufferers; as well as between migraine episodes in a single individual. Coupled with a lack of lab based diagnostics, this phenotypic heterogeneity presents difficulties in obtaining a clear diagnosis for genetic studies.
The implications of poor diagnosis can be significant in genetic analysis studies. Interpreting and comparing results of studies can be difficult when different diagnostic criteria or stringencies have been applied, particularly when a case group can be stratified by subtype (e.g. MA or MO in migraine). Furthermore, uncertainty in diagnostic methods can also create havoc when replicating results in independent populations which may be maintained by external groups. Therefore the diagnostic strategies used for genetic studies should be well documented. In particular the method of collection of phenotypic data (i.e. survey, interview, questionnaire etc.) on which diagnosis is based should be carefully considered.
49 The quality of phenotypic data collected is also critical to the analysis of complex disorders that involve various subtypes or can be characterised by independent traits. For example, in a single individual migraine characteristics can vary over a lifetime. Thus when searching for a gene that may contribute to a particular phenotype e.g. photophobia, the clinical history and phenotypic data available must be of a high quality to ensure consistency of the presence or absence of the particular subtype or trait across the individual’s migraine history.
Overall the analysis of complex disorders such as migraine is far from easy. Numerous analysis methods may be applied to validate findings and consistency is required in collection of phenotypic data and diagnosis. This will aid in defining more homogenous populations in which certain characteristics of complex disease may be studied with greater reliability. Furthermore defining phenotypic characteristics in cohorts will enable selection of appropriate populations for replication and validation studies.
50 CHAPTER THREE: Molecular Genetics of Migraine
Migraine is an episodic, neurological disorder that presents with variable clinical phenotypes. The exact causes and mechanisms that underlie common migraine (MA and MO) have not been easily identified. Migraine presents with numerous associated symptoms and phenotypes that can be heterogeneous at both a population level as well as between attacks occurring in a single individual.
Family and twin studies have clearly demonstrated that both the rare and common forms of migraine have a significant genetic basis (Gervil et al. 1999; Mulder et al. 2003; Stewart et al. 2006); however approaches to understanding this genetic basis have had varying degrees of success. Analysis of the severe monogenic migraine subtype FHM in affected families has provided the best avenue for identification of migraine genes. In contrast, the most prevalent forms MA and MO are largely accepted to be polygenic and consensus on the key genetic contributors is elusive. This is further complicated by a multifactorial mode of inheritance and environmental interactions which create a phenotypic spectrum associated with expression of the disorder.
Approaches to genetic studies of the common migraine subtypes, MA and MO include genome-wide methods such as linkage analysis and more recently GWAS. Numerous key susceptibility loci have been identified through these methods leading to subsequent candidate gene analysis. An alternative approach has been to identify and analyse candidate genes directly. These genes are selected on the basis of information provided by clinical and other genetic studies of the pathophysiology of migraine. Consequently neurotransmitters, hormones and vascular genes are of particular interest.
3.1 Molecular Genetics of Severe Migraine Subtypes
Hemiplegic migraine is a very severe, rare monogenic subtype of MA that when found in families (FHM) displays autosomal dominant transmission. This form of migraine can also occur as Sporadic Hemiplegic Migraine (SHM) where sufferers
51 have no first or second degree relative that share the aura with motor weakness that is characteristic of FHM and SHM. Linkage studies of FHM families have identified numerous genetic variants in independent genes that cause the disorder.
FHM1 is caused by mutations in the CACNA1A gene at 19p13 (Ophoff et al. 1996) and approximately 21 different causal missense mutations have been identified in this gene (de Vries et al. 2009). CACNA1A codes for the α1A subunit of the Cav2.1 channels. This subunit is involved in voltage sensitivity and as a result mutations lead to uptake of Ca2+ ions into the neuron in response to smaller depolarisations than wildtype channels. This in turn causes excess release of the neurotransmitter glutamate (Wessman et al. 2007).
The FHM2 gene is ATP1A2 (De Fusco et al. 2003), located at 1q21-31 with over 30 different causal mutations identified to date (de Vries et al. 2009). The gene encodes the Na+/K+-ATPase α2 subunit. The final known FHM gene (FHM3) is SCN1A encoding the voltage gated sodium channel gene on chromosome 2q24 (Dichgans et al. 2005) with 5 known FHM mutations (Vanmolkot et al. 2007; Castro et al. 2009; Vahedi et al. 2009).
The identification of FHM genes has provided insight into the pathophysiology of this severe form of MA. It has been suggested that mutations in all three genes may lead to increased efflux of glutamate and potassium in the synapse consequently resulting in increased susceptibility to CSD. CSD is a self-propagating depolarisations of neurons associated with disturbance of ionic gradients and neurotransmitter release (Moskowitz 2007) that may ultimately trigger the migraine aura. Further evidence implicating glutamate dysfunction in CSD and the aura includes the identification of a de novo mutation in the SLC1A3 gene that codes the glutamate transporter; Excitatory Amino Acid Transporter (EAAT1) 1. The mutation was identified in a single patient with episodes of ataxia, migraine, hemiplegia and seizures. The authors concluded from the study that the mutation led to reduced transporter function and consequently decreased glutamate uptake potentially contributing to neuronal hyperexcitability and resulting in the neurological disturbances described in the patient (Jen et al. 2005).
52 Functional studies of FHM mutations in cellular and animal models also support the view that increased levels of glutamate in the synaptic cleft can lead to CSD causing the aura. FHM1 mutant mice have been used to study the functional consequences of a number of FHM causing mutations in CACNA1A. In particular the R192Q FHM-1 mouse has been observed to have a gain of function effect leading to a lowered threshold for CSD (van den Maagdenberg et al. 2004). In vitro assays have also been employed to study numerous other FHM mutations, demonstrating altered channel activity for both FHM2 and FHM3 (Kahlig et al. 2008; Tavraz et al. 2008).
However, the role of the FHM genes in the headache phase of MA is still debatable. There is evidence to suggest that CSD may trigger the trigeminovascular system (TGVS) and downstream pain pathways leading to the migraine headache (Ayata 2010; Eikermann-Haerter and Ayata 2010), however the exact mechanisms that lead to the TGVS activation are still to be established (Messlinger 2009).
Genetic studies that support the theory that hemiplegic and common migraines share at least some genetic basis include the potential identification of a 4th FHM locus by Cuenca-Leon and colleagues. The locus at 14q32 was identified in a Spanish family with FHM, MA and MO (Cuenca-Leon et al. 2009) suggesting a possible shared locus for hemiplegic migraine and common subtypes. Another study also identified a number of different mutations in the SLC4A4 gene in migraine families. This study determined that homozygotes for any of 5 different mutations in this gene suffered either hemiplegic migraine, MA or MO depending on the particular mutation (Suzuki + - et al. 2010). This gene encodes an Na -HCO3 cotransporter NBCe1 that can affect neuronal excitability through regulation of pH in the brain.
However, investigation of the known FHM mutations has provided little evidence to suggest that these are causative in common migraine subtypes (Terwindt et al. 2001; Jen et al. 2004; Todt et al. 2005; Nyholt et al. 2008) or in many cases of SHM indicating that other genes are involved in this complex disorder. Therefore it is possible that the known FHM genes may only influence the aura or hemiplegia symptoms in the FHM sufferers, particularly as the CACNA1A gene (FHM1) is known to cause other neurological disorders such as episodic ataxia that are not
53 associated with migraine. Nonetheless the pathways in which the FHM genes act remain top candidates for common migraine studies.
3.2 Common Migraine Subtypes (MA and MO)
Studies of complex diseases such as common migraine pose many difficulties. Heritability studies of migraine have firstly shown that migraine is influenced by environmental factors which can alter the phenotypic expression of the disorder consequently affecting diagnosis which can be critical to genetic studies. Furthermore significant evidence points to the fact that common migraine is a polygenic multifactorial disorder and gene-gene interactions may thus also play a critical role.
There is significant debate also as to whether MA and MO are distinct disorders or a spectrum of migraine. The headache phase of these subtypes share the same clinical features, however MA is associated with reversible neurological symptoms that occur just prior to the onset of the headache phase (IHS 2004). Recent evidence through the use of alternative phenotyping methods such as Latent Class Analysis (LCA) suggests that MA and MO are not distinct entities. LCA recognises 3 major headache classes; Mild, Moderate or Severe and one asymptomatic class where groups are based on the combination and severity of symptoms (Nyholt et al. 2004). Therefore it is expected that a few key susceptibility genes will underlie both disorders and these may be alternatively influenced by other genetic and/or environmental factors. Genetic studies have supported this theory with many studies providing conflicting results as to whether specific loci or variants are associated with MA, MO or both (see Table 3-1).
A number of different approaches to identifying candidate genes have therefore been employed to overcome these problems. These may be divided into two categories, genome-wide approaches to identify susceptibility loci (including linkage and GWAS) and direct candidate gene analysis.
3.2.1 Linkage Analysis Analysis of migraine inheritance in affected pedigrees is a frequently used approach that has identified numerous migraine susceptibility loci. Table 3-1 outlines the
54 known loci; however as may be observed many are yet to be replicated. There are a number of reasons for this, firstly rare family specific markers may have a significant impact within subsets of families and consequently over represent a linkage signal (Anttila et al. 2008). These may not then be replicated in other pedigrees or case- control populations. Additionally, as migraine is highly prevalent in the population (12%) (Lipton et al. 2007), interference may occur from migraine sufferers married into the family where their genes influence the outcome of the linkage study. A final factor that may inhibit the linkage studies is difficulty in accurate diagnosis and the heterogeneity of migraine manifestation.
Studies by Nyholt, Anttila and colleagues have attempted to overcome this through the use of LCA (Nyholt et al. 2004) and Trait Component Analysis (TCA) (Anttila et al. 2006) to break down the MA/MO classification into more homogeneous groups for genetic analysis. As can be seen from Table 3-1 only a handful of migraine studies have made use of either TCA or LCA however results are promising. Both methods have been shown to identify linkage regions previously not seen through the ICHD-II classification. Using the LCA method Nyholt and colleagues identified the 5q21 region that is predominantly associated with pulsating headache (Nyholt et al. 2005). Similarly the LCA method identified linkage on 18p11 (Lea et al. 2005).
A study by Anttila and colleagues used all three methods and showed consistent linkage to 10q22-23 for 5 TCA phenotypes, the MA ICHD-II classification and the LCA class migrainous headache in a Finnish population (Anttila et al. 2008). This was also replicated in Australian populations. Furthermore, the 10q22-23 region is one that has been identified previously (Nyholt et al. 2005). This strongly suggests that these methods can assist in providing replication to confirm independent studies and be used to identify new regions of interest in their own right (Anttila et al. 2008).
Linkage studies have yielded a number of other susceptibility loci, and despite a number of these lacking replication many have analysed candidate genes identified within these loci. Some examples include the 19p13, 15q11-13 and the 10q25 loci.
55 Table 3-1 Summary of linkage studies (Part I)
Locus Migraine subtype Families Population Genotyping Method Reference 1q31 MA & MO 85 Australian Loci specific microsatellite markers (Lea et al. 2002) 2p12 TCA-pulsation, MA & LCA-migraine 58 Finnish Genome-wide scan (Anttila et al. 2008) 3qter LCA severe 21 Australian Genome-wide scan (Lea et al. 2005) 4q21 MO 103 Icelandic Genome-wide scan (Bjornsson et al. 2003) 4q24 MA 50 Finnish Genome-wide scan (Wessman et al. 2002) TCA – age at onset, photophobia, 50 Finnish Genome-wide scan (Anttila et al. 2006) phonophobia, pain intensity, unilaterality, pulsation 5q21 LCA Twins Australian Genome-wide scan (Nyholt et al. 2005) 6p12.2-p21.1 MO & MA 1 Swedish Genome-wide scan (Carlsson et al. 2002) Activity prohibiting headache and Twins Australian Genome-wide scan (Nyholt et al. 2005) photophobia 9q21-22 Visual migraine aura 36 Finnish Genome-wide scan (Tikka-Kleemola et al. 2010) 10q22-23 LCA 756 Australian Twins Genome-wide scan (Nyholt et al. 2005) MA, TCA – Unilaterality, pulsation, 210 Finnish and Genome-wide scan (Anttila et al. 2008) pain/intensity, nausea/vomiting, Australian photophobia & phonophobia. LCA – migrainous headache TCA-Phonophobia 50 Finnish Genome-wide scan (Anttila et al. 2006) LCA migraine, phonophobia, Twins Australian Genome-wide scan (Nyholt et al. 2005) photophobia
56 Table 3-1 Summary of linkage studies (Part II)
Locus Migraine subtype Families Population Genotyping Method Reference 11q24 MA 43 Canadian Genome-wide scan (Cader et al. 2003) 14q21.2-q22.3 MO 1 Italian Genome-wide scan (Soragna et al. 2003) TCA –pain intensity 125 Australian (Anttila et al. 2008) 15q11-q13 MA 10 - Loci specific microsatellite markers (Russo et al. 2005) 17p13.1 TCA-pulsation 50 Finnish Genome-wide scan (Anttila et al. 2006) 18p11 LCA severe 92 Australian Genome-wide scan (Lea et al. 2005) 18q12 TCA – attack length 58 Finnish Genome-wide scan (Anttila et al. 2008) TCA – aggravation by physical 50 Finnish Genome-wide scan (Anttila et al. 2006) exercise, attack length MO 103 Icelandic Genome-wide scan (Bjornsson et al. 2003)
19p13 MA 1 Australian Loci specific microsatellite markers (Nyholt et al. 1998b) MA 16 North American Loci specific microsatellite markers (Jones et al. 2001) Xp22 TCA pulsation, MA & LCA severe 58 Finnish Genome-wide scan (Anttila et al. 2008) Migraine – mixed 61 European descent Loci specific microsatellite markers (Wieser et al. 2010) Xq24-28 MA and MO 2 Australian Loci specific microsatellite markers (Nyholt et al. 2000) (Nyholt et al. 1998a)
57 3.2.1.1 Candidate Genes and 19p13 (MGR5, MIM ID: 607508) This locus was identified in two independent linkage studies as segregating with the MA phenotype (Nyholt et al. 1998b; Jones et al. 2001). Although the peak linkage regions identified do not overlap, the locus contains a number of interesting candidates. In particular the FHM1 gene CACNA1A is coded in the locus however this gene has not been consistently found to contribute to common migraine (Lea et al. 2001; Terwindt et al. 2001; Jen et al. 2004). Other genes within the region that have been of particular interest include the insulin receptor (INSR), notch homolog 3 (NOTCH3) and the low density lipoprotein receptor (LDLR) genes.
Analysis of the INSR gene showed 5 SNPs with positive association to migraine in 2 independent populations (McCarthy et al. 2001). This was also confirmed in a large study by Netzer and colleagues in a German population (Netzer et al. 2008a) where one of the five SNPs, rs2860174, showed significant allelic association P=0.005. While the exact role of the insulin receptor in migraine is still to be elucidated there is epidemiological evidence of migraine and diabetes co-morbidity. Furthermore fasting may also be considered a trigger for migraine occurrence (Netzer et al. 2008a).
The NOTCH3 gene also resides at 19p13.2-13.1 and encodes a large single pass transmembrane protein expressed in arterial vascular smooth muscle cells. Mutations in this gene are known to cause the rare autosomal dominant disorder cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy (CADASIL) which presents with recurrent subcortical ischemic strokes and MA. Therefore analysis of SNPs within exon 3 and 4 (where the majority of CADASIL mutations are found) of this gene have been undertaken in migraine populations. Schwaag and colleagues sequenced these exons in 97 patients; no new mutations were identified but associations were found with 2 SNPs (G864A genotypes P=0.008 and C381T allelic P=0.032) and the MO subtype (Schwaag et al. 2006). We recently undertook analysis of the same 2 polymorphisms in 2 independent Australian populations. The results confirmed the association in the first population, however upon replication the C381T SNP failed to show association suggesting further investigation of the variant is required to confirm its involvement in migraine. The 2nd SNP G864T however, showed strong association particularly with the MA subtype (population 1 genotype P=0.004, population 2 genotype P=0.005), however as this
58 SNP is synonymous it is still unclear how it may affect the function of NOTCH3 (Menon et al. 2011).
Finally the LDLR gene has also shown conflicting evidence for involvement in migraine. Mochi et al. (2003) identified a microsatellite in exon 18 that showed positive association with MO; however this association was not replicated in an Australian population (Curtain et al. 2004). Furthermore no relationship was found between cholesterol levels and migraine diagnosis in the Norfolk Island genetic isolate casting additional doubt on the role of this gene in migraine (Curtain et al. 2004).
3.2.1.2 Candidate Genes and 15q11-q13 (MGR5, MIM ID: 609179) Linkage studies have also implicated the 15q11-q13 region which contain the GABA- A subunits GABRB3, GABRA5 and GABRG3 (Russo et al. 2005). Sequencing of the coding regions of these genes identified only one polymorphism in GABRB3 in the affected individuals in only one (of five) families in the linkage studies. Further studies by Netzer et al. (2008b) and Oswell et al. (2008) in independent case-control cohorts also showed no evidence for involvement of this cluster in migraine. This suggests that if these genes are the causative ones in the region, then another polymorphism perhaps in a regulatory region may be playing a role.
While genetic association studies of GABA receptor subunits both at Xq24-28 and 15q11-13 have not conclusively confirmed their role in migraine, there is significant clinical evidence suggesting their involvement particularly as GABA-A receptor agonists such as topiramate and gabapentin are commonly used migraine prophylactics (Fernandez et al. 2008). Further research is required on both the subunits already considered and those not yet investigated.
3.2.1.3 Candidate Genes and 10q22-23 (MGR12 MIM ID: 611706) As this region has been identified in a number of studies (see Table 3-1) Anttila and colleagues (2008) focused candidate gene studies on the narrow region shared by these studies. 1323 SNPs were genotyped across the region and a few key genes were identified as possible candidates. KCNMA1 is a Maxi-K calcium level detecting potassium channel involved in ion transport similar to the existing FHM genes. NRG3
59 is a growth and differentiation factor involved in oligodendrocyte survival and finally GRID1 which is a glutamate receptor ionotropic delta 1 subunit. Unfortunately none of the SNPs tested reached the significance threshold. However the highest observation was identified in the KCNMA1 gene and 3 other regions outside coding genes were identified for potential follow-up. The authors suggest that the region may contain a number of susceptibility variants but a larger sample is required in order to detect these (Anttila et al. 2008).
3.2.1.4 Tresk and 10q25 (MG13 MIM ID: #613656) Lafreniere and colleagues recently identified the first functional variant in a gene to show linkage and causal involvement in familial MA (Lafreniere et al. 2010). The KCNK18 gene encodes the TRESK K2P channel involved in neuronal excitability. This study sequenced the coding region in 110 unrelated migraineurs and 80 controls and identified 2 variants that were found only in migraine sufferers. 1 variant was synonymous while the second was a 2bp deletion resulting in a prematurely truncated protein. Subsequent gene sequencing analysis in Australian samples also identified 9 other variants.
Linkage analysis of a large multigenerational family with MA identified a region from 10q25.2-25.3 (9.7Mb) in which KCNK18 was the only ion channel gene of 52 known genes encoded at the locus. Further analysis of this family identified the same 2bp deletion variant that segregated with all affected individuals within the family (Lafreniere et al. 2010).
To further investigate TRESK in migraine pathophysiology Lafreniere and colleagues also investigated the expression of TRESK in neuronal tissue using western blot and immunohistochemistry. This assay showed that the protein was strongly expressed in the trigeminal ganglion neurons of humans supporting a role for TRESK in neuronal excitability. In addition, similar mouse studies showed expression of TRESK in the spinal cord and brain regions of adult mice. Further analysis of the functional consequence of the identified 2bp deletion also showed that the truncated protein is non-functional and was not activated by the calcium ionaphore ionomycin which induces a robust and reversible action of the WT TRESK channel. Furthermore, it was demonstrated that the mutant channel causes a dominant-negative downregulation of
60 wildtype channel activity when co-expressed. This was determined through characterisation of K+ currents in Xenopus oocytes expressing both wild-type and mutant TRESK channels (Lafreniere et al. 2010). This study concluded that migraine risk may therefore increase as TRESK activity decreases due to genetic mutation.
This research clearly demonstrated how a linkage study can be used to confirm involvement of a specific chromosomal region as well as a specific candidate gene. The family analysed shared similar migraine episodes suggesting that a homogenous group for analysis potentially aided the identification of this gene. Therefore the results of this study indicate that further analysis of KCNK18 is required in migraine groups of similar phenotype and that TRESK may be an effective target for new therapeutics in a subgroup of migraineurs.
3.2.2 GWAS Genome-wide association studies are a relatively new approach that has been successful in aiding the current understanding of many complex disorders, including various cancers, alzheimers, inflammatory bowel disease and diabetes. This approach requires large case-control cohorts and genotyping of 100,000’s of SNPs generally using commercially available array techniques. Common variants with large effect size should be relatively easy to identify using this approach, however as is the case for most complex disorders, it is expected that risk alleles with smaller effect sizes are likely to contribute and these would then require large cohorts or meta-analysis of a number of studies to identify. Alternatively rare variants with larger effect sizes may also be implicated however these are not commonly well covered in the arrays used (Seng and Seng 2008). Despite these limitations GWAS have proved useful in both confirming existing susceptibility loci and identifying new regions for further investigation.
Unfortunately to date only one GWAS has been published focusing on migraine. The GWAS analysed over 3000 Finnish, German and Dutch migraineurs recruited from headache clinics and compared these to age and sex matched population based controls. The study classified migraineurs according to the ICHD-II criteria. The study identified only one marker at the appropriate level of significance on
61 chromosome 8q22.1. The marker is situated between the genes MTDH and PGCP, both of which are in pathways thought to regulate glutamate accumulation in the synaptic cleft. The association of this marker with MA was also replicated in Danish and Icelandic populations strongly suggesting the involvement of this locus in MA (Anttila et al. 2008). This finding potentially strengthens the evidence for a mechanism involving excess glutamate in the synaptic cleft contributing to the occurrence of a migraine aura and/or the headache phase as has been suggested in studies of hemiplegic migraine.
3.2.3 Candidate Genes In addition to genome-wide approaches such as a GWAS and linkage studies, many studies have taken a candidate gene approach using case-control association studies. These studies have generally either attempted to cover the specific gene using tagging SNPs or have selected known functional variants that may have been associated with disorders related to migraine. The genes that have been focused on primarily include genes with neurological, hormonal or vascular functions. Tables 3-2, 3-3 and 3-4 list a number of association studies that have shown positive association between migraine and genes in these categories respectively.
Replication studies have been undertaken on a significant number of these, however for the vast majority, results are conflicting or inconclusive. This may be due to any number of reasons including under-powered studies, ethnic differences, or the genetic and phenotypic heterogeneity of the disorder which could vary significantly between independent case-control cohorts, contributing to the difficulty in replicating associations with variants of small effect size.
3.2.3.1 Neurological Candidate Genes The trigeminovascular system is believed to be integral to the onset of migraine (Moskowitz 2008). The neurotransmitters, peptides, receptors and channels located in various components of this system may trigger vascular dysfunction and downstream pain signals and are therefore key candidates. The serotonergic system is of particular interest as 5-HT is a neurotransmitter involved in a plethora of biological functions including information processing and nociception. Furthermore 5-HT receptor
62 Table 3-2 Positive migraine association Studies: neurological genes Gene Locus Reference Ethnicity Cases Controls # SNPs Associated SNPs P value Serotonin related Genes HTR1E 6q14-q15 (Corominas et al. 2010) Spanish 528 528 8 rs828358 P=00018* rs1581774 P=0.016* (MA) HTR2A 13q14-q21 Spanish 528 528 24 rs7984966 P=0.037* (MO) rs7322347 P=0.07* (MO) rs9534511 P=0.012* (MA) rs6561332 P=0.016* (MA) HTR2C+ Xq24 Spanish 528 528 9 rs4911871 P=0.029* rs2428721 P=0.036* (MA) HTR3A 11q23.1 Spanish 528 528 4 rs1176717 P=0.042* (MA) HTR3B 11q23.1 Spanish 528 528 9 rs11214775 P=0.025* (MO) HTR4 5q31-q33 Spanish 528 528 17 rs7721747 P=0.034* (MO) HTR7 10q21-q24 Spanish 528 528 11 rs1298056 P=0.0058* (MA) DDC+ 7p12.2 Spanish 528 528 15 rs1982406 P=0.0035* (MA) rs6944090 P=0.021* (MA) MAOA+ Xp11.3 Spanish 528 528 2 rs3027400 P=0.0093* (MO) rs2072743 P=0.043* (MO) Dopamine related Genes DBH 9q34 (Fernandez et al. 2006) Australian 275 275 2 19bp in/del P=0.003 (MA) (Fernandez et al. 2009) Australian 200 200 rs16111115 P=0.012 300 300 rs1611115 P=0.031 (Todt et al. 2009) German 650 2937 1 rs2097629 P=5.57x10-8 (Corominas et al. 2009) Spanish 263 274 11 rs1611131 P=0.04 SLC6A3+ 5p15.3 (Todt et al. 2009) German 650 2937 1 rs40184 P=6.36x10-7 DRD2+ 11q23 (Todt et al. 2009) German 650 2937 1 rs7131056 P=0.034 DRD3+ 3q13.3 (Corominas et al. 2009) Spanish 263 274 10 rs12363125 P=0.03 rs22832265 P=0.008 Glutamate Receptors GRIA1 5q31.1 (Formicola et al. 2010) Italian 250 260 6 rs2195450 P=0.00002 (MA) rs548294 P=0.0003 (MO) GRIA3 Xq25 (Formicola et al. 2010) Italian 250 260 8 rs3761555 P=0.0001 (MA) *Authors note that after applying a Bonferroni correction for multiple comparisons the significant threshold for the study was set to 1.24x10-4 therefore none of the P-values remain significant. + Other studies have failed to show association between migraine and these genes.
63 agonists (Triptans) originally developed as vasoconstrictive agents, have been observed to mitigate migraine attacks. While the exact mechanisms of the serotonergic system in migraine are still unknown one theory considers a deficiency of central 5-HT associated with sensitivity to an increase in 5-HT release to be a basis for migraine aetiology (Hamel 2007; Panconesi 2008). Extensive research has been carried out to identify genetic variants that may alter the functions of a number of genes involved in 5-HT function and regulation. These genes include the 5-HT1B, 1D and 2C receptors, the 5-HT transporter (SLC6A4), Tryptophan Hydroxylase (TPH2), Monoamine Oxidase A enzyme (MAOA), and CGRP genes.
The serotonin transporter gene SLC6A4 in particular has been extensively studied. This gene on chromosome 17q11.2 encodes the integral membrane protein that transports serotonin into and out of the synaptic cleft in a sodium dependent manner. In this gene 2 polymorphisms have been of particular interest. The first is an insertion deletion polymorphism in intron 2 known as STin2, with 2 common variants designated STin2.10 and STin2.12. Analysis of this polymorphism has provided conflicting results. Schurks recently conducted a meta-analysis of 5 studies considering this polymorphism and found that the non-STin2.12 alleles appear to provide a protective effect against migraine in the populations studied (Schurks et al. 2010a). The second polymorphism is a 44-bp insertion/deletion in the promoter region known as 5-HTTLPR. The shorter allele is associated with slower clearing of serotonin from the synaptic cleft (Schurks et al. 2010a). Early studies provided evidence of an association with migraine or MA (Juhasz et al. 2003; Marziniak et al. 2005). In contrast Todt et al. (2006) conducted a study including this variation and a number of SNPs across the gene and found no association to MA and a recent study by Corominas et al. (2010) also confirmed this finding. A meta-analysis of 10 studies considering this polymorphism also determined no overall association, although the authors noted that migraine type and gender may modify the influence this gene has on migraine (Schurks et al. 2010d).
Similarly the dopaminergic system has been implicated in the pathogenesis of migraine due to the presence of dopamine driven processes that occur prior to or during a migrainous episode (Sicuteri 1977), the incidence of which are increased by dopamine and dopamine agonists. It is therefore hypothesised that migraineurs may
64 be hypersensitive to dopamine and it may act as a trigger for the migraine attack (Goadsby et al. 2002; Akerman and Goadsby 2007). However, the exact mechanism through which the dopaminergic system influences migraine remains unclear. Dopamine receptors are present in the trigeminocervical complex and administration of dopamine agonists in rat inhibits neuronal firing and consequent nociceptive transmission (Bergerot et al. 2007). Yet, dopamine antagonists are also known to be effective in relieving the migraine therefore it is uncertain whether these function through the dopamine receptor or another path (Charbit et al. 2010). Despite the apparently conflicting roles for dopamine and dopamine antagonists in migraine there is significant evidence to suggest there is an influence on the pathogenesis of this disease and therefore investigations into dopamine receptors, transporters and the dopamine beta hydroxylase (DBH) gene have been undertaken.
DBH converts dopamine into noradrenaline which is also a key neurotransmitter. In an Australian population a promoter insertion/deletion polymorphism that is associated with reduced plasma enzyme activity has been studied. The homozygous del/del genotype was shown to increase migraine risk in males up to three times (Fernandez et al. 2006). Further analysis of the promoter region in the Australian population considered a SNP that is responsible for 31-52% of enzyme activity. This SNP, rs161115, showed significant association in the migraine cohorts tested (population 1 P=0.012, population 2 P=0.031) and particularly associates with MA (Fernandez et al. 2009). Corominas and colleagues similarly analysed this SNP in a Spanish population. They found association in migraine only in their first cohort but this was not replicated in their second, therefore it was dropped from further analysis (Corominas et al. 2009a). A final SNP was also recently considered by Todt and colleagues and showed significant association in a German MA population using ~650 cases and an enlarged control group (rs2097629 allelic P=0.0116). This SNP is not in LD with the highly functional SNP analysed in the Australian population (Todt et al. 2009) suggesting that there may be two different functional variants that influence this enzyme and its role in MA.
3.2.3.2 Hormone Candidate genes The observations that migraine generally increases in women at puberty (Lipton et al. 2001a), and may be altered with reproductive milestones such as menstruation,
65 pregnancy, menopause and hormone therapy (Maggioni et al. 1997; MacGregor 2009) strongly indicates that fluctuating hormones, particularly estrogen levels may play a role in triggering a migraine attack. It has previously been hypothesised that prolonged exposure to high levels of estrogen prior to a drop in concentrations i.e. ‘estrogen withdrawal’ may precipitate the migraine event (MacGregor 2004). Genetic studies have focused on hormone receptors such as estrogen receptor 1 (ESR1) and the progesterone receptor.
ESR1 is located on chromosome 6q25.1 and is expressed in a number of areas of the brain and other tissues. ESR1 is believed to be involved in gene expression and may also be involved in modulation of neurotransmitters such as CGRP, glutamate and serotonin. In addition steroid hormones are thought to have vascular effects such as influencing NO production thereby affecting vascular tone (Gupta et al. 2007).
An early study considered a SNP in ESR1 known to be associated with breast cancer. The G594A SNP in exon 8 was found to be positively associated with migraine in two independent Australian case-control populations (population 1 genotypic P=0.008, population 2 genotypic P=4x10-5) (Colson et al. 2004). A separate study of the progesterone receptor in the same populations also showed association (Population 1 genotypic P=0.04, Population 2 genotypic P=0.019). Furthermore, analysis of both hormonal genes together determined that the interaction of the PROGINS Alu insertion allele in intron 7 combined with the 594A ESR1 allele increased migraine risk by 3.2 (Colson et al. 2005). However follow-up studies of 2 further SNPs within the ESR1 gene in intron 1 and exon 4 (G325C) found no association in the same population (Colson et al. 2006a). This is in contrast to a Spanish study which identified positive association with the G325C but not the original G594A polymorphism (Oterino et al. 2006). Negative results were similarly found in a larger Finnish cohort that looked at 26 SNPs across the gene (Kaunisto et al. 2006), and another Spanish study that found no association across 3 SNPs (Corominas et al. 2009b).
Recently Schurks and colleagues conducted a meta-analysis of sex hormone receptor genes and migraine in order to summarise the existing data on these variants and their respective associations with migraine. This study analysed the previously discussed
66
Table 3-3 Positive migraine association Studies: hormone related genes Gene Locus Reference Ethnicity Cases Controls # SNPs Associated SNPs P value Progesterone 11q22 (Colson et al. 2005) Australian 275 275 1 P=0.02 Receptor 300 300 1 P=0.003 ESR1+ 6q25.1 (Colson et al. 2004) Australian 224 224 1 rs2228480 P=0.003
260 260 1 rs2228480 P=8x10-6 (Oterino et al. 2006) Spanish 240 160 rs1801132 P=0.008 (females) +Other studies have failed to show association between migraine and these genes
Table 3-4 Positive migraine association studies: vascular related genes Gene Locus Reference Ethnicity Cases Controls # SNPs Associated SNPs P value ACE 17q23.3 (Joshi et al. 2009) Indian 150 150 rs4646994 P=0.04 (MA) (Paterna et al. 1997) 191 201 rs4646994 P<0.05 (Kowa et al. 2005) 176 248 rs4646994 P<0.01 (MA) MTHFR+ 1p36.3 (Kowa et al. 2000) Japanese 74 261 1 rs1801133 P<0.01 (Kara et al. 2003) Turkish 102 136 2 rs1801133 P=0.015 (Lea et al. 2004) Australian 270 270 1 rs1801133 P=0.017 (MA) NOTCH3 19p13.2-13.1 (Schwaag et al. 2006) Caucasian 97 97 2 rs1043994 P=0.005 (Menon et al. 2011) Australian 275 275 2 rs3815188 P=0.002 (MO) rs1043994 P=0.001 (MA) 300 300 2 rs3815188 P=0.06 (MO) rs1043994 P=0.003 (MA) EDNRA 4q31.22 (Tikka-Kleemola et al. 2009) Finnish 850 900 13 rs2048894 P=0.015 (MA) + Other studies have failed to show association between migraine and these genes
67 ESR1 G594A and C325G SNPs as well as the PROGINS Alu insertion. An additional ESR1 Pvu II C>T SNP was also considered. The authors identified an association between migraine and both the ESR1 G594A and C325G SNPs that followed dominant and recessive models respectively. In contrast no associations were identified for the remaining variants tested (Schurks et al. 2010c).
In order to overcome some of the conflicting reports on ESR1 and to further investigate the gene-gene interactions a study was conducted by Oterino and colleagues considering a multilocus analysis of 5 estrogen related genes. Nominal association was observed in the ESR1, ESR2 and FSHR genes and further analysis of gene-gene interactions suggested these loci were significantly associated with MA/MO and MA alone (Oterino et al. 2008). These studies support the role of hormones and/or hormone related genes in migraine. However it is still unclear whether the genes analysed so far are the key contributors. If so, the relationship between hormone levels and their specific influences on migraine pathophysiology in the central nervous system and/or the vascular system still needs to be defined.
3.2.3.3 Vascular Candidate genes Migraine was once thought to be a vascular disorder due to observations of increased blood flow prior to and during a migraine episode. This theory hypothesised that the initiating event in a migraine episode occurred in the perivascular nerves of the cerebral vasculature (Parsons and Strijbos 2003). However development of new effective migraine drugs has shown that vasoconstriction is not required for migraine treatment and therefore vasodilation of cranial vessels may only be a secondary phenomenon caused through activation of the Trigeminovascular system (Goadsby et al. 2002). Despite this finding, the involvement of the vasculature in migraine pathophysiology is clear and the effects of vasoactive drugs in migraine treatment cannot be ignored. Consequently numerous genetic studies of vascular genes that can alter vascular endothelial function have been undertaken. These include NOS, CGRP, ACE, MTHFR and the NOTCH3 gene.
Homocysteine is an important regulator in vascular disease and is thought to also play a role in migraine. Homocysteine is an intermediate metabolite of methionine, and MTHFR catalyses the reduction of 5,10-methyltetrahydrofolate to 5-
68 methyltetrahydrofolate, the predominant circulatory form of folate, which is the carbon donor required for methylation of homocysteine to methionine (Lea et al. 2009). A SNP in the MTHFR gene C677T causes an alanine to be replaced with a valine within the catalytic domain of the enzyme reportedly reducing the enzymatic capacity by up to 50% (Frosst et al. 1995). This in turn may lead to mild hyperhomocysteinemia which has been associated with endothelial cell injury (Hering-Hanit et al. 2001), reduced production of NO (reviewed by Colson et al. 2006b), oxidative stress and may contribute to the activation of the trigeminal fibres.
The C677T SNP has been studied in a number of migraine populations with conflicting results. Kowa and colleagues considered a Japanese cohort of 74 cases and 261 controls and found the TT genotype was significantly associated with migraineurs, particularly MA (Kowa et al. 2000) and this has been confirmed in a number of independent studies of various ethnicities. (Kara et al. 2003; Lea et al. 2004; Scher et al. 2006; Liu et al. 2010). In contrast though, Schurks recently conducted a large study using data from the Women’s health study in the US where it was found that the TT genotype conferred a modest protective effect on MA (Schurks et al. 2010e). Alternatively, Oterino did not identify an association in a Spanish population, however it was noted that the T allele was higher in the MA cohort than in the MO cohort. A lack of association has also been found in a number of other studies (Kaunisto et al. 2006; Ferro et al. 2008; Joshi et al. 2009).
Due to the number of studies conducted on this particular variant it is also interesting to consider the results of a large meta-analysis. Rubino and colleagues were the first to consider the MTHFR SNP using the meta-analysis approach and analysed 8 studies involving 2961 migraineurs predominantly of Caucasian populations. This analysis showed a significant association between MA and the TT genotype (Rubino et al. 2009). Interestingly, the second and more recent analysis by Schurks et al. (2010b) also found a similar result, however their analysis determined that the result appeared to be driven by the non-Caucasian populations included in the study. The authors noted the inclusion of a number of recent studies and variations in methodology that may contribute to the variations in results (Schurks et al. 2010b). However, overall, the Rubino et al. (2009) meta-analysis and the Schurks recent analysis showed a significant involvement of this gene in MA.
69 The role of the MTHFR gene is further supported by observations that supplementation with folic acid (Vitamin B9) combined with vitamins B12 and B6 can not only lower homocysteine levels but can also affect migraine symptoms. This was demonstrated in a recent pilot study that showed reduction in homocysteine levels also reduced migraine frequency, severity and disability in MA sufferers. Furthermore, this response was associated with the C677T SNP where individuals with at least one C allele showed improved response over the TT genotype (Lea et al. 2009).
3.3 Migraine and the X Chromosome
Epidemiological evidence and the female preponderance of this disorder strongly suggest a hormonal influence on migraine. However, families with an excess of affected females and lack of male to male transmission may also suggest an X chromosomal component. Therefore there have been a number of studies using a linkage analysis approach that considered the X chromosome. To date three susceptibility loci had been identified on the X chromosome at Xp22, Xq13 and Xq24-q28.
Xp22 In 2002 Wessman showed nominal linkage for Xp in MA in a Finnish population (Wessman et al. 2002). Antilla’s study in 2008 comparing the TCA, LCA and ICHD- II classification methods also provided further evidence for the Xp22 locus, demonstrating linkage for pulsation, the LCA severe class and the ICHD-II MA class in the Finnish population used (Anttila et al. 2008). In addition, a study of 61 families of European descent by Weiser has helped to confirm this region; however this study did not consider any particular subtype or characteristic of migraine (Wieser et al. 2010).
Xq13 Oterino and colleagues (2001) reported a linkage analysis of a large multigenerational pedigree displaying co-segregating MA with Charcot-Marie-Tooth disease (CMTX1) caused by a mutation in the connexin 32 gene located in the Xp13. In this analysis
70 eleven markers from Xp12 – Xq28 were genotyped in 42 members of the pedigree. Results demonstrated linkage of MA to the Xp13 marker effectively excluding the other known susceptibility loci on the X chromosome.
Xq24-q28 (MGR2, MIM ID: 300125) Research conducted in our Genomics Research Centre laboratory by Nyholt and colleagues reported evidence for an X-linked component in migraine (Nyholt et al. 1998a) that was later localised to the Xq24-28 region (Nyholt et al. 2000). These studies identified excess allele sharing in 2 large multigenerational Australian pedigrees (designated MF7 and 14) at the region between DXS1001 and DXS1206 at Xq24 with a peak combined non-parametric LOD (LOD*) score of 2.388 (p=0.0005) calculated using the GENEHUNTER-PLUS program. As well as peak combined LOD* scores >2 at Xq27 between markers DXS984-DXS1123, and at Xq28 between markers DXS8091 - Xqter (Figure 3-1). Further analysis of these families using the HOMOG program supported genetic heterogeneity of the migraine in these families (P=0.0065) suggesting that multiple loci are contributing to migraine (Nyholt et al. 2000).
Figure 3-1 Xq24-q28 analysis (from (Nyholt et al. 2000)) Non-parametric analysis of MF7 and MF14 conducted in GENEHUNTER-PLUS
71 The region from Xq24-Xq28 harbours a number of candidate genes that have been investigated including 5-hydroxytryptamine (serotonin) receptor 2C (5HT2C), Glutamate Receptor ionatropic AMPA3 (GRIA3), gamma-aminobutyric acid A receptor epsilon (GABRE), gamma-aminobutyric acid receptor theta (GABRQ) and gamma-aminobutyric acid A receptor 3 (GABRA3).
The serotonin receptor 5HT2C at Xq24 has been analysed in a number of populations, however these studies have shown no indication that variants in this gene are involved in migraine (Burnet et al. 1997; Johnson et al. 2003; Racchi et al. 2004). Similarly, the GABRE and GABRQ genes at Xq28 that code for 2 of the 19 subunits of the GABA-A receptors have shown no association in an Australian population (Fernandez et al. 2008). However GABRA3, also at Xq28 is yet to be analysed. This gene therefore remains a potential candidate that should be investigated due to the role of GABA as the main inhibitory neurotransmitter in the brain.
In contrast to GABA, glutamate is the main excitory neurotransmitter and functions through the Alpha-amino-3- hydroxy-5-methyl-4-isoxazole-propionin acid (AMPA) ionatropic receptor. Glutamate is believed to be required for cortical spreading depression and to activate the TNS and central sensitization, thus inhibition of glutamate release culminates in an anti-nociceptive effect (Vikelis and Mitsikostas 2007; Neeb et al. 2010). GRIA3 at Xq24 codes for 1 of 4 subunits for the AMPA receptor.
Formicola and colleagues analysed a number of SNPs in each GRIA gene in an Italian population of 250 migraineurs and 260 controls. Their results indicate positive association with 2 SNPs in GRIA1 (5q33.2, rs548294 MO allelic P=0.008, rs2195450 MA allelic P=0.0005) and 1 SNP in GRIA3 (rs3761555 MA Females allelic P=0.003) (Formicola et al. 2010). These results suggest that GRIA3 may contribute to the linkage signal at Xq24, however further research into the glutamate system is warranted to confirm its role in migraine susceptibility and pathophysiology.
While GRIA3 appears a prime candidate for migraine susceptibility at Xq24, there are numerous genes in the Xq27-Xq28 loci that still vie for attention as candidate genes for migraine studies. In particular these include genes that have shown expression in
72 various regions of the brain such as the Chromosome X Open Reading Frame 1 gene (CXorf1) that is expressed in the hippocampus (Redolfi et al. 1998); Orphan G protein coupled receptor 50 (GPR50) that is expressed in the hypothalamus (Sidibe et al. 2010); Cyclic Nucleotide Gated Ion Channel 2 (CNGA2) which is expressed in the trigeminal ganglion and cerebral and basilar arteries (Kruse et al. 2006; Qu et al. 2006; Podda et al. 2008); and GDP dissociation inhibitor 1 (GDI1) that is primarily expressed in neural and sensory tissues.
Other genes have known involvement in signalling and neurite activity such as SLIT and NTRK-like family member 2 (SLITRK2) (Aruga and Mikoshiba 2003; Aruga et al. 2003); ATPase Ca2+ transporting plasma membrane 3 (ATP2B3) (Burette and Weinberg 2007); Chloride Intracellular Channel 2 (CLIC2) (Board et al. 2004).
Examination of the role of the X chromosome in migraine is far from complete. A number of susceptibility loci have been identified, yet apart from GRIA3, no evidence has been published that identifies the causative factor. The X chromosome factor therefore remains an elusive key that may contribute to our understanding of the predisposition of women to this disorder.
3.4 Conclusion
Common migraine is a polygenic multifactorial disorder that is likely influenced by multiple genes and environmental triggers. Variants are likely to involve gene- environment and gene-gene interactions increasing the genetic variability of the disorder and perhaps explaining why many single gene studies are conflicting in their outcomes.
The search for migraine genes remains complicated by the fact that many susceptibility variants are likely to have modest contributions to the phenotypic expression of the disorder. Linkage studies are a sound approach for identifying contributing loci. However, confirmation in independent studies and possibly through future GWAS is required in case rare family-specific mutations of larger effect size distort linkage signals which may therefore not translate to analysis in case-control
73 populations. An additional influence on the outcome of many of the genetic studies is the spectrum of phenotypes among sufferers. Approaches such as LCA and TCA in addition to the use of specific ICHD-II criteria may assist to improve concordance amongst the studies being undertaken as has been seen in a handful of linkage studies.
The approaches used in migraine genetics, be it genome-wide or candidate gene, each have a role to play in identifying new regions and confirming existing studies. Overall, current linkage, GWAS and candidate gene studies provide tantalising insights into the pathophysiology of migraine. There have been some successes such as the recent studies implicating a functional mutation in the TRESK gene, and others that warrant further investigation at genetic and/or clinical levels such as ESR1 and MTHFR. These genes and others considered in this review are promising migraine candidates that require further investigation particularly of gene-gene interactions to assist in building a gene profile of this complex disorder.
In relation to this PhD the identification of the Xq24-q28 susceptibility is the foundation on which the work here expands. In order to do this, additional markers and Caucasian families have been genotyped in the Xq27 and Xq28 regions and analysed to assist in narrowing the identified loci. Furthermore, key markers have been analysed in a Caucasian population to determine if these susceptibility loci contribute to migraine generally or are pedigree specific. Finally the above mentioned candidate genes are also analysed for association with migraine in a Caucasian migraine association population. An X chromosome scan in the Norfolk Island genetic isolate is presented to identify new X chromosomal loci and to determine if Xq susceptibility loci contribute to migraine in independent populations. The design and methods employed are presented in the following chapter. Results have been documented for publication with manuscripts presenting results in chapters 5-7 and additional work described in chapter 8.
74 CHAPTER FOUR: Materials and Methodology
The purpose of the research described herein was to identify and analyse genes on the X chromosome for association with migraine. Due to the complex nature of migraine numerous complementary techniques were employed to reach this goal. As migraine is highly heritable, pedigree analysis was employed to identify new susceptibility regions on the X chromosome, and to refine existing regions. This approach required screening of DNA markers that were genotyped using various techniques described here. Analysis of this data also involved a number of approaches including both simple linkage analysis (accounting for the hemizygous nature of males due to the X chromosome) and the application of novel approaches to account for complex pedigree structures. The pedigree analysis provided a basis through which the study design progressed to a candidate gene approach using migraine association populations, again involving various genotyping and analysis techniques that are described in this chapter.
4.1 Sample Ascertainment, Structure and Epidemiology
A number of different cohorts were used throughout this research. The ascertainment and classification of the samples within each cohort as well as advantages and disadvantages for the use of each cohort type are explained here. Ethical clearance was granted by the Griffith University Human Research Ethics Committee prior to the initiation of this project. All participants provided signed informed consent.
4.1.1 Norfolk Island Population Norfolk Island is located in the South Pacific Ocean off the east coast of Australia. The present day Norfolk Island population can trace their family histories back to the original founders of the population who consisted of 9 British Bounty mutineers, 6 Tahitian women who originally inhabited the nearby Pitcairn Island in 1790 (Hoare 1999). Detailed genealogical databases have allowed reconstruction of the full Norfolk Island pedigree that incorporates ~6000 individuals and spans 11 generations.
75 The Norfolk Island population represents a unique cohort in which to study complex disorders. Geographical isolation as well as strict quarantine and immigration laws have resulted in a more homogeneous lifestyle then is typically seen in outbred populations. Furthermore it is expected that these factors in combination with relatively few genetic founders contribute to a reduction in genetic diversity simplifying genetic models and minimising environmental effects. These factors synergistically increase the possibility of identifying susceptibility genes particularly for complex disorders.
For this study 600 participants were initially recruited based on permanent resident status to ensure sampling from a shared genealogical background. A comprehensive medical questionnaire was used to obtain phenotypic data including migraine information regarding family history, symptoms, triggers and medication. Migraine diagnosis was in accordance with ICHD-II guidelines (IHS 2004). Migraine prevalence has been determined at 25.5% in the Norfolk population (Cox et al. 2011).
Venous blood samples were obtained from the 600 participants (261 males, 339 females) with a mean age of 50.8 years (standard deviation 16.4 years). Blood samples were collected in EDTA tubes and DNA extracted from 10-20 ml using a standard salting out procedure (Miller S.A et al. 1988) explained in detail in the next section. Concentration and purity of the DNA yields were determined spectrophotometrically using the NanoDrop ND-1000 (NanoDrop Technologies, Inc. see section 4.2.4 DNA quantification).
To ensure familial links between all participants in the study genealogical data was obtained via questionnaire as well as municipal and historical records. 377 of the 600 individuals collected have been identified as descendants of the Norfolk Island genetic founders (9 Isle of Man ‘Bounty Mutineers,’ 6 Tahitian women and 2 Caucasian sailors that joined the population in the 19th century). 288 of the 377 individuals were selected as highly informative individuals within the pedigree structure and were therefore used for genotyping. This subset consisted of 136 males and 152 females including 76 migraine cases consisting of 22 males and 54 females.
76 4.1.2 Migraine Pedigrees Eight Migraine Families (MF) were used for the studies outlined in this thesis. Pedigrees were designated MF6, MF7, MF14, MF47, MF55, MF541, MF878 and MF879. Pedigree structures are shown in Figures 4-1 to 4-8. These families were selected as there was no father to son transmission of migraine observed, except in the case of individual III:5 to IV:6 in MF14 and individual I:1 to II:3 in MF879. MF7 and MF14 have previously been reported to show linkage to the susceptibility region Xq24-28 (Nyholt et al. 1998a; Nyholt et al. 2000) and are further fine-mapped in this study. Similarly, MF6, MF47 and MF55 have been identified previously and have had a limited number of markers analysed in the region of interest (unpublished data). These families were therefore interrogated to a greater extent in this study. MF541, MF878 and MF879 have been collected recently and have not been used for any prior analysis.
These pedigrees include a total of 180 individuals including 125 for whom DNA was available. Individuals were diagnosed for migraine (MA or MO) according to ICHD- II criteria (IHS 2004). A questionnaire was used to obtain specific information from study participants on a number of relevant factors including age of onset, frequency, duration, associated symptoms, medication response and triggers in accordance with IHS guidelines. A total of 96 individuals were classified as affected. Venous blood was collected from participants and DNA was isolated from lymphocytes using a standard salting out procedure (Miller S.A et al. 1988) as described below. In some cases participants under the age of 15 provided saliva samples instead of blood.
4.1.3 Migraine Association Populations The migraine population used in the candidate gene studies outlined in this thesis consisted of 1000 Caucasian Australians recruited from the east coast of Australia. Cases include 366 MA (289 female, 77 male) and 134 MO (97 female, 37 male) sufferers. A higher proportion of female migraineurs have been included in the population for consistency with the female preponderance of the disorder (3:1). Migraineurs were interviewed and completed a detailed questionnaire prior to diagnosis according to ICHD-II criteria as described previously (IHS 2004; Colson et al. 2005). The control group was matched to cases for age (+/- 5 years), sex and
77 ethnicity (Caucasian origin) to reduce the possibility of population stratification. Venous whole blood samples were collected and genomic DNA was extracted using a standard salting out procedure (Miller S.A et al. 1988).
Figure 4-1 MF6
78
Figure 4-2 MF7 See Figure 4-1 for legend
79
Figure 4-3 MF14 See Figure 4-1 for legend
80
Figure 4-4 MF47 See Figure 4-1 for legend
Figure 4-5 MF55 See Figure 4-1 for legend)
81
Figure 4-6 MF541 See Figure 4-1 for legend
82
Figure 4-7 MF878 See Figure 4-1 for legend
83
Figure 4-8 MF879 See Figure 4-1 for legend
84 4.2 DNA Extractions
DNA has previously been extracted and prepared for use for the Norfolk Island and the migraine association populations; however DNA preparations were required for the Migraine Families. MF6, MF7, MF14, MF47 and MF55 (‘existing families’) had previously been extracted and DNA was available as TE stocks, therefore these only required the clean-up (ethanol precipitation) and quantification steps outlined below. However DNA for MF541, MF878 and MF879 (‘new families’) had to be extracted from blood or saliva prior to clean-up and quantification. The extraction process used for blood samples was a standard salting out procedure (Miller S.A et al. 1988). Where a saliva sample had been obtained (as the participant was under the age of 15) extraction was as per the manufacturer’s instructions for the Oragene®DNA kits.
4.2.1 Blood Extractions DNA was extracted from 10 ml whole blood samples that had been stored at -80°C. The extraction procedure required the following reagents:
NKM Buffer (0.14 M NaCl, 30 mM KCl, 3 mM MgCl2)
RSB Buffer (10 mM Tris pH 7.5, 10 mM NaCl, 3 mM MgCl2) Proteinase K Lympholysis Buffer 6 M NaCl 1x TE Buffer 100% (absolute) ethanol
The protocol required blood to be thawed at room temperature and blood samples transferred into 50 ml falcon tubes. The final volumes of all samples were made up to 25 ml by addition of NKM buffer and the contents were vigorously shaken. Samples were then centrifuged at 4800 rpm for 25 minutes at 4°C and the supernatant was discarded. 5 ml RSB buffer was then added to the pellet which was broken up using a transfer pipette. Once the pellet was resuspended the final volumes were then made up to 25 ml by adding RSB buffer in order to lyse remaining red blood cells. A second centrifugation step followed at 4000 rpm for 15 minutes at 4°C. The process of
85 removing the supernatant, breaking up the pellet, addition of RSB buffer and centrifugation was repeated a second time to ensure removal of red blood cells. After the third centrifugation the supernatant was again disposed of and the pellet resuspended in 1 ml of RSB buffer, 4 ml Lympholysis buffer and 250 μl Proteinase K to lyse the white blood cells. Samples were then sealed with the lid and parafilm and placed in a shaking waterbath at 37°C overnight.
Following overnight digestion 2 ml of 6 M NaCl was added to each tube and centrifuged at 2500 rpm for 15 minutes at 4°C. The supernatant was then decanted into a 10 ml tube and centrifuged again at 2500 rpm for 15 minutes 4°C. Finally the supernatant was then transferred to a fresh 50 ml tube and 2 volumes chilled 100% ethanol was added. Upon gentle swirling the DNA precipitates and was transferred to a tube containing 1 ml 1x TE buffer using an inoculating loop. The DNA samples were then incubated at 37°C overnight to resuspend DNA.
4.2.2 Saliva Extractions Four participants all under the age of 15 gave saliva samples instead of blood for DNA extraction. Saliva was collected using the Oragene®DNA kits. These kits seal the saliva sample with a lid that upon closing releases the Oragene DNA solution from the cap and mixes with the saliva to stabilise the DNA. Samples are then stable at room temperature for a number of years. This procedure required the following reagents:
Oragene DNA Purifier 100% Ethanol 70% Ethanol TE Ice
Purification of DNA from the saliva sample was conducted according to the Oragene®DNA kits instructions. This firstly required incubation of the sample at 50°C in a water incubator for an hour. 500ul of the mixed sample was then transferred into a 1.5 ml eppendorf tube and 20 μl (1//25th volume) of Oragene DNA purifier was
86 added to the tube and vortexed. Samples were then incubated on ice for 10 minutes prior to centrifugation at room temp for 5 minutes at 13000 rpm. The clear supernatant was transferred to a fresh eppendorf tube and an equal volume of room temperature 100% ethanol was added and mixed by inversion. DNA was allowed to precipitate by standing at room temperature for 10 minutes. DNA was then pelleted by centrifugation in a known orientation for 2 minutes at 13000 rpm and the supernatant was discarded. The pellet was washed in 250 μl 70% ethanol and allowed to stand at room temp for 1 minute before the ethanol was again removed. The pellet was then dissolved in 100ul of 1x TE buffer and stored at 4°C for use.
4.2.3 Ethanol Precipitation TE stocks were obtained for all migraine families (from either pre-existing extractions or those outlined above). 100 μl aliquots from each stock sample were used for ethanol precipitation and resuspending in H2O as working stocks. The ethanol precipitation procedure required the following reagents:
Sodium Acetate 100% Ethanol 70% Ethanol
H20 Ice
The ethanol precipitation procedure required 100 μl TE stock with 10 μl Sodium Acetate and 300 μl chilled 100% Ethanol to be added to a 1.5 ml eppendorf tube. The contents of the tube were then vortexed and centrifuged at 4°C for 25 minutes at 10000 rpm. The supernatant was discarded and 100 μl chilled 70% ethanol added to the remaining DNA pellet. Samples were again vortexed and centrifuged at 4°C for 25 minutes at 10000 rpm. As much supernatant as possible was removed with care not to remove the DNA pellet which was then air dried prior to resuspension of the DNA in
100 μl H2O. Samples were then incubated at 37°C to ensure resuspension.
87 4.2.4 DNA Quantification
After the DNA for the migraine families had been extracted and resuspended in H20 the samples were quantified to determine the concentration of DNA obtained from the extraction and clean-up process.
Quantification used the Thermo Scientific Nanodrop 1000. Initial concentrations obtained from the extraction methods ranged from 80.51 ng/ul to 634.37 ng/ul. Purity varied with the 260/280 ratio ranging from 1.62 to 1.96. These readings were used to dilute the DNA samples to 20 ng/ul ready for PCR.
4.3 Selection of Polymorphisms
4.3.1 Norfolk Island Study The Norfolk Island study analysed an X chromosome scan which was completed as part of a GWAS. This study used the Illumina Infinium High Density (HD) Human 610-Quad DNA analysis BeadChip version 1. 17,861 X chromosome SNPs are included on this chip.
4.3.2 Pedigrees The pedigree analysis sought to expand previous work that identified particularly the Xq27 and Xq28 susceptibility regions. At the start of this project a number of dinucleotide microsatellite markers had been analysed in the existing families, however these were not consistently genotyped across all families (Table 4-1).
Table 4-1 Existing genotyping in migraine families Marker MF7 MF6 MF14 MF47 MF55 DXS8064 X X X DXS1001 X X X DXS1206 X X X X X DXS984 X X X DXS8106 X X X X X DXS8043 X DXS297 X X X X DXS8091 X X X DXS1123 X X DXS8061 X X X X X DXS15 X X X X DXS1073 X DXS1108 X X
88 To build consistency for this study, the markers not previously genotyped in all existing families were filled in and all 13 markers were genotyped in the new families.
4.3.3 Candidate Gene Studies 4.3.3.1 Selection of Candidate Genes The GRIA3 gene codes a subunit of the AMPA glutamate receptor and is located at Xq24. This gene has previously shown association to migraine in an Italian cohort (Formicola et al. 2010). Therefore candidate gene studies at the Xq24 locus were confined to this gene to determine if the previously observed association could be replicated. In addition the previous study also identified association at the related GRIA1 gene on 5q33; consequently an analysis of this gene has also been included in this study.
In order to select candidate genes in the Xq27 and Xq28 loci gene maps of both regions were downloaded from NCBI (Build 37.2). The Xq28 map spanned from DXS1123 to Xqtel. A total of 188 genes are known to be encoded in this region (See Appendix B), of these 128 are protein coding, 18 encode microRNAs or small nucleolar RNAs, 31 are pseudogenes and the remaining 11 are hypothetical locations.
The aim of this study was to determine if a migraine susceptibility gene could be identified, therefore pseudogenes and RNAs were filtered out and 10 candidate genes were chosen for further analysis based on known gene function and or expression data when it was available. It must be noted that many genes have been identified in this region, however, the majority have limited functional information available. As a consequence the choice of candidate genes may have been biased towards genes with known function. 1 candidate gene was later removed from the study due to limitations in study design and an inability to multiplex any SNPs with frequency data into the genotyping method.
Similarly the NCBI gene map (Build 37.2) was downloaded for the Xq27 region from DXS8043 to DXS297 (See Appendix C). This region was found to be particularly gene poor with only 4 genes, 4 pseudogenes, 6 microRNAs and 2 hypothetical locations in the region. Consequently only 2 genes were selected for analysis in this
89 region. A brief summary of the function of the candidate genes chosen for this study is provided in Appendix D.
4.3.3.2 Selection of SNPs Prior to selecting SNPs for analysis it was determined that the genotyping method would be MALDI-TOF MS (explained in section 4.4.2.2) therefore the number of SNPs selected was limited to the maximum that could be multiplexed using this approach (i.e. 31 SNPs). Therefore candidate genes at the Xq27 and Xq28 loci were interrogated using a tag SNP based approach to obtain a preliminary list of SNPs that would cover the greatest variation across the selected genes. HapMap CEU SNP genotype data was downloaded for each candidate gene and analysed for LD in Haploview v4.2 (Barrett et al. 2005; Barrett 2009). The tagger algorithm was used to select for the least number of SNPs required to capture the greatest proportion of variation across the gene. In the first instance r2 was set to 0.8 and a pairwise approach was employed. Upon completion of this analysis it was clear that 3 genes could not have SNPs selected using this approach as the number of SNPs genotyped within these genes in the HapMap CEU population was n ≤ 2. These genes were GDI1, SLITRK2 and CXorf1.
The GDI1 gene spans 6.292kb and has 11 exons in the Xq28 region. There is only one reported SNP that has been genotyped in the CEU population in this gene. The single SNP is located in exon 3 and was included in the final analysis. Similarly only 1 SNP in the SLITRK2 gene has been genotyped in the HapMap CEU population. This SNP is located in exon 1 of the 8kb gene that codes 5 exons in total. A nearby 5’ SNP was therefore also chosen for analysis. Finally, only 2 SNPs have been genotyped in the CXorf1 gene at Xq27. This gene is 2.4kb and codes only 1 exon, therefore both SNPs were also chosen for inclusion in the study.
Figures 4-9 to 4-15 below detail the CEU LD plots generated in Haploview for each of the remaining candidate genes and the position of the chosen SNPs for these genes. A final selection of 29 SNPs across 9 candidate genes were chosen for analysis. One 12bp deletion was also selected for in-house genotyping in the GPR50 gene and the remaining gene was excluded from the study as no polymorphisms could be identified that would fit within the constraints of the multiplex.
90 Legend - Figures 4-9 to 4-15 Intron/Exon structure of gene shown with exons identified as vertical black bars Positions of Genotyped SNPs shown with arrow LD plot for each gene was generated in Haploview using HapMap CEU data LD of tagged SNPs determined in Haploview using tagger algorithm nb. Results of tagger analysis in GABRA3 (Fig 4-11) are presented in Table 4-2
Figure 4-9 CNGA2 gene structure and polymorphisms
91
Figure 4-10 ABCD1 gene structure and polymorphisms
Figure 4-11 GABRA3 gene structure and polymorphisms
92 Table 4-2 GABRA3 polymorphisms Tagging SNP r2 SNPs Tagged r2 SNPs Tagged r2 SNPs Tagged rs3848926 rs5970220 0.96 rs4828684 0.96 rs17320283 1.00 rs5925128 rs1491792 rs1907600 rs5970223 rs1491793 rs7049894 rs11094547 rs1491791 rs6653441 rs5970221 rs4828692 0.90 rs7062484 rs12833553 0.89 rs11796848 1.00 rs2131190 0.89 rs11796556 rs6627221 0.86 rs5925134 1.00 rs5970232 1.00 rs994424 rs5925140 rs5970242 rs994423
rs5925139 rs7064104 rs389292 0.89 rs10218364 0.96 rs5970293 1.00 rs4828589 rs4612533 1.00 rs6627588 rs6627595 0.94 rs1112122 1.00 rs6627594 1.00 rs6627237 rs970417 0.89 rs6526102 1.00 rs6653443 1.00 rs5970284 0.95 rs6627574 rs1602622 rs1602619 0.95 rs1492294 rs6653479 rs5970269 1.00 rs1009386 rs6526100 rs12688628 rs1602624 rs5970265 rs1565610 rs1492304 rs2201169 rs11094571 rs5969877 rs1492295 rs1492293 rs5970247 rs5970281 rs5969888 rs6526103 0.93 rs7061691 1.00 rs11798711 1.00 rs11798075 1.00 rs1492302 rs11798735 rs11795489 rs11796898 rs11797166 rs17326848 rs34624298 rs11795523
Figure 4-12 NSDHL gene structure and polymorphisms
93
Figure 4-13 ATP2B3 gene structure and polymorphisms
Figure 4-14 FLNA gene structure and polymorphisms
94
Figure 4-15 CLIC2 gene structure and polymorphisms
4.4 Genotyping
A number of different genotyping methods were employed throughout this study. RFLP was used for genotyping of the migraine association population for the candidate gene study of the gene GRIA1. In addition, microsatellite sizing by capillary electrophoresis as well as genotyping of deletion polymorphisms in the candidate genes GPR50 and GRIA3 was also performed in these populations. For genotyping of dinucleotide markers in the pedigree study, capillary electrophoresis was employed. Genotyping was also undertaken using outsourced array and mass spectrometry methods. All methods and their application are discussed in further detail below.
4.4.1 In-house Methods 4.4.1.1 PCR PCR was used to amplify microsatellites for the pedigree analysis, as well as an insertion polymorphism for the candidate gene GPR50 and polymorphisms within the GRIA1 and GRIA3 genes.
95 Primers had either been used in previous studies or were designed for optimisation. New primers were designed by downloading the sequence surrounding the polymorphism and entering this into NCBI primer-blast (available at: www.blast.ncbi.nlm.nih.gov/Blast.cgi). Primer sequences with a Tm close to 60°c GC content>55% and a length 17-25bps were determined and primer pairs were checked for unintended amplicons. In some cases Primer Express v3.0 was also used to obtain or check primers to detect complementary sequences in primer pairs or within a single primer. Table 4-3 lists primer sequences used for all polymorphisms that were genotyped in-house.
PCRs generally followed the protocol outlined in Table 4-4 and were run on Applied Biosystems Veriti® Themal Cyclers or Corbett Robotics Palm Cyclers. If the protocol varied from that shown in Table 4-4 this variation was determined by firstly running a PCR using a temperature gradient to confirm the best annealing temperature (final annealing temperatures are shown in Table 4-3). The temperature gradient covered temperatures from 55-65°Cc at 1°C increments. Where further optimisation of the
PCR protocol was required an MgCl2 gradient was also performed with MgCl2 concentrations ranging from 1.25-3.0mM at increments of ~0.25mM.
The PCR cycle included a 10 minute denaturing step at 95°C, followed by 30-40 cycles of 95°C for 45 seconds, X°C for 45 seconds and 72°C for 45 seconds (where X°C is the annealing temperature specified in Table 4-3). The PCR cycle concluded with a final extension step of 72°C for 7 minutes and products were stored at 4°C prior to analysis. PCR products were checked for amplification using agarose gel electrophoresis using 2% gels and ethidium bromide for DNA detection under UV light. 100bp DNA ladders (New England BioLabs® Inc.) were also run with the PCR product to confirm correct amplification at the approximate expected size range.
4.4.1.2 RFLP RFLP was used to genotype polymorphisms in the GRIA1 gene. The first SNP rs548294 was digested using the enzyme MWO-I. This enzyme was obtained from New England Biolabs® Inc., Genesearch. The SNP coded at rs548294 is a G to A change which removes the MWO-I recognition site therefore allowing a genotype to be determined. After amplification 10.0 μl of PCR product was incubated in a reaction
96 Table 4-3 Primer sequences Marker Primers Amplicon Annealing Temp (°C) DXS8064 F: FAM - AGAATCGCTTGACCCTTG 3’ 208-214bp 57 R:CTGATGGCTGCCAACTC 3’ DXS1001 F: FAM - TGTACAAGTAACCCTCGTGACACG 3’ 376-386bp 62 R: TAGTGGCTGGCAGAGAGATTCC 3’ DXS1206 F: FAM - TGCCATAGGTAGTCATAGCATAGCC 3’ 277-291bp 62 R: CAGAGCATGGGACTTCTCAACC 3’ DXS984 F: NED - TGGAGGTCTGATTTAATGGCAGC 3’ 127-141bp 64 R: GCCCTACTCCATTCCACACTGG 3’ DXS8106 F: FAM - CTCCTTGCACTTGCTGTGG 3’ 387-405bp 61 R: TGCTTGCACCCTGTGAAGTC 3’ DXS8043 F: FAM - AGTTCTCAGAAACATTTGGTTAGGC 3’ 166-184bp 62 R: AATTATTGGCAAAGAGTACAGGCAG 3’ DXS297 F: NED - TTGGACTTCCCAAGCCTCCACAA 3' 191-197bp 65 R: TTCTGAGTCTGTGCAGTGTATTTGTCAG 3' DXS8091 F: FAM - CCACATTCAGGTTCCACAGGTACC 3’ 90-108bp 65 R: TGCAAGATCCAGGCAAAAGTCTC 3’ DXS1123 F: FAM - TGCCTAAATGTTCGCAAGCCCATTC 3’ 168-176bp 69 R: ACAAACAGCTGCCTCCTAGAAACCC 3’ DXS8061 F: VIC - GCAAGCTTGAAGTGTCCATGAGG 3’ 141-155bp 64 R: AGAAGCTGATGTGCTCCCTGC 3’ DXS15 F: PET - AGCACATGGTATAATGAACCTCCACG 3' 154-166bp 65 R: CAGTGTGAGTAGCATGCTAGCATTTG 3' DXS1073 F: VIC - TTGGGTGGAATTCCGTGACC 3’ 197-211bp 65 R: CCAAAGAATGCCCTCTCCGA 3’ DXS1108 F: PET - GGAGTGAATTCATCATATGTGATTTCC 3’ 168-184bp 63 R: ACTAGGCGACTAATACAGTGGTGCTC 3’ GPR50 F: VIC – GCCTGACTCTGTTCATTTCAAGCCT 3’ 196-208bp 65 R: CTTAGGGTGGCTGGTAGTGGCA 3’ GRIA3 F: FAM-CTAGTGTGGGGTGGAAAGGA3’ 341bp 58 rs10704237 R: GGGACCCACCGCAGGGAAAC3’ GRIA3 F: FAM-GACATGCAAGATTCCAGTATGC3’ 280-292bp 58 CA repeat R: GGAAATACCAAAAACATCTGGTC3’ GRIA1 F: AGATGAAGAAACAGAGGTC3’ 312bp 56 rs548294 R: CCCCAGGTACTATTCAAAG3’ GRIA1 F: TCTAAGAGGAGGGGGCAAGG3’ 367bp 60 rs2195450 R: GCTTGGTAGATGGTGCTTGA3’
Table 4-4 PCR protocol Reagent [Stock] [Final] 1X (μl) PCR Buffer 5x 1x 3.0 F Primer 5μM 100nM 0.3 R Primer 5μM 100nM 0.3 dNTPs 5mM 200μM 0.6 MgCl 25mM 1.75mM 1.05 DNA 20ng/ul 50ng 2.5 TAQ polymerase 5U/ul 0.5U 0.1 H2O 7.15 Total 15.0
97 mix with 0.08ul of MWOI enzyme (5000 u/ml), 2.0 μl of buffer 3 (New England
BioLabs® Inc.) and made up to 20.0 μl with H20 for a period of 4 hours at 60°C. Digestion product was then run on 3% agarose gels for manual genotyping. Expected band sizes were 312bp for the A allele, and 125bp and 187bp for the G allele.
Similarly the rs2195450 SNP in GRIA1 was also genotyped by RFLP using the enzyme TAQI also obtained from New England Biolabs® Inc., Genesearch. The SNP coded at rs2195450 is a C to T change which removes the TAQI recognition site therefore allowing a genotype to be determined. After amplification, 12.0 μl of PCR product was incubated with 0.21 μl of TAQ1 enzyme 2.4 μl of buffer 3 and 0.2 μl of BSA (New England BioLabs® Inc.) for a period of 2.5 hours at 65°C. The reaction was then heat inactivated at 80°c for 20 minutes as per manufacturers instructions. The digestion product was run on 3% agarose gels for manual genotyping. Expected band sizes were 367bp for the T allele, and 217bp and 150bp for the C allele.
4.4.1.3 Microsatellite All remaining polymorphisms genotyped in-house were completed by sizing through CE using the Applied Biosystems (Life Technologies, Australia) ABI 3130 Genetic Analyser and genotypes were called using the Applied Biosystems (Life Technologies, Australia) GENEMAPPER® software version 4.0. These included the microsatellite markers used in the pedigree study and 3 microsatellite markers (DXS8043, DXS8061 and the GRIA3 CA repeat) and 2 deletion polymorphisms (in GPR50 and GRIA3) in the migraine association population.
In order to reduce time and reagent costs microsatellites used in the pedigree analysis were run as 3 multiplex reactions. The markers that were multiplexed for each genotyping run are listed in Table 4-5. These multiplexes were developed ensuring that markers with overlapping amplicon sizes were labelled with different fluorescent labels; and that sufficient size differences were allowed where two markers with the same label were included in a single multiplex. 0.1 – 0.3 μl of each amplified marker (depending on the strength of the PCR) were added to the electrophoresis plate with 0.25 μl of 1:10 diluted GeneScanTM 500 LIZ® size standard. The final volume was then made up to 10 μl with HI-DI Formamide (Applied Biosystems).
98 Table 4-5 Microsatellite multiplexes Multiplex # Marker Label Expected Amplicon Observed Amplicons* size 1 DXS8091 FAM 92 - 114 90 102 - 108 DXS8043 FAM 155 -189 166 174 178 - 184 DXS15 PET 150 -166 150 - 166 DXS297 NED 179 - 198 187 - 197 DXS8061 VIC 129 - 157 141 143 147 -155 2 Amellogenin FAM 106 and 112 106 112 DXS1001 FAM 373 – 393 376 - 382 386 DXS1206 FAM 279 – 297 277 279 285 289 291 3 DXS1108 PET 168 – 180 162 - 168 176 - 184 DXS984 NED 152 – 176 127 137 - 141 DXS1123 FAM 166 – 178 166 168 172 - 178 DXS8064 VIC 129 – 157 208 - 212 214 DXS1073 VIC 195 – 221 197 - 201 211 213 DXS8106 FAM 346 - 412 387 - 391 399 401 405 *all markers are dinucleotide repeats
A raw data file for each sample is obtained from the ABI3130 Genetic Analyser showing fluorescent intensities of each marker in the sample as shown in Figure 4-16 and Figure 4-17. Stutter peaks were seen for all markers. These peaks are of smaller fragments size and lower intensity to the major allele peaks and represent the presence of polymerase skipping. This process occurs when the polymerase skips or deletes nucleotides within the microsatellite (Ziegle et al. 1992) during the PCR process. Consequently heterozygotes will always show greater signal intensity for the smaller allele as this represents both the allele and the stutter created from the larger allele. Figure 4-16 shows a raw data file from multiplex 1 and is a male sample, therefore all markers appear homozygous. Figure 4-17 is also from multiplex 1 however results are from a female sample and a number of markers are showing heterozygous genotypes.
The data is analysed using the GENEMAPPER software where fragments are sized by reference to the LIZ labelled size standard run with each sample. Bins of expected alleles sizes are setup within the software for each marker, and genotypes are assigned based on the peak size corresponding to a particular predetermined bin. Genotyping quality is automatically determined using various parameters that include the quality of the size standard, the presence of stutter peaks and the height ratio of allele peaks in heterozygote samples. The automatic genotype call was manually checked against the sized electrophoretogram created by the software and samples were rerun if the result
99 was not clear. Examples of the automatic genotype calling for markers DXS8043 and DXS1108 each for 4 samples are shown in Figures 4-18 and 4-19 respectively.
Figure 4-16 Electrophoretogram raw data male Example of a raw data file generated by genotyping using the ABI 3130 and obtained from the GENEMAPPER software. This figure shows results from a male sample with markers in Multiplex 1 (See Table 4-5)
Figure 4-17 Electrophoretogram raw data female Example of a raw data file generated by genotyping using the ABI 3130 and obtained from the GENEMAPPER software. This figure shows results from a female sample with markers in Multiplex 1 (See Table 4-5)
100
Figure 4-18 DXS8043 GENEMAPPER genotyping electrophoretograms Example of GENEMAPPER software automatic genotype calling of 4 independent samples for the DXS8043 dinucleatide marker.
Figure 4-19 DXS1108 GENEMAPPER genotyping electrophoretograms Example of GENEMAPPER software automatic genotype calling of 4 independent samples for the DXS1108 dinucleatide marker.
101 4.4.2 Outsourced Genotyping Methods 4.4.2.1 Illumina Genotyping of the X chromosome scan in the Norfolk population was conducted using the Illumina Infinium High Density (HD) Human 610-Quad DNA analysis BeadChip version 1 using 200ng of DNA per sample. Samples were scanned on the Illumina BeadArray 500GX Reader and the Illumina BeadScan image data acquisition software (version 2.3.0.13) was used to collect raw data. Preliminary analysis of raw data was undertaken in Illumina GenomeStudio software (V2010.1) with the recommended parameters for the Infinium assay. Genotype cluster files were generated based on clustering of genotypes within the Norfolk Island population and results were provided as genotype data in spreadsheet format.
4.4.2.2 MALDI-TOF MS Genotyping of a number of polymorphisms was performed at the Australian Genome Research Facility (AGRF) using the Sequenom MassArray on an Autoflex Spectrometer and iPLEX GOLD chemistry (Millis 2011). Polymorphisms were genotyped in 3 different multiplexes using the Australian migraine association population and results were provided as raw genotyping data in an excel spreadsheet.
The first multiplex consisted of a series of SNPs from various migraine susceptibility regions in the genome. Of relevance to this study were 2 pairs of SNPs each flanking the key microsatellite markers that were genotyped in the migraine association population. These SNPs included: rs5970389, rs6653488 located 5’ and 3’ of the DXS8061 Xq28 marker respectively, and rs3213562 and rs5920369 located 5’ and 3’ of the DXS8043 Xq27 marker respectively.
The second multiplex consisted of 29 polymorphisms covering 9 candidate genes as discussed in section 4.3.3.2. Finally the third multiplex included 28 SNPs that were identified as top ranking SNPs prioritised by P value in the Xq12, Xq27 and Xq28 regions in the Norfolk Island study for replication studies.
102 4.5 Analysis
Once genotype data had been obtained a variety of statistical techniques were employed for analysis. For all X chromosome markers raw data was checked for male heterozygotes and where possible these individuals were regenotyped or genotypes were excluded from the analysis as errors. Other quality control methods were employed using specific programs and are explained in detail below.
4.5.1 Input Files The bioinformatics programs used in this study share similar input files. This section will describe the basic components of each input file. Where a program varies in its inclusions or format of these files this will be explained with the application of the program in further detail below.
The PED file is the primary file used for most bioinformatics programs that were included in this study. The PED file describes the relationship of the individuals tested and generally also includes the phenotype of the trait of interest (migraine) and genotype data for all tested markers. The PED file is a text file without headings that consists of at least the first 6 columns which are described in Table 4-6
Table 4-6 PED file Column Description Notes 1 Family ID Unique if population unrelated 2 Individual ID Unique ID number 3 Paternal ID 0 = unknown/founder individual 4 Maternal ID 0 = unknown/founder individual 5 Sex 1 = Male, 2 = Female i.e. Migraine Status 0 = Unknown, 1 = No Migraine 2 – 6 Affectation/Phenotype Migraine 7 Marker 1 Genotype of marker 1, coding of genotypes varies between programs. 8 Marker 2 Males were always genotyped as homozygous as per the ... software instructions. 100 Marker 100 etc.
For association analysis where individuals are not related each individual was given a unique pedigree number and Paternal and Maternal IDs where set to unknown.
103 The MAP file is a second file commonly used in statistical software. This file lists names and the positions of genetic markers. Some MAP files require 4 columns including chromosome, SNP name, genetic distance (cM) and base-pair positions. Others only require SNP name and base-pair positions and the program assumes the map is autosomal unless a separate command is used to tell the program the markers are sex-linked.
The DAT file is third most commonly used file in these programs. This file describes the columns of the PED file that come after the standard first five columns describing the pedigree. For example if the PED file contains an affectation status in column 6 then the DAT file will list the name of the associated disease, similarly the subsequent list will name the order of markers that are have genotypes in the PED file. This is a simple form DAT file, an alternative more complicated DAT file will also list the frequencies of the alleles for each marker and will include the map file as well. This complicated format is described in more detail below.
A final file used in this study is the COVAR file. This file lists covariates to be included in a specific analysis. The file consists of at least 3 columns listing family ID and individual ID (that match the PED file) and are followed by columns listing specific covariates, including for example the age and kinship coefficients used in the Norfolk Island analysis.
4.5.2 Power The Australian migraine association population used in these studies consisted of 500 cases and 500 controls as stated in section 4.1.3. The power of this cohort to detect association was determined using the Genetic Power Calculator (Purcell et al. 2003). Input parameters are outlined in Table 4-7 below. It is estimated that the cohort has 83% power to detect allelic associations conferring an increased risk of 1.5 or greater, at the 0.05 significance level. It is noted however, that this analysis is an estimate as it assumes n=2000 alleles, which is not the case given that the majority of this research focuses on X chromosomal polymorphisms and the cohort includes male samples.
104 It is also noted that analysis has been conducted on subgroups – separated by gender or migraine subtype. These groups are smaller cohorts and consequently power to detect analysis is reduced to <80%.
Table 4-7 Power parameters Number of cases 500 Number of controls 500 High risk allele frequency (A) 0.1 Prevalence 0.12 Genotypic relative risk Aa 1.5 Genotypic relative risk AA 1.5 Genotypic risk for aa (baseline) 0.1096
4.5.3 Chi Square All SNPs genotyped in the migraine association population were analysed using contingency tables and a χ2 analysis. This analysis was completed in the software PLINK v1.07 which is a free whole genome association toolset (Purcell et al. 2007). For association analysis PLINK requires a PED file and the 4 column MAP input file as described above. Genotypes were included in the PED file and alleles were coded as A G C or T. Males were coded as homozygotes for X chromosome polymorphisms. PLINK will recognise the input data as X chromosomal from the MAP files and therefore will identify any heterozygote males and set these to missing genotypes. Similarly when determining allele counts PLINK will read a homozygous male coded as e.g. AA as a single A allele whereas the corresponding female genotype is read as 2 A alleles.
Association analysis was run in PLINK using the command line:
>plink --ped x500.ped --map x500.map --assoc --ci 0.95
Contingency tables were generated through adding the --counts command which provide allele counts rather then frequencies in both the case and control cohorts. The --assoc command generates a plink.assoc output file that contains the results of the analysis under the following column headings:
105 CHR Chromosome SNP SNP ID BP Physical position (base-pair) A1 Minor allele name (based on whole sample) F_A Frequency of this allele in cases F_U Frequency of this allele in controls A2 Major allele name CHISQ Basic allelic test chi-square (1df) P Asymptotic p-value for this test OR Estimated odds ratio (for A1, i.e. A2 is reference) SE Standard Error L95 Lower bound of 95% confidence interval for odds ratio U95 Upper bound of 95% confidence interval for odds ratio
The significance level for each study analysed through this approach was adjusted for multiple testing using a Bonferroni correction (see section 4.5.1.2 below).
4.5.3.1 HWE Hardy-Weinberg equilibrium (HWE) calculations were a control measure used in association studies presented in this thesis to gauge the possibility of genotyping errors. The HWE calculation determines expected genotype frequencies and therefore can be used to indicate if there has been a deviation from these frequencies (Painter et al.). While Hardy-Weinberg deviations can be caused by a number of factors the samples used in this study were selected to be unrelated, random mating individuals, from a Caucasian background and collected from a similar geographic area. This was to reduce the effects that migration, ethnic variances and related individuals may have on causing Hardy-Weinberg deviations. Therefore any observed deviations may be attributable to genotyping error. The HWE formula is as follows:
p2 + 2pq +q2 = 1
Where p is the frequency of the ‘A’ allele and q similarly corresponds to the frequency of the ‘a’ allele. For X chromosome studies this analysis was undertaken in the female control group cohort. Analysis was restricted to controls as it is expected that allele frequencies in the case group will differ from controls where association is observed and therefore this may be the cause of deviation from HWE.
106 4.5.3.2 Significance level The significance level for association tests in this study was initially set at a threshold of P = 0.05 i.e. 1 in 20 chance that rejecting the null hypothesis (no association) is incorrect due to chance (Type I error). This level was then adjusted for multiple testing using a Bonferroni correction. This approach divides the P value by the number of tests in order to adjust the significance threshold as the chance of incorrectly rejecting the null hypothesis increases when multiple comparisons are performed on the same data set (Bland and Altman 1995; Painter et al. 2011).
4.5.4 CLUMP The microsatellites DXS8043 at Xq27 and DXS8061 at Xq28 were genotyped in the migraine association population as described previously. As these markers are dinucleotide repeats there is significant allelic heterogeneity in the population and therefore a number of genotypes and alleles were observed including many at low frequencies. Therefore χ2 analysis could not be used as the statistics become unreliable when an observed count is less then 5. In this case the CLUMP program (Sham and Curtis 1995) was therefore employed. This program uses a Monte Carlo approach to test significance by performing simulations. The program produces 4 statistics T1-T4. T1 is the raw data, T2 is created by grouping columns of small frequencies together, T3 is determined by comparing each column against the total of all other columns and finally T4 is a 2 x 2 contingency table that tests whether certain alleles are more common in cases then controls. This is achieved through clumping of allele groups that are more common then expected in the one group and creating a second group where the second allele is higher (Sham and Curtis 1995).
For the analysis of the DXS8043 and DXS8061 markers 1000 simulations were run and the T1 statistics have been reported. Input files are unique to this program and are text files consisting of 5 lines:
Line 1: Number of alleles for the marker being tested Line 2: Allele counts for cases separated by spaces Line 3: Allele counts for controls separated by spaces Line 4: Number of simulations Line 5: A value to seed the random number generator used by CLUMP.
107 Input files were created for DXS8043 and DXS8061 and were run using the commands:
>clump DXS8043_total.inp DXS8043_total.out and >clump DXS8061_total.inp DXS8061_total.out
Example input files and resulting outputs are shown for the CLUMP analysis of total migraine and DXS8043/DXS8061 in Figures 4-20 and 4-21 respectively.
4.5.5 Linkage Analysis Analysis of the pedigrees required a number of steps involving checking the input files were accurate and described the pedigrees correctly, checking for mendelian errors in the genotyping, analysis of the raw data and determination of haplotypes. For all linkage analyses affectation status was either migraine affected (combining MA and MO), unaffected or unknown. PED files for all families have been included as Appendix E. These files were used for all analysis explained in this section.
4.5.5.1 Error Checking Pedstats is a program that will check that the PED and DAT files are being read together properly (Wigginton and Abecasis 2005). Pedstats will check for the number of pedigrees in the PED file, determine any errors in pedigree structure e.g. if any unrelated individuals have been included, and will list any inconsistencies in mendelian inheritance. All PED and DAT files were run through Pedstats prior to analysis.
Pedstats was run using the command:
>Pedstats –p x.ped –d MINX_8MF.DAT --chromosomeX
The pedigree analysis program MERLIN was also used for error detection (Abecasis et al. 2002). This analysis will flag mendelian inheritance issues as well as unlikely genotypes based on gene flow in the pedigree. MINX is the X-linked version of MERLIN and was used for all analysis. Error detection in MINX is run using the following commands:
108
Figure 4-20 CLUMP DXS8043 files Figure 4-21 CLUMP DXS8061 files Example input files are shown in top 6 lines of figures 4-20 and 4-21, followed by output generated in the CLUMP software
109 >minx –d x.ped –d MINX_8MF.DAT –m MINX_8MF.MAP –error
The corrected PED files are shown in Appendix E. The marker positions were taken from Duffy’s interpolated Rutger’s Map (Duffy 2006) and the final MAP and DAT files were as follows:
MINX_8MF.DAT A migraine M DXS1206 M DXS984 M DXS8106 M DXS8043 M DXS297 M DXS8091 M DXS1123 M DXS8061 M DXS15 M DXS1073 M DXS1108
MINX_8MF.MAP CHROMOSOME MARKER POSITION X DXS1206 0.0 X DXS984 13.81 X DXS8106 23.39 X DXS8043 28.6 X DXS297 32.8 X DXS8091 35.55 X DXS1123 37.22 X DXS8061 47.68 X DXS15 48.24 X DXS1073 51.09 X DXS1108 51.57 Unlikely genotypes are listed in a file named merlin.err and were manually checked in the pedigrees to determine if these were data input errors or true genotyping inconsistencies that were then excluded from the analysis.
4.5.5.2 GENEHUNTER-PLUS and ASM Pedigree analysis was undertaken in the program GENEHUNTER-PLUS - a modified version of GENEHUNTER 1.3 (version 1.2) (Kruglyak et al. 1996), which includes the X-linked version of the program. GENEHUNTER-PLUS (Kong and Cox 1997) was chosen over other pedigree analysis programs such as MERLIN for consistency with previous studies.
110 The analysis used a non-parametric multipoint linkage analysis employing the Sall scoring function and an exponential model. A non-parametric analysis was undertaken due to difficulties in defining the mode of inheritance in the pedigrees. The Sall statistic measures identity by descent (IBD) allele sharing among all individuals simultaneously in contrast to the Spairs statistic that considers pairs of affected relatives. Finally an exponential model was preferred over a linear model to detect large increases in allele sharing in a small number of families. The auxiliary program Allele Sharing Models (ASM) is used to for the exponential model when calculating the LOD scores (Kong and Cox 1997).
The analysis was undertaken using the following commands:
>xghp.solaris >photo x.out >run XGHP_11.in >scan pedigree x.PED
The ‘photo’ command creates a text file of the analysis as it runs to be saved as a record for reference. The ‘run’ command allows a series of commands to be loaded from a text file (XGHP_11.in); this helps to ensure the same analysis is run across all pedigrees. The XGHP_11.in file included the following commands:
Load markers XGHP_8MF_all.DAT discard OFF single point OFF score ALL analysis NPL max bits 20
The DAT file used for analysis in this program is the extended version. This version requires the input of allele frequencies for each marker as well as marker positions (shown as recombination fractions). Allele frequencies were determined using MINX and the files listed previously for error checking. The following command was used and frequencies were determined using all individuals.
>minx –d x.ped –d MINX_8MF.DAT –m MINX_8MF.MAP --frequencies
Figure 4-22 shows the final DAT file that was used for analysis of all pedigrees in GENEHUNTER-PLUS. An additional advantage of using GENEHUNTER-PLUS
111 over other programs is allowance in the DAT file for penetrance information. This is crucial to complex diseases where different genetic factors may play a role. In addition, this allowed the use of families such as MF14 and MF879 where male to male transmission is observed which is not strictly consistent with an X-linked theory, but may represent incomplete penetrance or sporadic cases. Thus the linkage analysis used in this study allowed for incomplete penetrance as shown in lines 7 and 8 of the DAT file (Figure 4-22).
Figure 4-22 GENEHUNTER-PLUS DAT file
The Map used for the linkage studies is in the second last line of this file and corresponds to the same MAP positions as are specified in the MINX_8MF.map file previously shown.
To fit the exponential model the ASM program was applied. This is done following the scan completed in GENEHUNTER-PLUS and is activated with the command:
>asm EXP
112 This command finds the nullprobs.dat and probs.dat output files from GENEHUNTER-PLUS and determines LOD scores based on the specified model. The output file has 5 columns which specify the following:
Column 1: Position (cM) Column 2: Weighted NPL score Column 3: Zlr (Zlr = sign (dhat)*sqrt(2.0*ln(10.0)*LOD)) Column 4: Maximised LOD* for the allele-sharing model selected (LOD) Column 5: delta which produced the maximised lodscore
In addition to determining linkage for each family individually, combined (global) analysis was also conducted. The global analysis was run on 4 groups of families:
1. all eight migraine families 2. MF7, MF17, MF47 and MF879 as these families showed linkage to either the Xq27 or Xq28 loci 3. MF7, MF14 and MF879 as these families showed linkage to the Xq27 locus 4. MF7, MF14 and MF47 as these families showed linkage to the Xq28 locus
This analysis required all families to be combined in a single PED file and the ‘total scan’ command was run after the ‘scan pedigree x.ped’ analysis was complete. Subsequently the ASM program was invoked as specified above. The ASM outputs from all analysis are included in Appendix F and Appendix G. To assist interpretation of results graphs of the maximised LOD scores were created using Microsoft Excel.
4.5.6 Pedigree-based Association Incorporating Logistic Regression The Norfolk Island pedigree is extremely complex with many consanguineous marriages and inbreeding loops. While this may increase the likelihood of detecting susceptibility variants through a reduction in genetic diversity, it creates significant problems with analysis. Many pedigree analysis programs will not handle such a complex structure for a straight linkage based analysis. This project had an additional hurdle in that the hemizygous nature of the X chromosome in males requires a program to handle male and female samples differently during analysis. Therefore the analysis of the X chromosome scan in the Norfolk Island population was limited in the options and available software for analysis of this data.
113 The approach used employed a scan of the X chromosome data using a logistic regression analysis adjusted for age, sex and relatedness of the cohort. The relatedness was measured by a kinship coefficient calculated for each member of the pedigree, which was estimated by IBD-matrices, generated using the SOLAR program (Almasy and Blangero 1998). This approach used an approximation of the relatedness in order to highlight associations that could then be followed up in candidate gene studies.
The software PLINK was used to run the logistic regression analysis. A PED file as described previously was used to input the genotype data for 14,124 SNPs and the migraine status of the 288 genotyped individuals. This file was created in the same way as a file used for a typical association study i.e. all individuals were entered as if they were unrelated. Similarly a 4 column map file listing the SNPs and chromosomal positions of all SNPs was created as previously described. Finally a COVAR file listing all genotyped individuals and their age and predetermined kinship coefficient was created.
The logistic regression was performed in PLINK using the command
>plink --ped NI_X.PED --map NI_X.MAP --covar NI_X_Covar.txt --logistic --ci 0.95
This creates an output file plink.assoc.logistic that includes the following information listed in columns:
CHR Chromosome SNP SNP identifier BP Physical position (base-pair) A1 Tested allele (minor allele by default) TEST Code for the test (see below) NMISS Number of non-missing individuals included in analysis BETA/OR Regression coefficient (--linear) or odds ratio (--logistic) STAT Coefficient t-statistic P Asymptotic p-value for t-statistic L95 Lower bound of 95% confidence interval for odds ratio U95 Upper bound of 95% confidence interval for odds ratio
For each SNP tested under an allelic model PLINK returned 4 lines of data detailing firstly the test of the SNP while controlling for the three covariates age, sex and relatedness and secondly the effect of each covariate. Adding the command ‘--hide-
114 covar,’ returns only the analysis of the SNP while controlling for the covariates. Importantly PLINK will set all heterozygote male genotypes to missing during the analysis.
This output file was then read into WGAviewer v1.26 for creation of a Manhattan plot and annotation of the top prioritised 100 SNPs using Ensembl.
4.5.6.1 SOLAR The bioinformatic software program SOLAR (Sequential Oligogenic Linkage Analysis Routines) (Almasy and Blangero 1998) was used to perform an exact analysis of the top associated SNPs identified in the Norfolk Island population after the preliminary pedigree-based association using logistic regression based scan had been performed. This approach allowed the results to be adjusted using an exact analysis of the pedigree rather then the approximate approach used in the initial full chromosome scan.
The analysis undertaken in SOLAR was a variance components-based linkage model to determine the polygenic heritability of migraine and the proportion of the variance caused by the covariates. In this case, each SNP was coded as a covariate and included in the tested model with age and sex. SNPs were coded 0, 1, 2 for female genotypes AA, AB, BB respectively. Males were coded 0, 2 for genotypes A and B respectively.
The files used for the SOLAR program vary slightly from those previously described. The PED file contains only 4 columns – id, paternal id, maternal id and sex. A phenotype (PHEN) then describes the covariates age, migraine diagnosis (mamo) and each SNP recoded as above. Files are csv format and where genotypes or phenotypes were unknown samples were left blank instead of using ‘0’ or ‘x’ as may be applied in other programs.
The following commands were entered sequentially to run this analysis in SOLAR.
115 >load pedigree 1078_NI.ped >load phenotypes NI_TOP25.phen >model new >trait mamo >covar age sex snp_rsXXX >polygenic -screen
Upon completion of this analysis a P value of the effect of the SNP on the heritability is obtained. Additional commands such as >parameter bage and >parameter bage se will provide the beta coefficient for the relevant covariate and the standard error.
4.5.7 Haplotype Analysis Haplotype analysis was used to identify risk haplotypes segregating with migraine at Xq28 in the migraine association population and at Xq12 and Xq27 in the Norfolk Island population. In both cases it was observed that a series of consecutive SNPs at specific loci were showing suggestive association. The respective PED files for these studies were loaded into Haploview to look for variations in the regions of LD between the groups (Barrett et al. 2005). However due to the high number of SNPs genotyped in the Norfolk population the input files were restricted to 42 SNPs and 28 SNPs in the Xq12 and Xq27 regions respectively. These SNPs correspond to the clustering of suggestive SNPs seen in the association analysis.
Haplotype analysis at the Xq28 region used an association analysis implemented in Haploview v4.2 (Barrett et al. 2005). For the Xq27 and Xq12 regions Haploview was used only to determine the length of the LD blocks. The files were then analysed in PLINK through a haplotype based association test using logistic regression so that the covariates of age, sex and relatedness could be adjusted for. The command used in PLINK was as follows:
>plink --ped NI_X.PED --map NI_X.MAP --covar NI_X_Covar.txt –hap-logistic -- hap-window x --ci 0.95
Where –hap-window specifies the number of SNPs to include within the haplotype block as had been predetermined based on the LD analysis from Haploview. The PLINK output consists of a file with the following headings.
116 NSNP Number of SNPs in this haplotype NHAP Number of common haplotypes (threshold determined by --mhf, 0.01 default) CHR Chromosome code BP1 Physical position of left-most (5') SNP (base-pair) BP2 Physical position of right-most (3') SNP (base-pair) SNP1 SNP ID of left-most (5') SNP SNP2 SNP ID of left-most (3') SNP HAPLOTYPE Haplotype F Frequency in sample OR Estimated odds ratio STAT Test statistic (T from Wald test) P Asymptotic p-value
This row is repeated for all haplotypes present in the population for the haplotype window specified.
4.6 Methods Conclusions
This study investigated the X chromosome in relation to migraine susceptibility. As outlined in this chapter a number of migraine pedigrees including the Norfolk pedigree and migraine families were used to identify and refine migraine susceptibility regions. In particular linkage analysis was employed to identify regions of excess allele sharing in migraine families and haplotype analysis allowed these regions to be refined to narrow regions that were common to a number of families. Association studies in case-control cohorts were used to determine if the susceptibility regions identified in these pedigrees translated to the general population. Finally key SNPs within genes coded in these regions were analysed in migraine cohorts using an association study approach. The following chapters outline the results of these studies with chapter five describing the analysis of the Xq27 and Xq28 susceptibility regions. This includes the analysis of the migraine families using a linkage study approach, the investigation of key microsatellite markers in the general migraine association population and finally the genotyping and analysis of SNPs within key genes located in the Xq27 and Xq28 regions.
Chapter six describes the analysis of the candidate gene GRIA3 located at Xq24 that was conducted using an association study method. Chapter seven details the X
117 chromosome scan conducted on the Norfolk Island pedigree using the multi-step approach incorporating logistic regression for the initial scan followed by the analysis in SOLAR that accounts for the pedigree structure exactly. Finally results chapter eight describes the analysis of the SNPs identified in the Norfolk study (chapter seven) in the Caucasian migraine association population.
The final chapter summarises these results and discusses the significance of the results in relation to migraine molecular genetics.
118 CHAPTER FIVE: Investigation of the Xq27 and Xq28 Susceptibility Regions
Migraine is a complex neurological disorder that is characterised by debilitating headache and temporary neurological disturbances. Investigations into migraine genetics have also shown that male probands have a higher proportion of affected first degree relatives and migraine has a female preponderance of 3:1. These findings collectively indicate that an X-linked dominant form of inheritance may be playing a role in migraine.
We have previously implicated the Xq24-28 locus in migraine using 2 independent multigenerational Australian pedigrees that demonstrated excess allele sharing at the Xq24, Xq27 and Xq28 loci. Here, we expand this work to investigate a further 6 independent migraine pedigrees using 11 microsatellite markers spanning the Xq27- 28 region. Furthermore, 11 candidate genes are investigated in an Australian case- control cohort consisting of 500 cases and 500 controls.
Microsatellite analysis showed evidence of excess allele sharing to the Xq27 marker DXS8043 (LOD* 1.38 p=0.005) in MF879 whilst a second independent pedigree showed excess allele sharing to DXS8061 at Xq28 (LOD* 1.5 p = 0.004). Furthermore, analysis of these key markers in a case-control cohort showed significant association to migraine in females at the DXS8043 marker (T1 P=0.009) and association with MO at DXS8061 (T1 P=0.05). Further analysis of 11 key genes across these regions showed significant association of a 3 marker risk haplotype in the NSDHL gene at Xq28 (P=0.0082).
The results of this study, combined with previous evidence supporting an X-linked mode of inheritance suggest that susceptibility loci influencing migraine exist independently on Xq27 and Xq28 and SNP haplotype analyses indicate the Xq28 causative factor is relatively common in the general population. The NSDHL gene should be further investigated to elucidate their functional implications in this disorder.
119 5.1 Introduction
Migraine is a common episodic disorder displaying a complex aetiology which is often inherited in families. The disorder affects approximately 18% of adult females and 6% of adult males globally (Stovner et al. 2006; Lipton et al. 2007). Migraine is characterised by recurrent headache and neurological disturbances such as photophobia and phonophobia. Numerous studies have demonstrated the genetic contribution to migraine through pedigree and twin studies with heritability estimates ranging between 34 and 57% (Mulder et al. 2003).
Genetic studies have identified three ion channel genes that independently cause the rare migraine subtype Familial Hemiplegic Migraine (FHM) (Ophoff et al. 1996; De Fusco et al. 2003; Dichgans et al. 2005). However, identification of genes contributing to the common forms of migraine; MA and MO, have been more difficult. Genes that have been implicated in MA and/or MO include hormone receptors (Estrogen (Colson et al. 2006a) and Progesterone (Colson et al. 2005)), serotonin genes (Ogilvie et al. 1998; Erdal et al. 2001; Yilmaz et al. 2001), DBH (Lea et al. 2000; Fernandez et al. 2009) and the KCNK18 gene encoding the TRESK potassium channel (Lafreniere et al. 2010) among others. Despite these efforts by our group and others, it is still unknown how many genes may be interacting and contributing to common migraine pathogenesis. The focus in present genetic research largely centres on neurological, vascular or hormonal genes based on the current understanding of migraine pathophysiology.
Linkage studies have also been used to identify a number of migraine susceptibility loci on various chromosomes. Results from our laboratory have implicated 1q31 (Lea et al. 2002); 3qter, 18p11 (Lea et al. 2005); 19p13 (Nyholt et al. 1998b) and the Xq24- 28 locus (Nyholt et al. 1998a; Nyholt et al. 2000) in Australian migraine pedigrees. Others have also identified 2 independent loci on chromosomes 4; 4q21 using Icelandic pedigrees (Bjornsson et al. 2003) and 4q24 in Finnish pedigrees (Wessman et al. 2002); as well as 5q21 using studies of Australian twins (Nyholt et al. 2005); 6p12.2-21.1 in a Swedish pedigree (Carlsson et al. 2002) and 15q11-q13 (Russo et al. 2005). Susceptibility loci have also been identified on 10q22-23 (Nyholt et al. 2005;
120 Anttila et al. 2006; Anttila et al. 2008), 18q12 (Bjornsson et al. 2003; Anttila et al. 2006; Anttila et al. 2008) and Xp22 (Anttila et al. 2008; Wieser et al. 2010) using traditional migraine classification methods as well as TCA or LCA. However, the majority of these susceptibility loci have not been confirmed in independent populations and many are yet to have susceptibility genes identified within them. This lack of consensus in identification of susceptibility loci and genes highlights the likely heterogeneity of this disorder.
Due to the genetic heterogeneity the mode of transmission of migraine remains largely unclear. The biased preponderance of migraine in women, as well as studies that demonstrate that a high proportion of affected males have a greater number of affected first degree relatives (Stewart et al. 1997) suggests an X-linked dominant form of migraine inheritance. Families transmitting migraine through this mode of inheritance would show father to daughter transmission and an absence of father to son transmission. Therefore an excess of affected females is likely to be observed within these families.
The X chromosome has been previously identified by our laboratory as potentially harbouring a susceptibility locus at Xq24-28 (Nyholt et al. 1998b; Nyholt et al. 2000). Studies of two Australian pedigrees identified three specific loci with excess allele sharing in the region. The first was at Xq24 between markers DXS1001 and DXS1206, a second at Xq27 between markers DXS984-DXS1123, and the third at Xq28 between marker DXS8091 and Xqter. The Xq24 locus contains the candidate gene GRIA3 which has been investigated and results described elsewhere (Formicola et al. 2010). In this study we sought to further investigate the Xq27 and Xq28 locus through studies of six new independent Australian migraine pedigrees as well as candidate gene studies in an Australian case-control cohort.
121 5.2 Materials and Methods
5.2.1 Pedigree Study 5.2.1.1 Pedigree Collection and Phenotyping This study examined eight Australian Caucasian pedigrees, six new pedigrees designated MF6, MF47, MF55, MF541, MF878 and MF879, as well as the previously reported pedigrees MF7 and MF14 (Nyholt et al. 1998a; Nyholt et al. 2000). These pedigrees include 180 individuals including 125 for whom DNA was available. Ethical clearance was granted prior to the commencement of the study by the Griffith University Human Research Ethics Committee. All participants signed informed consent statements prior to inclusion in the study and were interviewed by a clinical neurologist. Individuals were diagnosed for migraine (MA or MO) according to ICHD-II criteria (IHS 2004). A questionnaire was used to obtain specific information from study participants on a number of relevant factors including age of onset, frequency, duration, associated symptoms, medication response and triggers in accordance with IHS guidelines. A total of 96 individuals were classified as affected. Venous blood was collected from participants and DNA was isolated from lymphocytes using a standard salting out procedure (Miller S.A et al. 1988).
5.2.1.2 Pedigree Genotyping Eleven dinucleotide microsatellite markers were tested in this study. Nine of the 11 markers were examined in the initial study to identify this locus and were therefore repeated in the new pedigrees for consistency. The additional 2 markers, DXS8043 and DXS1073, were used to aid in finemapping the identified region.
Marker order and positions were acquired from Duffy’s interpolated Rutgers Map (Duffy 2006) with markers spaced approximately 5.5cM apart spanning a 66cM region. Marker names and cumulative positions in Kosambi centimorgans (cM) are:
DXS1206 (0.0) - DXS984 (13.8) - DXS8106 (23.39) - DXS8043 (28.6) - DXS297 (32.8) - DXS8091 (35.54) - DXS1123 (37.22) - DXS8061 (47.68) - DXS15 (48.24) - DXS1073 (51.09) - DXS1108 (51.57)
122 Microsatellites were amplified using standard PCR with 5’ labelled forward primers where the labels were either FAM, PET, NED or VIC. Tables 4-3 and 4-4 in section 4.4.1.1 list primer sequences and PCR conditions. Where possible, due to overlapping fragment size and limited labels, PCR products were multiplexed (as outlined in Table 4-5) before sizing by capillary electrophoresis using the Applied Biosystems (Life Technologies, Australia) ABI 3130 Genetic Analyser. Genotypes were called using the Applied Biosystems (Life Technologies, Australia) GENEMAPPER® software Version 4.0.
5.2.1.3 Statistical Analysis of Pedigrees Pedigrees were selected as there was no father to son transmission of migraine observed, except in the case of individual III:5 to IV:6 in MF14 and individual I:1 to II:3 in MF879. These cases of male to male transmission may be a result of sporadic cases. Alternatively, due to the high prevalence of migraine in the population, it is possible that parents married into the family have contributed new susceptibility variants coded in other regions of the genome, as would beconsistent with a heterogeneous polygenic basis for migraine. To allow for incomplete penetrance and sporadic cases, penetrance values in the DAT file used for linkage analysis was set to 0.007, 0.7, 0.7 in females and to 0.007, 0.7 in males (see Figure 4-22).
Prior to statistical analysis genotype data was screened for mendelian and relationship inconsistencies using Pedcheck (O'Connell and Weeks 1998). Error checking was also performed in the MINX program and unlikely genotypes were resolved or removed. The Kong and Cox GENEHUNTER PLUS (X-Linkage version 1.2) statistical software (Kong and Cox 1997) was used to carry out non-parametric multipoint linkage analysis of these pedigrees using the Sall scoring function. Non-parametric analysis was chosen to overcome difficulties in defining unknown inheritance parameters. The Sall statistic was chosen as it measures IBD allele sharing among all individuals simultaneously rather then pairs of affected relatives therefore assigning a higher score as more individuals share the same allele by descent. The nonparametric LOD scores (LOD*) were determined implementing the exponential model. Previously reported pedigrees MF7 and MF14 were reanalysed with additional markers and an updated marker map. Linkage peaks yielding LOD* scores ≥ 1 are suggestive of linkage for X chromosome markers whilst LOD* scores ≥ 2 indicate
123 significant linkage. LOD scores were converted to P values using the ConvertingLODtoPvalue spreadsheet (Nyholt 2000). Haplotypes were determined from genotype and pedigree information using the X-linked version of Merlin, Minx (Abecasis et al. 2002) and diagrams were created using the Haplopainter program (Thiele and Nurnberg 2005) .
5.2.2 Association study 5.2.2.1 Population Collection and Phenotyping The study protocol was approved by Griffith University’s ethics committee for experimentation on humans. All individuals were of Caucasian origin and gave informed consent before participating in the research. Migraineurs were diagnosed as either MA or MO, based strictly on criteria specified by the IHS. All individuals were grouped together and phenotyped as being affected with typical migraine, as well as being diagnosed separately as MA or MO subgroups. The case-control study cohort was comprised of 500 migraineurs (360 MA/140 MO; 393 Female/107 Male) and 500 unrelated control individuals. To minimise potential bias from population stratification, the control group was matched for sex, age (+/- 5 years) and ethnicity (Caucasian origin) to the migraineurs.
5.2.2.2 Microsatellite Markers Two key microsatellite markers identified in the Xq27 (DXS8043) and Xq28 (DXS8061) loci were genotyped in the association study population to determine if the association was pedigree specific or could translate to the general population. Genotyping was carried out using the same process described for the pedigree study.
5.2.2.3 Candidate Gene Studies Gene maps of the Xq27 and Xq28 regions were obtained from NCBI (Homo sapiens Genome Build 37.2) and public databases e.g. Pubmed, were interrogated for functional gene information in order to select a final list of 2 genes in the Xq27 locus and 9 genes in the Xq28 locus (Appendix D).
HapMap CEU data was downloaded for each candidate gene and the tagger algorithm in Haploview (Barrett 2009) was used to identify 29 SNPs across 9 of the 12 genes
124 that captured the maximum degree of variation attainable while being constrained by multiplexing genotyping techniques. The SNPs were genotyped at Australian Genome Research Facility (AGRF) using the Sequenom MassArray on an Autoflex Spectrometer and iPLEX GOLD chemistry
Five additional polymorphisms were genotyped across the 2 regions. One is an insertion/deletion polymorphism in GPR50 amplified using standard PCR with 5’ FAM labelled forward primer (see Table 4-3 and 4-4 in section 4.4.1.1). Genotypes were determined by capillary electrophoresis using the Applied Biosystems (Life Technologies, Australia) ABI 3130 Genetic Analyser. Genotypes were called using the Applied Biosystems (Life Technologies, Australia) GENEMAPPER® software Version 4.0. The remaining 4 additional SNPs flanked the microsatellite markers tested in the case-control cohort and were also genotyped at AGRF as specified above.
5.2.2.4 Statistical Analysis of Case-control Cohort Microsatellites were tested for association using the CLUMP analysis program (Sham and Curtis 1995). To detect association between the SNP markers and migraine, we performed chi-square (χ²) analysis to test for significant differences in allele and genotype frequencies (females only for associated SNPs) in case versus control results. χ² provides the likelihood of a deviation in the distribution of the same attributes in different classes (e.g. allelic frequencies in controls versus affected subjects). Power calculations determined that this cohort has ~80% power to detect allelic associations conferring an increased risk of 1.5 or greater at the 0.05 significance level (see section 4.5.2). However, it should be noted that power to detect association is reduced to <80% in subgroups divided by gender or migraine subtype. Hardy-Weinberg Equilibrium (HWE) for females control genotypes was also calculated using PLINK (Purcell et al. 2007). For the candidate gene study locus- specific correction set the significance level at 0.008 for Xq27 and 0.001 for Xq28. Haplotype frequency estimates and analysis of the haplotype block consisting of markers rs5970389, rs6653488 and rs2071256 in the MO and control cohorts were performed using Haploview (Barrett et al. 2005).
125 5.3 Results
The current study presents an analysis of the previously identified migraine susceptibility locus, Xq27 - Xq28, using six new multigenerational Australian migraine pedigrees. In addition two of these markers were also analysed in a case- control migraine cohort to investigate this signal in a general migraine population. Finally, this study also interrogates a number of candidate genes within the identified susceptibility region.
5.3.1 Pedigree Analysis We reported previously the identification of a migraine susceptibility locus on chromosome Xq24-28 (Nyholt et al. 2000). Significant excess allele sharing was indicated by nonparametric linkage analysis producing maximum global LOD* scores of 2.31 between DXS8106 and DXS091; and LOD* 2.18 at DXS8061.
For the present study, we included new and combined analysis of the two previously published pedigrees and six new migraine pedigrees confirming the excess allele- sharing across this region. Analysis was performed in the pedigrees with individuals classed as affected, unaffected or unknown, combining the MA and MO phenotypes as a spectrum of the migraine disorder (See Appendix F for raw results). An analysis was also undertaken with individuals classed as either unknown or affected in case migraine had not manifested in the individual due to a lack of environmental triggers. However, this analysis did not significantly vary the outcome and therefore results are not included.
Analysis of the pedigrees verified and refined the previously localised susceptibility regions. Migraine pedigrees MF6, MF55, MF541 and MF878 did not show evidence of excess allele sharing to the Xq27-28 region. However, analysis of the remaining migraine families confirmed the previous observation at Xq27 and Xq28 (Figure 5-1). At the Xq27 loci maximum LOD* scores of 1.32 (P=0.0068), 1.35 (P=0.0063) and 1.38 (P=0.0058) were identified for MF7, MF14 and MF879 respectively. Similarly at the Xq28 locus maximum LOD* scores of 1.33-1.50 spanned the locus from DXS8061 to the final marker DXS1108 in MF7, MF14 and MF47 (Figure 5-1).
126
DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 1.6
1.4
1.2
1
0.8
0.6 LOD*score 0.4
0.2
0 0 5 10 15 20 25 30 35 40 45 50 Position (cM) MF6 MF47 MF55 MF541 MF878 MF879 MF7 MF14 Figure 5-1 Independent analysis of migraine families Results of non-parametric GENEHUNTER-PLUS analysis (LOD*) of each migraine family across the Xq27 and Xq28 susceptibility loci.
An initial global scan of all migraine families included in the study was suggestive of linkage across Xq27 and Xq28 with maximum LOD* scores peaking at 1.63 (P=0.003) at DXS8043 (Figure 5-2, See Appendix G for raw results). To determine the effect the 4 families that did not show excess allele sharing to the Xq27 or Xq28 region were having on this result a further analysis was conducted removing these families (Figure 5-2, See Appendix G for raw results).
Previous analysis of MF7 and MF14 has also shown evidence to support genetic heterogeneity of the disorder in these families (Nyholt et al. 2000) potentially suggesting that 2 independent loci exist in this region. Analysis of the new migraine families supports this theory as the individual analysis of MF879 only shows excess allele sharing at Xq27 and conversely MF47 only shows excess allele sharing at Xq28. A global analysis was therefore performed on each region using only those pedigrees that implicated the respective loci in order to refine the linkage regions for further analysis (See Appendix G for raw results). This analysis revealed a maximum global LOD score of 3.69 (P=1.88x10-5) between markers DXS8043 and DXS297 at Xq27 for the combined analysis of MF7, MF14 and MF879. At Xq28 the combined analysis of MF7, MF14 and MF47 showed a LOD* of 3.65 (P=2.07X10-5) at DXS8061. However it should be reiterated that this analysis has been employed solely for the purpose of refining the linkage regions, rather then to increase the statistical
127 significance of the linkage peaks which would require heterogeneity analysis across all 8 families prior to conducting a global analysis.
DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 4
3.5
3
2.5
2 LOD*score
1.5
1
0.5
0 0 5 10 15 20 25 30 35 40 45 50 Position (cM) 4 Migraine Families 8 Migraine Families Figure 5-2 Global analysis of migraine families Results of non-parametric GENEHUNTER-PLUS analysis (LOD*) of combined migraine families across the Xq27 and Xq28 susceptibility loci.
Haplotype analysis confirmed a narrowed Xq27 region from the previous study through critical recombination events in the new MF879 pedigree. Individuals IV:4 and IV:5 in MF879 show recombination at DXS8043xDXS297, the markers that bound the Xq27 locus, suggesting that a susceptibility candidate in this region is between these markers (Figure 5-3). However, it is noted that individual IV:4 is a double recombinant around marker DXS297 suggesting some caution in interpreting this results. Genotyping in this individual was therefore confirmed by repeat genotyping to ensure accuracy of the genotype calling.
The new pedigree MF47 was also shown to display excess allele sharing to the Xq28 region, however, haplotype analysis did not narrow the region that was identified previously (Figure 5-4) (Nyholt et al. 2000).
The Xq27 region implicated by these studies is bordered by markers DXS8043 and DXS297 translating to at most a 2.4Mb region. Similarly the Xq28 region is limited by DXS8061 to the Xqter equating to an approximate 4Mb region.
128
Figure 5-3 MF879 haplotypes
129
Figure 5-4 MF47 haplotypes
130 5.3.2 Investigation of Key Microsatellite Markers in a Case-control Cohort In order to test if the signals identified at the Xq27 and Xq28 loci were pedigree specific or translated to the general population two key microsatellite markers in the regions were genotyped in an Australian migraine case-control cohort. The markers DXS8043 and DXS8061 (Xq27 and Xq28 respectively) showed evidence of association in the tested female population (P=0.009 and P=0.07 respectively). While DXS8043 appears not to be specific to a particular migraine subtype, DXS8061 is clearly associated with the MO class (P=0.05, Table 5-1).
Table 5-1 Microsatellite analysis in case-control cohort Marker Position Locus CLUMP analysis P value# All Cases MA v All MO v All Male v Female v v Controls Controls Controls Controls* Controls+ DXS8043 143836250 Xq27 0.014 0.019 0.012 0.7 0.009 DXS8061 151772260 Xq28 0.25 0.5 0.05 0.61 0.07 * Pooled male MA and MO samples v male controls + Pooled female MA and MO samples v female controls # P value for T1 statistic reported
Two SNPs in close proximity to each marker were also genotyped in this population to verify the association seen by the microsatellites. rs5920369 located 720bp downstream to DXS8043 confirmed the association to the female population (genotype P=0.01). While rs6653488, 2.8Kb from DXS8061 also showed association to females (P=0.02).
5.3.3 Xq27 Candidate Gene Study Interrogation of public databases for gene annotations in the ~2.4Mb Xq27 loci bordered by markers DXS8043 and DXS297 identified 4 annotated protein coding genes. Consequently only 2 candidate genes were selected for further analysis, SLITRK2 and CXorf1 (See section 4.3.3.1 for detailed candidate gene selection methodology and see Appendix D for a summary of gene function). Four SNPs were identified for analysis across these 2 genes. All SNPs were in HWE however none showed association (Table 5-2).
131 Table 5-2 Xq27 Candidate gene study SNP Gene Position (bp) Position in Gene MAF Cases v Controls^ rs3213562* 143645815 0.49 0.8 DXS8043+ 143836250 0.014+ rs5920369* 143836970 0.20 0.05 rs4827657 SLITRK2 144705965 5' near gene 0.26 0.64 rs3810704 SLITRK2 144711944 exonic 0.16 0.58 rs2474401 CXorf1 1.445E+09 5' near gene 0.12 0.91 rs5919909 CXorf1 144717834 3' UTR 0.34 0.99 ^Allelic χ2 P value reported *These two SNPs were not part of the candidate gene study but were tested as they flanked the Xq27 microsatellite marker DXS8043 +CLUMP analysis P value for T1 statistic reported
5.3.4 Xq28 Candidate Gene Study In contrast to the Xq27 locus, the Xq28 region harbours numerous potential candidate genes. In this region 29 SNPs were chosen across 9 candidate genes. These genes were selected based on known or implied functional roles in hormonal, vascular or neurological systems (See section 4.3.3.1 for detailed candidate gene selection methodology and see Appendix D for a summary of gene function).
One SNP was found to be in HWD (P<0.001) and was not used for further studies. Chi square analysis identified a SNP in CNGA2 with positive association, P=0.048, which was also stronger in females (genotype P=0.0038) when analysed by gender, however, this did not remain significant after corrections. No other SNPs were associated in the entire cohort or when stratified by gender (Table 5-3).
However, to further investigate the association to MO observed at the microsatellite marker DXS8061, the SNPs around this marker were also tested for association to this migraine class. This analysis identified 3 SNPs around the DXS8061 marker that independently showed a trend towards association with MO (P≤0.07). Haplotype analysis of this series of SNPs revealed a 3 marker risk haplotype at rs5970389, rs6653488 and rs2071256 (P=0.0082, OR=1.53 CI=1.11-2.11) in the NSDHL gene (Table 5-4).
132
Table 5-3 Xq28 Candidate gene study Position in Cases v SNP Gene Position (bp) Gene MAF Controls^ rs75945486 GRP50 exonic 0.44 0.31 rs5925018 CNGA2 150653608 intronic 0.49 0.048 rs733871 CNGA2 150659252 intronic 0.14 0.70 rs5925021 CNGA2 150664255 3'utr 0.08 0.44 rs3848926 GABRA3 151072099 3' near gene 0.24 0.84 rs5970220 GABRA3 151083507 3' near gene 0.24 0.56 rs4828692 GABRA3 151122473 intronic 0.12 0.61 rs12833553 GABRA3 151161270 intronic 0.12 0.64 rs6627221 GABRA3 151166530 intronic 0.17 0.13 rs970417 GABRA3 151233139 intronic 0.01 0.16 rs389292 GABRA3 151292110 intronic 0.31 0.82 rs4612533 GABRA3 151342723 intronic 0.26 0.93 rs6627595 GABRA3 151368659 intronic 0.39 0.45 rs3827416 NSDHL 151753107 intronic 0.23 0.13 rs3788741 NSDHL 151753989 intronic 0.16 0.71 rs12836478 NSDHL 151770371 intronic 0.17 0.62 rs5970389* NSDHL 151771743 intronic 0.35 0.21 DXS8061+ NSDHL 151772260 intronic 0.25+ rs6653488* NSDHL 151775061 intronic 0.23 0.02 rs2071256 NSDHL 151787108 intronic 0.40 0.30 rs2285034 ATP2B3 152467567 intronic 0.36 0.50 rs2269415 ATP2B3 152476922 intronic 0.45 0.94 rs5945145 ATP2B3 152484242 intronic 0.35 0.16 rs4898421 ATP2B3 152487354 intronic 0.41 0.84 rs4898424 ATP2B3 152492677 intronic 0.27 0.96 rs6571291 ABCD1 152650291 intronic 0.10 0.09 rs2239470 FLNA 153251607 intronic 0.20 0.80 rs4834 GDI1 153667183 exonic 0.11 0.57 rs499428 CLIC2 154167240 intronic 0.21 0.23 rs547724 CLIC2 154196319 intronic 0.26 0.54 ^Allelic χ2 P value reported *These two SNPs were not initially tested as part of the candidate gene study but were tested as they flanked the Xq28 microsatellite marker DXS8061 + CLUMP analysis P value for T1 statistic reported
Table 5-4 Xq28 MO Risk haplotype rs5970389, rs6653488, rs2071256 Haplotype Freq (MO) Freq (Cases) χ2 P OR 95% CI TTC 0.689 0.591 6.99 0.008 1.53 1.11-2.11 CCG 0.148 0.229 6.68 0.009 0.58 0.38-0.88 CTG 0.112 0.120 0.01 0.746 0.92 0.57-1.48 TTG 0.035 0.051 0.86 0.351 0.69 0.31-1.51
133 5.4 Discussion
The genetic basis of common migraine, MA and MO is yet to be deciphered. To date, loci on many chromosomes as well as a number of genes have been implicated yet we still lack replication studies to validate these findings. Furthermore, the current understanding of migraine pathophysiology is fragmented with the vascular, hormonal and neural systems all thought to play a role.
The unequal sex distribution in the general population strongly implicates a hormonal influence on migraine pathophysiology. Epidemiological data supports this as prevalence significantly increases in females at the time of puberty and decreases around the time of menopause (Lipton et al. 2001b). Furthermore, 7-8% of migraineurs experience the ICHD classified, Pure Menstrual Migraine, where MO occurs only on day 1±2 days of menstruation in every 2 of 3 cycles. While it is not specifically known how hormones contribute to migraine it has been demonstrated that some hormones, particularly estrogen, interact with other genes that have been implicated in migraine such as serotonin receptors and calcitonin gene related peptide (CGRP) (Gupta et al. 2007). However, to date studies of these genes have failed to provide conclusive evidence that hormones are the sole factor causing the female preponderance of migraine. It is therefore a distinct possibility that genes on the X chromosome may also be contributing to susceptibility to the disorder.
In this study we have expanded on previous work to narrow known susceptibility regions on the X chromosome in order to reduce the number of potential candidate genes for testing in common migraine. The initial scan by this laboratory was performed using 28 markers across the X chromosome in two Australian pedigrees. Results demonstrated significant NPL scores at markers DXS1001 (Xq24) and DXS1123 (Xq28) (Nyholt et al. 1998a). A follow-up study of these two pedigrees focused on the Xq region using 16 markers that identified three distinct loci. The first was at Xq24 spanning from markers DXS1001 to DXS1206 the second was at Xq27 from markers DXS984 to DXS1123 and the third at Xq28 from DXS8091 to Xqter (Nyholt et al. 2000).
134 Here, we have expanded on these findings in six new Australian migraine pedigrees. In the initial study the Xq27 locus was identified as spanning an 11Mb region from markers DXS984 to DXS1123. Through genotyping of an additional marker in this region (DXS8043) we have reduced the locus shared by 3 pedigrees to a 2.4Mb region, assuming that all pedigrees are segregating to the same susceptibility gene. In addition the genotyping of additional migraine pedigrees has identified 2 previously unreported pedigrees that segregate to the 2 susceptibility regions providing additional support for an X-linked genetic influence in migraine susceptibility.
This study also provides further evidence to indicate that the identified regions can be considered two separate loci, independently harbouring susceptibility variants. While MF7 and MF14 show excess allele sharing at both regions MF879 only associated with the Xq27 locus, similarly MF47 only associated with the Xq28 locus. In addition, the Xq27 locus has now been identified in three separate studies. Oedegaard and colleagues (Odegaard et al. 2009) conducted a genome-wide linkage study of bi- polar disorder and co-morbid migraine using pedigrees derived from the National Institute of Mental Health (NIMH) Genetics Initiative for Bipolar Disorder. They identified a linkage peak with a LOD score of 1.6 (P=0.003) in the Xq27 region (marker DXS9908 approx 1Mb from DXS8043) that segregated with only the migraine and not the bipolar phenotype.
Our finding that the key markers identifying the X chromosome susceptibility regions are also associated in a case-control population provides evidence that these loci are not pedigree specific and may be contributing to migraine in the general population. This result provides the impetus to identify the causative variants through candidate gene studies.
The Xq27 loci appears to be a particularly ‘gene poor’ region. To date few annotations in this region are available on public databases. Consequently only 2 genes (SLITRK2 and CXorf1) were examined in the region with neither showing association.
In contrast the Xq28 region contains numerous potential candidate genes such as the neurotransmitter genes SLC6A8, GABRE, GABR3 and GABRQ and ion channel genes
135 CLIC2 and CNGA2. In particular GABRA3 was of interest due to the plethora of evidence implicating a role for glutamate in migraine particularly in increasing susceptibility to cortical spreading depression (Jen et al. 2005; Vikelis and Mitsikostas 2007). The GABRE and GABRQ genes located in the Xq28 region were not selected as these have been analysed and reported previously as showing no association to migraine in our Australian populations (Fernandez et al. 2008). The ATP2B3 gene was also selected as this gene belongs to the ATPase family similar to the FHM2 gene ATP1A2, and furthermore, ATP2B3 is thought to play a role in calcium homeostasis as does the FHM1 gene CACNA1A. Other genes selected for this study are known ion channels thought to be involved in neurological processes. The ion channel genes were of particular interest as it has been established that FHM can be caused by independent mutations in three different ion channels. Additionally ion channels may play a role in neuronal excitability, potentially contributing to the precipitation of migraine attack.
While this study did not strongly implicate any of the tested genes in migraine, suggestive evidence of a 3 marker haplotype in the MO subtype may be seen in the NSDHL gene. This gene is involved in the synthesis of cholesterol and loss-of- function mutations are typically associated with CHILD syndrome, a severe disorder that is lethal in males. However numerous studies have linked CVD with migraine and some studies suggest an influence of cholesterol levels on both disorders (Silberstein 2005; Bigal 2011; Gruber et al. 2011). Nonetheless the results of this study nominally suggest an involvement of NSDHL in migraine, therefore if this gene is involved, significant further research into the functional implications of genetic variants in this gene need to be investigated.
5.5 Conclusion
The unequal sex distribution of migraine is a unique facet of this disorder that may be explained by hormones and/or underlying genetic variations on the X chromosome that potentially cause migraine predisposition. The polygenic nature of migraine suggests that an interaction of both these mechanisms is also a possibility. The results of this study suggest that two distinct susceptibility variants may exist at Xq27
136 (marked by DXS8043 – DXS297) and Xq28 (DXS8061 – Xqter) and that these susceptibility loci may extend to association in the general population. Further studies are warranted to identify the pathological relevance of these loci to migraine aetiology.
137 138 CHAPTER SIX: Investigation of the Xq24 candidate gene GRIA3
The excitory neurotransmitter glutamate has been implicated in both the hyper- excitability required for CSD as well as activation of the trigeminovascular system required for the allodynia associated with migraine. Polymorphisms in the GRIA1 and GRIA3 genes that code for two of four subunits of the glutamate receptor AMPA have been previously associated with migraine. Furthermore the GRIA3 gene is coded within a previously identified migraine susceptibility locus at Xq24. This study investigated polymorphisms in both these genes in an Australian case-control population.
Two variants in GRIA1 and three variants in GRIA3 were genotyped in 500 unrelated migraine cases and 500 matched controls and data was analysed for association. Analysis showed no association between migraine and the GRIA1 gene. However, association was seen with the GRIA3 SNP rs3761555 (P=0.03).
The results of this study provide evidence for the translation of a pedigree specific susceptibility locus into the general population with the GRIA3 gene implicated as a candidate for contributing to migraine susceptibility. Furthermore, this study supports the plethora of evidence suggesting that glutamate dysfunction may contribute to migraine susceptibility, warranting further investigation of the glutamatergic system and particularly of the GRIA3 gene.
6.1 Introduction
Migraine is a common neurovascular disorder inducing severe pain and neurological symptoms that cause temporary incapacitation in the migraineur. This disorder interrupts all aspects of daily life including social, family and work commitments significantly affecting the sufferers overall quality of life. Epidemiological studies estimate prevalence at 12% (Lipton et al. 2007); consequently migraine is a significant burden to the individual, families, employers and healthcare systems.
139 Migraine is diagnosed according the ICHD-II (IHS 2004). Two common forms of migraine are recognised by this classification; MA and MO which vary depending on the presence or absence of accompanying temporary neurological disturbances known as the aura.
The pathophysiological mechanisms that underlie the common forms of migraine are still to be deciphered. Current research suggests that the trigeminovascular system plays a significant role in migraine (Parsons and Strijbos 2003; Lambert and Zagami 2008; Messlinger 2009) and therefore neurotransmitters and their associated receptors are of particular interest (D'Andrea and Leon 2010). Glutamate is the major excitory neurotransmitter in the CNS and therefore genes involved in its regulation and function are strong candidates for involvement in migraine pathophysiology and subsequent genetic analyses. Glutamate interacts with metabotropic and ionatropic receptors including kainate receptors, Alpha-amino-3-hydroxy-5-methyl-4-isoxazole- propionin acid (AMPA) ionatropic receptors and N-methyl-D-aspartate (NMDA) receptors. In addition to these receptors the Excitatory Amino Acid Transporters (EAAT) 1 is involved in the reuptake of excess glutamate from the synaptic cleft.
Previous work has implicated glutamate in both CSD and trigeminovascular activation (Vikelis and Mitsikostas 2007; Andreou and Goadsby 2009). CSD is characterised by a self-propagating depolarisations of the neurons associated with disturbances in ionic gradients and neurotransmitter release (Moskowitz 2007) that are thought to trigger the aura associated with MA. Excess glutamate is thought to contribute to the neuronal hyper-excitability, a mechanism that is implicated in both common and monogenic forms of MA. Mutations in the FHM genes; CACNA1A and ATP1A2, have been shown to cause excess release or slow uptake of glutamate from the synaptic cleft respectively, resulting in a lowering of the threshold for CSD susceptibility (Moskowitz et al. 2004). Jen et al (Jen et al. 2005) examined a de novo mutation in the SLC1A3 gene encoding EAAT1, in a patient with episodes of ataxia, migraine, hemiplegia and seizures. The authors concluded that the mutation contributed to neuronal hyperexcitability through decreased transporter function resulting in the hemiplegia and other neurological symptoms. Further evidence hinting at a role for the glutamatergic system in migraine includes the recent GWAS reported by Anttila et al (2010) demonstrating association at a marker localised to 8q22.1. This
140 marker is between 2 genes involved in glutamate homeostasis; Astrocyte Elevated Gene 1 and Plasma Glutamate Carboxypeptidase.
In this study we investigate polymorphisms in the ionatropic AMPA receptors. These receptors are comprised of 4 subunits coded by the glutamate receptor ionatropic AMPA (GRIA) 1 to 4 genes at chromosomal loci 5q33, 4q32, Xq24 and 11q24 respectively. Interestingly the GRIA3 gene is coded within the Xq24-28 migraine susceptibility locus previously identified by our laboratory (Nyholt et al. 1998a; Nyholt et al. 2000).
Formicola and colleagues (Formicola et al. 2010) recently analysed a number of SNPs in each of these genes in an Italian population of 250 migraineurs and 260 controls. Their results indicated positive association with 2 SNPs in GRIA1 (rs548294 MO allelic P=0.008, rs2195450 MA allelic P=0.0005) and 1 SNP in GRIA3 (rs3761555 MA Females allelic P=0.003). The aim of this study was to determine if mutations in the GRIA1 and GRIA3 genes previously identified by Formicola et al. (2010) also contribute to migraine susceptibility in an Australian case-control cohort. In addition, 2 further GRIA3 variants are tested, due to preliminary analysis in our laboratory that had shown an indication of association in a small population.
6.2 Methods
6.2.1 Population A total of 1000 Caucasian Australians were recruited for this study from the east coast of Australia. Migraineurs were interviewed and completed a detailed questionnaire prior to diagnosis according to ICHD-II criteria as described previously (IHS 2004; Colson et al. 2005). 500 migraineurs were age (+/- 5 years), gender and ethnicity (Caucasian origin) matched to 500 controls. Whole blood samples were collected and genomic DNA was extracted using a standard salting out procedure (Miller et al. 1988). Participants gave informed consent prior to participation in the study and the study protocol was approved by the Griffith University Research Ethics Committee.
141 6.2.2 Genotyping 6.2.2.1 GRIA1 Two SNPs that showed positive association to migraine in an Italian population were selected for this study; rs548294 and rs2195450. rs548294 was amplified by standard PCR using the primers: forward - 5’AGATGAAGAAACAGAGGTC3’and reverse - 5’CCCCAGGTACTATTCAAAG3’. The PCR conditions used were: 95ºc initial denaturation for 5 minutes, followed by 35 cycles of 95ºc for 45 seconds, 56ºc for 45 seconds and 72ºc for 45 seconds, with a final extension step of 72ºc for 7 minutes. The PCR products were examined on a 3% agarose gel for an expected fragment size of 311bp. The samples were then genotyped by restriction digest using MWO-I (New England Biolabs, Gensearch, Australia) as per the manufacturers’ instructions and visualised on a 3% agarose gel.
The rs2195450 SNP was amplified by standard PCR using the primers: forward - 5’TCTAAGAGGAGGGGGCAAGG3’ and reverse - 5’GCTTGGTAGATGGTGCTTGA3’. The PCR conditions used were: 95ºc initial denaturation for 5 minutes, followed by 35 cycles of 95ºc for 45 seconds, 60ºc for 45 seconds and 72ºc for 45 seconds, with a final extension step of 72ºc for 7 minutes. The PCR products were examined on a 3% agarose gel for an expected fragment size of 367bp. The samples were then genotyped by restriction digest using TAQI (New England Biolabs, Gensearch, Australia) as per the manufacturers’ instructions and visualised on a 3% agarose gel.
6.2.2.2 GRIA3 Three polymorphisms were selected for this study with rs3761555 previously demonstrating association to migraine in an Italian population (Formicola et al. 2010). The remaining polymorphisms included a 10bp exonic insertion (rs10704237) and a CA repeat in the 3’UTR. The rs3761555 SNP was genotyped at the Australian Genome Research Facility (AGRF, Brisbane, Australia) using the Sequenom MassArray on an Autoflex Spectrometer and iPLEX GOLD chemistry. The rs10704237 SNP and the CA 3’ UTR repeat were both genotyped by capillary electrophoresis using the Applied Biosystems ABI 3130 Genetic Analyser (Life Technologies, Australia). Samples were amplified using standard PCR with 5’ FAM
142 labelled forward primers. Primer sequences for rs10704237 were: forward - 5’FAM- CTAGTGTGGGGTGGAAAGGA3’, and reverse - 5’GGGACCCACCGCAGGGAAAC3’. Primer sequences for the CA repeat were: forward - 5’FAM-GACATGCAAGATTCCAGTATGC3’, and reverse - 5’GGAAATACCAAAAACATCTGGTC3’. PCR conditions for both reactions were: 95ºc initial denaturation for 5 minutes, followed by 35 cycles of 95ºc for 45 seconds, 58ºc for 45 seconds and 72ºc for 45 seconds, with a final extension step of 72ºc for 7 minutes. PCR products along with Hi-Di ™ Formamide/GeneScan ™ 500 LIZ ® size standard were examined on the 3130 Genetic Analyser and genotypes were automatically called using the Applied Biosystems GENEMAPPER® software Version 4.0 (Life Technologies, Australia).
6.2.3 Statistical Analysis Individuals genotyped for less than 3 of the 5 polymorphisms were also excluded from the study resulting in a population total of 845 (413 migraineurs and 432 controls). Hardy Weinberg equilibrium (HWE) was determined for all SNPs using controls samples (females only for rs3761555). Allele and genotype frequencies were determined using standard contingency tables and a χ2 analysis was performed. CLUMP analysis was employed for analysis of the GRIA3 microsatellite repeat (Sham and Curtis 1995). Bonferroni correction was performed to account for multiple testing, setting the significance level at P=0.01. Power calculations determined that this cohort has ~80% power to detect allelic associations conferring an increased risk of 1.5 or greater at the 0.05 significance level (see section 4.5.2). However, it should be noted that power to detect association is reduced to <80% in subgroups divided by gender or migraine subtype.
6.3 Results
6.3.1 GRIA1 Two SNPs in the GRIA1 gene that have previously demonstrated positive association to migraine in an Italian population were selected and tested for association in an Australian case-control cohort. Both SNPs showed comparable minor allele frequencies (rs548294 MAF: 0.37, rs2195450 MAF: 0.24) to those reported in
143 HapMap release #28 for Caucasian populations (0.36 and 0.27 respectively). When examined for Hardy-Weinberg Equilibrium (HWE), rs548294 was found to be in HWE while rs2195450 was not. rs548294 is located at the 5’ end of the gene at position -2010 C/T. This SNP was successfully genotyped in 375 controls and 351 cases (85.9%). When divided for gender and subtype the migraine group consisted of 63 male and 288 female migraineurs and included 253 MA and 98 MO sufferers. Chi square analysis revealed no significant association with migraine nor was association seen when the cohort was divided into gender or migraine subtype (see Table 6-1).
Table 6-1 GRIA1 rs548294 analysis Genotype Allele N TT CT CC T C Control 375 49 (13.0%) 186 (49.6%) 140 (37.3%) 284 (37.8%) 466 (62.1%) Male 82 11 (13.4%) 39 (47.5%) 32 (39.0%) 61 (37.2%) 103 (62.8%) Female 293 38 (12.9%) 147 (50.1%) 108 (36.8%) 223 (38.0%) 363 (61.9%) Case 351 45 (12.8%) 171 (48.7%) 135 (38.4%) 261 (37.1%) 441 (62.8%) Male 63 8 (12.70%) 28 (44.4%) 27 (42.8%) 44 (34.9%) 82 (65.0%) Female 288 37 (12.8%) 143 (49.6%) 108 (37.5%) 217 (37.6%) 359 (62.3%) MA 253 34 (13.4%) 13 (44.6%) 106 (41.9%) 181 (35.7%) 325 (64.2%) MO 98 11 (11.2%) 58 (59.1%) 58 (59.1%) 153 (36.6%) 116 (59.1%)
GRIA1 rs548294 χ2 Results Allele χ2 (P value) Total Cases v Controls 0.07 (0.78) M. Cases v M. Controls 0.16 (0.69) F. Cases v F. Controls 0.22 (0.63) MA v Controls 0.56 (0.45) MO v Controls 0.57 (0.45)
HWE Controls P=0.29
The second SNP genotyped, rs2195450 is located in the first intron at position +561G/A in the GRIA1 gene. This SNP was successfully genotyped in 379 controls and 322 cases (82.9%). The migraine group consisted of 68 male and 254 female migraineurs and included 254 MA and 68 MO sufferers. Similar to rs548294, analysis of rs2195450 did not show significant association with migraine as a whole or with migraine subtypes or a particular gender (see Table 6-2).
144 Table 6-2 GRIA1 rs2195450 analysis Genotype Allele N AA GA GG A G Control 379 33 (8.7%) 120 (31.6%) 226 (59.3%) 186 (24.5%) 572 (75.4%) Male 94 10 (10.6%) 29 (30.8%) 55 (58.5%) 49 (26.0%) 139 (73.9%) Female 285 23 (8.0%) 91 (31.9%) 171 (60.0%) 137 (24.0%) 433 (75.9%) Case 322 28 (8.7%) 117 (36.3%) 177 (54.9%) 173 (26.8%) 471 (73.1%) Male 68 6 (8.8%) 22 (32.3%) 40 (58.5%) 34 (25.0%) 102 (75.0%) Female 254 22 (8.6%) 95 (37.4%) 137 (53.9%) 139 (27.3%) 369 (72.6%) MA 254 24 (9.4%) 96 (37.8%) 134 (52.7%) 144 (28.3%) 364 (71.6%) MO 68 4 (5.8%) 21 (30.8%) 43 (63.2%) 29 (21.3%) 107 (78.6%) rs2195450 χ2 Results Allele χ2 (P value) Total Cases v Controls 0.98 (0.32) M. Cases v M. Controls 0.04 (0.82) F. Cases v F. Controls 1.56 (0.21) MA v Controls 2.28 (0.13) MO v Controls 0.65 (0.41)
HWE Controls P=0.004
6.3.2 GRIA3 The rs3761555 SNP, located at position -1952 C/T in the GRIA3 gene, has previously been associated with migraine in the study by Formicola and colleagues in an Italian population (Formicola et al. 2010). In particular this study showed association with females and the MA subtype. In our Australian cohort 413 controls and 396 cases consisting of 73 male and 323 female migraineurs were successfully genotyped (95.7%). When broken down by subtype 293 MA and 103 MO sufferers were included. The minor allele frequency was demonstrated to be similar to the reported frequency in HapMap (0.27 cf. 0.30) and analysis in the female control cohort indicated that this SNP was in HWE. Chi square analysis revealed a positive association with the MA subtype (Allele P=0.01) with an overrepresentation of the T allele, a trend towards association was also seen in the female gender (Allele P=0.04). This data is summarised in Table 6-3.
As some association was found within the GRIA3 gene, we next examined a 10bp insertion (rs10704237) in exon 1 and a microsatellite marker located in the 3’UTR region for association within our population. Frequency data was not available for rs10704237 prior to genotyping and the Australian population showed no variation at this locus, therefore no further analysis was performed on this marker. In contrast, eight VNTR alleles were observed for the 3’UTR microsatellite marker that consists
145 of a dinucleotide CA repeat. 349 controls and 301 cases were genotyped for this analysis (76.9%) with the cohort consisting of 51 male and 250 female migraineurs, including 221 MA and 80 MO sufferers. CLUMP analysis of this polymorphism revealed no association to migraine.
Table 6-3 GRIA3 rs3761555 analysis rs3761555 Genotype and allele frequencies rs3761555 χ2 Results Allele N C T Allele χ2 (P value) Control 413 201 (27.3%) 535 (72.6%) Total Cases v Controls 4.43 (0.03) Male 90 21 (23.3%) 69 (76.7%) M. Cases v M. Controls 0.74 (0.38) Female 323 180 (27.9%) 466 (72.1%) F. Cases v F. Controls 3.91 (0.04) Case 396 162 (22.6%) 557 (77.4%) MA v Controls 6.1 (0.01) Male 73 13 (17.8%) 60 (82.2%) MO v Controls 0.09 (0.76) Female 323 149 (23.1%) 497 (76.9%) Genotype MA 293 113 (21.3%) 419 (78.7%) F. Cases v F. Controls 4.91 (0.08) MO 103 49 (26.3%) 138 (73.7%) Genotype Female only N CC CT TT HWE Controls P=0.17 Control 323 30 (9.3%) 120 (37.2%) 173 (53.6%) (Females only) Case 323 21 (6.5%) 107 (33.1%) 195 (60.4%)
6.4 Discussion
Glutamate is the major excitory neurotransmitter in the brain. Abnormalities in the glutamatergic system are implicated in neuronal hyperexcitability and this is thought to increase susceptibility to CSD, manifesting as the symptoms of aura (D'Andrea and Leon 2010). Glutamate also plays a role in activation of the trigeminovascular system (Ma 2001; Vikelis and Mitsikostas 2007) that is required for the nociceptive transmission and head pain associated with migraine. Glutamate functions through a number of receptors including the AMPA receptors. Two genes that code for subunits of the ionatropic AMPA receptors, GRIA1 and GRIA3 have previously been implicated in migraine in an association analysis of an Italian population (Formicola et al. 2010). The GRIA3 gene is also located within a susceptibility locus at Xq24 that was identified in 2 large, multigenerational, independent, Caucasian Australian families (Nyholt et al. 1998a; Nyholt et al. 2000). The present study investigated the association of these genes to migraine in an Australian case-control cohort.
146 The results of this study did not support previous evidence implicating a role for GRIA1 in migraine susceptibility. Although the population examined was age (+/- 5yrs), gender and ethnicity (Caucasian origin) matched to control for population stratification, and internal controls including repeat samples and negative controls were used to control for genotyping error, it should be noted that in this study, the rs2195450 SNP was in Hardy Weinberg disequilibrium, suggesting caution when interpreting this result.
In contrast, association with migraine was demonstrated in the GRIA3 promoter polymorphism examined. This variant indicated stronger association with the female gender, consistent with the previous study. However, it is possible that these variants may also be associated in males as this study is limited by a small male cohort. This finding should therefore be analysed in a larger study.
Our study also confirmed the previous report of association at the rs3761555 SNP within the MA subgroup of migraineurs. In addition, the remaining two variants tested within this gene did not confirm the preliminary analysis conducted by our laboratory in the smaller population. However, this result is consistent with the preliminary analysis conducted by Formicola et al (2010) that selected 8 SNPs for complete coverage across the GRIA3 gene, based on LD, to test in a small population (100 cases v 100 controls). From this preliminary analysis, rs3761555 SNP was also the only SNP to show a trend of allelic difference between cases and controls, suggesting no association across other regions of the gene.
Despite the consistency in these findings, our results also show an interesting variation from those previously reported by Formicola and colleagues. Analysis in their Italian Caucasian population indicated a significant over-representation of the C allele in migraineurs compared to controls (34% cf. 22% respectively). In the Australian population the opposite was observed with a significant under- representation of the C allele in migraineurs compared to controls (22% cf. 27% respectively). This variation may reflect different population demographics and suggest that rs3761555 is not the causative SNP but is in linkage disequilibrium in both populations with another SNP that is. Alternatively, different variants may be contributing to migraine within this gene in each population.
147
Nonetheless, the rs3761555 GRIA3 SNP has been previously hypothesised to be a putative transcription factor binding site. Formicola et al. (2010) demonstrated that the T allele, over-represented in our migraine cohort, has shown increased binding strength and reduced promoter activity. Therefore they suggest that this SNP may affect the expression of this gene, contributing to migraine susceptibility. Further functional and expression studies are required to more fully examine this proposed mechanism to fully clarify the role of GRIA3 in migraine particularly relating to MA.
6.5 Conclusion
This study supports current research that implicates dysfunction of the glutamatergic system in migraine pathophysiology. Furthermore, it provides evidence for the translation of a pedigree specific linkage signal into the general population with the GRIA3 gene implicated as a strong candidate for contributing to migraine susceptibility. Further research is required to elucidate the mechanism through which GRIA3 may contribute to migraine and to determine if the rs3761555 variant is causal or if other mutations may be contributing.
148 CHAPTER SEVEN: Investigation of the X chromosome in the Norfolk Island Isolated Population
Migraine is a common and debilitating neurovascular disorder with a complex envirogenomic aetiology. Numerous studies have demonstrated a preponderance of women affected with migraine and previous pedigree linkage studies in our laboratory have identified susceptibility loci on chromosome Xq24-Xq28. In this study we have used the genetic isolate of Norfolk Island to further analyse the X chromosome for migraine susceptibility loci.
Genotype data from an X chromosome scan consisting of 14,124 SNPs was ascertained for 288 individuals comprising a large core-pedigree of which 76 were affected with migraine. For computational efficiency a multi-step analysis approach was employed. This involved firstly a pedigree-based association approach incorporating logistic regression controlling for age, sex and the relatedness of the cohort. The chromosome-wide significance threshold was set at 110-5 based on number of effective test calculations.
Prioritization of results from this step implicated 3 clusters of SNPs, the first including a novel migraine susceptibility locus at Xq12 and the second and third mapped to the previously identified migraine susceptibility loci at Xq27 and Xq28. SNPs at these clusters were analysed in the second step using a linkage-based probit regression model adjusted for age and sex to account exactly for the relatedness of the cohort. The strongest association at Xq12 was seen at rs599958 (OR=1.75, P=8.92x10-4), while at the Xq27 locus the strongest association was shown at rs6525667 (OR=1.53, P=1.65x10-4). Further analysis of SNPs at these loci in two independent population cohorts provided evidence for replication at Xq12 and Xq27.
Overall, this study provides compelling evidence for a novel susceptibility locus on Xq12. The strongest SNP (rs102834) yielded a combined P-value of 1.610-5, and is located within the 5’UTR of the HEPH gene, which is involved in iron homeostasis in the brain and may represent a novel pathway for involvement in migraine pathogenesis.
149 7.1 Introduction
Many approaches have been employed to investigate and identify the genes that influence migraine pathophysiology. These include case-control association studies, GWAS and linkage studies from which many susceptibility loci have been implicated. However, the complex nature of the disorder and the heterogeneity in its phenotypic manifestation make it difficult to identify shared genetic components in outbred population studies. One method that has been employed to overcome this issue is to analyse migraine in a genetic isolate.
The Norfolk Island population is a genetic isolate that is descendant from 9 British ‘Bounty’ mutineers, 2 additional Caucasian sailors and 6 Tahitian women. The population is ideal for genetic studies as family histories can be traced through detailed genealogical databases to the original founders allowing reconstruction of the 11 generation pedigree. In addition, geographical isolation as well as strict quarantine and immigration laws have also resulted in a more homogenous lifestyle than is seen in an outbred population (Hoare 1999). Overall it is expected that the genetic diversity is reduced, consequently simplifying genetic models and minimising environmental effects, synergistically increasing the possibility of identifying susceptibility genes (Bellis et al. 2008b). The unique characteristics of this population therefore make it amenable to the study of complex disorders, such as migraine, where it is suspected that multifactorial inheritance may be playing a role in disease manifestation.
The prevalence of migraine in the Norfolk Island population is estimated at 25.5%, which is substantially higher than is seen in outbred populations. Heritability of the disorder in the pedigree has been analysed using SOLAR v4.0.6 to determine the ratio of migraine variance that is explained by additive polygenic effects compared to total phenotypic variance. This was estimated at H2=0.54 (Cox et al. 2011) and is consistent with other populations with heritability estimates for the disorder ranging between 0.34-0.57 (Mulder et al. 2003; Svensson et al. 2003). To interrogate the genetic basis for migraine in this population a linkage analysis interrogating the autosomal data from a GWAS has recently been reported. The results of this study identified regions with maximum LOD scores of 1.6 (P=0.003) and 1.26 (P=0.008) at
150 chromosomes 13q33.1 and 9q22.32, respectively - regions that have also been linked to highly comorbid disorders such as epilepsy, bipolar and schizophrenia (Cox et al. 2011). In addition, this population has also been used to implicate the KCNN3 potassium channel gene in migraine, with 4 SNPs from a gene-wide analysis showing significant association (Cox et al. 2011b). Combined, these results suggest a strong genetic basis for the occurrence of migraine in this population.
Of interest to the work described in this thesis is the observation that migraine in the Norfolk Island population displays a similar epidemiological trend to outbred populations with women affected ~3:1 over men. This chapter therefore expands the interrogation of the X chromosome to analysis of X chromosome-wide data obtained from the previously mentioned GWAS conducted in the isolated Norfolk Island population.
To search for X chromosomal loci for migraine in the Norfolk Island population a multi-step analysis approach was employed. This included a pedigree-based association approach incorporating logistic regression, to identify regions of interest followed by a linkage-based probit regression model adjusted for age and sex that accounts exactly for the relatedness of the cohort. This approach was employed as the complexity of the pedigree structure combined with the X chromosomal data precluded a linkage analysis in available pedigree analysis software. The strongest associations were then analysed for validation in two independent populations including prospective cohort from the Women’s Genome Health Study (WGHS) and an Australian migraine association population.
7.2 Methods
7.2.1 Norfolk Island Study 7.2.1.1 Study Population The ascertainment and population structure of the Norfolk Island cohort have been explained in detail in previous chapters (see section 4.1.1). In summary, 377 individuals that have been identified as descendants of the Norfolk Island founders (9 Isle of Man ‘Bounty Mutineers,’ 6 Tahitian women and 2 Caucasian sailors that
151 joined the population in the 19th century), have provided blood samples for DNA extraction and undertaken a comprehensive medical questionnaire to obtain phenotypic data including migraine information regarding family history, symptoms, triggers and medication. These individuals have been diagnosed for migraine in accordance with ICHD-II guidelines (IHS 2004). Collection and phenotypic characterisation of the Norfolk Island cohort have been described in detail previously (Bellis et al. 2008a; Bellis et al. 2008b). 288 of the 377 individuals were then selected as highly informative individuals within the pedigree structure and were therefore used for genotyping in this study.
7.2.1.2 Genotyping The subset of 288 related individuals (consisting of 136 males and 152 females), genotyped for this study include 76 migraine cases consisting of 22 males and 54 females. 17,681 X chromosome-wide SNPs were genotyped on the Illumina Infinium High Density (HD) Human 610-Quad DNA analysis BeadChip version 1 using 200ng of DNA per sample, as part of a GWAS. Samples were scanned on the Illumina BeadArray 500GX Reader and the Illumina BeadScan image data acquisition software (version 2.3.0.13) was used to collect raw data. Preliminary analysis of raw data was undertaken in Illumina GenomeStudio software (v2010.1) with the recommended parameters for the Infinium assay. Genotype cluster files were generated based on clustering of genotypes within the Norfolk Island population.
SNPs with a minor allele frequency of less than 1% were excluded from the analysis and genotypic data was analysed for discrepancies including male heterozygosity using PLINK (Purcell et al. 2007). Discrepant genotypes were blanked prior to analysis.
7.2.1.3 Statistical Analysis For computational efficiency a logistic regression approach implemented in the PLINK software (Purcell et al. 2007) was employed as a preliminary analysis for SNP associations. Migraine affection status was set as a binary outcome variable and additive effects of SNPs were calculated after adjusting for covariates age, gender and relatedness. The relatedness variable is defined as a kinship coefficient calculated for each member of the pedigree, which was estimated by IBD-matrices generated using
152 the SOLAR program (Almasy and Blangero 1998). A total of 14,124 SNPs spanning the X chromosome were analysed. We estimated the chromosome-wide statistical significance threshold based on empirical calculations performed by the simpleM program. Considering inter-SNP dependence across the chromosome we estimated the number of effective tests to be ~4000 and thus set the significance threshold to 110-5 (Gao et al. 2010). SNP results were annotated using the Whole Genome Association Study Viewer (WGAViewer) program (Ge et al. 2008) and NCBI build 37.1. A secondary analysis was also performed on the top 25 SNPs prioritised by P value. This analysis was undertaken using polygenic analysis in the SOLAR software package (Almasy and Blangero 1998). This analysis uses a variance components- based linkage model to determine the polygenic heritability and the proportion of variance caused by the covariates, therefore handling the pedigree structure of the cohort in an exact fashion. SNPs were coded 0, 1, 2 for genotypes AA AB BB in females respectively and 0, 2 for genotypes A, B in males.
Haplotype analysis was carried out firstly using Haploview (Barrett et al. 2005) to identify regions of linkage disequilibrium across key loci. Haplotype based association was then determined using PLINK to obtain haplotype frequencies and associations while controlling for age, gender and relatedness.
7.2.2 Replication Populations SNPs identified in the haplotype blocks at the Xq12 and Xq27 loci as well as significant SNPs in the Xq28 locus were investigated in two independent populations - a prospective cohort from the WGHS and an Australian migraine association population.
Ascertainment and ethical approval of the WGHS cohort are described elsewhere (Ridker et al. 2008). The WGHS is derived from participants of the Women’s Health Study (WHS) and consists of 23,294 unrelated women of European descent. A total of 5122 women reported migraine during the study. Of those reporting active migraine at baseline, 1177 could be classified as having MA and 1826 as having MO (Chasman et al. 2011).
153 The ascertainment and characteristics of the Australian migraine association population used for this study are described in detail in section 4.1.3. In short, this population consists of 500 migraineurs age (+/- 5years), sex and ethnicity (Caucasian origin) matched to 500 controls. Cases included 366 MA and 134 MO samples. Power calculations determined that this cohort has ~80% power to detect allelic associations conferring an increased risk of 1.5 or greater at the 0.05 significance level (see section 4.5.2). However, it should be noted that power to detect association is reduced to <80% in subgroups divided by gender or migraine subtype.
7.2.2.1 Genotyping The Australian migraine association population samples were genotyped for 31 SNPs across the Xq12, Xq27 and Xq28 regions at AGRF using MALDI-TOF MS using the Sequenom MassArray on an Autoflex Spectrometer and iPLEX GOLD chemistry.
Genotyping of the WGHS cohort was undertaken by collaborators and has been described in detail elsewhere (Ridker et al. 2008).
7.2.2.2 Statistical Analysis Genotyping in the Australian migraine association population was checked for male heterozygosity using the software PLINK (Purcell et al. 2007). Statistical analysis was undertaken using an association approach and χ2 analysis also in this software. Females were coded as AA, AB, BB and males were coded as homozygotes for their respective alleles i.e. Genotype A = AA, B = BB. Bonferroni correction calculates the significance level at 0.001 and HWE was determined in female controls only.
Statistical analysis of the WGHS cohort was undertaken by collaborators and the methodology has been described in detail elsewhere (Chasman et al. 2011).
7.3 Results
To investigate the possible involvement of an X chromosome susceptibility gene in migraine we conducted a chromosome-wide SNP analysis in the isolated Norfolk Island population consisting of 288 related individuals including 76 migraine cases
154 (22 male, 54 female). From the 17,681 SNPs genotyped 14,124 were used for analysis, with the remainder discarded due to low genotyping call rates or low minor allele frequency.
7.3.1 Single Marker Analysis Preliminary analysis of the X chromosome scan was undertaken using a logistic regression analysis that adjusted for age, gender and relatedness. This approach used a kinship coefficient as an approximate adjustment for the relatedness in the cohort. A Manhattan plot of P values from this analysis is depicted in Figure 7-1. SNPs were prioritised according to P value and although no SNPs remained significant after correction for multiple testing, analysis of the 25 top ranked SNPs (see Table 7-1 – results presented as “Approximate P”) clearly shows a cluster of 11 SNPs localised to the previously identified loci on Xq27 (strongest association at rs6525667 (Approximate P=6.23x10-5) and Xq28 (strongest association at rs6627483 Approximate P=1.81x10-4). Interestingly 10 of the 25 top ranked SNPs also mapped to a new 377Kb locus at Xq12, with the strongest association seen at rs670546 (Approximate P=9.67x10-5). Analysis by migraine subtype at these loci also showed association in both MA and MO (MA n=51, MO n=26). However neither resulted in an increased association compared to the combined analysis (data not shown).
Figure 7-1 Manhattan plot of PLINK logistic regression analysis * + -Log(P) based on Approximate P, Corrected significance threshold set to P=1x10-5 (dotted line)
The top 25 SNPs were reanalysed using SOLAR to account for relatedness exactly. SNP associations were tested within a linkage-based probit regression model adjusted for age and sex. Results of the exact analysis (Exact P) are included in Table 7-1. This
155 analysis caused a minor reshuffling of the SNP prioritisation (strongest association at Xq12 was seen at rs599958 (OR=1.75, Exact P=8.92x10-4), Xq27 remained the same) and, as expected, a reduction in significance due to the effects of the pedigree structure. However the results of the approximate and exact analysis were found to be correlated (r=0.399 P=0.048).
Table 7-1 Single marker analysis results: Top 25 SNPs after prioritization MAF Approximate CI Locus SNP cases controls P * OR L95 U95 Exact P + Xq27 rs6525667 0.28 0.47 6.23x10-5 1.53 1.23 1.91 1.65x10-4 Xq27 rs910618 0.29 0.48 8.94x10-5 1.49 1.19 1.86 3.67x10-4 Xq12 rs670546 0.16 0.05 9.67x10-5 1.71 1.27 2.29 1.34x10-3 Xq27 rs1998005 0.32 0.53 1.08x10-4 1.43 1.15 1.77 6.13x10-4 Xq12 rs599958 0.16 0.05 1.09x10-4 1.75 1.21 2.53 8.92x10-4 Xp11 rs5918294 0.55 0.36 1.24x10-4 1.48 1.20 1.81 1.67x10-4 Xq12 rs5918974 0.16 0.05 1.48x10-4 1.64 1.19 2.26 2.48x10-3 Xq12 rs1028348 0.15 0.05 1.61x10-4 1.71 1.22 2.41 2.33x10-3 Xq12 rs5965083 0.16 0.06 1.67x10-4 1.65 1.19 2.28 2.41x10-3 Xq12 rs5918577 0.16 0.06 1.67x10-4 1.73 1.24 2.40 1.13x10-3 Xq12 rs760867 0.16 0.06 1.67x10-4 1.68 1.23 2.29 1.63x10-3 Xq28 rs6627483 0.25 0.10 1.81x10-4 1.47 1.13 1.91 3.87x10-3 Xq28 rs12843815 0.52 0.31 2.57x10-4 1.31 1.07 1.61 8.81x10-3 Xq27 rs5920070 0.34 0.53 2.84x10-4 1.43 1.16 1.78 5.65x10-3 Xq27 rs1339482 0.35 0.54 3.88x10-4 1.42 0.18 11.57 7.39x10-4 Xq27 rs5920197 0.38 0.21 4.30x10-4 1.35 1.08 1.68 8.14x10-3 Xq27 rs12014291 0.37 0.53 4.44x10-4 1.40 1.13 1.73 1.69x10-3 Xq12 rs6525038 0.16 0.06 4.70x10-4 1.58 1.18 2.13 4.10x10-3 Xq12 rs5964480 0.16 0.06 4.70x10-4 1.62 1.18 2.23 3.09x10-3 Xq12 rs6525037 0.16 0.06 5.55x10-4 1.57 1.15 2.15 4.87x10-3 Xp22 rs2071201 0.2 0.09 5.89x10-4 1.38 1.05 1.81 2.12x10-2 Xq27 rs5920067 0.32 0.50 6.07x10-4 1.41 1.14 1.74 1.13x10-3 Xq27 rs12555969 0.32 0.50 6.97x10-4 1.42 1.16 1.74 1.00x10-3 Xq27 rs5920061 0.37 0.52 8.06x10-4 1.37 1.10 1.70 3.23x10-3 Xq27 rs5919666 0.55 0.38 9.31x10-4 1.32 1.07 1.63 8.00x10-3 *Approximate P – P values generated in the preliminary analysis using a logistic regression analysis that adjusted for age, gender and relatedness. This approach used a kinship coefficient as an approximate adjustment for the relatedness in the cohort + Exact P – P values generated in SOLAR within a linkage-based probit regression model adjusted for age and sex
7.3.2 Haplotype Analysis The novel Xq12 SNP cluster was considered for further haplotype analysis. The 377kb Xq12 locus included 21 SNPs incorporating 10 of the top 25 ranked SNPs from the study. A linkage disequilibrium plot of the region was generated in Haploview and 2 major haplotype blocks were identified (Figure 7-2). Block 1 consisted of 11 analysed SNPs. Analysis of this block revealed 3 haplotypes with haplotype 2
156
Figure 7-2 LD Plot Xq12 in Norfolk population *Annotated gene positions from NCBI build 37.3
157 significantly over-represented in migraineurs (OR=4.39 P=1.1x10-4). Similarly block 2 included 5 genotyped SNPs representing 5 haplotypes. Of these, haplotype 1 was significantly over-represented in migraineurs (OR=4.48 P=1.6x10-4). The effect of this haplotype block appears to be due to the presence of the minor allele at rs1028348, located within the 5’UTR of the HEPH gene. Table 7-2 includes haplotype and frequency data for this locus.
Table 7-2 Xq12 Haplotype analysis Block 1 (rs6525037|rs6525038|rs5964480|rs5965083|rs5964486|rs5964488| rs1977106|rs670546|rs599958|rs5918974|rs5918577) Frequency T statistic HAPLOTYPE Cases Controls OR (Wald Test) P GAGGTTCCCTC 0.76 0.85 1.47 1.89 0.16 TGTACCCTACA 0.17 0.05 4.39 15.0 1.1x10-4 GAGGTTTCCTC 0.07 0.10 1.61 1.29 0.25 Block 2 (rs760867|rs5919015|rs1028348|rs7054364|rs1011526) Frequency T statistic HAPLOTYPE Cases Controls OR (Wald Test) P GCAAT 0.15 0.05 4.48 14.2 1.6x10-4 GCGAT 0.02 0.01 1.67 0.27 0.60 ACGAT 0.06 0.06 1.01 4.4x10-4 0.98 ATGCC 0.24 0.26 1.14 0.31 0.57 ACGCC 0.53 0.62 1.40 2.50 0.11 Haplotype analysis using PLINK software adjusting for age, gender and relatedness.
The 2nd SNP cluster identified includes 10 the top 25 ranked SNPs that localise to a previously identified Xq27 migraine locus. Haplotype analysis in the Xq27 region revealed 2 haplotype blocks (Figure 7-3). Block 1 consisted of 6 SNPS and 5 haplotypes were identified. Of these haplotype 1 was significantly underrepresented in migraineurs (OR=2.63 P=1.3x10-4), while haplotype 5 was significantly overrepresented (OR=2.3 P=4x10-4). Similarly the second block also consisted of 6 SNPs with 4 different haplotypes recognised. Haplotype 1 was again underrepresented in cases (OR=2.22 P=6.9x10-4) and haplotype 3 significantly overrepresented in migraineurs (OR=2.06 P=1.6x10-3). Table 7-3 includes haplotype and frequency data for this locus.
158
Figure 7-3 LD Plot Xq27 in Norfolk population *No annotated genes in this region from NCBI build 37.3
159 Table 7-3 Xq27 Haplotype analysis Block 1: (rs4827700|rs4263905|rs5920061|rs12014291|rs910618|rs6525667) Frequency T statistic HAPLOTYPE Cases Controls OR (Wald Test) P AGTCGA 0.29 0.47 2.63 14.5 1.3x10-4 AGTCAG 0.07 0.03 2.42 3.06 0.08 GGTCAG 0.007 0.01 1.40 0.09 0.77 GTTCAG - 0.02 NA NA NA GTCTAG 0.63 0.47 2.30 12.30 4.6x10-4 Block 2: (rs4827703|rs5920067|rs12555969|rs1339482|rs5920070|rs1998005) Frequency T statistic HAPLOTYPE Cases Controls OR (Wald Test) P TCTGTG 0.33 0.51 2.22 11.5 6.9x10-4 TTCGTG - 0.02 NA NA NA CTCAGA 0.57 0.40 2.06 9.93 1.6x10-3 TTCAGA 0.10 0.07 1.54 1.15 0.283 Haplotype analysis using PLINK software adjusting for age, gender and relatedness of the cohort.
7.3.3 Replication Analysis The Xq12, Xq27 and Xq28 loci were assessed in the independent WGHS cohort and an Australian migraine association population. Thirty-one SNPs across the three loci were analysed in both populations consisting of 21 of the top 25 prioritised SNPs in the Norfolk population and 10 additional SNPs that ranked within the top 100 that also map to these loci.
Analysis in the WGHS cohort displayed evidence of association at the Xq12 locus at several SNPs. The strongest association was observed at rs1028348 (OR=4.479 P=7x10-3), which is located within the 5’UTR of the Hephaestin (HEPH) gene. Association was strengthened at 10 SNPs in the MO subtype (P<0.05). There was no association observed for the MA subtype, nor was association observed at markers analysed in the Xq27 or Xq28 loci (Table 7-4) in the tested WGHS cohort.
In the Australian migraine association population two SNPs failed genotyping quality control standards at AGRF and therefore were not available for analysis. A further SNP was completely homozygous in our population. Of the remaining 27 SNPs eleven were tested in the Xq12 locus. None of these SNPs showed association to migraine (see Table 7-5) however, it was noted that five of the eleven were not in HWE when tested in the female controls. Similarly the 3 SNPs tested in the Xq28 locus also showed no association to migraine, although all were in HWE (Table 7-5). These SNPs are located approx 85kb 3’ of the CNGA2 gene. This gene was
160 Table 7-4 WGHS Replication analysis (Part I) MAF+ Migraine MO* Locus SNP Position Cases Controls P OR CI(L95-U95) P OR CI(L95-U95) Xq12 rs6525037 65092185 0.102 0.106 0.239 1.04 (0.97-1.12) 0.035 1.13 (1.01-1.27) rs6525038 65109743 0.096 0.100 0.235 1.05 (0.97-1.13) 0.015 1.16 (1.03-1.31) rs5964480 65131862 0.099 0.103 0.193 1.05 (0.98-1.13) 0.025 1.14 (1.02-1.29) rs5965083 65147151 0.130 0.135 0.158 1.05 (0.98-1.12) 0.009 1.15 (1.03-1.28) rs5964486 65160788 0.150 0.158 0.068 1.06 (1.00-1.13) 0.019 1.12 (1.02-1.24) rs5964488 65170494 0.160 0.167 0.049 1.06 (1.00-1.13) 0.018 1.13 (1.02-1.25) rs670546 65188841 0.132 0.138 0.063 1.06 (1.00-1.14) 0.003 1.17 (1.06-1.31) rs5918974 65191060 0.132 0.138 0.062 1.06 (1.00-1.14) 0.003 1.18 (1.06-1.31) rs760867 65297008 0.132 0.139 0.047 1.07 (1.00-1.14) 0.002 1.18 (1.06-1.31) rs1028348 65300888 0.106 0.116 0.007 1.10 (1.03-1.18) 0.002 1.20 (1.07-1.34) rs7054364 65309361 0.212 0.221 0.076 1.05 (0.99-1.11) 0.116 1.07 (0.98-1.16) rs1011526 65332812 0.212 0.221 0.077 1.05 (0.99-1.11) 0.117 1.07 (0.98-1.16) rs1264216 65351982 0.214 0.221 0.123 1.04 (0.99-1.10) 0.132 1.07 (0.98-1.16) Xq27 rs4827700 145154227 0.417 0.411 0.278 1.02 (0.98-1.07) 0.951 1.00 (0.93-1.07) rs4263905 145155129 0.417 0.411 0.269 1.02 (0.98-1.07) 0.971 1.00 (0.93-1.07) rs5920061 145160245 0.417 0.411 0.280 1.02 (0.98-1.07) 0.861 1.01 (0.94-1.08) rs12014291 145163240 0.451 0.445 0.259 1.03 (0.98-1.07) 0.656 1.02 (0.95-1.09) rs910618 145175718 0.375 0.373 0.618 1.01 (0.97-1.06) 0.054 1.07 (1.00-1.15) rs6525667 145175719 0.375 0.373 0.624 1.01 (0.97-1.06) 0.054 1.07 (1.00-1.15) rs4827703 145188598 0.498 0.493 0.422 1.02 (0.97-1.06) 0.186 1.05 (0.98-1.12) rs5920067 145190854 0.423 0.420 0.657 1.01 (0.96-1.05) 0.237 1.04 (0.97-1.11) rs12555969 145191584 0.381 0.379 0.796 1.01 (0.96-1.05) 0.320 1.04 (0.96-1.12) rs1339482 145193916 0.440 0.436 0.578 1.01 (0.97-1.06) 0.173 1.05 (0.98-1.12) *Analysis of the MA subtype showed no statistically significant results. + HapMap MAF for the CEU population are listed in Table 7-5
161 Table 7-4 WGHS Replication analysis (Part II) MAF Migraine MO* Locus SNP Position Cases Controls P OR CI(L95-U95) P OR CI(L95-U95) Xq27 rs5920070 145198977 0.441 0.437 0.542 1.01 (0.97-1.06) 0.155 1.05 (0.98-1.12) rs1998005 145223893 0.441 0.437 0.529 1.01 (0.97-1.06) 0.148 1.05 (0.98-1.12) rs5919666 145415492 0.486 0.91 0.353 1.02 (0.98-1.07) 0.120 1.06 (0.99-1.13) rs5920197 145498892 0.405 0.401 0.581 1.01 (0.97-1.06) 0.340 1.03 (0.99-1.11) rs11094364 150739787 0.211 0.209 0.637 1.01 (0.96-1.07) 0.479 1.03 (0.94-1.23) Xq28 rs5970126 150773056 0.108 0.107 0.721 1.01 (0.94-1.08) 0.774 1.02 (0.90-1.14) rs12843815 150789378 0.455 0.453 0.654 1.01 (0.97-1.05) 0.626 1.02 (0.95-1.08) rs6627483 150789378 0.125 0.122 0.216 1.07 (0.96-1.20) 0.658 1.04 (0.86-1.25) *Analysis of the MA subtype showed no statistically significant results. + HapMap MAF for the CEU population are listed in Table 7-5
162 previously tested for association with migraine in this population and also showed no association.
The remaining 13 SNPs mapped to the Xq27 region. All SNPs were in HWE and χ2 analysis identified 4 at P≤0.05, with the strongest association observed at rs4827703 (P=0.006 OR=1.3 95%CI 1.07-1.58 see Table 7-5). Further analysis of this SNP by subtype showed that it was driven by the MA subtype (P=0.007 cf. P=0.17) but was not gender specific.
Table 7-5 Australian migraine population replication Analysis
MAF Expected Locus SNP Position Cases Controls MAF* P^ OR CI Xq12 rs6525037 65092185 na na 0.14 na na na rs6525038 65109743 0.09 0.11 0.12 0.17 0.80 (0.57-1.09) rs5964480 65131862 0.10 0.11 0.11 0.32 0.85 (0.62-1.16) rs5965083 65147151 0.14 0.14 0.15 0.96 0.99 (0.75-1.30) rs5964486 65160788 na na 0.14 na na na rs5964488 65170494 0.16 0.17 0.18 0.50 0.92 (0.70-1.18) rs670546 65188841 0.13 0.14 0.15 0.85 0.97 (0.73-1.28) rs5918974 65191060 0.14 0.14 0.15 0.94 0.99 (0.74-1.30) rs760867 65297008 0.14 0.14 0.12 0.95 0.99 (0.75-1.30) rs1028348 65300888 0.12 0.12 0.10 0.78 0.96 (0.71-1.29) rs7054364 65309361 0.22 0.24 0.25 0.18 0.85 (0.68-1.07) rs1011526 65332812 0.22 0.24 0.19 0.23 0.87 (0.69-1.09) rs1264216 65351982 0.22 0.24 0.25 0.28 0.88 (0.70-1.10) Xq27 rs4827700 145154227 0.42 0.44 0.38 0.55 0.94 (0.77-1.14) rs4263905 145155129 0.42 0.44 0.36 0.44 0.92 (0.76-1.12) rs5920061 145160245 0.42 0.43 0.39 0.49 0.93 (0.77-1.13) rs12014291 145163240 0.47 0.47 0.44 0.97 0.99 (0.82-1.20) rs6525667 145175718 0.37 0.38 0.37 0.46 0.92 (0.76-1.13) rs6525667 145175719 na na 0.38 na na na rs4827703 145188598 0.53 0.46 0.49 0.006 1.30 (1.07-1.58) rs5920067 145190854 0.44 0.38 0.35 0.01 1.26 (1.04-1.54) rs12555969 145191584 0.44 0.38 0.26 0.02 1.25 (1.03-1.52) rs1339482 145193916 0.46 0.40 0.44 0.03 1.23 (1.01-1.50) rs1998005 145198977 0.47 0.42 0.43 0.05 1.21 (0.99-1.46) rs5919666 145223893 0.46 0.50 0.48 0.08 0.84 (0.69-1.02) rs5920197 145415492 0.43 0.41 0.40 0.31 1.10 (0.90-1.34) rs11094364 145498892 0.24 0.22 0.21 0.25 1.14 (0.90-1.43) Xq28 rs5970126 150739787 0.09 0.10 0.11 0.32 0.85 (0.62-1.17) rs12843815 150773056 0.46 0.47 0.53 0.69 0.96 (0.79-1.16) rs6627483 150789378 0.14 0.13 0.17 0.90 1.01 (0.77-1.33) * HapMap reported MAF in CEU population. na – SNP failed genotype QC ^Allelic χ2 P value reported
163 7.4 Discussion
In complex disease such as migraine the interplay between environmental and genetic factors can make it difficult to distinguish the role each factor plays in the pathophysiology of disease. In order to account for this, the unique genetic isolate population of Norfolk Island was used to conduct an X-chromosome wide pedigree based association study. This population was also used due to its inimitable pedigree structure in which relationships can be traced through genealogical data to 17 original founders. This familial nature of the population leads to a decrease in genetic variation which, when combined with homogeneity of environmental influences should improve the likelihood of detecting genetic contributors to disease.
We analysed the X chromosomal data from a recent GWAS using a two-step method. This involved a preliminary scan of the data using a logistic regression pedigree-based association study approach controlling for age, gender and relatedness. This preliminary step was carried out for computational efficiency due to a lack of accessible software that could deal with firstly the complexity of the pedigree structure, combined with the hemizygote nature of X chromosome data in males and finally also allowed for automating of the analysis of ~14,000 SNPs. A secondary analysis of a reduced number of SNPs was then conducted using the program SOLAR that will account for the relatedness of the cohort exactly. Using this method we observed associations at the previously identified regions on Xq27 and Xq28 as well as a cluster of associated SNPs at a novel Xq12 locus. Follow-up haplotype analysis at the Xq12 and Xq27 regions also confirmed 2 haplotype blocks within both regions and identified migraine risk haplotypes.
7.4.1 Replication Studies Replication studies were subsequently undertaken in two additional independent cohorts to assess the strength of the observed associations in outbred populations. The results of these analyses showed no evidence of associations at the Xq28 locus in either cohort examined. In contrast associations were observed at both the Xq12 and Xq27 loci however the cohorts used showed conflicting results.
164 Associations were observed for most of the SNPs examined in the novel Xq12 region in the WGHS cohort. However the results in the Australian migraine population were at odds with both the WGHS and Norfolk isolate findings. There could be a number of possible reasons for this.
Firstly, it should be noted that a number of SNPs were not in HWE suggesting either genotyping errors (although as this genotyping was undertaken by a service provider stringent QC measures were applied) or potential for underlying population stratification that requires further investigation.
In addition, the WGHS migraine cohort (Chasman et al. 2011) is ten times larger (n~5000) then the Australian migraine association population. Furthermore the highest odds ratio observed was only 1.1. Therefore it is possible that the effect size of the variant contributing to migraine in this locus is too small to be seen in a population of n=500.
Finally, analysis in the WGHS cohort suggested stronger association in the MO cohort. Both the Norfolk Island and the Australian association population were too small to detect effects in the MO subtype. However, this result may suggest that the susceptibility variant may be affecting only a subtype of migraine or particular traits. Therefore further studies at this locus could consider alternate approaches to classification such as TCA and/or LCA to ensure the phenotypic homogeneity of migraine in the cohort.
The follow-up studies of markers at the Xq27 locus observed in the Norfolk Island isolate also showed similar contradictory findings. In this case several markers identified in the Norfolk Island study showed evidence of association to migraine in the Australian migraine association population however not in the WGHS cohort. While the conflicting nature of this finding is concerning, the association in the Australian population supports previous results reported in this thesis that showed evidence for association at the Xq27 microsatellite marker DXS8043 in this same Australian migraine population. The DXS8043 marker was translated from the pedigree analysis that initially identified the Xq27 susceptibility region that spans ~2.4Mb and is bordered by DXS8043 and DXS297.
165 The Xq27 SNPs showing association presented in this chapter are located within the 2.4Mb region identified in the pedigree analysis yet significantly downstream (~430Kb) of the candidate genes - SLITRK2 and CXorf1, described in chapter 5 that did not show association to migraine in the same Australian migraine population. Further interrogation of the NCBI database Build 37.2 failed to identify any annotations in the regions that these associated SNPs map to (see figure 7-3). Consequently progressing studies at this locus may require bioinformatics approaches to identify new genes, regulatory sites or microRNAs that may be coded here.
While the lack of association in the WGHS cohort sheds doubt on the involvement of the Xq27 locus in migraine susceptibility, it should be noted that involvement of this locus has been observed in 3 independent pedigrees, the Australian migraine population and the genetic isolate of Norfolk Island.
Finally, it should be noted that a contributing factor to the conflicting data between the Norfolk Island population and the replication populations used may be ethnic differences. The WGHS cohort consists of American women with a confirmed European ancestry. The Australian migraine population was collected from the east coast of Australia and consists of individuals of Caucasian origin. In contrast, the Norfolk population has a unique mix of European and Polynesian ancestry. It is possible therefore, that LD patterns across our identified regions vary between cohorts due to these ethnic differences. Therefore SNPs linked to causal variants in Norfolk may not be linked in the outbred populations resulting in the conflicting findings.
7.4.2 The Novel Xq12 Locus The novel 377kb Xq12 locus identified by SNP prioritization contains 2 genes (see Figure 7-2), Hephaestin (HEPH) and V-set and immunoglobin domain containing 4 (VSIG4) as well as 4 pseudogenes and 1 microRNA - mir223.
VSIG4 is coded in the middle of the first LD region identified in our haplotype analysis. This gene is part of the complement receptor family and appears to play an important role in regulating innate and adaptive immune response through clearance of autologous and pathogenic cells (He et al. 2008). Four SNPs encoded in this gene
166 were genotyped in the Norfolk population, however only 2 of the 4 were informative and were followed-up in the replication studies. Both SNPs (rs5964486 and rs5964488) are located within intron 1 of the gene.
Similarly the 5’ end of HEPH gene is coded in the 2nd LD block. Hephaestin is an iron transport protein involved in cellular iron export through oxidising ferrous to ferric iron for uptake by transferrin or other iron carriers. Hephaestin expression has been identified throughout the human gastrointestinal tract as well as in pancreatic islets and the enteric nerves (Hudson et al. 2010). Qian and colleagues (2007) also determined that this protein is expressed in the cortex, hippocampus, striatum and substrantia nigra of rats and that development and iron status have a significant effect on the expression of the HEPH gene. Furthermore, it has been shown in mice that Hephaestin is required for iron homeostasis in the CNS (Hahn et al. 2004).
Eight SNPs encoded in this gene were analysed in the Norfolk study with 4 of the 8 suggesting association at P<0.003. Three of these 4 SNPs are located within the 2nd LD block identified at Xq12. Two of these 3 SNPs are intronic (rs1028348 and rs7054364), however the final SNP that also showed the strongest association (rs1028348 - P=2.33x10-3 in the Norfolk study, P=7x10-3 in the WGHS), is in the 5’UTR of the HEPH gene. Currently the functional effect of this SNP is unknown. Therefore further investigation of the effect of this SNP on HEPH expression may be warranted in addition to further sequencing of this locus to identify any rare function- altering variants that may have been tagged by associated SNPs in this study.
It is possible that the causative genetic variation in this region does not reside in either of the two encoded genes - VSIG4 or HEPH and could plausibly involve regulatory regions for other genes coded nearby, or could affect the function of the microRNA mir223. However, determining this would involve bioinformatic approaches to identify new regulatory sites in the region and/or an analysis of mir223 function. While both analyses are worthwhile areas of investigation these are beyond the scope of this thesis. Therefore, further discussion will focus on the candidate genes coded here.
167 Both VSIG4 and HEPH could plausibly play a role in migraine pathology. However, recent studies considering the relationship between migraine and iron provide a stronger case for prioritising HEPH gene in future research. Two studies have reported elevated iron concentrations in the periaqueductal grey matter (Welch et al. 2001) and the putamen, globus pallidus and red nucleus (Kruit et al. 2009) in migraineurs, with higher iron concentrations seemingly associated with longer migraine history. Interestingly neither study identified a distinction between MA and MO suggesting that a dysfunction of iron homeostasis may contribute to the migraine per se rather than the aura. The location of the excess iron deposition, and the role these CNS areas are known to play in central pain processing, may support this hypothesis.
However, at this stage it is unclear how the iron accumulation may contribute to migraine. Kruit and colleagues (2009) consider that the iron accumulation could be a result of frequent activation of the nuclei involved in central pain. Alternatively the excess iron potentially contributes to the frequency of migraine through damage to these structures due to oxidative stress (Kruit et al. 2009). Another proposed mechanism is that the increased iron deposition may be a key to dysfunction of the nociceptive network in general which increases susceptibility to a migraine trigger rather than causing the migraine directly. Welch (2009) postulates that this may be the case as common migraine is episodic and generally decreases with age, whereas the increased iron concentration does not reflect this change. Therefore an external trigger that increases migraine risk, but also becomes less available with age may be involved.
Despite the current lack of understanding of the exact mechanisms through which iron homeostasis may contribute to migraine these studies and previous findings that show expression of Hephaestin in the gastrointestinal tract (nausea and vomiting are common symptoms associated with migraine), support the need for further investigation of the HEPH gene to provide insight into these aspects of migraine pathophysiology.
168 7.5 Conclusion
We have provided compelling evidence for a new migraine susceptibility locus at Xq12 in a pedigree from the genetic isolate of Norfolk Island and shown the association to extend to the general migraine population. Furthermore analysis in the Norfolk Island cohort provides additional evidence to support an existing migraine locus at Xq27. The Xq12 locus contains the HEPH gene which potentially plays a critical role in iron homeostasis in the brain. HEPH provides a promising new candidate gene that should be further investigated in the migraine context to determine if variants in this gene contribute to migraine pathophysiology.
169 170 CHAPTER EIGHT: Discussion and Future Directions
8.1 Research Overview
This thesis has investigated new and previously identified migraine susceptibility regions that map to the q arm of the X chromosome. The research has confirmed that the Xq27 and Xq28 regions play a role in migraine susceptibility not only in rare migraine families but also in the general population. Furthermore, this research has provided additional evidence to support GRIA3 at Xq24 as a candidate gene in migraine, providing further support for a role of the neurotransmitter glutamate in migraine pathophysiology.
Finally, this thesis has also presented the results of an X chromosome scan arising from a GWAS undertaken in a genetic isolate - the first of its kind to use this unique approach in a founder population. This analysis identified a novel migraine locus at Xq12 that was also replicated in an independent case-control cohort. The results of the scan also implicated the Xq27 locus substantiating the involvement of this region in migraine susceptibility.
The Xq27 and Xq28 regions were investigated for excess allele sharing in six new migraine families, with the identification of two families that independently showed evidence of excess allele sharing to opposite regions (MF879 at Xq27 - LOD 1.38 p=0.005; MF47 at Xq28 - LOD 1.5 p = 0.004). In combination with the previously reported migraine families, this analysis provided evidence to implicate these two distinct migraine susceptibility regions. Therefore key microsatellite markers from both regions were subsequently analysed in case-control cohorts. Results suggested that these regions may also influence migraine in the general population, particularly in the female cohort (Xq27 – DXS8043 P = 0.009, Xq28 – DXS8061 P = 0.07). Follow-up analyses of key SNPs in genes with potential pathophysiological roles in migraine subsequently identified a risk haplotype in the NSDHL gene (P=0.0082) that may warrant further investigation.
171 Analysis of the Xq24 region was restricted to the previously implicated candidate gene GRIA3 that encodes a subunit of the glutamate AMPA receptor. The analysis of this gene focused on three polymorphisms located in the promoter region, exon 1 and the 3’UTR respectively. Results from this analysis indicated a significant association of the promoter SNP, rs3761555, in the migraine case-control population (P = 0.03).
Finally, the X chromosome scan in the Norfolk Island pedigree identified a novel migraine susceptibility locus at Xq12 with risk haplotypes identified around the HEPH and VSIG4 genes. Importantly, SNPs located in these haplotypes also showed replication in an independent population. Furthermore, this scan provided additional replication evidence for the migraine susceptibility locus at Xq27.
Specifically, analysis has approached the investigation of these loci as four distinct regions each harbouring a gene(s) that contributes to migraine susceptibility in both migraine families and migraine cases in the general population. A discussion of the reasons for this approach and the results of the analysis by chromosomal region are given here.
8.1.1 Xq12 Although the primary aim of this research was to investigate previously implicated X chromosomal migraine regions this study also identified a novel locus for migraine on Xq12 through undertaking the first X chromosome scan for migraine in a genetic isolate. In conducting the chromosome-wide scan of the Norfolk Island population a cluster of associated SNPs was identified at a previously unreported Xq12 locus. Haplotype analysis in this population at the Xq12 cluster provided evidence to suggest the candidate genes HEPH and VSIG4 that are coded at this site may merit further investigation. A literature review suggests a potential role for HEPH in iron homeostasis in the brain (Hahn et al. 2004; Qian et al. 2007), particularly in regions associated with the current understanding of migraine pathophysiology. Furthermore, previous studies have shown excess iron deposits in the brains of migraineurs (Welch et al. 2001; Kruit et al. 2009). Combined these suggest a plausible role for the HEPH gene in migraine that warrants further investigation.
172 Alternatively, a role for VSIG4 is more tenuous. VSIG4 is thought to be involved in inflammation and the immune system possibly by acting as a receptor for component 3c and regulating T cell activation (Vogt et al. 2006; He et al. 2008). While there has been evidence put forward suggesting a role for neurogenic inflammation in migraine (Durham 2006) it is not clear how this directly affects migraine onset or progression. Furthermore, there is no evidence to suggest a direct role of VSIG4 in neurogenic inflammation. Further investigation of the function of this gene is required to establish if a link to migraine susceptibility exists.
In order to determine the extent of the influence that this susceptibility locus is having on migraine, we sought to replicate the finding in two independent populations. These populations included the WGHS and an Australian migraine association population. Eleven SNPs across the locus were therefore investigated in all three populations (see Figure 9-1)
Interestingly the results were contradictory in their findings. The association at this locus was confirmed in the WGHS cohort and was strengthened in the MO subtype but was not associated in the Australian migraine association population. These results may have been caused by a small effect size that is increased in the Norfolk population due to its genetic make-up, but too small to be seen in an outbred population of the n=500 size such as the Australian population (WGHS n~5000 cases). Alternatively, the conflicting results at this locus may reflect heterogeneity in the genetic basis of migraine where different traits are caused by different genetic loci. Therefore if the migraine cohorts in the WGHS and Australian populations differ in phenotypic expression or traits associated with migraine, association may be seen in one cohort and not the other.
This represents a key limiting factor in many current genetic studies of migraine. While homogeneity in the populations used in this study has been sought through age, sex and ethnicity (Caucasian origin) matching controls as well as strict adherence to ICHD-II criteria it is becoming increasingly apparent that migraine is characterised by many traits and expressions each of which may be influenced by different subsets of genes. Furthermore a variety of interacting facets may combine to lower an as yet undefined threshold causing some people to be more susceptible to migraine onset.
173
Figure 8-1 Xq12 Summary – P values of analysis in NI, WGHS and MAP populations
174 The mixture of facets reducing this threshold may also be different for subsets of migraineurs particularly if the traits that characterise their migraine episode vary. The variable response to treatments provides credence to this argument suggesting that the same mechanisms are not occurring in all migraine sufferers despite the outcome appearing phenotypically similar. Therefore investigation of homogenous cohorts defined by symptoms, traits or study aspects e.g. treatment efficacy, may provide stronger associations.
8.1.2 Xq24 The Xq24 susceptibility locus was originally identified in MF7 and MF14 and is bordered by DXS1001-DXS1206 which represents an approx 7Mb region. This region codes for 54 genes including 20 pseudogenes, 1 hypothetical location, 1 microRNA and 32 genes - 12 of which belong the cancer/testis gene family expressed in germ cells. Of the remaining genes only GRIA3 stands out as a migraine candidate as it codes for a subunit of the ionatropic glutamate receptor AMPA. Therefore this was the only susceptibility gene analysed in this region and analysis of further migraine pedigrees for linkage to this locus was not undertaken.
The analysis of the GRIA3 gene was also undertaken to replicate a previous study of this gene and migraine carried out in an Italian population (Formicola et al. 2009). This study also found that the related GRIA1 gene (chromosome 5) was associated with migraine; therefore GRIA1 was also included in the study presented here. Polymorphisms were analysed in an Australian case-control cohort and association was only seen at 1 of 3 GRIA3 polymorphisms (rs3761555) and none of the GRIA1 variants tested.
While this result confirms the previous analysis of the GRIA3 gene it was noted that the opposite allele was over-represented in the case cohorts. This may be an ethnic difference suggesting that the rs3761555 SNP is not itself causative but is in LD with a causative SNP in both populations. However, the previous study also provided evidence suggesting that the SNP is in a promoter binding site and the T allele, as was over-represented in the Australian case cohort, reduces the promoter activity and therefore affects the expression of the gene.
175 While the studies to date provide tantalising glimmers of a role for GRIA3 in migraine, it is far from established. Further studies of this gene are required to fully elucidate the effects of its functional variants to provide a substantive link to migraine pathophysiology.
8.1.3 Xq27 The Xq27 susceptibility locus was similarly first identified in a study of MF7 and MF14 that also implicated the Xq24 and Xq28 regions. This study conducted an analysis that suggested linkage heterogeneity and provided the initial basis for dividing these susceptibility loci into distinct regions.
The Xq27 locus has been further interrogated within this PhD through a number of approaches. Firstly excess allele sharing at this locus was identified in a 3rd migraine family. This family did not show linkage to Xq28 verifying the division of susceptibility loci into distinct regions. Analysis of this family refined the region to a 2Mb locus bordered by DXS8043 and DXS297. The DXS8043 marker was then genotyped in the Australian migraine association population to give an indication of whether the locus was specific to the families and familial migraine or related to migraine generally. This analysis provided evidence of an association and therefore a candidate gene approach was employed to identify causative factors in this region using the migraine association population. However, the region is particularly gene poor, therefore only 2 genes were identified for analysis and neither showed association.
Subsequently, the X chromosome scan in the Norfolk Island population identified a cluster of SNPs showing association that mapped to the Xq27 region identified in the initial pedigree study. This cluster was within the bounds of the linkage signal but was located approximately 430Kb downstream of the candidate genes previously assessed. Therefore this cluster of SNPs was also analysed in the migraine association population and association was observed at 4 of the 13 SNPs tested.
However, similar to the analysis of the Xq12 SNP cluster, analysis in the second and larger replication cohort (WGHS) failed to identify association at this locus. The reasons for this conflicting result have been discussed previously. In particular the
176
Figure 8-2 Xq27 Summary – P values of analysis in NI, WGHS and MAP populations
177 need for homogenous groups for replications studies are pertinent to this case. The interpretation of this result could suggest that it is possibly a false positive because association was only observed in the smaller population and therefore effect size is seemingly not an issue. Alternatively, as discussed previously the cohorts may differ substantially and this may also contribute to the variable results.
This study has provided some evidence to suggest that a susceptibility variant may be located within the Xq27 locus. However, proceeding with analysis in this region is recommended with caution due to the current state of conflicting data. In addition the lack of currently annotated genes in the region precludes further candidate gene studies. A bioinformatics approach considering possible binding sites or new microRNAs is likely required to progress research at this locus.
8.1.4 Xq28 The final susceptibility locus spanned from DXS8061 to Xqter. Similar to the Xq27 analysis a 3rd migraine family was identified that showed excess allele sharing in this region, but not in the Xq27 locus. Follow-up analysis of the DXS8061 microsatellite marker in the migraine association population indicated that this region may be contributing to migraine generally particularly in the MO subtype and in females.
The Xq28 region contains 188 genes therefore these were prioritised to consider genes with known or implicated neurological, vascular or hormonal functions. A final selection of 9 genes was analysed in the candidate gene study, none of which showed association to migraine. However investigation of the SNPs surrounding the DXS8061 marker in the MO cohort did identify a haplotype that was associated with disease in this group. This haplotype is in the NSDHL gene which is involved in cholesterol synthesis and known variants in this gene cause a rare disorder known as CHILD syndrome that is lethal in males.
Nonetheless high co-morbidity of migraine with cardiovascular disorders could provide a tentative link for a role for this gene in migraine. However further functional studies of genetic variants are required to give support to this mechanism.
178
Figure 8-3 Xq28 Summary – P values of analysis in NI, WGHS and MAP populations
179 Alternatively, other transcription factors or microRNAs in this gene may be playing an unseen role.
The Xq28 region was also considered in the Norfolk Island results. The clustering of SNPs at this site was not as strong as that observed at Xq12 and Xq27. Nonetheless, three SNPs mapping to an intergenic site between the candidate genes CNGA2 and GABRA3 were identified in the top 25 SNPs prioritised by P value. Not surprisingly these SNPs were not associated with migraine in the Australian association population given that the adjacent genes were also not associated. Furthermore no evidence for association was identified in the WGHS cohort.
Finally it is worth noting that the SNPs that identified the NSDHL haplotype were not genotyped in the Norfolk population however nearby SNPs were. None of these showed association to migraine generally or to the MO subtype. However, a limiting factor to this analysis is that the MO subgroup in the Norfolk Island cohort only consists of 25 individuals. Nonetheless a previous study has also found no relationship between cholesterol levels and migraine diagnosis in this population (Curtain et al. 2004). Combined these results suggest that further investigation of the relationship between migraine and cholesterol is required to understandhow the NSDHL gene may contribute to migraine generally, or to specific migraine subtypes.
8.2 Research Summary
Overall, the research described in this thesis has employed a number of analysis techniques to investigate four X chromosomal loci that may independently contribute to migraine susceptibility. Results of this research have implicated all four regions in migraine in the general population not just in rare migraine families. In addition, evidence has been presented to implicate the candidate genes GRIA3, HEPH, VSIG4 and NSDHL for further studies. Limitations to the research described herein have also been discussed and should be addressed in the analysis of these genes and the further interrogation of regions such as Xq27. Strategies to address these limitations, to further interrogate these genes and regions and finally to elucidate pathological mechanisms need to be explored.
180 8.3 Future Directions
The result of this research has been the identification of potential candidate genes for further studies. These genes include HEPH and VSIG4 at Xq12, GRIA3 at Xq24 and NSDHL at Xq28. Furthermore, the Xq27 region has now been identified in three different cohorts, these include three migraine families, the Norfolk Island population and the migraine association population. This presents an intriguing locus, due to its apparent lack of genes for further investigation. Finally the Xq28 region has by no means been interrogated to its full extent with a substantial number of other possible genes waiting further studies.
Strategies for future research arising from the work described in this thesis will be outlined subsequently. In particular it is recognised that technology is progressing at an amazing rate, with many emerging techniques that will provide a more comprehensive examination of the genes identified so far, as well as others including microRNAs and regulatory sites in these regions. The application of some of these techniques to progress this field of migraine research will be discussed.
8.3.1 Phenotypic Heterogeneity of Migraine A common theme that has been described throughout this thesis has been the difficulties in replicating genetic studies, which may be in part due to the heterogeneity of the migraine disorder. There are fundamental debates that have existed for long periods of time regarding whether MA and MO represent a spectrum of a common disorder or are distinct (Rapoport 1996; Russell et al. 2002; Nyholt et al. 2004; Ligthart et al. 2006). In many cases, including the research herein, the former is accepted and MA and MO are considered together assuming a common etiological basis. However classification of migraine according to these subtypes, which themselves can present with a variety of overlapping symptoms may in fact be hindering research.
Migraine is accepted as a polygenic disorder, therefore it is plausible that genetic variants contribute to particular traits rather then migraine as a whole. Furthermore, these traits may underpin the high co-morbidity that migraine has with other complex
181 polygenic disorders. The results of studies that use TCA and LCA have shown that approaches such as these may be required in tandem with classification by ICHD-II (Lea et al. 2005; Nyholt et al. 2005; Anttila et al. 2006; Anttila et al. 2008) to determine the particular characteristic that a susceptibility locus is influencing within the overall migraine phenotype.
Consequently as the need for large study cohorts develops to enable the use of GWAS approaches it should be recognised that a concurrent need for high quality phenotypic data is also required. This will allow analysis of homogenous subsets or common traits in these large populations, which may not have been feasible in the populations of the previous century due to low sample size. These traits may include efficacy of medications, migraine severity, frequency, duration, triggers and symptoms such as photo- or phonophobia to name a few. Correlations between genotypic variants and phenotypic subtypes may provide stronger associations that are more readily replicable.
In relation to the research described here further characterisation of migraine in the pedigrees and the association population may assist to identify particular subgroups in which the associations already seen may be strengthened. It is entirely plausible that associations with particular traits have also been missed already by not investigating this approach.
Adjusting our analysis approach to include phenotypic subgroups may have implications particularly relevant to the understanding of the variable efficacy of migraine treatments.
8.3.2 Sequencing Ultimately genetic research will involve sequencing of entire genomes of affected and unaffected individuals to directly identify disease causing variants. Currently the cost of sequencing is prohibitive to this approach. However, with the progression of technology and emerging high-throughput techniques, this cost is steadily decreasing.
Emerging technologies that are becoming amenable to the budgets of genetic research include exome or targeted sequencing. Exome sequencing focuses only on the
182 protein-coding regions of the genome and provides a means to identify rare coding variants where the GWAS approach lacks. However, while it has been estimated that 85% of mutations with large effects on disease-related traits occur within protein coding regions (Johansson and Feuk, 2012; Mestan et al.2011) exome studies are limited by the assumption that variations outside coding regions have little to no functional consequence. In contrast targeted sequencing that interrogates genetic susceptibility loci that have previously been identified in linkage or GWAS studies (Huebner et al. 2011) have an advantage in allowing for the possibility of functional consequences caused by variants within non-coding regions that may influence gene transcription and expression.
These sequencing methods both require enrichment of the exome or target region and subsequent sequencing of fragments such that most nucleotides at any given locus are sequenced at a minimum of 20 times for reliability. Already a number of kits are available to facilitate these methods including the Nimblegen 2.1M human exome sequence capture microarray that is capable of enriching for 180,000 coding exons and 550 microRNAs. The Agilent SureSelect is a competing method with various kits, the most comprehensive is the SureSelect Human All Exon 50Mb kit that targets 50Mb of DNA incorporating both exons and non-coding RNAs, using only 3μg of DNA. Currently it is thought that greater then 96% of RefSeq genes can be captured using these techniques (Bras and Singleton 2011).
Sequencing can be used to identify causal genetic variants through a number of different approaches. Firstly an analysis can undertake a sequencing approach of a number of unrelated affected individuals and look for variants present in the individuals that are not in a reference sequence. However studies using this technique have noted that any given individual may have between 5,000- 20,000 single nucleotide variants in exons (Bras and Singleton 2011). Therefore processing of the data produced by sequencing of large regions such as the entire exome can be problematic. To date, most approaches firstly filter out common variants through comparisons to databases such as dbSNP and HapMap. Variants that are computationally determined unlikely to be pathogenic i.e. are benign, are also filtered out. Remaining variants are screened for those in common among the affected individuals. In theory, if enough affected individuals are compared this should narrow
183 to the causative variant. This approach has been applied successfully in mendelian disorders (Robinson et al. 2011). However, the success of this approach is yet to be determined on polygenic complex disorders that are likely to have causative variants missed where they are not shared by all affected samples.
An alternative approach is sequencing in combination with pedigree analysis (Bras and Singleton 2011; Robinson et al. 2011). In this approach a susceptibility region may have already been identified and the sequencing may therefore be targeted either across the entire region, therefore including the potential for identification of causal non-coding variants. Alternatively, if the region is of considerable size then exons from the genes in the region may be sequenced in a targeted exome approach. Analysis in a large pedigree which presents with a phenotypically homogenous example of a complex disease may result in identification of rare family specific variants that are only present in all affected individuals.
However, a limitation to this approach; particularly when considering analysis of large regions such as a whole exome, may be that numerous candidate variants will be identified due to the high proportion of variants shared by the individuals as a result of their relatedness. Possibly some variants may represent the polygenic aspect of the complex disease while others are likely to be coincidental and result from the shared genetic background of the pedigree. Discerning these will require function and expression data of the genes to determine pathological relevance. Narrowing the region of interest first may therefore assist in identifying the gene of interest. Other strategies to determine whether the shared alleles are identical by descent in the affected individuals may also assist to filter out alleles that are not associated with the disease therefore narrowing the list of potentially causative variants (Rodelsperger et al. 2011).
Exome sequencing has been used to identify numerous causative genes particularly in recessive disorders (Nikopoulos et al. 2010; Rehman et al. 2010; Sun et al. 2010; Volpi et al. 2010; Becker et al. 2011; Johnston et al. 2011; Kalay et al. 2011). However, its application to complex disease is still in its infancy and it remains to be seen how this technology can best be applied (Lyon et al. 2011). In contrast it is likely that targeted sequencing will find a niche in the analysis of complex disease due to the
184 potential for involvement of non-coding variants that affect regulatory pathways. Overall both these strategies are highly efficient in reducing the overall volume (and cost) of sequencing and consequently the number of rare variants identified for follow-up when compared to a whole genome or exome sequencing strategy. It is envisaged that in particular targeted sequencing is a potential technology that could be applied to further progress the work described in this thesis. In combination with the above phenotyping strategies this may provide a powerful means for identifying susceptibility variants.. This approach will also provide advantage over exome sequencing because it is not limited to coding sequences and therefore may provide better insight to the role of ‘gene poor’ susceptibility regions such as Xq27; as well as allowing for the possibility that disease causing mechanisms may exist in regulatory regions.
8.3.3 Bioinformatics Developing bioinformatics approaches to analyse and interpret genetic data is an ongoing challenge in the analysis of complex disease. As research evolves from a candidate gene approach that requires prior knowledge of gene function, to a whole genome approach that simply identifies regions of association it is becoming common to implicate loci with no obvious connection to disease (Tung and Yeo 2011). While some of these may reflect false positives and poor study design, many reflect a lack of understanding of the functions of various features coded within our genome.
Programs such as GRAIL and ALAMUT are examples that will assist in the short- term to help prioritize the results of large scale studies. GRAIL looks for gene relationships between implicated loci and disease. This is achieved through assessing abstracts in PubMed and assigning statistical significance to gene and gene pathways associated with the loci and implicated through the literature (Raychaudhuri et al. 2009). Alternatively ALAMUT (Interactive Biosoftware) is dedicated to mutation diagnostics. This software integrates data sources and performs in-silico predictions regarding the pathogenicity of genetic variants. These programs and others like them will be critical to quickly interpreting the plethora of data that is becoming available in increasing quantities. However, it is becoming increasingly apparent that these programs must also evolve so that data can be interpreted and candidates prioritised in combination with the phenotypic data of the individual, cohort or pedigree studied, as
185 well as databases of neutral and disease causing variants (Lindblom and Robinson 2011).
Large scale projects are currently underway to assist in bridging the knowledge gap of understudied, or more likely, underreported regions of the genome and create the integrated databases of genetic variations that are required. Some of these projects include the human variome project (Cotton et al. 2009) and GEN2PHEN (Webb et al. 2011). These projects ultimately aim to create a database of human variations that are associated with disease and, importantly, variations that have previously not shown association with disease. This will conteract the effects of publication bias where negative associations are often underreported so that research can be directed to truly new targets. Overall this will assist to reduce redundancy in global research efforts and accelerate progression towards uncovering the mechanisms that underpin complex disorders such as migraine. Furthermore, as pharmacogenomics becomes an increasingly likely scenario for disease treatment, these projects will be vital to filtering out false positives to discern the exact genetic variants that cause disease and influence response to treatment.
Bioinformatics is a growing field that will be critical to the success of genetic research studies. Collection, storage, analysis and access to data are key areas where consensus systems are required and standards need to be imposed (Lindblom and Robinson 2011). This will enable the development of new software tools that mine vast databases of available phenotypic and genotypic data. Furthermore, interpretation of these results will require new bioinformatic approaches to assist in developing decision making frameworks relevant to the researcher and the clinician.
8.4 Conclusion
The research described in this thesis has been carried out using numerous techniques for both genotyping and analysis. Through these methods a number of candidate genes have been implicated in various susceptibility loci on the X chromosome. This finding is consistent with the theory that common migraine is a polygenic disorder and it is likely that a number of variants of small effect size are contributing to either
186 lowering the threshold for susceptibility to a migraine episode and/or are contributing to individual traits that characterise the migraine phenotype. Further investigation of these genes using new and emerging technologies is required to confirm their role in migraine and the exact mechanisms that variants in these genes may be altering.
187
188
APPENDICES
189 Appendix A ICHD-II Part One: The Primary Headaches 1. Migraine 1.1-1.2 IHS 1.1 Migraine without Aura Previously used terms: Common migraine, hemicrania simplex Description: Recurrent headache disorder manifesting in attacks lasting 4-72 hours. Typical characteristics of the headache are unilateral location, pulsating quality, moderate or severe intensity, aggravation by routine physical activity and association with nausea and/or photophobia and phonophobia Diagnostic criteria: A. At least 5 attacks1 fulfilling criteria B-D B. Headache attacks lasting 4-72 hours (untreated or unsuccessfully treated)2;3;4 C. Headache has at least two of the following characteristics: 1. unilateral location 5;6 2. pulsating quality7 3. moderate or severe pain intensity 4. aggravation by or causing avoidance of routine physical acti`vity (eg, walking or climbing stairs) D. During headache at least one of the following: 1. nausea and/or vomiting 2. photophobia and phonophobia8 E. Not attributed to another disorder9 Notes: 1. Differentiating between 1.1 Migraine without aura and 2.1 Infrequent episodic tension- type headache may be difficult. Therefore at least 5 attacks are required. Individuals who otherwise meet criteria for 1.1 Migraine without aura but have had fewer than 5 attacks should be coded 1.6.1 Probable migraine without aura. 2. When the patient falls asleep during migraine and wakes up without it, duration of the attack is reckoned until the time of awakening. 3. In children, attacks may last 1-72 hours (although the evidence for untreated durations of less than 2 hours in children requires corroboration by prospective diary studies). 4. When attacks occur on ≥15 days/month for >3 months, code as 1.1 Migraine without aura and as 1.5.1 Chronic migraine. 5. Migraine headache is commonly bilateral in young children; an adult pattern of unilateral pain usually emerges in late adolescence or early adult life. 6. Migraine headache is usually frontotemporal. Occipital headache in children, whether unilateral or bilateral, is rare and calls for diagnostic caution; many cases are attributable to structural lesions. 7. Pulsating means throbbing or varying with the heartbeat. 8. In young children, photophobia and phonophobia may be inferred from their behaviour. 9. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
IHS 1.2: Migraine with Aura Previously used terms: Classic or classical migraine, ophthalmic, hemiparaesthetic, hemiplegic or aphasic migraine, migraine accompagnée, complicated migraine Description: Recurrent disorder manifesting in attacks of reversible focal neurological symptoms that usually develop gradually over 5-20 minutes and last for less than 60 minutes. Headache with the features of migraine without aura usually follows the aura symptoms. Less commonly, headache lacks migrainous features or is completely absent. Diagnostic criteria: A. At least 2 attacks fulfilling criterion B B. Migraine aura fulfilling criteria B and C for one of the subforms 1.2.1-1.2.6 C. Not attributed to another disorder1 Notes:
190 1. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
IHD 1.2.1 Typical Aura with Migraine Headache Description: Typical aura consisting of visual and/or sensory and/or speech symptoms. Gradual development, duration no longer than one hour, a mix of positive and negative features and complete reversibility characterise the aura which is associated with a headache fulfilling criteria for 1.1 Migraine without aura. Diagnostic criteria: A. At least 2 attacks fulfilling criteria B-D B. Aura consisting of at least one of the following, but no motor weakness: 1. fully reversible visual symptoms including positive features (eg, flickering lights, spots or lines) and/or negative features (ie, loss of vision) 2. fully reversible sensory symptoms including positive features (ie, pins and needles) and/or negative features (ie, numbness) 3. fully reversible dysphasic speech disturbance C. At least two of the following: 1. homonymous visual symptoms1 and/or unilateral sensory symptoms 2. at least one aura symptom develops gradually over ≥5 minutes and/or different aura symptoms occur in succession over ≥5 minutes 3. each symptom lasts ≥5 and ≤60 minutes D. Headache fulfilling criteria B-D for 1.1 Migraine without aura begins during the aura or follows aura within 60 minutes E. Not attributed to another disorder2 Notes: 1. Additional loss or blurring of central vision may occur. 2. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
IHS 1.2.2 Typical Aura with non-migraine headache Description: Typical aura consisting of visual and/or sensory and/or speech symptoms. Gradual development, duration no longer than one hour, a mix of positive and negative features and complete reversibility characterise the aura which is associated with a headache that does not fulfil criteria for 1.1 Migraine without aura. Diagnostic criteria: A. At least 2 attacks fulfilling criteria B-D B. Aura consisting of at least one of the following, but no motor weakness: 1. fully reversible visual symptoms including positive features (eg, flickering lights, spots or lines) and/or negative features (ie, loss of vision) 2. fully reversible sensory symptoms including positive features (ie, pins and needles) and/or negative features (ie, numbness) 3. fully reversible dysphasic speech disturbance C. At least two of the following: 1. homonymous visual symptoms1 and/or unilateral sensory symptoms 2. at least one aura symptom develops gradually over ≥5 minutes and/or different aura symptoms occur in succession over ≥5 minutes 3. each symptom lasts ≥5 and ≤60 minutes D. Headache that does not fulfil criteria B-D for 1.1 Migraine without aura begins during the aura or follows aura within 60 minutes E. Not attributed to another disorder2 Notes: 1. Additional loss or blurring of central vision may occur. 2. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
191
IHS 1.2.3 Typical aura without headache Description: Typical aura consisting of visual and/or sensory symptoms with or without speech symptoms. Gradual development, duration no longer than one hour, a mix of positive and negative features and complete reversibility characterise the aura which is not associated with headache. Diagnostic criteria: A. At least 2 attacks fulfilling criteria B-D B. Aura consisting of at least one of the following, with or without speech disturbance but no motor weakness: 1. fully reversible visual symptoms including positive features (eg, flickering lights, spots or lines) and/or negative features (ie, loss of vision) 2. fully reversible sensory symptoms including positive features (ie, pins and needles) and/or negative features (ie, numbness) C. At least two of the following: 1. homonymous visual symptoms1 and/or unilateral sensory symptoms 2. at least one aura symptom develops gradually over ≥5 minutes and/or different aura symptoms occur in succession over ≥5 minutes 3. each symptom lasts ≥5 and ≤60 minutes D. Headache does not occur during aura nor follow aura within 60 minutes E. Not attributed to another disorder2 Notes: 1. Additional loss or blurring of central vision may occur. 2. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
IHS 1.2.4 Familial Hemiplegic Migraine (FHM) Description: Migraine with aura including motor weakness and at least one first- or second-degree relative has migraine aura including motor weakness. Diagnostic criteria: A. At least 2 attacks fulfilling criteria B and C B. Aura consisting of fully reversible motor weakness and at least one of the following: 1. fully reversible visual symptoms including positive features (eg, flickering lights, spots or lines) and/or negative features (ie, loss of vision) 2. fully reversible sensory symptoms including positive features (ie, pins and needles) and/or negative features (ie, numbness) 3. fully reversible dysphasic speech disturbance C. At least two of the following: 1. at least one aura symptom develops gradually over ≥5 minutes and/or different aura symptoms occur in succession over ≥5 minutes 2. each aura symptom lasts ≥5 minutes and < 24 hours 3. headache fulfilling criteria B-D for 1.1 Migraine without aurabegins during the aura or follows onset of aura within 60 minutes D. At least one first- or second-degree relative has had attacks fulfilling these criteria A-E E. Not attributed to another disorder1 Notes: 1. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
IHS 1.2.5 Sporadic Hemiplegic Migraine (SHM) Description: Migraine with aura including motor weakness but no first- or second-degree relative has aura including motor weakness. Diagnostic criteria: A. At least 2 attacks fulfilling criteria B and C B. Aura consisting of fully reversible motor weakness and at least one of the following:
192 1. fully reversible visual symptoms including positive features (eg, flickering lights, spots or lines) and/or negative features (ie, loss of vision) 2. fully reversible sensory symptoms including positive features (ie, pins and needles) and/or negative features (ie, numbness) 3. fully reversible dysphasic speech disturbance C. At least two of the following: 1. at least one aura symptom develops gradually over ≥5 minutes and/or different aura symptoms occur in succession over ≥5 minutes 2. each aura symptom lasts ≥5 minutes and <24 hours 3. headache fulfilling criteria B-D for 1.1 Migraine without aura begins during the aura or follows onset of aura within 60 minutes D. No first- or second-degree relative has attacks fulfilling these criteria A-E E. Not attributed to another disorder1 Note: 1. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
IHS 1.2.6 Basilar-type migraine Previously used terms: Basilar artery migraine, basilar migraine Description: Migraine with aura symptoms clearly originating from the brainstem and/or from both hemispheres simultaneously affected, but no motor weakness. Diagnostic criteria: A. At least 2 attacks fulfilling criteria B-D B. Aura consisting of at least two of the following fully reversible symptoms, but no motor weakness: 1. dysarthria 2. vertigo 3. tinnitus 4. hypacusia 5. diplopia 6. visual symptoms simultaneously in both temporal and nasal fields of both eyes 7. ataxia 8. decreased level of consciousness 9. simultaneously bilateral paraesthesias C. At least one of the following: 1. at least one aura symptom develops gradually over ≥5 minutes and/or different aura symptoms occur in succession over ≥5 minutes 2. each aura symptom lasts ≥5 and ≤60 minutes D. Headache fulfilling criteria B-D for 1.1 Migraine without aura begins during the aura or follows aura within 60 minutes E. Not attributed to another disorder1
Notes: 1. History and physical and neurological examinations do not suggest any of the disorders listed in groups 5-12, or history and/or physical and/or neurological examinations do suggest such disorder but it is ruled out by appropriate investigations, or such disorder is present but attacks do not occur for the first time in close temporal relation to the disorder.
193 Appendix B NCBI Build 37.2 Gene Map DXS1123 - Xqtel
Start (bp) Stop (bp) Gene 147582139 148082193 AFF2 AF4/FMR2 family, member 2 148560295 148586884 IDS hCG_1993933 148622519 148632086 CXorf40A chromosome X open reading frame 40A 148663309 148669116 MAGEA9B melanoma antigen family A, 9 -like 148678216 148713487 TMEM185A hCG_1999996 148769903 148798928 MAGEA11 hCG_1644200 148850483 148854484 LOC100506205 transmembrane protein 185A-like 148856264 148858525 LOC100506164 heat shock transcription factor, X-linked-like 148863600 148869399 MAGEA9 hCG_1644199 149009941 149014609 MAGEA8 hCG_1756364 149097863 149099786 LOC100132460 similar to heat shock transcription factor family, X linked 1 149100415 149106716 CXorf40B protein CXorf40B 149399291 149438002 LOC389901 similar to ATP-dependent DNA helicase II, 70 kDa s ubunit 149531551 149682448 MAMLD1 hCG_1811536 149737047 149841616 MTM1 hCG_39236 149861869 149933575 MTMR1 hCG_1640369 149934809 150067289 CD99L2 CD99 molecule-like 2 150151763 150159248 HMGB3 hCG_39238 150345056 150349937 GPR50 G protein-coupled receptor 50 150565657 150577836 VMA21 VMA21 vacuolar H+-ATPase homolog (S. cerevisiae) 150732007 150845211 PASD1 PAS domain containing 1 150863730 150870063 PRRG3 hCG_1642068 150884508 150891666 FATE1 fetal and adult testis expressed 1 150903218 150914036 CNGA2 cyclic nucleotide gated channel alpha 2 151081361 151093642 MAGEA4 hCG_38709 151121596 151143151 GABRE hCG_38708 151282526 151286444 MAGEA5 hCG_1757057 151302891 151307025 MAGEA10 hCG_38711 151335634 151619831 GABRA3 gamma-aminobutyric acid (GABA) A receptor, alpha 3 151806637 151821825 GABRQ gamma-aminobutyric acid (GABA) receptor, theta 151867245 151870814 MAGEA6 melanoma antigen family A, 6 151876743 151877747 CSAG3 CSAG family, member 3 151885384 151887080 MAGEA2B melanoma antigen family A, 2B 151896280 151896822 LOC100130935 similar to CSAG family, member 2 151899293 151903184 MAGEA12 hCG_1737748 151903228 151909518 CSAG1 chondrosarcoma associated gene 1 151918388 151922364 MAGEA2 melanoma antigen family A, 2 151927734 151928738 CSAG2 CSAG family, member 2 151934652 151938240 MAGEA3 melanoma antigen family A, 3 151995871 151999301 CETN2 centrin, EF-hand protein, 2 151999511 152037907 NSDHL NAD(P) dependent steroid dehydrogenase-like 152082986 152142025 ZNF185 hCG_39570 152157368 152162671 PNMA5 hCG_2001741 152197130 152200901 LOC100128960 similar to Putative paraneoplastic antigen-like protein 6B-like protein 152220309 152224938 LOC100129956 similar to hCG2040565 152224766 152228827 PNMA3 hCG_39569 152240819 152243402 LOC100287428 similar to paraneoplastic antigen-like protein 6A 152244152 152246070 LOC100287466 similar to Paraneoplastic antigen-like protein 6A
194 Start (bp) Stop (bp) Gene 152338301 152340107 PNMA6A hCG_1792470 152341614 152342813 LOC100287500 similar to paraneoplastic antigen like 6A 152481522 152486116 MAGEA1 hCG_1742531 152584221 152587591 LOC649201 paraneoplastic antigen like 6A-like 152599613 152618384 ZNF275 zinc finger protein 275 152662364 152663269 LOC649238 similar to hCG1645335 152683781 152687086 ZFP92 zfp-92|zinc finger protein 92 homolog|zinc finger protein homologous to Zfp92 in mouse 152710178 152711945 TREX2 three prime repair exonuclease 2 152713123 152736603 HAUS7 HAUS augmin-like complex, subunit 7 152746989 152748274 LOC100507326 extracellular matrix protein 2-like 152751320 152752546 LOC100507352 extracellular matrix protein 2-like 152760347 152775004 BGN biglycan 152801580 152848387 ATP2B3 ATPase, Ca++ transporting, plasma membrane 3 152853383 152864632 FAM58A family with sequence similarity 58, member A 152871972 152875745 LOC100131652 similar to Mof4 family associated protein 1 152907897 152916781 DUSP9 dual specificity phosphatase 9 152935188 152939816 PNCK hCG_39226 152953752 152962048 SLC6A8 solute carrier family 6 (neurotransmitter transporter, creatine), member 8 152965947 152990201 BCAP31 B-cell receptor-associated protein 31 152990323 153010216 ABCD1 ATP-binding cassette, sub-family D (ALD), member 1 153029651 153044801 PLXNB3 hCG_39224 153046456 153051187 SRPK3 SFRS protein kinase 3 153051221 153059967 IDH3G isocitrate dehydrogenase 3 (NAD+) gamma 153060094 153063954 SSR4 hCG_39222 153067623 153096003 PDZD4 PDZ domain containing 4 153126971 153141399 L1CAM L1 cell adhesion molecule 153167985 153172620 AVPR2 arginine vasopressin receptor 2 153172830 153191714 ARHGAP4 Rho GTPase activating protein 4 153195377 153200468 NAA10 N(alpha)-acetyltransferase 10, NatA catalytic subunit 153200722 153210232 RENBP hCG_39215 153213008 153236819 HCFC1 hCG_39225 153237991 153248646 TMEM187 hCG_39211 153275957 153285342 IRAK1 hCG_39214 153287264 153363188 MECP2 hCG_37678 153409725 153424507 OPN1LW opsin 1 (cone pigments), long-wave-sensitive 153448085 153462352 OPN1MW opsin 1 (cone pigments), medium-wave-sensitive 153485203 153498755 OPN1MW2 GOP 153498932 153523438 TEX28 hCG_41349 153524027 153558713 TKTL1 hCG_41350 153576900 153603006 FLNA filamin A, alpha 153607597 153609883 EMD emerin 153626571 153630680 RPL10 hCG_2008007 153629579 153640427 DNASE1L1 deoxyribonuclease I-like 1 153639877 153650065 TAZ hCG_41340 153656978 153664862 ATP6AP1 ATPase, H+ transporting, lysosomal accessory protein 1 153665259 153671814 GDI1 GDP dissociation inhibitor 1 153672485 153679002 FAM50A family with sequence similarity 50, member A 153686623 153701985 PLXNA3 hCG_41342 153705241 153707596 LAGE3 L antigen family, member 3
195 Start (bp) Stop (bp) Gene 153712056 153714932 UBL4A hCG_41352 153715650 153719002 SLC10A3 hCG_41353 153733323 153744566 FAM3A family with sequence similarity 3, member A 153759606 153775787 G6PD glucose-6-phosphate dehydrogenase 153770459 153793261 IKBKG hCG_2003089 153813418 153815075 CTAG1A cancer/testis antigen 1A 153845865 153847522 CTAG1B cancer/testis antigen 1B 153880246 153881853 CTAG2 cancer/testis antigen 2 153903526 153979348 GAB3 hCG_18119 153991031 154005964 DKC1 dyskeratosis congenita 1, dyskerin 154006959 154033802 MPP1 hCG_18116 154064063 154250998 F8 coagulation factor VIII, procoagulant component 154113317 154113833 H2AFB1 H2A histone family, member B1 154114635 154116336 F8A1 coagulation factor VIII-associated (intronic transcript) 1 154255064 154285191 FUNDC2 FUN14 domain containing 2 154289897 154299547 MTCP1NB mature T-cell proliferation 1 neighbor 154292309 154299547 MTCP1 hCG_1640796 154299710 154351349 BRCC3 BRCA1/BRCA2-containing complex, subunit 3 154444550 154468098 VBP1 hCG_17616 154487526 154493852 RAB39B hCG_18045 154505500 154563986 CLIC2 chloride intracellular channel 2 154610436 154610783 H2AFB2 H2A histone family, member B2 154611764 154612879 F8A2 factor VIII intron 22 protein 154687146 154688261 F8A3 factor VIII intron 22 protein 154689080 154689596 H2AFB3 H2A histone family, member B3 154718672 154842622 TMLHE hCG_17611 154997451 155012117 SPRY3 hCG_1796314 155110943 155173433 VAMP7 hCG_19055 155227246 155240482 IL9R hCG_1993757 155244602 155246576 hCG_1742852 hCG_1742852 155255326 155257848 DDX11L16 DEAD/H (Asp-Glu-Ala-Asp/His) box polypeptide 11 like 16
196 Appendix C NCBI Build 37.2 Gene Map DXS8043 - DXS297
Start (bp) Stop (bp) Gene 144258102 144258421 CYCSP44 cytochrome c, somatic pseudogene 44 144329107 144337728 SPANXN1 SPANX family, member N1 TRM1 tRNA methyltransferase 1 homolog (S. 144711637 144713958 LOC347422 cerevisiae) pseudogene 144899347 144907360 SLITRK2 SLIT and NTRK-like family, member 2 144899861 144901849 LOC100506096 hypothetical protein LOC100506096 144908928 144911374 CXorf1 chromosome X open reading frame 1 145075793 145075869 MIR890 microRNA 890 145076302 145076378 MIR888 microRNA 888 145078187 145078261 MIR892A microRNA 892a 145078716 145078792 MIR892B microRNA 892b 145082571 145082649 MIR891B microRNA 891b 145109312 145109390 MIR891A microRNA 891a elongation factor, RNA polymerase II, 2 145636674 145637284 LOC100128690 pseudogene 145701108 145702280 LOC100419906 ankyrin repeat domain 11 pseudogene 145891308 145891927 LOC100133053 hypothetical LOC100133053 145895624 145896249 CXorf51 chromosome X open reading frame 51
197 Appendix D Xq27 and Xq28 Candidate Genes
Xq27 Candidate Genes Gene Product Function # SNPs analysed SLITRK2 SLIT and NTRK-like family, member 2 May play a role in modulating neurite activity 2 CXorf1 chromosome X open reading frame 1 Exact function is unknown however is expressed in the hippocampus 2 Xq28 Candidate Genes Gene Product Function # SNPs analysed GRP50 Orphan G-protein coupled receptor 50 Expressed in the hypothalamus and the pituitary and has shown positive 1 association between bipolar affective disorder and major depressive disorder in women. CNGA2 Cyclic nucleotide gated ion channel Expressed in the middle cerebral and basilar arteries and the trigeminal 3 ganglion. Rat studies have shown that cyclic nucleotide gated channels affect membrane potential suggesting that they may have a role in regulating excitability in the CNS. GABRA3 Gamma-aminobutyric acid (GABA) A receptor, alpha 3 GABA, the major inhibitory neurotransmitter in the vertebrate brain 9 NSDHL NAD(P) Dependant Steroid Dehydrogenase-like Involved in cholesterol biosynthesis 7 ATP2B3 ATPase Ca2+ transporting plasma membrane 3 Brain specific receptor that potentially has a role in calcium homeostasis 5 (PMCA3) and signalling. ABCD1 ATP-binding cassette, sub-family D (ALD), member 1 Involved in peroxisomal import of fatty acids and/or fatty acyl-CoAs in 1 the organelle FLNA filamin A, alpha (actin binding protein 280) This actin-binding protein interacts with integrins, transmembrane 1 receptor complexes, and second messengers GDI1 GDP dissociation inhibitor 1 Primarily expressed in neural and sensory tissues, GDI1 Regulates the 1 GDP/GTP exchange reaction of most Rab proteins CLIC2 Chloride Intracellular Channel 2 Voltage-gated chloride channel activity, that may modulate Ryanodine 2 Receptor (RyR) calcium release channels
198 Appendix E PED files MF6 MF7 MF14 MF47 MF55 MF541 MF878 MF879
199 Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF6 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF6 2 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF6 3 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF6 4 1 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF6 5 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF6 6 1 2 2 2 2 2 0 0 2 8 3 3 1 2 1 2 2 2 2 3 9 2 3 3 0 0 MF6 7 1 2 1 2 2 2 6 6 2 2 3 3 2 2 1 1 2 2 3 3 2 2 3 3 1 1 MF6 8 1 2 2 2 2 2 0 0 2 8 3 3 2 2 1 2 0 0 2 2 9 5 0 0 0 0 MF6 9 4 3 2 2 2 4 0 0 7 7 2 3 7 1 1 1 1 2 3 3 0 0 0 0 0 0 MF6 10 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF6 11 5 6 2 2 1 2 6 4 2 2 0 0 1 1 1 5 0 0 2 3 2 2 0 0 0 0 MF6 12 5 6 2 2 1 2 1 4 2 8 3 4 1 2 2 5 2 3 2 3 2 2 3 3 0 0 MF6 13 10 11 2 2 2 2 6 4 2 2 2 3 1 1 1 2 2 2 3 3 2 2 3 3 0 0
200 Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF7 1 0 0 2 1 2 2 3 7 0 0 2 7 3 4 4 4 1 6 0 0 2 7 4 6 1 3 MF7 2 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF7 3 2 1 2 2 2 2 3 7 2 7 0 0 6 4 1 4 4 1 1 5 4 7 2 4 6 1 MF7 4 2 1 2 0 2 2 3 7 0 0 3 7 6 3 1 4 4 6 0 0 4 2 2 6 6 3 MF7 5 0 0 1 2 2 2 5 5 2 2 8 8 0 0 4 4 4 4 4 4 4 4 0 0 1 1 MF7 6 5 3 2 1 2 2 5 3 2 2 3 8 4 6 1 4 4 4 4 5 4 4 0 0 0 0 MF7 7 5 3 2 1 2 2 5 3 2 2 3 8 4 6 1 4 4 4 4 5 4 4 2 4 1 6 MF7 8 0 0 1 1 2 2 3 3 2 2 4 4 0 0 4 4 2 2 1 1 5 5 0 0 0 0 MF7 9 5 3 2 2 2 2 5 3 2 2 3 8 4 6 1 4 4 4 4 5 4 4 2 4 1 6 MF7 10 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF7 11 5 3 2 1 2 2 5 7 2 2 0 0 4 6 1 4 4 4 4 5 4 4 0 0 0 0 MF7 12 0 0 1 2 2 2 7 7 2 2 8 8 6 6 1 1 4 4 4 4 5 5 0 0 2 2 MF7 13 5 3 1 2 2 2 7 7 2 2 3 3 6 6 1 1 4 4 5 5 4 4 2 2 0 0 MF7 14 0 0 2 2 2 5 3 4 2 2 5 5 0 0 1 4 3 3 2 5 4 5 2 4 2 3 MF7 15 5 3 2 1 2 2 5 3 2 7 2 8 4 4 4 4 4 1 1 4 4 7 0 0 0 0 MF7 16 0 0 1 2 2 2 3 3 2 2 0 0 5 5 1 1 4 4 1 1 7 7 0 0 2 2 MF7 17 5 3 2 2 2 2 5 7 2 2 3 8 4 6 1 4 4 4 4 5 4 4 2 4 1 6 MF7 18 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF7 19 8 7 1 1 2 2 5 5 2 2 8 8 4 4 1 1 4 4 4 4 4 4 4 4 1 1 MF7 20 8 7 2 1 2 2 3 3 2 2 0 0 3 4 1 4 2 4 1 5 5 4 4 4 0 0 MF7 21 10 9 2 2 2 2 6 3 2 2 3 5 4 6 1 1 2 4 3 5 7 4 2 2 3 6 MF7 22 12 11 1 0 2 2 7 7 2 2 0 0 6 6 1 1 4 4 4 4 0 0 4 4 1 1 MF7 23 12 11 2 0 2 2 7 7 2 2 3 8 6 6 1 1 4 4 4 4 5 4 0 0 2 1 MF7 24 13 14 2 2 2 5 7 4 2 2 3 5 6 3 1 4 4 3 5 5 4 5 2 4 6 2 MF7 25 13 14 2 2 2 5 7 4 2 2 3 5 6 3 1 4 4 3 2 5 4 4 2 2 6 3 MF7 26 16 15 1 0 2 2 3 3 2 2 2 2 4 4 4 4 1 1 1 1 7 7 4 4 1 1 MF7 27 18 17 1 2 2 2 7 7 2 2 0 0 6 6 4 4 4 4 5 5 4 4 0 0 6 6 MF7 28 18 17 1 1 2 2 7 7 2 2 0 0 6 6 1 1 4 4 4 4 4 4 0 0 0 0
201
Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF14 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 2 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 3 1 2 1 1 2 2 6 6 2 2 0 0 4 4 1 1 3 3 4 4 5 5 4 4 2 2 MF14 4 0 0 2 1 2 2 3 7 2 2 0 0 4 5 4 6 4 4 3 3 3 7 0 0 2 6 MF14 5 1 2 1 2 2 2 6 6 2 2 5 5 4 4 1 1 3 3 4 4 5 5 4 4 0 0 MF14 6 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 7 1 2 2 2 1 2 3 6 2 2 0 0 4 5 1 6 3 6 4 4 5 5 0 0 2 2 MF14 8 3 4 2 2 2 2 6 7 2 2 0 0 4 5 1 4 3 4 3 4 5 7 0 0 0 0 MF14 9 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 12 5 6 2 1 2 2 6 3 2 7 5 5 4 3 1 6 3 3 4 4 5 7 2 4 2 2 MF14 13 0 0 1 1 2 2 3 3 2 2 2 2 3 3 6 6 2 2 5 5 4 4 4 4 0 0 MF14 15 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 16 15 7 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 17 0 0 2 0 0 0 0 0 0 0 2 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF14 18 15 7 1 2 2 2 6 6 2 2 5 5 4 4 1 1 3 3 4 4 5 5 4 4 2 2 MF14 19 0 0 2 1 2 2 3 11 6 7 2 6 3 4 4 6 4 5 2 4 5 7 0 0 2 6 MF14 20 9 8 2 2 2 2 4 6 1 2 5 5 4 4 1 1 3 4 3 4 3 5 4 4 2 6 MF14 26 13 12 2 2 2 2 3 6 2 2 2 5 3 4 1 6 2 3 4 5 4 5 4 4 0 0 MF14 28 16 17 2 2 2 2 6 6 2 4 2 5 3 4 1 1 3 4 4 4 5 7 1 4 2 6 MF14 30 18 19 2 2 2 2 6 11 2 6 2 5 4 3 1 6 3 5 4 4 5 5 2 4 2 6 MF14 31 18 19 2 2 2 2 6 3 2 6 2 5 4 3 1 6 3 5 4 4 0 0 2 4 2 6 MF14 32 18 19 2 2 2 2 0 0 2 6 2 5 4 3 1 6 0 0 4 4 0 0 4 3 0 0 MF14 36 18 19 1 2 2 2 3 3 7 7 0 0 4 4 4 4 4 4 2 2 7 7 0 0 0 0
202
Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF47 1 0 0 1 1 2 2 2 2 2 2 5 5 8 8 4 4 2 2 2 2 9 9 5 5 1 1 MF47 2 0 0 2 2 2 2 0 0 2 8 1 5 1 2 1 5 2 4 1 4 8 1 3 4 3 4 MF47 3 1 2 1 1 0 0 0 0 0 0 1 1 0 0 1 1 4 4 0 0 0 0 4 4 3 3 MF47 4 1 2 1 2 2 2 0 0 2 2 1 1 1 1 5 5 2 2 4 4 1 1 3 3 4 4 MF47 5 1 2 2 2 2 2 1 2 2 8 5 5 8 2 4 5 2 2 2 4 9 1 3 5 1 4 MF47 6 0 0 1 0 0 0 4 4 0 0 1 1 0 0 3 3 3 3 0 0 0 0 1 1 4 4 MF47 7 1 2 2 2 2 2 2 4 2 2 0 0 8 1 4 5 2 2 2 4 0 0 3 5 1 4 MF47 8 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF47 9 6 5 1 2 2 2 1 1 8 8 5 5 2 2 5 5 2 2 4 4 1 1 3 3 4 4 MF47 10 8 7 1 2 2 2 2 2 2 2 1 1 1 1 5 5 2 2 4 4 0 0 3 3 4 4 MF47 11 8 7 2 2 1 2 1 4 2 2 1 3 0 0 0 0 2 2 0 0 8 1 2 3 4 4
Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF55 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF55 2 0 0 2 2 2 2 8 7 2 1 4 1 1 3 7 5 0 0 5 2 0 0 6 1 1 4 MF55 3 1 2 1 2 2 2 7 7 2 2 1 1 1 1 7 7 0 0 5 5 3 3 1 1 4 4 MF55 4 1 2 2 2 2 2 8 7 5 2 7 1 1 4 4 7 0 0 5 5 3 4 3 1 1 4 MF55 5 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF55 6 1 2 2 2 2 2 8 8 5 1 7 1 3 4 4 5 0 0 5 5 2 4 0 0 1 1 MF55 7 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF55 8 0 0 1 0 2 2 8 8 5 5 7 7 4 4 5 5 0 0 5 5 4 4 6 6 4 4 MF55 9 5 6 2 2 2 2 4 8 2 1 4 1 0 0 5 4 0 0 2 5 2 4 3 6 0 0 MF55 10 7 6 1 2 2 2 8 8 0 0 0 0 4 4 0 0 0 0 5 5 0 0 3 3 0 0 MF55 11 8 9 1 0 2 2 8 8 1 1 1 1 0 0 0 0 0 0 5 5 0 0 0 0 0 0 MF55 12 8 9 1 2 2 2 8 8 1 1 1 1 3 3 5 5 0 0 5 5 2 2 6 6 1 1 MF55 13 8 9 1 2 2 2 8 8 1 1 1 1 0 0 5 5 0 0 5 5 4 4 6 6 0 0
203
Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF541 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 2 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 3 1 2 1 1 2 2 1 1 3 3 1 1 2 2 1 1 0 0 2 2 3 3 0 0 5 5 MF541 4 1 2 1 1 2 2 1 1 7 7 3 3 4 4 2 2 3 3 5 5 7 7 3 3 4 4 MF541 5 0 0 1 1 2 2 2 2 0 0 4 4 3 3 1 1 0 0 2 2 2 2 4 4 3 3 MF541 6 1 2 2 2 2 3 1 4 3 3 1 4 1 2 1 4 0 0 2 6 3 3 0 0 3 5 MF541 7 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 8 1 2 2 2 2 3 1 4 3 3 1 4 1 2 1 4 0 0 2 6 3 3 2 2 3 5 MF541 9 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 10 5 6 2 2 2 2 1 2 3 3 1 4 2 3 1 1 2 4 2 6 2 3 2 4 3 3 MF541 11 5 6 1 2 2 2 1 1 3 3 1 1 2 2 1 1 2 2 2 2 3 3 2 2 5 5 MF541 12 0 0 1 1 2 2 2 2 4 4 5 5 4 4 3 3 0 0 4 4 2 2 0 0 4 4 MF541 13 5 6 2 2 2 2 2 4 3 3 4 4 1 3 1 4 3 4 2 6 2 3 2 4 3 3 MF541 14 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 15 5 6 2 2 2 2 1 2 3 3 1 4 2 3 1 4 0 0 2 6 2 3 2 4 3 3 MF541 16 7 8 1 1 0 0 4 4 3 3 1 1 2 2 1 1 2 2 2 2 3 3 2 2 5 5 MF541 17 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 18 7 8 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF541 19 7 8 2 2 2 3 5 4 3 3 1 5 1 2 1 5 2 3 2 2 2 3 2 2 0 0 MF541 20 9 10 2 1 2 2 1 2 0 0 2 4 3 4 1 4 2 4 2 5 2 2 3 4 2 3 MF541 21 9 10 1 2 2 2 2 2 0 0 1 1 2 2 1 1 2 2 6 6 3 3 0 0 3 3 MF541 22 9 10 1 1 2 2 2 2 0 0 4 4 2 2 1 1 2 2 6 6 3 3 0 0 3 3 MF541 23 9 10 1 2 2 2 1 1 3 3 4 4 3 3 1 1 4 4 2 2 2 2 4 4 3 3 MF541 24 12 13 1 2 2 2 4 4 0 0 4 4 3 3 1 1 0 0 2 2 2 2 4 4 3 3 MF541 25 12 13 2 1 2 2 2 4 3 4 4 5 3 4 1 3 3 4 2 4 2 2 1 4 3 4 MF541 26 12 13 1 1 2 2 4 4 3 3 4 4 1 1 4 4 4 4 6 6 3 3 2 2 3 3 MF541 27 14 15 1 2 2 2 1 1 3 3 1 1 2 2 4 4 4 4 6 6 0 0 2 2 3 3 MF541 28 16 17 2 1 2 3 4 4 3 3 1 5 2 5 1 4 2 2 2 5 3 4 2 3 4 5 MF541 29 16 17 1 1 4 4 1 1 6 6 1 1 4 4 4 4 2 2 7 7 3 3 2 2 1 1
204 Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF878 1 0 0 1 1 2 2 1 1 7 7 5 5 3 3 3 3 4 4 3 3 2 2 2 2 2 2 MF878 2 0 0 2 1 2 3 1 4 3 3 3 5 3 4 1 4 2 3 4 5 3 4 3 3 3 3 MF878 3 27 28 2 2 2 2 1 2 3 7 1 5 2 4 1 2 2 3 5 5 2 3 3 3 2 2 MF878 4 27 28 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 5 29 30 2 2 2 2 1 1 0 0 5 2 4 4 1 3 2 3 4 5 3 3 0 0 1 3 MF878 6 29 30 1 1 2 2 1 1 0 0 5 5 2 2 4 4 2 2 4 4 3 3 0 0 3 3 MF878 7 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 8 1 2 1 1 2 2 1 1 3 3 5 5 3 3 1 1 2 2 4 4 4 4 3 3 3 3 MF878 9 4 5 2 2 2 2 1 1 3 7 2 5 4 4 2 3 2 2 4 5 2 3 1 3 1 2 MF878 10 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 11 4 5 2 2 2 2 1 1 3 7 2 5 4 4 2 3 2 2 5 5 2 3 3 3 2 3 MF878 12 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 13 4 5 2 2 2 2 1 1 0 0 5 5 0 0 1 2 2 3 5 5 2 3 0 0 2 3 MF878 14 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 15 4 5 2 2 2 2 1 1 3 7 5 5 4 4 1 2 2 3 4 5 2 3 1 3 1 2 MF878 16 0 0 1 2 1 1 2 2 0 0 5 5 4 4 4 4 1 1 2 2 2 2 2 2 1 1 MF878 17 6 7 2 1 2 2 1 1 1 3 1 5 2 4 1 4 0 0 4 4 3 4 0 0 0 0 MF878 18 6 7 2 1 2 2 1 1 1 3 1 5 2 4 1 4 2 2 2 4 1 3 0 0 2 3 MF878 19 8 9 1 2 2 2 1 1 3 3 2 2 4 4 3 3 2 2 4 4 3 3 1 1 1 1 MF878 20 8 9 1 2 2 2 1 1 3 3 2 2 4 4 3 3 2 2 4 4 3 3 1 1 1 1 MF878 21 8 9 2 2 2 2 1 1 3 7 5 5 3 4 1 2 2 2 4 5 2 4 0 0 2 3 MF878 22 10 11 1 2 2 2 1 1 3 3 2 2 4 4 3 3 2 2 5 5 2 2 3 3 2 2 MF878 23 12 13 2 1 2 3 1 1 3 7 4 5 4 4 2 2 1 2 5 5 3 4 3 3 3 3 MF878 24 12 13 2 1 2 3 1 1 3 3 4 5 4 4 1 2 1 3 5 5 3 4 3 3 3 3 MF878 25 14 15 1 1 2 2 1 1 3 3 5 5 4 4 1 1 3 3 4 4 3 3 1 1 1 1 MF878 26 16 15 1 1 2 2 1 1 7 7 5 5 0 0 2 2 2 2 4 4 3 3 1 1 0 0 MF878 27 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 28 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 29 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF878 30 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
205
Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108
MF878 31 29 30 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Family Individual Paternal Maternal Migraine ID ID ID ID Sex Status DXS1206 DXS984 DXS8106 DXS8043 DXS297 DXS8091 DXS1123 DXS8061 DXS15 DXS1073 DXS1108 MF879 1 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF879 2 0 0 2 1 2 3 1 2 6 7 3 4 2 4 1 5 1 4 0 0 0 0 2 3 1 4 MF879 3 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF879 4 1 2 2 2 2 2 2 4 3 7 1 3 3 4 2 5 2 4 3 4 3 4 2 3 1 4 MF879 5 1 2 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF879 6 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF879 7 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MF879 8 3 4 2 2 2 2 2 3 2 7 3 4 0 0 4 5 1 4 3 4 2 4 3 3 4 6 MF879 9 3 4 1 2 2 2 2 2 7 7 3 3 4 4 5 5 4 4 3 3 4 4 3 3 4 4 MF879 10 3 4 2 2 2 2 2 3 2 7 0 0 0 0 4 5 1 4 4 4 0 0 0 0 4 6 MF879 11 0 0 1 1 2 2 3 3 5 5 6 6 4 4 1 1 2 2 1 1 3 3 3 3 4 4 MF879 12 3 4 2 2 2 2 2 3 2 7 3 4 3 4 4 5 1 4 3 4 2 4 3 3 4 6 MF879 13 5 6 2 1 2 2 1 1 3 6 2 3 3 4 5 5 2 4 4 5 4 6 3 3 1 4 MF879 14 5 6 2 2 2 2 1 1 0 0 2 3 3 4 5 5 2 4 4 5 4 7 2 3 1 4 MF879 15 7 8 2 2 1 2 2 3 2 7 3 4 3 4 4 4 1 2 2 4 2 2 3 3 4 6 MF879 16 11 12 1 2 2 2 0 0 7 7 3 3 0 0 5 5 4 4 3 3 4 4 3 3 4 4 MF879 17 11 12 1 1 2 2 2 2 7 7 3 3 4 4 5 5 4 4 3 3 4 4 3 3 4 4 MF879 18 11 12 2 2 2 2 3 3 0 0 4 6 4 4 1 4 1 2 0 0 2 3 3 3 4 6 MF879 19 11 12 2 2 2 2 2 3 0 0 3 6 3 4 1 4 1 2 1 4 2 3 3 3 4 6
206 Appendix F Individual Analysis of Migraine Families Column 1: Position (cM) Column 2: Weighted NPL score Column 3: Zlr (Zlr = sign (dhat)*sqrt(2.0*ln(10.0)*LOD)) Column 4: Maximised LOD* for the allele-sharing model selected (LOD) Column 5: delta which produced the maximised lodscore
MF6 DXS1206 1.49999 1.74449 0.660834 1.4016 2.761 1.27887 1.55351 0.524063 1.1894 5.523 1.07588 1.29377 0.363469 0.932299 8.284 0.88757 0.959644 0.199974 0.643166 11.046 0.710998 0.671315 0.0978604 0.430869 DXS984 0.543686 0.470418 0.0480533 0.302521 15.724 0.449721 0.402867 0.0352434 0.275179 17.641 0.357172 0.328576 0.0234437 0.239087 19.558 0.265424 0.248528 0.0134124 0.193212 21.475 0.173838 0.164032 0.00584266 0.136544 DXS8106 0.0817554 0.0768512 0.0012825 0.0680734 24.434 0.048281 0.0470283 0.000480256 0.0440365 25.475 0.0141654 0.0143421 4.46662E-05 0.014206 26.517 -0.0205406 -0.0217107 0.000102353 -0.0235205 27.558 -0.0557884 -0.0619225 0.000832628 -0.0737705 DXS8043 -0.091531 -0.107641 0.00251599 -0.145354 29.441 -0.157629 -0.211413 0.0097055 -0.406735 30.282 -0.223211 -0.375301 0.0305854 -1.54848 31.123 -0.288236 -0.647579 0.0910625 -5 31.964 -0.352659 -0.8347 0.151292 -5 DXS297 -0.416434 -0.970808 0.204654 -5 33.354 -0.368556 -0.894002 0.173553 -5 33.903 -0.320164 -0.803823 0.140306 -5 34.451 -0.271265 -0.693839 0.104537 -5 35.000 -0.221861 -0.550355 0.0657719 -5 DXS8091 -0.171956 -0.328226 0.0233938 -5 35.884 -0.124535 -0.171033 0.00635209 -0.333421 36.218 -0.0749485 -0.0945973 0.00194317 -0.138683 36.553 -0.0231912 -0.0275301 0.000164578 -0.0340608 36.887 0.0307429 0.0347413 0.000262087 0.0376783 DXS1123 0.0868592 0.094125 0.00192382 0.0923032 39.314 0.309145 0.304491 0.0201327 0.236948 41.407 0.556446 0.497834 0.0538176 0.329395 43.499 0.828587 0.664293 0.095824 0.38616 45.592 1.12519 0.794882 0.137202 0.411927 DXS8061 1.44554 0.887136 0.170897 0.411422 47.796 1.44255 0.886105 0.1705 0.411266 47.907 1.43956 0.885057 0.170097 0.411073 48.019 1.43659 0.883994 0.169689 0.410821 48.130 1.43363 0.882914 0.169274 0.410605 DXS15 1.43068 0.881818 0.168854 0.410353 48.812 1.42542 0.922253 0.184695 0.444677 49.383 1.41992 0.966772 0.202956 0.484075 49.953 1.41423 1.01588 0.224097 0.529578 50.524 1.40838 1.06978 0.248508 0.582339 DXS1073 1.40241 1.12802 0.276303 0.642914 51.189 1.39903 1.13699 0.280718 0.653454 51.285 1.39565 1.14603 0.285197 0.664187 51.380 1.39226 1.15511 0.289735 0.675136 51.475 1.38888 1.16423 0.294327 0.686181 DXS1108 1.38549 1.17337 0.298967 0.697442
207 MF7 0 -0.0650662 1.05084 0.239788 5 2.7614 -0.106232 0.918971 0.183382 5 5.5228 -0.156041 0.683019 0.101303 5 8.2842 -0.216319 -0.414593 0.0373248 -0.961877 11.0456 -0.289261 -0.4446 0.0429233 -0.896556 13.807 -0.377518 -0.477709 0.0495542 -0.837243 15.724 0.257616 1.59338 0.551306 5 17.641 0.947932 1.99666 0.865689 5 19.558 1.69924 2.20581 1.05655 5 21.475 2.51804 2.34851 1.19767 5 23.392 3.41157 2.45821 1.31218 5 24.4336 3.40818 2.45803 1.31198 5 25.4752 3.40855 2.45845 1.31243 5 26.5168 3.41267 2.45947 1.31352 5 27.5584 3.42057 2.46109 1.31525 5 28.6 3.43226 2.46331 1.31763 5 29.441 3.43112 2.46391 1.31827 5 30.282 3.43244 2.46491 1.31933 5 31.123 3.43624 2.46629 1.32082 5 31.964 3.44251 2.46807 1.32272 5 32.805 3.45128 2.47024 1.32505 5 33.3538 2.90561 2.37802 1.22796 5 33.9026 2.36114 2.2538 1.10302 5 34.4514 1.8174 2.06626 0.927098 5 35.0002 1.27395 1.69845 0.62641 5 35.549 0.730341 0.604218 0.0792761 0.435556 35.8836 0.803227 0.715518 0.111172 0.551164 36.2182 0.876008 0.84724 0.155872 0.710879 36.5528 0.948707 1.02077 0.226259 5 36.8874 1.02135 1.25997 0.344728 5 37.222 1.09395 1.41559 0.43514 5 39.3144 1.53307 1.84935 0.742666 5 41.4068 1.96989 2.0218 0.887628 5 43.4992 2.40962 2.1364 0.991103 2.3113 45.5916 2.85738 2.22043 1.0706 2.06672 47.684 3.31825 2.28444 1.13322 1.9233 47.7956 3.318 2.28441 1.13318 1.9233 47.9072 3.31778 2.28437 1.13315 1.9233 48.0188 3.31761 2.28434 1.13312 1.9233 48.1304 3.31747 2.28432 1.1331 1.9233 48.242 3.31738 2.28429 1.13307 1.9233 48.8124 3.31537 2.28409 1.13287 1.92371 49.3829 3.31441 2.284 1.13278 1.92371 49.9533 3.3145 2.28402 1.1328 1.92371 50.5238 3.31564 2.28414 1.13292 1.92371 51.0942 3.31782 2.28437 1.13315 1.9233 51.1894 3.31785 2.28438 1.13316 1.9233 51.2846 3.3179 2.28439 1.13317 1.9233 51.3798 3.31799 2.28441 1.13319 1.9233 51.475 3.31811 2.28442 1.1332 1.9233 51.5702 3.31825 2.28444 1.13322 1.9233
208 MF14 0 5.7445 2.66147 1.53815 0.904957 2.7614 5.87541 2.66178 1.5385 0.890197 5.5228 6.06514 2.66463 1.5418 0.875475 8.2842 6.31902 2.66995 1.54796 0.860716 11.0456 6.64417 2.67764 1.55689 0.846053 13.807 7.04977 2.68762 1.56852 0.831487 15.724 6.86237 2.66486 1.54206 0.831799 17.641 6.70142 2.63573 1.50853 0.824564 19.558 6.56416 2.59813 1.4658 0.807353 21.475 6.44791 2.54927 1.4112 0.779001 23.392 6.35004 2.4855 1.34147 0.739291 24.4336 6.35212 2.48741 1.34353 0.74071 25.4752 6.3627 2.48939 1.34568 0.741683 26.5168 6.38184 2.49146 1.34792 0.742404 27.5584 6.40961 2.49362 1.35025 0.742753 28.6 6.44611 2.49585 1.35267 0.742693 29.441 6.44121 2.4954 1.35218 0.74256 30.282 6.44188 2.495 1.35174 0.742092 31.123 6.44811 2.49465 1.35137 0.74149 31.964 6.45993 2.49435 1.35104 0.740577 32.805 6.47734 2.4941 1.35077 0.739447 33.3538 6.47158 2.494 1.35066 0.7397 33.9026 6.46819 2.49391 1.35057 0.73976 34.4514 6.46716 2.49385 1.3505 0.73976 35.0002 6.46851 2.49381 1.35046 0.7397 35.549 6.47222 2.49379 1.35043 0.739447 35.8836 6.4645 2.49339 1.35 0.739544 36.2182 6.45766 2.493 1.34958 0.739603 36.5528 6.45171 2.49262 1.34917 0.739603 36.8874 6.44663 2.49224 1.34876 0.739544 37.222 6.44244 2.49187 1.34836 0.739544 39.3144 6.33986 2.48888 1.34512 0.742344 41.4068 6.27307 2.48622 1.34225 0.743822 43.4992 6.24107 2.48393 1.33978 0.743666 45.5916 6.2434 2.48202 1.33772 0.742188 47.684 6.28002 2.48045 1.33603 0.739195 47.7956 6.29627 2.48627 1.34231 0.741646 47.9072 6.31261 2.49201 1.34851 0.744038 48.0188 6.32905 2.49768 1.35465 0.74643 48.1304 6.34558 2.50328 1.36073 0.748762 48.242 6.36221 2.50881 1.36675 0.750998 48.8124 6.43973 2.54342 1.40472 0.768929 49.3829 6.51691 2.57347 1.43812 0.783532 49.9533 6.59392 2.59959 1.46745 0.795238 50.5238 6.67093 2.62229 1.49319 0.804108 51.0942 6.74808 2.64202 1.51575 0.810406 51.1894 6.71146 2.63987 1.51328 0.810562 51.2846 6.67506 2.63772 1.51082 0.810718 51.3798 6.63889 2.63557 1.50835 0.810874 51.475 6.60295 2.63342 1.50589 0.81103 51.5702 6.56722 2.63127 1.50344 0.811127
209 MF47 0 -0.267962 -0.969458 0.204085 -5 2.7614 -0.304079 -1.0233 0.227384 -5 5.5228 -0.343243 -1.07953 0.253062 -5 8.2842 -0.385207 -1.13812 0.281275 -5 11.0456 -0.429488 -1.199 0.312172 -5 13.807 -0.475264 -1.2621 0.34589 -5 15.724 -0.428093 -1.17084 0.297679 -5 17.641 -0.380378 -1.06077 0.244339 -5 19.558 -0.334543 -0.920229 0.183885 -5 21.475 -0.292999 -0.722295 0.113288 -5 23.392 -0.258207 -0.41524 0.0374415 -1.61233 24.4336 -0.231964 -0.332848 0.0240573 -0.867386 25.4752 -0.206437 -0.27377 0.0162752 -0.546692 26.5168 -0.181805 -0.227084 0.0111977 -0.378479 27.5584 -0.158252 -0.187959 0.00767154 -0.275179 28.6 -0.135969 -0.154295 0.00516958 -0.204666 29.441 -0.134226 -0.153132 0.005092 -0.203596 30.282 -0.133546 -0.152808 0.00507045 -0.203752 31.123 -0.133923 -0.153325 0.00510481 -0.204666 31.964 -0.135356 -0.15468 0.00519547 -0.206396 32.805 -0.137847 -0.156865 0.00534326 -0.209197 33.3538 0.697588 1.78451 0.691503 5 33.9026 1.64342 2.17031 1.02282 5 34.4514 2.70035 2.37741 1.22733 5 35.0002 3.86925 2.52009 1.37907 5 35.549 5.1511 2.62945 1.50136 5 35.8836 5.14607 2.62904 1.50088 5 36.2182 5.14168 2.62867 1.50047 5 36.5528 5.13794 2.62835 1.5001 5 36.8874 5.13484 2.62809 1.4998 5 37.222 5.13237 2.62787 1.49956 5 39.3144 5.09211 2.62476 1.49601 5 41.4068 5.07698 2.62366 1.49475 5 43.4992 5.08678 2.62456 1.49578 5 45.5916 5.12166 2.62747 1.49909 5 47.684 5.18207 2.63237 1.50469 5 47.7956 5.17854 2.63214 1.50443 5 47.9072 5.17509 2.63192 1.50418 5 48.0188 5.17171 2.6317 1.50393 5 48.1304 5.16841 2.63149 1.50369 5 48.242 5.16517 2.63128 1.50345 5 48.8124 5.16563 2.63128 1.50345 5 49.3829 5.16798 2.63143 1.50362 5 49.9533 5.1722 2.63173 1.50396 5 50.5238 5.17832 2.63217 1.50447 5 51.0942 5.18633 2.63277 1.50515 5 51.1894 5.18623 2.63276 1.50514 5 51.2846 5.18617 2.63276 1.50514 5 51.3798 5.18617 2.63276 1.50514 5 51.475 5.18623 2.63276 1.50514 5 51.5702 5.18633 2.63277 1.50515 5
210 MF55 0 0.087305 0.275428 0.0164729 0.57067 2.7614 0.0558518 0.117121 0.0029787 0.200543 5.5228 0.016779 0.0292512 0.000185798 0.048315 8.2842 -0.0310833 -0.0482392 0.000505305 -0.0825801 11.0456 -0.0889729 -0.128294 0.00357409 -0.245192 13.807 -0.158145 -0.220166 0.0105258 -0.518028 15.724 -0.162892 -0.235083 0.0120004 -0.607327 17.641 -0.170255 -0.255472 0.0141724 -0.73089 19.558 -0.180171 -0.281937 0.0172607 -0.892842 21.475 -0.192582 -0.314778 0.0215161 -1.0917 23.392 -0.207436 -0.353862 0.0271908 -1.32269 24.4336 0.114018 0.247626 0.0133152 0.367278 25.4752 0.417556 1.02359 0.227512 1.40506 26.5168 0.703174 1.26189 0.345779 1.44065 27.5584 0.970698 1.35965 0.401431 1.3175 28.6 1.21977 1.37841 0.412581 1.13528 29.441 0.883042 1.0238 0.227608 0.780071 30.282 0.569288 0.605108 0.0795098 0.438861 31.123 0.27791 0.291241 0.0184188 0.234652 31.964 0.00840383 0.00995965 2.15398E-05 0.0115617 32.805 -0.239641 -0.453194 0.0445988 -1.94315 33.3538 -0.239026 -0.451516 0.044269 -1.93236 33.9026 -0.238546 -0.450128 0.0439973 -1.92396 34.4514 -0.238202 -0.449031 0.0437832 -1.91663 35.0002 -0.237994 -0.448227 0.0436264 -1.91103 35.549 -0.237923 -0.447714 0.0435268 -1.90583 35.8836 -0.205583 -0.388234 0.0327296 -1.73589 36.2182 -0.173178 -0.319382 0.0221501 -1.49996 36.5528 -0.140709 -0.237637 0.0122626 -1.10356 36.8874 -0.108175 -0.155627 0.00525924 -0.380871 37.222 -0.075573 -0.0996394 0.00215584 -0.165617 39.3144 0.129884 0.136423 0.00404136 0.123107 41.4068 0.338282 0.314119 0.0214261 0.227068 43.4992 0.549897 0.460908 0.0461299 0.289842 45.5916 0.764924 0.579898 0.0730227 0.327977 47.684 0.983406 0.671234 0.0978367 0.345849 47.7956 0.555362 0.448092 0.0436003 0.274674 47.9072 0.208585 0.203302 0.00897502 0.163069 48.0188 -0.0572459 -0.224537 0.0109479 -5 48.1304 -0.242452 -0.80876 0.142035 -5 48.242 -0.347353 -1.06233 0.245062 -5 48.8124 -0.325209 -0.97658 0.207095 -5 49.3829 -0.303331 -0.873411 0.16565 -5 49.9533 -0.281793 -0.743406 0.120007 -5 50.5238 -0.260666 -0.576392 0.0721424 -3.64885 51.0942 -0.240021 -0.454954 0.0449457 -1.95781 51.1894 -0.239437 -0.453398 0.0446389 -1.95048 51.2846 -0.238859 -0.451855 0.0443356 -1.94422 51.3798 -0.238289 -0.450324 0.0440357 -1.93689 51.475 -0.237724 -0.448806 0.0437393 -1.93022 51.5702 -0.237166 -0.4473 0.0434463 -1.92289
211 MF541 0 1.59008 2.24722 1.09659 5 2.7614 1.29622 2.11732 0.97348 5 5.5228 0.971192 1.91707 0.798049 3.54102 8.2842 0.609326 1.55348 0.524044 5 11.0456 0.206216 0.39128 0.0332452 0.30114 13.807 -0.239738 -0.71806 0.111963 -5 15.724 -0.225005 -0.703534 0.107479 -5 17.641 -0.216585 -0.691448 0.103818 -5 19.558 -0.214778 -0.681851 0.100956 -5 21.475 -0.21996 -0.674757 0.0988665 -5 23.392 -0.232596 -0.670132 0.0975157 -5 24.4336 0.240932 -0.208504 0.00944027 -5 25.4752 0.661872 0.489931 0.0521224 0.206144 26.5168 1.03202 0.632327 0.0868236 0.223859 27.5584 1.3529 0.7169 0.111602 0.227707 28.6 1.62576 0.766095 0.127444 0.225315 29.441 2.22511 2.41628 1.2678 3.54102 30.282 2.42217 2.52837 1.38815 3.54102 31.123 2.28295 2.47005 1.32485 3.54102 31.964 1.87092 2.22502 1.07503 3.54102 32.805 1.24795 0.627103 0.0853949 0.20552 33.3538 1.01916 0.543577 0.0641617 0.192863 33.9026 0.809581 0.458767 0.0457023 0.177792 34.4514 0.619109 0.373934 0.030363 0.160113 35.0002 0.447689 0.29008 0.0182721 0.139307 35.549 0.295291 0.207701 0.00936763 0.114357 35.8836 0.295203 0.207656 0.00936365 0.114297 36.2182 0.295159 0.207634 0.00936167 0.114297 36.5528 0.295159 0.207634 0.00936167 0.114297 36.8874 0.295203 0.207656 0.00936365 0.114297 37.222 0.295291 0.207701 0.00936763 0.114357 39.3144 0.126549 0.105424 0.00241343 0.0741791 41.4068 -0.0156245 -0.0168778 6.18569E-05 -0.0191456 43.4992 -0.133481 -0.509404 0.0563481 -5 45.5916 -0.228755 -0.859475 0.160406 -5 47.684 -0.30266 -1.05146 0.24007 -5 47.7956 -0.262736 -0.946618 0.194583 -5 47.9072 -0.221012 -0.812771 0.143447 -5 48.0188 -0.177487 -0.626049 0.0851081 -5 48.1304 -0.132164 -0.281976 0.0172655 -5 48.242 -0.0850407 -0.116812 0.00296298 -0.23226 48.8124 -0.0311334 -0.0346412 0.000260579 -0.0423058 49.3829 0.0284073 0.0274119 0.000163167 0.0249987 49.9533 0.0936486 0.081202 0.00143182 0.061247 50.5238 0.164661 0.130774 0.00371361 0.0857294 51.0942 0.241518 0.177782 0.00686323 0.104166 51.1894 0.247636 0.181332 0.00714005 0.105235 51.2846 0.253747 0.18485 0.00741983 0.106773 51.3798 0.259849 0.188334 0.00770217 0.107976 51.475 0.265945 0.191786 0.00798711 0.109202 51.5702 0.272033 0.195207 0.00827455 0.110331
212 MF878 0 0.00220266 1.08786 0.256978 5 2.7614 -0.0222779 0.967886 0.203424 5 5.5228 -0.0448939 0.806436 0.141219 5 8.2842 -0.0657431 0.556252 0.0671889 5 11.0456 -0.0848293 -0.188642 0.00772735 -0.514158 13.807 -0.102045 -0.211132 0.00967974 -0.557485 15.724 -0.115532 -0.229351 0.0114224 -0.597856 17.641 -0.127782 -0.244074 0.012936 -0.62652 19.558 -0.138657 -0.255402 0.0141646 -0.644392 21.475 -0.147983 -0.263359 0.0150608 -0.650906 23.392 -0.155542 -0.267888 0.0155833 -0.647192 24.4336 -0.186762 -0.30495 0.0201935 -0.711697 25.4752 -0.213974 -0.335365 0.0244224 -0.762608 26.5168 -0.23714 -0.360488 0.0282187 -0.804204 27.5584 -0.256198 -0.381197 0.0315539 -0.837652 28.6 -0.27106 -0.398086 0.0344119 -0.865247 29.441 -0.271035 -0.398313 0.0344512 -0.866978 30.282 -0.27095 -0.398333 0.0344545 -0.867386 31.123 -0.270809 -0.398144 0.0344219 -0.866978 31.964 -0.270613 -0.397748 0.0343535 -0.865499 32.805 -0.270361 -0.397145 0.0342494 -0.863108 33.3538 -0.270725 -0.397913 0.034382 -0.865908 33.9026 -0.271066 -0.398593 0.0344995 -0.869778 34.4514 -0.271382 -0.399185 0.0346021 -0.870439 35.0002 -0.271675 -0.399689 0.0346895 -0.872578 35.549 -0.271942 -0.400104 0.0347616 -0.873239 35.8836 -0.272281 -0.400522 0.0348344 -0.8739 36.2182 -0.272603 -0.400895 0.0348993 -0.874309 36.5528 -0.27291 -0.401226 0.0349568 -0.874309 36.8874 -0.273203 -0.401516 0.0350075 -0.874309 37.222 -0.273484 -0.401769 0.0350516 -0.874309 39.3144 -0.238067 -0.356913 0.0276618 -0.793664 41.4068 -0.213383 -0.318625 0.0220452 -0.698103 43.4992 -0.197824 -0.285879 0.0177468 -0.584671 45.5916 -0.188641 -0.2578 0.0144318 -0.4719 47.684 -0.182609 -0.232714 0.0117598 -0.374201 47.7956 -0.179807 -0.228092 0.0112973 -0.363408 47.9072 -0.177006 -0.223512 0.0108482 -0.353024 48.0188 -0.174205 -0.218973 0.010412 -0.34264 48.1304 -0.171404 -0.214473 0.00998853 -0.332508 48.242 -0.168603 -0.210014 0.00957746 -0.322629 48.8124 -0.168836 -0.210422 0.00961472 -0.323698 49.3829 -0.169046 -0.210775 0.00964704 -0.324612 49.9533 -0.169234 -0.211074 0.00967438 -0.325177 50.5238 -0.169399 -0.211317 0.0096967 -0.325682 51.0942 -0.169542 -0.211505 0.00971393 -0.32609 51.1894 -0.169676 -0.211719 0.00973363 -0.326499 51.2846 -0.169809 -0.211932 0.00975322 -0.326907 51.3798 -0.169941 -0.212144 0.00977269 -0.327412 51.475 -0.170072 -0.212354 0.00979205 -0.327821 51.5702 -0.170203 -0.212562 0.00981128 -0.328229
213 MF879 0 0.634594 1.4789 0.47493 5 2.7614 0.435003 1.04075 0.235205 5 5.5228 0.259357 0.261735 0.0148757 0.173669 8.2842 0.106919 0.103469 0.00232475 0.0804408 11.0456 -0.0217658 -0.0251908 0.000137797 -0.0320776 13.807 -0.124977 -0.344607 0.0257871 -2.80193 15.724 -0.0843211 -0.14852 0.0047899 -0.603865 17.641 -0.0431264 -0.0534157 0.000619573 -0.0832412 19.558 -0.00473681 -0.00518681 5.84192E-06 -0.00580495 21.475 0.0277528 0.0282319 0.000173076 0.0267294 23.392 0.0521012 0.0498713 0.000540077 0.0427143 24.4336 0.0591638 0.055179 0.000661152 0.0456707 25.4752 0.0631291 0.0575881 0.000720144 0.0465843 26.5168 0.0640492 0.0574722 0.000717249 0.0458636 27.5584 0.0619858 0.0551239 0.000659832 0.043784 28.6 0.0569484 0.0506473 0.000557015 0.0405751 29.441 0.0630562 0.0555059 0.000669009 0.0436875 30.282 0.0691652 0.0602756 0.000788928 0.0467404 31.123 0.0753283 0.0650019 0.0009175 0.0495407 31.964 0.0815987 0.0697282 0.00105577 0.0522814 32.805 0.0880298 0.0744958 0.00120509 0.0549853 33.3538 0.0402752 0.0369844 0.000297024 0.0312605 33.9026 -0.00743138 -0.00763884 1.26709E-05 -0.00804067 34.4514 -0.0551219 -0.069318 0.00104339 -0.119489 35.0002 -0.102828 -0.240566 0.0125668 -2.05831 35.549 -0.150583 -0.587444 0.0749354 -5 35.8836 -0.136446 -0.471655 0.0483062 -5 36.2182 -0.122367 -0.336002 0.0245153 -2.88044 36.5528 -0.108339 -0.238891 0.0123924 -1.34015 36.8874 -0.0943552 -0.172035 0.00642667 -0.687972 37.222 -0.0804083 -0.123326 0.00330268 -0.361016 39.3144 -0.10472 -0.219308 0.0104439 -1.11543 41.4068 -0.122984 -0.340209 0.0251331 -2.94082 43.4992 -0.135573 -0.459057 0.0457601 -5 45.5916 -0.142712 -0.521064 0.0589571 -5 47.684 -0.144474 -0.532899 0.0616657 -5 47.7956 -0.144403 -0.532331 0.0615343 -5 47.9072 -0.144332 -0.531766 0.0614039 -5 48.0188 -0.144262 -0.531205 0.0612743 -5 48.1304 -0.144193 -0.530646 0.0611456 -5 48.242 -0.144124 -0.530091 0.0610177 -5 48.8124 -0.143777 -0.527304 0.0603777 -5 49.3829 -0.143446 -0.524616 0.0597636 -5 49.9533 -0.143133 -0.522027 0.0591753 -5 50.5238 -0.142837 -0.519539 0.0586125 -5 51.0942 -0.14256 -0.517152 0.0580752 -5 51.1894 -0.142379 -0.51569 0.0577474 -5 51.2846 -0.142197 -0.514227 0.0574202 -5 51.3798 -0.142015 -0.512763 0.0570937 -5 51.475 -0.141833 -0.511298 0.0567678 -5 51.5702 -0.14165 -0.509831 0.0564425 -5
214 Appendix G Global Analysis of Migraine Families Column 1: Position (cM) Column 2: Weighted NPL score Column 3: Zlr (Zlr = sign (dhat)*sqrt(2.0*ln(10.0)*LOD)) Column 4: Maximised LOD* for the allele-sharing model selected (LOD) Column 5: delta which produced the maximised lodscore
1. All families: 0.00000e+00 3.04774e+00 4.26262e+00 3.94555e+00 5.00000e+00 2.76140e+00 2.78125e+00 3.86158e+00 3.23805e+00 5.00000e+00 5.52280e+00 2.53720e+00 3.25896e+00 2.30628e+00 5.00000e+00 8.28420e+00 2.31271e+00 2.46562e+00 1.32010e+00 6.16545e-01 1.10456e+01 2.10650e+00 1.79301e+00 6.98105e-01 3.02581e-01 1.38070e+01 1.91949e+00 1.50154e+00 4.89588e-01 2.45482e-01 1.57240e+01 1.96837e+00 1.51923e+00 5.01189e-01 2.46322e-01 1.76410e+01 2.03249e+00 1.54490e+00 5.18269e-01 2.48498e-01 1.95580e+01 2.11057e+00 1.57363e+00 5.37724e-01 2.50904e-01 2.14750e+01 2.20158e+00 1.59901e+00 5.55206e-01 2.51840e-01 2.33920e+01 2.30500e+00 1.61751e+00 5.68129e-01 2.50481e-01 2.44336e+01 2.52670e+00 1.74608e+00 6.62034e-01 2.64193e-01 2.54752e+01 2.73551e+00 1.83109e+00 7.28070e-01 2.69229e-01 2.65168e+01 2.93222e+00 1.88614e+00 7.72506e-01 2.69363e-01 2.75584e+01 3.11756e+00 1.92193e+00 8.02105e-01 2.67090e-01 2.86000e+01 3.29216e+00 1.94571e+00 8.22070e-01 2.63844e-01 2.94410e+01 3.55613e+00 2.74748e+00 1.63917e+00 3.49324e-01 3.02820e+01 3.68608e+00 2.83997e+00 1.75139e+00 3.50306e-01 3.11230e+01 3.70554e+00 2.79419e+00 1.69538e+00 3.48915e-01 3.19640e+01 3.63717e+00 2.60582e+00 1.47449e+00 3.43939e-01 3.28050e+01 3.50304e+00 1.92150e+00 8.01747e-01 2.57390e-01 3.33538e+01 3.70903e+00 2.08368e+00 9.42796e-01 2.73122e-01 3.39026e+01 3.96209e+00 2.20528e+00 1.05604e+00 2.80289e-01 3.44514e+01 4.26242e+00 2.29964e+00 1.14835e+00 2.83677e-01 3.50002e+01 4.61026e+00 2.37681e+00 1.22671e+00 2.85407e-01 3.55490e+01 5.00597e+00 2.44314e+00 1.29613e+00 2.86380e-01 3.58836e+01 5.02301e+00 2.44992e+00 1.30334e+00 2.86826e-01 3.62182e+01 5.04134e+00 2.45704e+00 1.31093e+00 2.87271e-01 3.65528e+01 5.06094e+00 2.46449e+00 1.31889e+00 2.87739e-01 3.68874e+01 5.08180e+00 2.47225e+00 1.32721e+00 2.88222e-01 3.72220e+01 5.10393e+00 2.48031e+00 1.33588e+00 2.88727e-01 3.93144e+01 4.96569e+00 2.46498e+00 1.31942e+00 2.89107e-01 4.14068e+01 4.86532e+00 2.45418e+00 1.30787e+00 2.89396e-01 4.34992e+01 4.79965e+00 2.44712e+00 1.30036e+00 2.89516e-01 4.55916e+01 4.76656e+00 2.44340e+00 1.29642e+00 2.89447e-01 4.76840e+01 4.76471e+00 2.44279e+00 1.29577e+00 2.89158e-01 4.77956e+01 4.78093e+00 2.44849e+00 1.30182e+00 2.89470e-01 4.79072e+01 4.79784e+00 2.45435e+00 1.30806e+00 2.89768e-01 4.80188e+01 4.81545e+00 2.46037e+00 1.31449e+00 2.90080e-01 4.81304e+01 4.83375e+00 2.46655e+00 1.32110e+00 2.90384e-01 4.82420e+01 4.85275e+00 2.47288e+00 1.32788e+00 2.90710e-01 4.88124e+01 4.84621e+00 2.48020e+00 1.33575e+00 2.91706e-01 4.93829e+01 4.84441e+00 2.48802e+00 1.34419e+00 2.92693e-01 4.99533e+01 4.84736e+00 2.49634e+00 1.35320e+00 2.93652e-01 5.05238e+01 4.85505e+00 2.50515e+00 1.36277e+00 2.94603e-01 5.10942e+01 4.86748e+00 2.51444e+00 1.37289e+00 2.95502e-01 5.11894e+01 4.86175e+00 2.51418e+00 1.37261e+00 2.95599e-01 5.12846e+01 4.85609e+00 2.51393e+00 1.37233e+00 2.95681e-01 5.13798e+01 4.85049e+00 2.51367e+00 1.37205e+00 2.95769e-01 5.14750e+01 4.84496e+00 2.51341e+00 1.37177e+00 2.95842e-01 5.15702e+01 4.83949e+00 2.51314e+00 1.37148e+00 2.95911e-01
215 2. Global analysis of MF7, MF14, MF47 and MF879 0.00E+00 3.20E+00 3.45E+00 2.58E+00 5.00E+00 2.76E+00 3.11E+00 3.26E+00 2.30E+00 5.00E+00 5.52E+00 3.03E+00 2.97E+00 1.91E+00 5.00E+00 8.28E+00 2.99E+00 2.70E+00 1.59E+00 7.12E-01 1.10E+01 2.96E+00 2.16E+00 1.01E+00 6.79E-01 1.38E+01 2.97E+00 1.78E+00 6.87E-01 3.07E-01 1.57E+01 3.23E+00 1.87E+00 7.58E-01 3.15E-01 1.76E+01 3.51E+00 1.96E+00 8.34E-01 3.26E-01 1.96E+01 3.81E+00 2.05E+00 9.10E-01 3.36E-01 2.15E+01 4.12E+00 2.12E+00 9.79E-01 3.43E-01 2.34E+01 4.46E+00 2.18E+00 1.04E+00 3.44E-01 2.44E+01 4.56E+00 2.20E+00 1.06E+00 3.43E-01 2.55E+01 4.67E+00 2.22E+00 1.07E+00 3.42E-01 2.65E+01 4.78E+00 2.24E+00 1.09E+00 3.39E-01 2.76E+01 4.89E+00 2.25E+00 1.10E+00 3.36E-01 2.86E+01 5.01E+00 2.26E+00 1.11E+00 3.33E-01 2.94E+01 5.26E+00 3.28E+00 2.33E+00 6.70E-01 3.03E+01 5.40E+00 3.37E+00 2.47E+00 6.72E-01 3.11E+01 5.46E+00 3.34E+00 2.42E+00 6.72E-01 3.20E+01 5.43E+00 3.18E+00 2.19E+00 6.70E-01 3.28E+01 5.32E+00 2.30E+00 1.15E+00 3.33E-01 3.34E+01 5.65E+00 2.53E+00 1.39E+00 3.62E-01 3.39E+01 6.04E+00 2.68E+00 1.55E+00 3.74E-01 3.45E+01 6.49E+00 2.78E+00 1.68E+00 3.78E-01 3.50E+01 7.00E+00 2.86E+00 1.77E+00 3.80E-01 3.55E+01 7.57E+00 2.92E+00 1.85E+00 3.81E-01 3.59E+01 7.57E+00 2.92E+00 1.85E+00 3.80E-01 3.62E+01 7.56E+00 2.92E+00 1.85E+00 3.80E-01 3.66E+01 7.56E+00 2.92E+00 1.85E+00 3.80E-01 3.69E+01 7.56E+00 2.92E+00 1.85E+00 3.80E-01 3.72E+01 7.56E+00 2.92E+00 1.85E+00 3.80E-01 3.93E+01 7.43E+00 2.90E+00 1.83E+00 3.80E-01 4.14E+01 7.34E+00 2.89E+00 1.82E+00 3.80E-01 4.35E+01 7.30E+00 2.89E+00 1.81E+00 3.79E-01 4.56E+01 7.28E+00 2.88E+00 1.80E+00 3.79E-01 4.77E+01 7.31E+00 2.88E+00 1.81E+00 3.79E-01 4.78E+01 7.31E+00 2.89E+00 1.81E+00 3.79E-01 4.79E+01 7.32E+00 2.89E+00 1.81E+00 3.80E-01 4.80E+01 7.33E+00 2.89E+00 1.82E+00 3.80E-01 4.81E+01 7.34E+00 2.89E+00 1.82E+00 3.80E-01 4.82E+01 7.35E+00 2.90E+00 1.82E+00 3.80E-01 4.88E+01 7.42E+00 2.91E+00 1.84E+00 3.81E-01 4.94E+01 7.49E+00 2.93E+00 1.86E+00 3.83E-01 5.00E+01 7.57E+00 2.94E+00 1.88E+00 3.84E-01 5.05E+01 7.65E+00 2.96E+00 1.90E+00 3.85E-01 5.11E+01 7.73E+00 2.97E+00 1.92E+00 3.86E-01 5.12E+01 7.72E+00 2.97E+00 1.92E+00 3.86E-01 5.13E+01 7.70E+00 2.97E+00 1.92E+00 3.86E-01 5.14E+01 7.69E+00 2.97E+00 1.91E+00 3.86E-01 5.15E+01 7.68E+00 2.97E+00 1.91E+00 3.86E-01 5.16E+01 7.67E+00 2.97E+00 1.91E+00 3.86E-01
216 3. Global analysis of MF7, MF14 and MF879 0 3.88316 3.91866 3.33448 5 2.7614 3.72115 3.85639 3.22937 5 5.5228 3.57196 3.76423 3.07686 5 8.2842 3.43313 3.61009 2.83002 5 11.0456 3.30371 3.2829 2.34029 5 13.807 3.18508 1.81352 0.714164 0.318483 15.724 3.31248 NaN -0.113711 5 17.641 3.45616 1.21817 0.322233 5 19.558 3.61748 1.42859 0.443172 5 21.475 3.79799 1.20292 0.314217 5 23.392 3.99949 1.97207 0.844494 0.324144 24.4336 4.39775 2.82594 1.73412 5 25.4752 4.77422 2.95382 1.89463 5 26.5168 5.13038 2.92905 1.86298 5 27.5584 5.46759 2.74386 1.63485 5 28.6 5.78708 2.40409 1.25503 0.354153 29.441 6.27156 4.05991 3.5792 1.14129 30.282 6.53094 4.12547 3.69574 1.08519 31.123 6.60368 4.08743 3.62789 1.03601 31.964 6.52676 3.94103 3.37266 0.992116 32.805 6.33623 2.4162 1.26771 0.348456 33.3538 6.19148 2.37813 1.22807 0.345752 33.9026 6.05979 2.34022 1.18924 0.34241 34.4514 5.94111 2.30439 1.1531 0.338963 35.0002 5.83538 2.27227 1.12117 0.335947 35.549 5.74259 2.24505 1.09448 0.333578 35.8836 5.74057 2.2446 1.09404 0.333555 36.2182 5.7393 2.24423 1.09368 0.333532 36.5528 5.73877 2.24395 1.0934 0.333504 36.8874 5.739 2.24375 1.0932 0.333481 37.222 5.73997 2.24362 1.09309 0.333458 39.3144 5.47015 2.19948 1.0505 0.331319 41.4068 5.24625 2.16038 1.01348 0.329299 43.4992 5.06483 2.12668 0.982106 0.327472 45.5916 4.92315 2.09872 0.956448 0.32586 47.684 4.81914 2.07676 0.936543 0.324515 47.7956 4.8384 2.08282 0.942013 0.325006 47.9072 4.85877 2.08911 0.947709 0.325489 48.0188 4.88025 2.09561 0.953622 0.325994 48.1304 4.90284 2.10233 0.959742 0.326521 48.242 4.92653 2.10923 0.96606 0.327041 48.8124 4.87212 2.10799 0.964918 0.327688 49.3829 4.82443 2.10761 0.964576 0.328326 49.9533 4.78341 2.10808 0.965005 0.329046 50.5238 4.74897 2.10936 0.966175 0.329767 51.0942 4.72109 2.11141 0.968051 0.330502 51.1894 4.71286 2.11077 0.967468 0.330562 51.2846 4.70472 2.11014 0.966885 0.330621 51.3798 4.69665 2.1095 0.966302 0.330672 51.475 4.68867 2.10886 0.965719 0.330718 51.5702 4.68076 2.10823 0.965136 0.330777
217 4. Global analysis of MF7, MF14 and MF47 0 3.38267 2.96727 1.91191 1.209 2.7614 3.42799 2.91682 1.84745 1.13651 5.5228 3.49916 2.84928 1.76289 1.04655 8.2842 3.59743 2.7575 1.65114 0.931326 11.0456 3.72464 2.63564 1.50843 0.802533 13.807 3.88335 2.48273 1.33848 0.682757 15.724 4.12047 2.9974 1.95094 1.1137 17.641 4.40046 3.13298 2.13142 1.1535 19.558 4.72439 3.13584 2.13532 1.08783 21.475 5.09365 3.03866 2.00502 0.85461 23.392 5.50996 2.89636 1.82162 0.665064 24.4336 5.51421 2.90723 1.83533 0.671987 25.4752 5.52502 2.91669 1.84729 0.677624 26.5168 5.5424 2.92462 1.85735 0.681687 27.5584 5.5664 2.9309 1.86534 0.684065 28.6 5.59706 2.93542 1.8711 0.684644 29.441 5.59544 2.93452 1.86994 0.683693 30.282 5.59805 2.93234 1.86716 0.68064 31.123 5.6049 2.92884 1.86271 0.677996 31.964 5.61597 2.92397 1.85653 0.673249 32.805 5.63126 2.91769 1.84855 0.6673 33.3538 5.64806 3.53376 2.71162 1.03966 33.9026 5.67118 3.65053 2.89379 1.04501 34.4514 5.70067 3.6638 2.91486 1.03101 35.0002 5.7366 3.5911 2.80032 0.985193 35.549 5.77901 3.34078 2.42354 0.800178 35.8836 6.03525 3.43591 2.56353 0.842339 36.2182 6.29407 3.5196 2.68993 0.879405 36.5528 6.55557 3.59322 2.80363 0.910245 36.8874 6.81984 3.65823 2.90601 0.935099 37.222 7.08697 3.71604 2.99857 0.954954 39.3144 7.26015 3.86573 3.24501 1.02374 41.4068 7.46499 3.95574 3.3979 1.04944 43.4992 7.70386 4.01874 3.50698 1.05866 45.5916 7.97951 4.06672 3.59123 1.05825 47.684 8.29508 4.10526 3.65962 1.05086 47.7956 8.30618 4.11088 3.66965 1.05495 47.9072 8.3174 4.11641 3.67952 1.05892 48.0188 8.32873 4.12183 3.68922 1.06279 48.1304 8.34018 4.12716 3.69876 1.0665 48.242 8.35175 4.1324 3.70816 1.07012 48.8124 8.43919 4.17061 3.77706 1.1057 49.3829 8.52819 4.20224 3.83456 1.1311 49.9533 8.61888 4.22844 3.88253 1.14699 50.5238 8.7114 4.25021 3.9226 1.15498 51.0942 8.80588 4.26836 3.95619 1.15677 51.1894 8.78469 4.26715 3.95393 1.15718 51.2846 8.76368 4.26594 3.95169 1.15753 51.3798 8.74285 4.26473 3.94946 1.15803 51.475 8.72219 4.26354 3.94725 1.15844 51.5702 8.70171 4.26235 3.94505 1.15891
218
REFERENCES
219 References
Aamodt AH, Stovner LJ, Langhammer A, Hagen K, Zwart JA (2007) Is headache related to asthma, hay fever, and chronic bronchitis? The Head-HUNT Study. Headache 47(2):204-212. Abd-Elsalam KA (2003) Bioinformatic tools and guidelines for PCR primer design. African Journal of Biotechnology 2(5):91-95. Abecasis GR, Cherny SS, Cookson WO, Cardon LR (2002) Merlin--rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30(1):97-101. Alders EE, Hentzen A, Tan CT (1996) A community-based prevalence study on headache in Malaysia. Headache 36(6):379-384. Almasy L, Blangero J (1998) Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 62(5):1198-1211. Altshuler D, Daly MJ, Lander ES (2008) Genetic mapping in human disease. Science 322(5903):881-888. Andreou AP, Goadsby PJ (2009) Therapeutic potential of novel glutamate receptor antagonists in migraine. Expert Opin Investig Drugs 18(6):789-803. Antonaci F, Nappi G, Galli F, Manzoni GC, Calabresi P, Costa A (2011) Migraine and psychiatric comorbidity: a review of clinical findings. J Headache Pain 12(2):115-125. Anttila V, Kallela M, Oswell G, Kaunisto MA, Nyholt DR, Hamalainen E, Havanka H, Ilmavirta M, Terwilliger J, Sobel E, Peltonen L, Kaprio J, Farkkila M, Wessman M, Palotie A (2006) Trait components provide tools to dissect the genetic susceptibility of migraine. Am J Hum Genet 79(1):85-99. Anttila V, Nyholt DR, Kallela M, Artto V, Vepsalainen S, Jakkula E, Wennerstrom A, Tikka-Kleemola P, Kaunisto MA, Hamalainen E, Widen E, Terwilliger J, Merikangas K, Montgomery GW, Martin NG, Daly M, Kaprio J, Peltonen L, Farkkila M, Wessman M, Palotie A (2008) Consistently replicating locus linked to migraine on 10q22-q23. Am J Hum Genet 82(5):1051-1063. Anttila V, Stefansson H, Kallela M, Todt U, Terwindt GM, Calafato MS, Nyholt DR, Dimas AS, Freilinger T, Muller-Myhsok B, Artto V, Inouye M, Alakurtti K, Kaunisto MA, Hamalainen E, de Vries B, Stam AH, Weller CM, Heinze A, Heinze-Kuhn K, Goebel I, Borck G, Gobel H, Steinberg S, Wolf C, Bjornsson
220 A, Gudmundsson G, Kirchmann M, Hauge A, Werge T, Schoenen J, Eriksson JG, Hagen K, Stovner L, Wichmann HE, Meitinger T, Alexander M, Moebus S, Schreiber S, Aulchenko YS, Breteler MM, Uitterlinden AG, Hofman A, van Duijn CM, Tikka-Kleemola P, Vepsalainen S, Lucae S, Tozzi F, Muglia P, Barrett J, Kaprio J, Farkkila M, Peltonen L, Stefansson K, Zwart JA, Ferrari MD, Olesen J, Daly M, Wessman M, van den Maagdenberg AM, Dichgans M, Kubisch C, Dermitzakis ET, Frants RR, Palotie A (2010) Genome-wide association study of migraine implicates a common susceptibility variant on 8q22.1. Nat Genet 42:869-873. Anzola GP, Magoni M, Guindani M, Rozzini L, Dalla Volta G (1999) Potential source of cerebral embolism in migraine with aura: a transcranial Doppler study. Neurology 52(8):1622-1625. Aruga J, Mikoshiba K (2003) Identification and characterization of Slitrk, a novel neuronal transmembrane protein family controlling neurite outgrowth. Mol Cell Neurosci 24(1):117-129. Aruga J, Yokota N, Mikoshiba K (2003) Human SLITRK family genes: genomic organization and expression profiling in normal brain and brain tumor tissue. Gene 315:87-94. Ayata C (2010) Cortical spreading depression triggers migraine attack: pro. Headache 50(4):725-730. Bag B, Hacihasanoglu R, Tufekci FG (2005) Examination of anxiety, hostility and psychiatric disorders in patients with migraine and tension-type headache. Int J Clin Pract 59(5):515-521. Bandolier (2005) Patent Foramen Ovale and Migraine. Bandolier December 2005(142):142-144. Bank J, Marton S (2000) Hungarian migraine epidemiology. Headache 40(2):164- 169. Barbanti P, Aurilia C, Egeo G, Fofi L (2011) Migraine prophylaxis: what is new and what we need? Neurol Sci 32 Suppl 1:S111-115. Barrett JC (2009) Haploview: Visualization and analysis of SNP genotype data. Cold Spring Harb Protoc 2009(10):pdb ip71. Barrett JC, Fry B, Maller J, Daly MJ (2005) Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21(2):263-265.
221 Baumforth KR, Nelson PN, Digby JE, O'Neil JD, Murray PG (1999) Demystified ... the polymerase chain reaction. Mol Pathol 52(1):1-10. Bellis C, Cox HC, Dyer TD, Charlesworth JC, Begley KN, Quinlan S, Lea RA, Heath SC, Blangero J, Griffiths LR (2008a) Linkage mapping of CVD risk traits in the isolated Norfolk Island population. Hum Genet 124(5):543-552. Bellis C, Cox HC, Ovcaric M, Begley KN, Lea RA, Quinlan S, Burgner D, Heath SC, Blangero J, Griffiths LR (2008b) Linkage disequilibrium analysis in the genetically isolated Norfolk Island population. Heredity 100(4):366-373. Bellis C, Hughes RM, Begley KN, Quinlan S, Lea RA, Heath SC, Blangero J, Griffiths LR (2005) Phenotypical characterisation of the isolated norfolk island population focusing on epidemiological indicators of cardiovascular disease. Hum Hered 60(4):211-219. Berg J (2004) Economic evidence in migraine and other headaches: a review. Eur J Health Econ 5 Suppl 1:S43-54. Bergerot A, Storer RJ, Goadsby PJ (2007) Dopamine inhibits trigeminovascular transmission in the rat. Ann Neurol 61(3):251-262. Bigal ME, Lipton RB, Cohen J, Silberstein SD (2003) Epilepsy and migraine. Epilepsy Behav 4 Suppl 2:S13-24. Bilitewski U (2009) DNA microarrays: an introduction to the technology. Methods Mol Biol 509:1-14. Bjornsson A, Gudmundsson G, Gudfinnsson E, Hrafnsdottir M, Benedikz J, Skuladottir S, Kristjansson K, Frigge ML, Kong A, Stefansson K, Gulcher JR (2003) Localization of a gene for migraine without aura to chromosome 4q21. Am J Hum Genet 73(5):986-993. Bland JM, Altman DG (1995) Multiple significance tests: the Bonferroni method. BMJ 310(6973):170. Board PG, Coggan M, Watson S, Gage PW, Dulhunty AF (2004) CLIC-2 modulates cardiac ryanodine receptor Ca2+ release channels. Int J Biochem Cell Biol 36(8):1599-1612. Bolay H, Reuter U, Dunn AK, Huang Z, Boas DA, Moskowitz MA (2002) Intrinsic brain activity triggers trigeminal meningeal afferents in a migraine model. Nat Med 8(2):136-142.
222 Borsook D, Burstein R, Moulton E, Becerra L (2006) Functional imaging of the trigeminal system: applications to migraine pathophysiology. Headache 46 Suppl 1:S32-38. Bousser MG, Welch KM (2005) Relation between migraine and stroke. Lancet Neurol 4(9):533-542. Bras J, Singleton A (2011) Exome sequencing in Parkinson's disease. Clin Genet 80(2):104-109. Breslau N, Lipton RB, Stewart WF, Schultz LR, Welch KM (2003) Comorbidity of migraine and depression: investigating potential etiology and prognosis. Neurology 60(8):1308-1312. Breslau N, Schultz LR, Stewart WF, Lipton RB, Lucia VC, Welch KM (2000) Headache and major depression: is the association specific to migraine? Neurology 54(2):308-313. Burette A, Weinberg RJ (2007) Perisynaptic organization of plasma membrane calcium pumps in cerebellar cortex. J Comp Neurol 500(6):1127-1135. Burnet PW, Harrison PJ, Goodwin GM, Battersby S, Ogilvie AD, Olesen J, Russell MB (1997) Allelic variation in the serotonin 5-HT2C receptor gene and migraine. Neuroreport 8(12):2651-2653. Burton PR, Tobin MD, Hopper JL (2005) Key concepts in genetic epidemiology. Lancet 366(9489):941-951. Cader ZM, Noble-Topham S, Dyment DA, Cherny SS, Brown JD, Rice GP, Ebers GC (2003) Significant linkage to migraine with aura on chromosome 11q24. Hum Mol Genet 12(19):2511-2517. Carlsson A, Forsgren L, Nylander PO, Hellman U, Forsman-Semb K, Holmgren G, Holmberg D, Holmberg M (2002) Identification of a susceptibility locus for migraine with and without aura on 6p12.2-p21.1. Neurology 59(11):1804- 1807. Carod-Artal FJ, da Silveira Ribeiro L, Braga H, Kummer W, Mesquita HM, Vargas AP (2006) Prevalence of patent foramen ovale in migraine patients with and without aura compared with stroke patients. A transcranial Doppler study. Cephalalgia 26(8):934-939. Carroll JD (2008) Migraine Intervention With STARFlex Technology trial: a controversial trial of migraine and patent foramen ovale closure. Circulation 117(11):1358-1360.
223 Charbit AR, Akerman S, Goadsby PJ (2010) Dopamine: what's new in migraine? Curr Opin Neurol 23(3):275-281. Chasman DI, Schurks M, Anttila V, de Vries B, Schminke U, Launer LJ, Terwindt GM, van den Maagdenberg AM, Fendrich K, Volzke H, Ernst F, Griffiths LR, Buring JE, Kallela M, Freilinger T, Kubisch C, Ridker PM, Palotie A, Ferrari MD, Hoffmann W, Zee RY, Kurth T (2011) Genome-wide association study reveals three susceptibility loci for common migraine in the general population. Nat Genet 43(7):695-698. Cheung RT (2000) Prevalence of migraine, tension-type headache, and other headaches in Hong Kong. Headache 40(6):473-479. Chuang LY, Yang CH, Tsui KH, Cheng YH, Chang PL, Wen CH, Chang HW (2008) Restriction enzyme mining for SNPs in genomes. Anticancer Res 28(4A):2001-2007. Colson NJ, Lea RA, Quinlan S, Griffiths LR (2006a) No role for estrogen receptor 1 gene intron 1 Pvu II and exon 4 C325G polymorphisms in migraine susceptibility. BMC Med Genet 7:12. Colson NJ, Lea RA, Quinlan S, Griffiths LR (2006b) The role of vascular and hormonal genes in migraine susceptibility. Mol Genet Metab 88(2):107-113. Colson NJ, Lea RA, Quinlan S, MacMillan J, Griffiths LR (2004) The estrogen receptor 1 G594A polymorphism is associated with migraine susceptibility in two independent case/control groups. Neurogenetics 5(2):129-133. Colson NJ, Lea RA, Quinlan S, MacMillan J, Griffiths LR (2005) Investigation of hormone receptor genes in migraine. Neurogenetics 6(1):17-23. Connor KM, Shapiro RE, Diener HC, Lucas S, Kost J, Fan X, Fei K, Assaid C, Lines C, Ho TW (2009) Randomized, controlled trial of telcagepant for the acute treatment of migraine. Neurology 73(12):970-977. Corominas R, Ribases M, Camina M, Cuenca-Leon E, Pardo J, Boronat S, Sobrido MJ, Cormand B, Macaya A (2009) Two-stage case-control association study of dopamine-related genes and migraine. BMC Med Genet 10:95. Corominas R, Sobrido MJ, Ribases M, Cuenca-Leon E, Blanco-Arias P, Narberhaus B, Roig M, Leira R, Lopez-Gonzalez J, Macaya A, Cormand B (2010) Association study of the serotoninergic system in migraine in the Spanish population. Am J Med Genet B Neuropsychiatr Genet 153B(1):177-184.
224 Cox HC, Bellis C, Lea RA, Quinlan S, Hughes R, Dyer T, Charlesworth J, Blangero J, Griffiths LR (2009) Principal component and linkage analysis of cardiovascular risk traits in the Norfolk isolate. Hum Hered 68(1):55-64. Cox HC, Lea RA, Bellis C, Carless M, Dyer T, Blangero J, Griffiths LR (2011b) Variants in the human potassium channel gene (KCNN3) are associated with migraine in a high risk genetic isolate. J Headache Pain 12:603-608. Cox HC, Lea RA, Bellis C, Nyholt DR, Dyer TD, Haupt LM, Charlesworth J, Matovinovic E, Blangero J, Griffiths LR (2012) Heritability and genome-wide linkage analysis of migraine in the genetic isolate of Norfolk Island. Gene 494(1):119-123. Curtain R, Lea RA, Quinlan S, Bellis C, Tajouri L, Hughes R, Macmillan J, Griffiths LR (2004) Investigation of the low-density lipoprotein receptor gene and cholesterol as a risk factor for migraine. J Neurol Sci 227(1):95-100. D'Andrea G, Leon A (2010) Pathogenesis of migraine: from neurotransmitters to neuromodulators and beyond. Neurol Sci 31 Suppl 1:S1-7. Dahlof C, Linde M (2001) One-year prevalence of migraine in Sweden: a population- based study in adults. Cephalalgia 21(6):664-671. Dalla Volta G, Guindani M, Zavarise P, Griffini S, Pezzini A, Padovani A (2005) Prevalence of patent foramen ovale in a large series of patients with migraine with aura, migraine without aura and cluster headache, and relationship with clinical phenotype. J Headache Pain 6(4):328-330. Davey G, Sedgwick P, Maier W, Visick G, Strachan DP, Anderson HR (2002) Association between migraine and asthma: matched case-control study. Br J Gen Pract 52(482):723-727. Davies PT, Panayiotopoulos CP (2011) Migraine triggered seizures and epilepsy triggered headache and migraine attacks: a need for re-assessment. J Headache Pain 12(3):287-288. Dawn Teare M, Barrett JH (2005) Genetic linkage studies. Lancet 366(9490):1036- 1044. De Fusco M, Marconi R, Silvestri L, Atorino L, Rampoldi L, Morgante L, Ballabio A, Aridon P, Casari G (2003) Haploinsufficiency of ATP1A2 encoding the Na+/K+ pump alpha2 subunit associated with familial hemiplegic migraine type 2. Nat Genet 33(2):192-196.
225 De Simone R, Ranieri A, Marano E, Beneduce L, Ripa P, Bilo L, Meo R, Bonavita V (2007) Migraine and epilepsy: clinical and pathophysiological relations. Neurol Sci 28 Suppl 2:S150-155. de Vries B, Frants RR, Ferrari MD, van den Maagdenberg AM (2009) Molecular genetics of migraine. Hum Genet 126(1):115-132. Del Sette M, Angeli S, Leandri M, Ferriero G, Bruzzone GL, Finocchi C, Gandolfo C (1998) Migraine with aura and right-to-left shunt on transcranial Doppler: a case-control study. Cerebrovasc Dis 8(6):327-330. Dent W, Spiss H, Helbok R, Matuja W, Scheunemann S, Schmutzhard E (2004) Prevalence of migraine in a rural area in South Tanzania: a door-to-door survey. Cephalalgia 24(11):960-966. Dichgans M, Freilinger T, Eckstein G, Babini E, Lorenz-Depiereux B, Biskup S, Ferrari MD, Herzog J, van den Maagdenberg AM, Pusch M, Strom TM (2005) Mutation in the neuronal voltage-gated sodium channel SCN1A in familial hemiplegic migraine. Lancet 366(9483):371-377. Diener HC, Kurth T (2005) Is migraine a risk factor for stroke? Neurology 64(9):1496-1497. Domitrz I, Mieszkowski J, Kaminska A (2007) Relationship between migraine and patent foramen ovale: a study of 121 patients with migraine. Headache 47(9):1311-1318. Dowson A, Mullen MJ, Peatfield R, Muir K, Khan AA, Wells C, Lipscombe SL, Rees T, De Giovanni JV, Morrison WL, Hildick-Smith D, Elrington G, Hillis WS, Malik IS, Rickards A (2008) Migraine Intervention With STARFlex Technology (MIST) trial: a prospective, multicenter, double-blind, sham- controlled trial to evaluate the effectiveness of patent foramen ovale closure with STARFlex septal repair implant to resolve refractory migraine headache. Circulation 117(11):1397-1404. Ducros A, Romatet S, Saint Marc T, Allaf B (2011) Use of antimigraine treatments by general practitioners. Headache 51(7):1122-1131. Duffy DL (2006) An Integrated Genetic Map for Linkage Analysis. Behavior Genetics 36(1). Dufva M (2009) Introduction to microarray technology. Methods Mol Biol 529:1-22. Durham PL (2006) Calcitonin gene-related peptide (CGRP) and migraine. Headache 46 Suppl 1:S3-8.
226 Eikermann-Haerter K, Ayata C (2010) Cortical spreading depression and migraine. Curr Neurol Neurosci Rep 10(3):167-173. Eiland LS, Jenkins LS, Durham SH (2007) Pediatric migraine: pharmacologic agents for prophylaxis. Ann Pharmacother 41(7):1181-1190. Elston Lafata J, Moon C, Leotta C, Kolodner K, Poisson L, Lipton RB (2004) The medical care utilization and costs associated with migraine headache. J Gen Intern Med 19(10):1005-1012. Erdal ME, Herken H, Yilmaz M, Bayazit YA (2001) Association of the T102C polymorphism of 5-HT2A receptor gene with aura in migraine. J Neurol Sci 188(1-2):99-101. Etminan M, Takkouche B, Isorna FC, Samii A (2005) Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies. BMJ 330(7482):63. Fanciullacci MA, M; De Cesaris, F; Pietrini, U (2004) Arterial Hypertension and migraine: comorbidity or something else? J Headache Pain 5:S85-S87. Fernandez F, Colson N, Quinlan S, MacMillan J, Lea RA, Griffiths LR (2009) Association between migraine and a functional polymorphism at the dopamine beta-hydroxylase locus. Neurogenetics 10(3):199-208. Fernandez F, Esposito T, Lea RA, Colson NJ, Ciccodicola A, Gianfrancesco F, Griffiths LR (2008) Investigation of gamma-aminobutyric acid (GABA) A receptors genes and migraine susceptibility. BMC Med Genet 9:109. Fernandez F, Lea RA, Colson NJ, Bellis C, Quinlan S, Griffiths LR (2006) Association between a 19 bp deletion polymorphism at the dopamine beta- hydroxylase (DBH) locus and migraine with aura. J Neurol Sci 251(1-2):118- 123. Ferrari MD, Farkkila M, Reuter U, Pilgrim A, Davis C, Krauss M, Diener HC (2010) Acute treatment of migraine with the selective 5-HT1F receptor agonist lasmiditan--a randomised proof-of-concept trial. Cephalalgia 30(10):1170- 1178. Formicola D, Aloia A, Sampaolo S, Farina O, Diodato D, Griffiths LR, Gianfrancesco F, Di Iorio G, Esposito T (2010) Common variants in the regulative regions of GRIA1 and GRIA3 receptor genes are associated with migraine susceptibility. BMC Med Genet 11:103.
227 Gandhi S, Wood NW (2011) Genome-wide association studies: the key to unlocking neurodegeneration? Nat Neurosci 13(7):789-794. Gao X, Becker LC, Becker DM, Starmer JD, Province MA (2010) Avoiding the high Bonferroni penalty in genome-wide association studies. Genet Epidemiol 34(1):100-105. Gardner KL (2006) Genetics of migraine: an update. Headache 46 Suppl 1:S19-24. Gargus JJ, Tournay A (2007) Novel mutation confirms seizure locus SCN1A is also familial hemiplegic migraine locus FHM3. Pediatr Neurol 37(6):407-410. Ge D, Zhang K, Need AC, Martin O, Fellay J, Urban TJ, Telenti A, Goldstein DB (2008) WGAViewer: software for genomic annotation of whole genome association studies. Genome Res 18(4):640-643. Gervil M, Ulrich V, Kyvik KO, Olesen J, Russell MB (1999) Migraine without aura: a population-based twin study. Ann Neurol 46(4):606-611. Goadsby PJ (2009) Pathophysiology of migraine. Neurol Clin 27(2):335-360. Goadsby PJ, Lipton RB, Ferrari MD (2002) Migraine--current understanding and treatment. N Engl J Med 346(4):257-270. Goetzel RZ, Long SR, Ozminkowski RJ, Hawkins K, Wang S, Lynch W (2004) Health, absence, disability, and presenteeism cost estimates of certain physical and mental health conditions affecting U.S. employers. J Occup Environ Med 46(4):398-412. Goldstein DB (2009) Common genetic variation and human traits. N Engl J Med 360(17):1696-1698. Goode EL, Fridley BL, Sun Z, Atkinson EJ, Nord AS, McDonnell SK, Jarvik GP, de Andrade M, Slager SL (2007) Comparison of tagging single-nucleotide polymorphism methods in association analyses. BMC Proc 1 Suppl 1:S6. Grant SF, Hakonarson H (2008) Microarray technology and applications in the arena of genome-wide association. Clin Chem 54(7):1116-1124. Gray IC, Campbell DA, Spurr NK (2000) Single nucleotide polymorphisms as tools in human genetics. Hum Mol Genet 9(16):2403-2408. Gudmundsson LS, Thorgeirsson G, Sigfusson N, Sigvaldason H, Johannsson M (2006) Migraine patients have lower systolic but higher diastolic blood pressure compared with controls in a population-based study of 21,537 subjects. The Reykjavik Study. Cephalalgia 26(4):436-444.
228 Gundry CN, Vandersteen JG, Reed GH, Pryor RJ, Chen J, Wittwer CT (2003) Amplicon melting analysis with labeled primers: a closed-tube method for differentiating homozygotes and heterozygotes. Clin Chem 49(3):396-406. Gupta S, Mehrotra S, Villalon CM, Perusquia M, Saxena PR, MaassenVanDenBrink A (2007) Potential role of female sex hormones in the pathophysiology of migraine. Pharmacol Ther 113(2):321-340. Haan J, Terwindt GM, van den Maagdenberg AM, Stam AH, Ferrari MD (2008) A review of the genetic relation between migraine and epilepsy. Cephalalgia 28(2):105-113. Hagen K, Stovner LJ, Asberg A, Thorstensen K, Bjerve KS, Hveem K (2002) High headache prevalence among women with hemochromatosis: the Nord- Trondelag health study. Ann Neurol 51(6):786-789. Hagen PT, Scholz DG, Edwards WD (1984) Incidence and size of patent foramen ovale during the first 10 decades of life: an autopsy study of 965 normal hearts. Mayo Clin Proc 59(1):17-20. Hahn P, Qian Y, Dentchev T, Chen L, Beard J, Harris ZL, Dunaief JL (2004) Disruption of ceruloplasmin and hephaestin in mice causes retinal iron overload and retinal degeneration with features of age-related macular degeneration. Proc Natl Acad Sci U S A 101(38):13850-13855. Hamelsky SW, Lipton RB (2006) Psychiatric comorbidity of migraine. Headache 46(9):1327-1333. Hargreaves RJ, Shepheard SL (1999) Pathophysiology of migraine--new insights. Can J Neurol Sci 26 Suppl 3:S12-19. Hawkins K, Wang S, Rupnow M (2008) Direct Cost Burden Among Insured US Employees With Migraine. Headache 48(4):553-563. Hawkins K, Wang S, Rupnow MF (2007) Indirect cost burden of migraine in the United States. J Occup Environ Med 49(4):368-374. He JQ, Wiesmann C, van Lookeren Campagne M (2008) A role of macrophage complement receptor CRIg in immune clearance and inflammation. Mol Immunol 45(16):4041-4047. Henry P, Auray JP, Gaudin AF, Dartigues JF, Duru G, Lanteri-Minet M, Lucas C, Pradalier A, Chazot G, El Hasnaoui A (2002) Prevalence and clinical characteristics of migraine in France. Neurology 59(2):232-237.
229 Hering-Hanit R, Friedman Z, Schlesinger I, Ellis M (2001) Evidence for activation of the coagulation system in migraine with aura. Cephalalgia 21(2):137-139. Hirschhorn JN (2009) Genomewide association studies--illuminating biologic pathways. N Engl J Med 360(17):1699-1701. Ho TW, Ferrari MD, Dodick DW, Galet V, Kost J, Fan X, Leibensperger H, Froman S, Assaid C, Lines C, Koppen H, Winner PK (2008) Efficacy and tolerability of MK-0974 (telcagepant), a new oral antagonist of calcitonin gene-related peptide receptor, compared with zolmitriptan for acute migraine: a randomised, placebo-controlled, parallel-treatment trial. Lancet 372(9656):2115-2123. Hoare M (1999) Norfolk Island: A revised and enlarged History 1774-1998. Rockhampton: Central Queensland University Press. Hu XH, Markson LE, Lipton RB, Stewart WF, Berger ML (1999) Burden of migraine in the United States: disability and economic costs. Arch Intern Med 159(8):813-818. Huebner AK, Gandia M, Frommolt P, Maak A, Wicklein EM, Thiele H, Altmuller J, Wagner F, Vinuela A, Aguirre LA, Moreno F, Maier H, Rau I, Giesselmann S, Nurnberg G, Gal A, Nurnberg P, Hubner CA, del Castillo I, Kurth I (2011) Nonsense mutations in SMPX, encoding a protein responsive to physical force, result in X-chromosomal hearing loss. Am J Hum Genet 88(5):621-627. Humphrey PP (2008) The discovery and development of the triptans, a major therapeutic breakthrough. Headache 48(5):685-687. Hunter DJ, Altshuler D, Rader DJ (2008) From Darwin's finches to canaries in the coal mine--mining the genome for new biology. N Engl J Med 358(26):2760- 2763. IHS (2004) The International Classification of the Headache Disorders 2nd ed. Cephalalgia 24(Suppl 1):1-150. Jackson PE, Scholl PF, Groopman JD (2000) Mass spectrometry for genotyping: an emerging tool for molecular medicine. Mol Med Today 6(7):271-276. Jaillard AS, Mazetti P, Kala E (1997) Prevalence of migraine and headache in a high- altitude town of Peru: a population-based study. Headache 37(2):95-101. Jen JC, Kim GW, Dudding KA, Baloh RW (2004) No mutations in CACNA1A and ATP1A2 in probands with common types of migraine. Arch Neurol 61(6):926- 928.
230 Jen JC, Wan J, Palos TP, Howard BD, Baloh RW (2005) Mutation in the glutamate transporter EAAT1 causes episodic ataxia, hemiplegia, and seizures. Neurology 65(4):529-534. Jensen R, Stovner LJ (2008) Epidemiology and comorbidity of headache. Lancet Neurol 7(4):354-361. Jette N, Patten S, Williams J, Becker W, Wiebe S (2008) Comorbidity of migraine and psychiatric disorders--a national population-based study. Headache 48(4):501-516. Johansson AC, Feuk L (2012) Characterizing and interpreting genetic variation from personal genome sequencing. Methods Mol Biol 838:343-367. Johnson MP, Lea RA, Curtain RP, MacMillan JC, Griffiths LR (2003) An investigation of the 5-HT2C receptor gene as a migraine candidate gene. Am J Med Genet B Neuropsychiatr Genet 117B(1):86-89. Johnston MM, Rapoport AM (2010) Triptans for the management of migraine. Drugs 70(12):1505-1518. Jones KW, Ehm MG, Pericak-Vance MA, Haines JL, Boyd PR, Peroutka SJ (2001) Migraine with aura susceptibility locus on chromosome 19p13 is distinct from the familial hemiplegic migraine locus. Genomics 78(3):150-154. Joshi G, Pradhan S, Mittal B (2009) Role of the ACE ID and MTHFR C677T polymorphisms in genetic susceptibility of migraine in a north Indian population. J Neurol Sci 277(1-2):133-137. Juhasz G, Zsombok T, Laszik A, Gonda X, Sotonyi P, Faludi G, Bagdy G (2003) Association analysis of 5-HTTLPR variants, 5-HT2a receptor gene 102T/C polymorphism and migraine. J Neurogenet 17(2-3):231-240. Kan CW, Fredlake CP, Doherty EA, Barron AE (2004) DNA sequencing and genotyping in miniaturized electrophoresis systems. Electrophoresis 25(21- 22):3564-3588. Karger BL, Guttman A (2009) DNA sequencing by CE. Electrophoresis 30 Suppl 1:S196-202. Katsarava Z, Rabe K, Diener HC (2008) From migraine to stroke. Intern Emerg Med. Kaunisto MA, Kallela M, Hamalainen E, Kilpikari R, Havanka H, Harno H, Nissila M, Sako E, Ilmavirta M, Liukkonen J, Teirmaa H, Tornwall O, Jussila M, Terwilliger J, Farkkila M, Kaprio J, Palotie A, Wessman M (2006) Testing of variants of the MTHFR and ESR1 genes in 1798 Finnish individuals fails to
231 confirm the association with migraine with aura. Cephalalgia 26(12):1462- 1472. Kececi H, Dener S (2002) Epidemiological and clinical characteristics of migraine in Sivas, Turkey. Headache 42(4):275-280. Kelman L (2004) The premonitory symptoms (prodrome): a tertiary care study of 893 migraineurs. Headache 44(9):865-872. Kere J (2010) Genetics of complex disorders. Biochem Biophys Res Commun 396(1):143-146. Kong A, Cox NJ (1997) Allele-sharing models: LOD scores and accurate linkage tests. Am J Hum Genet 61(5):1179-1188. Kowa H, Fusayasu E, Ijiri T, Ishizaki K, Yasui K, Nakaso K, Kusumi M, Takeshima T, Nakashima K (2005) Association of the insertion/deletion polymorphism of the angiotensin I-converting enzyme gene in patients of migraine with aura. Neurosci Lett 374(2):129-131. Kowa H, Yasui K, Takeshima T, Urakami K, Sakai F, Nakashima K (2000) The homozygous C677T mutation in the methylenetetrahydrofolate reductase gene is a genetic risk factor for migraine. Am J Med Genet 96(6):762-764. Kraft P, Hunter DJ (2009) Genetic risk prediction--are we there yet? N Engl J Med 360(17):1701-1703. Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES (1996) Parametric and nonparametric linkage analysis: a unified multipoint approach. Am J Hum Genet 58(6):1347-1363. Kruit MC, Launer LJ, Overbosch J, van Buchem MA, Ferrari MD (2009) Iron accumulation in deep brain nuclei in migraine: a population-based magnetic resonance imaging study. Cephalalgia 29(3):351-359. Ku M, Silverman B, Prifti N, Ying W, Persaud Y, Schneider A (2006) Prevalence of migraine headaches in patients with allergic rhinitis. Ann Allergy Asthma Immunol 97(2):226-230. Kurth T, Slomke MA, Kase CS, Cook NR, Lee IM, Gaziano JM, Diener HC, Buring JE (2005) Migraine, headache, and the risk of stroke in women: a prospective study. Neurology 64(6):1020-1026. Kwok PY (2001) Methods for genotyping single nucleotide polymorphisms. Annu Rev Genomics Hum Genet 2:235-258.
232 Lafreniere RG, Cader MZ, Poulin JF, Andres-Enguix I, Simoneau M, Gupta N, Boisvert K, Lafreniere F, McLaughlan S, Dube MP, Marcinkiewicz MM, Ramagopalan S, Ansorge O, Brais B, Sequeiros J, Pereira-Monteiro JM, Griffiths LR, Tucker SJ, Ebers G, Rouleau GA (2010) A dominant-negative mutation in the TRESK potassium channel is linked to familial migraine with aura. Nat Med 16(10):1157-1160. Lambert GA, Zagami AS (2008) The Mode of Action of Migraine Triggers: A Hypothesis. Headache 49(2)253-275. Lampl C, Buzath A, Baumhackl U, Klingler D (2003) One-year prevalence of migraine in Austria: a nation-wide survey. Cephalalgia 23(4):280-286. Lampl C, Marecek S (2006) Migraine and stroke--why do we talk about it? Eur J Neurol 13(3):215-219. Lanteri-Minet M, Radat F, Chautard MH, Lucas C (2005) Anxiety and depression associated with migraine: influence on migraine subjects' disability and quality of life, and acute migraine management. Pain 118(3):319-326. Launer LJ, Terwindt GM, Ferrari MD (1999) The prevalence and characteristics of migraine in a population-based cohort: the GEM study. Neurology 53(3):537- 542. Laurell K, Larsson B, Eeg-Olofsson O (2004) Prevalence of headache in Swedish schoolchildren, with a focus on tension-type headache. Cephalalgia 24(5):380- 388. Lea R, Colson N, Quinlan S, Macmillan J, Griffiths L (2009) The effects of vitamin supplementation and MTHFR (C677T) genotype on homocysteine-lowering and migraine disability. Pharmacogenet Genomics 19(6):422-428. Lea RA, Nyholt DR, Curtain RP, Ovcaric M, Sciascia R, Bellis C, Macmillan J, Quinlan S, Gibson RA, McCarthy LC, Riley JH, Smithies YJ, Kinrade S, Griffiths LR (2005) A genome-wide scan provides evidence for loci influencing a severe heritable form of common migraine. Neurogenetics 6(2):67-72. Lea RA, Ovcaric M, Sundholm J, MacMillan J, Griffiths LR (2004) The methylenetetrahydrofolate reductase gene variant C677T influences susceptibility to migraine with aura. BMC Med 2:3.
233 Lea RA, Shepherd AG, Curtain RP, Nyholt DR, Quinlan S, Brimage PJ, Griffiths LR (2002) A typical migraine susceptibility region localizes to chromosome 1q31. Neurogenetics 4(1):17-22. Leniger T, Isbruch K, von den Driesch S, Diener HC, Hufnagel A (2001) Seizure- associated headache in epilepsy. Epilepsia 42(9):1176-1179. Leniger T, von den Driesch S, Isbruch K, Diener HC, Hufnagel A (2003) Clinical characteristics of patients with comorbidity of migraine and epilepsy. Headache 43(6):672-677. Leonardi M, Raggi A, Bussone G, D'Amico D (2010) Health-related quality of life, disability and severity of disease in patients with migraine attending to a specialty headache center. Headache 50(10):1576-1586. Levy D (2010) Migraine pain and nociceptor activation--where do we stand? Headache 50(5):909-916. Lewis D, Ashwal S, Hershey A, Hirtz D, Yonker M, Silberstein S (2004) Practice parameter: pharmacological treatment of migraine headache in children and adolescents: report of the American Academy of Neurology Quality Standards Subcommittee and the Practice Committee of the Child Neurology Society. Neurology 63(12):2215-2224. Ligthart L, Boomsma DI, Martin NG, Stubbe JH, Nyholt DR (2006) Migraine with aura and migraine without aura are not distinct entities: further evidence from a large Dutch population study. Twin Res Hum Genet 9(1):54-63. Lindblom A, Robinson PN (2011) Bioinformatics for human genetics: promises and challenges. Hum Mutat 32(5):495-500. Lipton RB, Bigal ME (2007) Ten lessons on the epidemiology of migraine. Headache 47 Suppl 1:S2-9. Lipton RB, Bigal ME, Diamond M, Freitag F, Reed ML, Stewart WF (2007) Migraine prevalence, disease burden, and the need for preventive therapy. Neurology 68(5):343-349. Lipton RB, Bigal ME, Kolodner K, Stewart WF, Liberman JN, Steiner TJ (2003a) The family impact of migraine: population-based studies in the USA and UK. Cephalalgia 23(6):429-440. Lipton RB, Diamond S, Reed M, Diamond ML, Stewart WF (2001a) Migraine diagnosis and treatment: results from the American Migraine Study II. Headache 41(7):638-645.
234 Lipton RB, Liberman JN, Kolodner KB, Bigal ME, Dowson A, Stewart WF (2003b) Migraine headache disability and health-related quality-of-life: a population- based case-control study from England. Cephalalgia 23(6):441-450. Lipton RB, Ottman R, Ehrenberg BL, Hauser WA (1994) Comorbidity of migraine: the connection between migraine and epilepsy. Neurology 44(10 Suppl 7):S28-32. Lipton RB, Scher AI, Kolodner K, Liberman J, Steiner TJ, Stewart WF (2002) Migraine in the United States: epidemiology and patterns of health care use. Neurology 58(6):885-894. Lipton RB, Stewart WF, Diamond S, Diamond ML, Reed M (2001b) Prevalence and burden of migraine in the United States: data from the American Migraine Study II. Headache 41(7):646-657. Ludvigsson P, Hesdorffer D, Olafsson E, Kjartansson O, Hauser WA (2006) Migraine with aura is a risk factor for unprovoked seizures in children. Ann Neurol 59(1):210-213. Lynch JR, Brown JM (1990) The polymerase chain reaction: current and future clinical applications. J Med Genet 27(1):2-7. Lyngberg AC, Rasmussen BK, Jorgensen T, Jensen R (2005) Has the prevalence of migraine and tension-type headache changed over a 12-year period? A Danish population survey. Eur J Epidemiol 20(3):243-249. Lyon GJ, Jiang T, Van Wijk R, Wang W, Bodily PM, Xing J, Tian L, Robison RJ, Clement M, Lin Y, Zhang P, Liu Y, Moore B, Glessner JT, Elia J, Reimherr F, van Solinge WW, Yandell M, Hakonarson H, Wang J, Johnson WE, Wei Z, Wang K (2011) Exome sequencing and unrelated findings in the context of complex disease research: ethical and clinical implications. Discov Med 12(62):41-55. MacClellan LR, Giles W, Cole J, Wozniak M, Stern B, Mitchell BD, Kittner SJ (2007) Probable migraine with visual aura and risk of ischemic stroke: the stroke prevention in young women study. Stroke 38(9):2438-2445. MacGregor EA (2004) Oestrogen and attacks of migraine with and without aura. Lancet Neurol 3(6):354-361. MacGregor EA (2009) Headache and hormone replacement therapy in the postmenopausal woman. Curr Treat Options Neurol 11(1):10-17.
235 Maggioni F, Alessi C, Maggino T, Zanchin G (1997) Headache during pregnancy. Cephalalgia 17(7):765-769. Magis D, Schoenen J (2011) Treatment of migraine: update on new therapies. Curr Opin Neurol 24(3):203-210. Mancia G, Rosei EA, Ambrosioni E, Avino F, Carolei A, Dacco M, Di Giacomo G, Ferri C, Grazioli I, Melzi G, Nappi G, Pinessi L, Sandrini G, Trimarco B, Zanchin G (2011) Hypertension and migraine comorbidity: prevalence and risk of cerebrovascular events: evidence from a large, multicenter, cross- sectional survey in Italy (MIRACLES study). J Hypertens 29(2):309-318. Manolio TA (2011) Genomewide association studies and assessment of the risk of disease. N Engl J Med 363(2):166-176. Manolio TA, Brooks LD, Collins FS (2008) A HapMap harvest of insights into the genetics of common disease. J Clin Invest 118(5):1590-1605. Mansfield ES, Robertson JM, Vainer M, Isenberg AR, Frazier RR, Ferguson K, Chow S, Harris DW, Barker DL, Gill PD, Budowle B, McCord BR (1998) Analysis of multiplexed short tandem repeat (STR) systems using capillary array electrophoresis. Electrophoresis 19(1):101-107. Manzoni G, C., Torelli P (2003) Epidemiology of migraine. J Headache Pain 4:S18- S22. Martin PR, MacLeod C (2009) Behavioral management of headache triggers: Avoidance of triggers is an inadequate strategy. Clin Psychol Rev 29(6):483- 495. Marziniak M, Mossner R, Schmitt A, Lesch KP, Sommer C (2005) A functional serotonin transporter gene polymorphism is associated with migraine with aura. Neurology 64(1):157-159. Matias-Guiu J, Porta-Etessam J, Mateos V, Diaz-Insa S, Lopez-Gil A, Fernandez C (2010) One-year prevalence of migraine in Spain: a nationwide population- based survey. Cephalalgia 31(4):463-470. Mattsson P, Svardsudd K, Lundberg PO, Westerberg CE (2000) The prevalence of migraine in women aged 40-74 years: a population-based study. Cephalalgia 20(10):893-899. Mayeux R (2005) Mapping the new frontier: complex genetic disorders. J Clin Invest 115(6):1404-1407.
236 McWilliams LA, Goodwin RD, Cox BJ (2004) Depression and anxiety associated with three pain conditions: results from a nationally representative sample. Pain 111(1-2):77-83. Mehle ME (2008) Allergy and migraine: is there a connection? Curr Opin Otolaryngol Head Neck Surg 16(3):265-269. Menon S, Cox HC, Kawahata M, Quinlan S, Macmillan JC, Haupt LM, Lea RA, Griffiths LR (2011) Association of a Notch 3 gene polymorphism with migraine susceptibility. Cephalalgia 31(3):264-270. Messlinger K (2009) Migraine: where and how does the pain originate? Exp Brain Res 196(1):179-193. Mestan KK, Ilkhanoff L, Mouli S, Lin S (2011) Genomic sequencing in clinical trials. J Transl Med 9:222. Miller S.A, Dykes D.D, H.F P (1988) A simple salting out procedure for extracting DNA from Human Nucleated cells. Nucleic Acid Reseach 16(3):1215. Millis MP (2011) Medium-throughput SNP genotyping using mass spectrometry: multiplex SNP genotyping using the iPLEX(R) Gold assay. Methods Mol Biol 700:61-76. Mongini F, Rota E, Deregibus A, Ferrero L, Migliaretti G, Cavallo F, Mongini T, Novello A (2006) Accompanying symptoms and psychiatric comorbidity in migraine and tension-type headache patients. J Psychosom Res 61(4):447-451. Moretti TR, Baumstark AL, Defenbaugh DA, Keys KM, Brown AL, Budowle B (2001) Validation of STR typing by capillary electrophoresis. J Forensic Sci 46(3):661-676. Morillo LE, Alarcon F, Aranaga N, Aulet S, Chapman E, Conterno L, Estevez E, Garcia-Pedroza F, Garrido J, Macias-Islas M, Monzillo P, Nunez L, Plascencia N, Rodriguez C, Takeuchi Y (2005) Prevalence of migraine in Latin America. Headache 45(2):106-117. Moschiano F, D'Amico D, Canavero I, Pan I, Micieli G, Bussone G (2011) Migraine and depression: common pathogenetic and therapeutic ground? Neurol Sci 32 Suppl 1:S85-88. Moskowitz MA (2007) Genes, proteases, cortical spreading depression and migraine: impact on pathophysiology and treatment. Funct Neurol 22(3):133-136. Moskowitz MA (2008) Defining a pathway to discovery from bench to bedside: the trigeminovascular system and sensitization. Headache 48(5):688-690.
237 Moskowitz MA, Bolay H, Dalkara T (2004) Deciphering migraine mechanisms: clues from familial hemiplegic migraine genotypes. Ann Neurol 55(2):276-280. Mueller LL (2007) Diagnosing and managing migraine headache. J Am Osteopath Assoc 107(10 Suppl 6):ES10-16. Mulder EJ, Van Baal C, Gaist D, Kallela M, Kaprio J, Svensson DA, Nyholt DR, Martin NG, MacGregor AJ, Cherkas LF, Boomsma DI, Palotie A (2003) Genetic and environmental influences on migraine: a twin study across six countries. Twin Res 6(5):422-431. Narbone MC, Gangemi S, Abbate M (2008) Migraine and stroke: from a questioned relationship to a supported comorbidity. Neurol Sci 29 Suppl 1:S7-11. Neeb L, Meents J, Reuter U (2010) 5-HT(1F) Receptor agonists: a new treatment option for migraine attacks? Neurotherapeutics 7(2):176-182. Netzer C, Freudenberg J, Heinze A, Heinze-Kuhn K, Goebel I, McCarthy LC, Roses AD, Gobel H, Todt U, Kubisch C (2008a) Replication study of the insulin receptor gene in migraine with aura. Genomics 91(6):503-507. Netzer C, Freudenberg J, Toliat MR, Heinze A, Heinze-Kuhn K, Thiele H, Goebel I, Nurnberg P, Ptacek LJ, Gobel H, Todt U, Kubisch C (2008b) Genetic association studies of the chromosome 15 GABA-A receptor cluster in migraine with aura. Am J Med Genet B Neuropsychiatr Genet 147B(1):37-41. Nyholt DR (2000) All LODs are not created equal. Am J Hum Genet 67(2):282-288. Nyholt DR, Curtain RP, Griffiths LR (2000) Familial typical migraine: significant linkage and localization of a gene to Xq24-28. Hum Genet 107(1):18-23. Nyholt DR, Dawkins JL, Brimage PJ, Goadsby PJ, Nicholson GA, Griffiths LR (1998a) Evidence for an X-linked genetic component in familial typical migraine. Hum Mol Genet 7(3):459-463. Nyholt DR, Gillespie NG, Heath AC, Merikangas KR, Duffy DL, Martin NG (2004) Latent class and genetic analysis does not support migraine with aura and migraine without aura as separate entities. Genet Epidemiol 26(3):231-244. Nyholt DR, LaForge KS, Kallela M, Alakurtti K, Anttila V, Farkkila M, Hamalainen E, Kaprio J, Kaunisto MA, Heath AC, Montgomery GW, Gobel H, Todt U, Ferrari MD, Launer LJ, Frants RR, Terwindt GM, de Vries B, Verschuren WM, Brand J, Freilinger T, Pfaffenrath V, Straube A, Ballinger DG, Zhan Y, Daly MJ, Cox DR, Dichgans M, van den Maagdenberg AM, Kubisch C, Martin NG, Wessman M, Peltonen L, Palotie A (2008) A high-density
238 association screen of 155 ion transport genes for involvement with common migraine. Hum Mol Genet 17(21):3318-3331. Nyholt DR, Lea RA, Goadsby PJ, Brimage PJ, Griffiths LR (1998b) Familial typical migraine: linkage to chromosome 19p13 and evidence for genetic heterogeneity. Neurology 50(5):1428-1432. Nyholt DR, Morley KI, Ferreira MA, Medland SE, Boomsma DI, Heath AC, Merikangas KR, Montgomery GW, Martin NG (2005) Genomewide significant linkage to migrainous headache on chromosome 5q21. Am J Hum Genet 77(3):500-512. O'Brien B, Goeree R, Streiner D (1994) Prevalence of migraine headache in Canada: a population-based survey. Int J Epidemiol 23(5):1020-1026. O'Connell JR, Weeks DE (1998) PedCheck: a program for identification of genotype incompatibilities in linkage analysis. Am J Hum Genet 63(1):259-266. Odegaard KJ, Greenwood TA, Lunde A, Fasmer OB, Akiskal HS, Kelsoe JR, Consortium NGIBD (2009) A genome-wide linkage study of bipolar disorder and co-morbid migraine: Replication of migraine linkage on chromosome 4q24, and suggestion of an overlapping susceptibility region for both disorders on chromosome 20p11. J Affect Disord. Ogilvie AD, Russell MB, Dhall P, Battersby S, Ulrich V, Smith CA, Goodwin GM, Harmar AJ, Olesen J (1998) Altered allelic distributions of the serotonin transporter gene in migraine without aura and migraine with aura. Cephalalgia 18(1):23-26. Olesen J, Ashina M (2011) Emerging migraine treatments and drug targets. Trends Pharmacol Sci 32(6):352-359. Olesen J, Friberg L, Olsen TS, Iversen HK, Lassen NA, Andersen AR, Karle A (1990) Timing and topography of cerebral blood flow, aura, and headache during migraine attacks. Ann Neurol 28(6):791-798. Ophoff RA, Terwindt GM, Vergouwe MN, van Eijk R, Oefner PJ, Hoffman SM, Lamerdin JE, Mohrenweiser HW, Bulman DE, Ferrari M, Haan J, Lindhout D, van Ommen GJ, Hofker MH, Ferrari MD, Frants RR (1996) Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca2+ channel gene CACNL1A4. Cell 87(3):543-552. Oswell G, Kaunisto MA, Kallela M, Hamalainen E, Anttila V, Kaprio J, Farkkila M, Wessman M, Palotie A (2008) No association of migraine to the GABA-A
239 receptor complex on chromosome 15. Am J Med Genet B Neuropsychiatr Genet 147B(1):33-36. Oterino A, Monton F, Cid C, Ruiz-Lavilla K, Gardner KL, Barmada M, Pascual J (2001) A new locus for migraine with aura on Xp13. Cephalalgia 21:346. Oterino A, Pascual J, Ruiz de Alegria C, Valle N, Castillo J, Bravo Y, Gonzalez F, Sanchez-Velasco P, Cayon A, Leyva-Cobian F, Alonso-Arranz A, Munoz P (2006) Association of migraine and ESR1 G325C polymorphism. Neuroreport 17(1):61-64. Oterino A, Toriello M, Cayon A, Castillo J, Colas R, Alonson-Arranz A, Ruiz-Alegria C, Quintela E, Monton F, Ruiz-Lavilla N, Gonzalez F, Pascual J (2008) Multilocus analyses reveal involvement of the ESR1, ESR2, and FSHR genes in migraine. Headache 48(10):1438-1450. Ottman R, Lipton RB (1996) Is the comorbidity of epilepsy and migraine due to a shared genetic susceptibility? Neurology 47(4):918-924. Painter JN, Nyholt DR, Montgomery GW (2011) Association mapping. Methods Mol Biol 760:35-52. Panoutsopoulou K, Zeggini E (2009) Finding common susceptibility variants for complex disease: past, present and future. Brief Funct Genomic Proteomic 8(5):345-352. Parsons AA, Strijbos PJ (2003) The neuronal versus vascular hypothesis of migraine and cortical spreading depression. Curr Opin Pharmacol 3(1):73-77. Patel NV, Bigal ME, Kolodner KB, Leotta C, Lafata JE, Lipton RB (2004) Prevalence and impact of migraine and probable migraine in a health plan. Neurology 63(8):1432-1438. Paterna S, Di Pasquale P, Cottone C, Seidita G, Cardinale A, Parrinello G, Ferrari G, Licata G (1997) Migraine without aura and ACE-gene deletion polymorphism: is there a correlation? Preliminary findings. Cardiovasc Drugs Ther 11(4):603- 604. Pearson TA, Manolio TA (2008) How to interpret a genome-wide association study. JAMA 299(11):1335-1344. Piccinelli P, Borgatti R, Nicoli F, Calcagno P, Bassi MT, Quadrelli M, Rossi G, Lanzi G, Balottin U (2006) Relationship between migraine and epilepsy in pediatric age. Headache 46(3):413-421.
240 Pierangeli G, Cevoli S, Zanigni S, Sancisi E, Monaldini C, Donti A, Ribani MA, Montagna P, Cortelli P (2004) The role of cardiac diseases in the comorbidity between migraine and stroke. Neurol Sci 25 Suppl 3:S129-131. Pietrini U, De Luca M, De Santis G (2005) Hypertension in headache patients? A clinical study. Acta Neurol Scand 112(4):259-264. Pringsheim T, Davenport WJ, Becker WJ Prophylaxis of migraine headache. CMAJ 182(7):E269-276. Prudenzano MP, Monetti C, Merico L, Cardinali V, Genco S, Lamberti P, Livrea P (2005) The comorbidity of migraine and hypertension. A study in a tertiary care headache centre. J Headache Pain 6(4):220-222. Purcell S, Cherny SS, Sham PC. (2003) Genetic Power Calculator: design of linkage and association genetic mapping studies of complex traits. Bioinformatics, 19(1):149-150. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ, Sham PC (2007) PLINK: a tool set for whole- genome association and population-based linkage analyses. Am J Hum Genet 81(3):559-575. Qian ZM, Chang YZ, Zhu L, Yang L, Du JR, Ho KP, Wang Q, Li LZ, Wang CY, Ge X, Jing NL, Li L, Ke Y (2007) Development and iron-dependent expression of hephaestin in different brain regions of rats. J Cell Biochem 102(5):1225- 1233. Racchi M, Leone M, Porrello E, Rigamonti A, Govoni S, Sironi M, Montomoli C, Bussone G (2004) Familial migraine with aura: association study with 5- HT1B/1D, 5-HT2C, and hSERT polymorphisms. Headache 44(4):311-317. Radat F, Swendsen J (2005) Psychiatric comorbidity in migraine: a review. Cephalalgia 25(3):165-178. Rapoport A, Edmeads J (2000) Migraine: the evolution of our knowledge. Arch Neurol 57(8):1221-1223. Rapoport AM (1996) MO and MA: distinct clinical entities? Cephalalgia 16(4):215. Rasmussen BK, Jensen R, Schroll M, Olesen J (1991) Epidemiology of headache in a general population--a prevalence study. J Clin Epidemiol 44(11):1147-1157. Raychaudhuri S, Plenge RM, Rossin EJ, Ng AC, Purcell SM, Sklar P, Scolnick EM, Xavier RJ, Altshuler D, Daly MJ (2009) Identifying relationships among
241 genomic disease regions: predicting genes at pathogenic SNP associations and rare deletions. PLoS Genet 5(6):e1000534. Redolfi E, Montagna C, Mumm S, Affer M, Susani L, Reinbold R, Hol F, Vezzoni P, Cimino M, Zucchi I (1998) Identification of CXorf1, a novel intronless gene in Xq27.3, expressed in human hippocampus. DNA Cell Biol 17(12):1009- 1016. Reisman M, Christofferson RD, Jesurum J, Olsen JV, Spencer MP, Krabill KA, Diehl L, Aurora S, Gray WA (2005) Migraine headache relief after transcatheter closure of patent foramen ovale. J Am Coll Cardiol 45(4):493-495. Ridker PM, Chasman DI, Zee RY, Parker A, Rose L, Cook NR, Buring JE (2008) Rationale, design, and methodology of the Women's Genome Health Study: a genome-wide association study of more than 25,000 initially healthy american women. Clin Chem 54(2):249-255. Robinson P, Krawitz P, Mundlos S (2011) Strategies for exome and genome sequence data analysis in disease-gene discovery projects. Clin Genet 80(2):127-132. Rodelsperger C, Krawitz P, Bauer S, Hecht J, Bigham AW, Bamshad M, de Condor BJ, Schweiger MR, Robinson PN (2011) Identity-by-descent filtering of exome sequence data for disease-gene identification in autosomal recessive disorders. Bioinformatics 27(6):829-836. Rodriguez-Murillo L, Greenberg DA (2008) Genetic association analysis: a primer on how it works, its strengths and its weaknesses. Int J Androl 31(6):546-556. Ross P, Hall L, Smirnov I, Haff L (1998) High level multiplex genotyping by MALDI-TOF mass spectrometry. Nat Biotechnol 16(13):1347-1351. Roux KH (1995) Optimization and troubleshooting in PCR. PCR Methods Appl 4(5):S185-194. Rubino E, Ferrero M, Rainero I, Binello E, Vaula G, Pinessi L (2009) Association of the C677T polymorphism in the MTHFR gene with migraine: a meta-analysis. Cephalalgia 29(8):818-825. Russell MB (2010) Genetics of menstrual migraine: the epidemiological evidence. Curr Pain Headache Rep 14(5):385-388. Russell MB, Olesen J (1995) Increased familial risk and evidence of genetic factor in migraine. BMJ 311(7004):541-544.
242 Russell MB, Ulrich V, Gervil M, Olesen J (2002) Migraine without aura and migraine with aura are distinct disorders. A population-based twin survey. Headache 42(5):332-336. Russo L, Mariotti P, Sangiorgi E, Giordano T, Ricci I, Lupi F, Chiera R, Guzzetta F, Neri G, Gurrieri F (2005) A new susceptibility locus for migraine with aura in the 15q11-q13 genomic region containing three GABA-A receptor genes. Am J Hum Genet 76(2):327-333. Sakai F, Igarashi H (1997) Prevalence of migraine in Japan: a nationwide survey. Cephalalgia 17(1):15-22. Schurks M, Rist PM, Kurth T (2010a) 5-HTTLPR polymorphism in the serotonin transporter gene and migraine: a systematic review and meta-analysis. Cephalalgia 30(11):1296-1305. Schurks M, Rist PM, Kurth T (2010b) MTHFR 677C>T and ACE D/I polymorphisms in migraine: a systematic review and meta-analysis. Headache 50(4):588-599. Schurks M, Rist PM, Kurth T (2010c) Sex hormone receptor gene polymorphisms and migraine: a systematic review and meta-analysis. Cephalalgia 30(11):1306- 1328. Schurks M, Rist PM, Kurth T (2010d) STin2 VNTR polymorphism in the serotonin transporter gene and migraine: pooled and meta-analyses. J Headache Pain 11(4):317-326. Schurks M, Zee R, Buring J, Kurth T (2010e) MTHFR 677C-T and ACE D/I polymorphisms and migraine attack frequency in women. Cephalalgia 30(4):447. Schwedt TJ, Demaerschalk BM, Dodick DW (2008) Patent foramen ovale and migraine: a quantitative systematic review. Cephalalgia 28(5):531-540. Schwedt TJ, Dodick DW (2006) Patent foramen ovale and migraine--bringing closure to the subject. Headache 46(4):663-671. Schwerzmann M, Nedeltchev K, Lagger F, Mattle HP, Windecker S, Meier B, Seiler C (2005) Prevalence and size of directly detected patent foramen ovale in migraine with aura. Neurology 65(9):1415-1418. Schwerzmann M, Wiher S, Nedeltchev K, Mattle HP, Wahl A, Seiler C, Meier B, Windecker S (2004) Percutaneous closure of patent foramen ovale reduces the frequency of migraine attacks. Neurology 62(8):1399-1401.
243 Seng KC, Seng CK (2008) The success of the genome-wide association approach: a brief story of a long struggle. Eur J Hum Genet 16(5):554-564. Sham PC, Curtis D (1995) Monte Carlo tests for associations between disease and alleles at highly polymorphic loci. Ann Hum Genet 59(Pt 1):97-105. Shih MC, Whittemore AS (2001) Allele-sharing among affected relatives: non- parametric methods for identifying genes. Stat Methods Med Res 10(1):27-55. Shivpuri D, Rajesh MS, Jain D (2003) Prevalence and characteristics of migraine among adolescents: a questionnaire survey. Indian Pediatr 40(7):665-669. Sicuteri F (1977) Dopamine, the second pututive protagonist in headache. . Headache 17:129-131. Smitherman TA, Walters AB, Maizels M, Penzien DB (2011) The Use of Antidepressants for Headache Prophylaxis. CNS Neurosci Ther. 17(5):462- 469 Soragna D, Vettori A, Carraro G, Marchioni E, Vazza G, Bellini S, Tupler R, Savoldi F, Mostacciuolo ML (2003) A locus for migraine without aura maps on chromosome 14q21.2-q22.3. Am J Hum Genet 72(1):161-167. Stam AH, de Vries B, Janssens AC, Vanmolkot KR, Aulchenko YS, Henneman P, Oostra BA, Frants RR, van den Maagdenberg AM, Ferrari MD, van Duijn CM, Terwindt GM (2010) Shared genetic factors in migraine and depression: evidence from a genetic isolate. Neurology 74(4):288-294. Stang PE, Carson AP, Rose KM, Mo J, Ephross SA, Shahar E, Szklo M (2005) Headache, cerebrovascular symptoms, and stroke: the Atherosclerosis Risk in Communities Study. Neurology 64(9):1573-1577. Stang PE, Crown WH, Bizier R, Chatterton ML, White R (2004) The family impact and costs of migraine. Am J Manag Care 10(5):313-320. Stevenson SB (2006) Epilepsy and migraine headache: is there a connection? J Pediatr Health Care 20(3):167-171. Stewart W, Breslau N, Keck PE, Jr. (1994) Comorbidity of migraine and panic disorder. Neurology 44(10 Suppl 7):S23-27. Stewart WF, Bigal ME, Kolodner K, Dowson A, Liberman JN, Lipton RB (2006) Familial risk of migraine: variation by proband age at onset and headache severity. Neurology 66(3):344-348.
244 Stewart WF, Lipton RB, Celentano DD, Reed ML (1992) Prevalence of migraine headache in the United States. Relation to age, income, race, and other sociodemographic factors. JAMA 267(1):64-69. Stewart WF, Lipton RB, Liberman J (1996) Variation in migraine prevalence by race. Neurology 47(1):52-59. Stewart WF, Staffa J, Lipton RB, Ottman R (1997) Familial risk of migraine: a population-based study. Ann Neurol 41(2):166-172. Stovner L, Hagen K, Jensen R, Katsarava Z, Lipton R, Scher A, Steiner T, Zwart JA (2007) The global burden of headache: a documentation of headache prevalence and disability worldwide. Cephalalgia 27(3):193-210. Stovner LJ, Zwart JA, Hagen K, Terwindt GM, Pascual J (2006) Epidemiology of headache in Europe. Eur J Neurol 13(4):333-345. Strauch K, Fimmers R, Baur MP, Wienker TF (2003) How to model a complex trait. 1. General considerations and suggestions. Hum Hered 55(4):202-210. Takeshima T, Ishizaki K, Fukuhara Y, Ijiri T, Kusumi M, Wakutani Y, Mori M, Kawashima M, Kowa H, Adachi Y, Urakami K, Nakashima K (2004) Population-based door-to-door survey of migraine in Japan: the Daisen study. Headache 44(1):8-19. Tan HJ, Suganthi C, Dhachayani S, Rizal AM, Raymond AA (2007) The coexistence of anxiety and depressive personality traits in migraine. Singapore Med J 48(4):307-310. Tatlidede AD, Oflazoglu B, Celik SE, Anadol U, Forta H (2007) Prevalence of patent foramen ovale in patients with migraine. Agri 19(4):39-42. Tekle Haimanot R, Seraw B, Forsgren L, Ekbom K, Ekstedt J (1995) Migraine, chronic tension-type headache, and cluster headache in an Ethiopian rural community. Cephalalgia 15(6):482-488. Tepper SJ, Sheftell FD, Bigal ME (2007) The patent foramen ovale-migraine question. Neurol Sci 28 Suppl 2:S118-123. Terwindt GM, Ophoff RA, van Eijk R, Vergouwe MN, Haan J, Frants RR, Sandkuijl LA, Ferrari MD (2001) Involvement of the CACNA1A gene containing region on 19p13 in migraine with and without aura. Neurology 56(8):1028-1032. Thiele H, Nurnberg P (2005) HaploPainter: a tool for drawing pedigrees with complex haplotypes. Bioinformatics 21(8):1730-1732.
245 Tietjen GE, Bushnell CD, Herial NA, Utley C, White L, Hafeez F (2007) Endometriosis is associated with prevalence of comorbid conditions in migraine. Headache 47(7):1069-1078. Tikka-Kleemola P, Artto V, Vepsalainen S, Sobel EM, Raty S, Kaunisto MA, Anttila V, Hamalainen E, Sumelahti ML, Ilmavirta M, Farkkila M, Kallela M, Palotie A, Wessman M (2010) A visual migraine aura locus maps to 9q21-q22. Neurology 74(15):1171-1177. Tikka-Kleemola P, Kaunisto MA, Hamalainen E, Todt U, Gobel H, Kaprio J, Kubisch C, Farkkila M, Palotie A, Wessman M, Kallela M (2009) Genetic association study of endothelin-1 and its receptors EDNRA and EDNRB in migraine with aura. Cephalalgia 29(11):1224-1231. Todd R, Donoff RB, Kim Y, Wong DT (2001) From the chromosome to DNA: Restriction fragment length polymorphism analysis and its clinical application. J Oral Maxillofac Surg 59(6):660-667. Todt U, Dichgans M, Jurkat-Rott K, Heinze A, Zifarelli G, Koenderink JB, Goebel I, Zumbroich V, Stiller A, Ramirez A, Friedrich T, Gobel H, Kubisch C (2005) Rare missense variants in ATP1A2 in families with clustering of common forms of migraine. Hum Mutat 26(4):315-321. Todt U, Freudenberg J, Goebel I, Heinze A, Heinze-Kuhn K, Rietschel M, Gobel H, Kubisch C (2006) Variation of the serotonin transporter gene SLC6A4 in the susceptibility to migraine with aura. Neurology 67(9):1707-1709. Todt U, Netzer C, Toliat M, Heinze A, Goebel I, Nurnberg P, Gobel H, Freudenberg J, Kubisch C (2009) New genetic evidence for involvement of the dopamine system in migraine with aura. Hum Genet 125(3):265-279. Tsimikas S (2005) Transcatheter closure of patent foramen ovale for migraine prophylaxis: hope or hype? J Am Coll Cardiol 45(4):496-498. Tung YC, Yeo GS (2011) From GWAS to biology: lessons from FTO. Ann N Y Acad Sci 1220:162-171. Van den Maagdenberg AM, Pietrobon D, Pizzorusso T, Kaja S, Broos LA, Cesetti T, van de Ven RC, Tottene A, van der Kaa J, Plomp JJ, Frants RR, Ferrari MD (2004) A Cacna1a knockin migraine mouse model with increased susceptibility to cortical spreading depression. Neuron 41(5):701-710.
246 Van Roijen L, Essink-Bot ML, Koopmanschap MA, Michel BC, Rutten FF (1995) Societal perspective on the burden of migraine in The Netherlands. Pharmacoeconomics 7(2):170-179. Vanmolkot KR, Babini E, de Vries B, Stam AH, Freilinger T, Terwindt GM, Norris L, Haan J, Frants RR, Ramadan NM, Ferrari MD, Pusch M, van den Maagdenberg AM, Dichgans M (2007) The novel p.L1649Q mutation in the SCN1A epilepsy gene is associated with familial hemiplegic migraine: genetic and functional studies. Mutation in brief #957. Online. Hum Mutat 28(5):522. Vikelis M, Mitsikostas DD (2007) The role of glutamate and its receptors in migraine. CNS Neurol Disord Drug Targets 6(4):251-257. Visscher PM, Hill WG, Wray NR (2008) Heritability in the genomics era--concepts and misconceptions. Nat Rev Genet 9(4):255-266. Vogt L, Schmitz N, Kurrer MO, Bauer M, Hinton HI, Behnke S, Gatto D, Sebbel P, Beerli RR, Sonderegger I, Kopf M, Saudan P, Bachmann MF (2006) VSIG4, a B7 family-related protein, is a negative regulator of T cell activation. J Clin Invest 116(10):2817-2826. Vossen RH, Aten E, Roos A, den Dunnen JT (2009) High-resolution melting analysis (HRMA): more than just sequence variant screening. Hum Mutat 30(6):860- 866. Voytas D (2001) Agarose gel electrophoresis. Curr Protoc Immunol Chapter 10:Unit 10 14. Wang SJ, Fuh JL, Young YH, Lu SR, Shia BC (2000) Prevalence of migraine in Taipei, Taiwan: a population-based survey. Cephalalgia 20(6):566-572. Webb AJ, Thorisson GA, Brookes AJ (2011) An informatics project and online "Knowledge Centre" supporting modern genotype-to-phenotype research. Hum Mutat 32(5):543-550. Welch KM (2009) Iron in the migraine brain; a resilient hypothesis. Cephalalgia 29(3):283-285. Welch KM, Nagesh V, Aurora SK, Gelman N (2001) Periaqueductal gray matter dysfunction in migraine: cause or the burden of illness? Headache 41(7):629- 637. Wessman M, Kallela M, Kaunisto MA, Marttila P, Sobel E, Hartiala J, Oswell G, Leal SM, Papp JC, Hamalainen E, Broas P, Joslyn G, Hovatta I, Hiekkalinna T, Kaprio J, Ott J, Cantor RM, Zwart JA, Ilmavirta M, Havanka H, Farkkila M,
247 Peltonen L, Palotie A (2002) A susceptibility locus for migraine with aura, on chromosome 4q24. Am J Hum Genet 70(3):652-662. Wessman M, Kaunisto MA, Kallela M, Palotie A (2004) The molecular genetics of migraine. Ann Med 36(6):462-473. Wessman M, Terwindt GM, Kaunisto MA, Palotie A, Ophoff RA (2007) Migraine: a complex genetic disorder. Lancet Neurol 6(6):521-532. WHO (2001) The World Health Report 2001 - Mental Health: New Understanding, New Hope. Wiehe M, Fuchs SC, Moreira LB, Moraes RS, Fuchs FD (2002) Migraine is more frequent in individuals with optimal and normal blood pressure: a population- based study. J Hypertens 20(7):1303-1306. Wieser T, Pascual J, Oterino A, Soso M, Barmada M, Gardner KL (2010) A novel locus for familial migraine on Xp22. Headache 50(6):955-962. Wigginton JE, Abecasis GR (2005) PEDSTATS: descriptive statistics, graphics and quality assessment for gene mapping data. Bioinformatics 21(16):3445-3447. Williams FM, Cherkas LF, Spector TD, MacGregor AJ (2004) A common genetic factor underlies hypertension and other cardiovascular disorders. BMC Cardiovasc Disord 4(1):20. Williams MA, Carson R, Passmore P, Silvestri G, Craig D (2010) Introduction to genetic epidemiology. Optometry 82(2):83-91. Wong TW, Wong KS, Yu TS, Kay R (1995) Prevalence of migraine and other headaches in Hong Kong. Neuroepidemiology 14(2):82-91. Wu DY, Ugozzoli L, Pal BK, Qian J, Wallace RB (1991) The effect of temperature and oligonucleotide primer length on the specificity and efficiency of amplification by the polymerase chain reaction. DNA Cell Biol 10(3):233-238. Yilmaz M, Erdal ME, Herken H, Cataloluk O, Barlas O, Bayazit YA (2001) Significance of serotonin transporter gene polymorphism in migraine. J Neurol Sci 186(1-2):27-30. Zencir M, Ergin H, Sahiner T, Kilic I, Alkis E, Ozdel L, Gurses D, Ergin A (2004) Epidemiology and symptomatology of migraine among school children: Denizli urban area in Turkey. Headache 44(8):780-785. Ziegle JS, Su Y, Corcoran KP, Nie L, Mayrand PE, Hoff LB, McBride LJ, Kronick MN, Diehl SR (1992) Application of automated DNA sizing technology for genotyping microsatellite loci. Genomics 14(4):1026-1031.
248 Zivadinov R, Willheim K, Jurjevic A, Sepic-Grahovac D, Bucuk M, Zorzon M (2001) Prevalence of migraine in Croatia: a population-based survey. Headache 41(8):805-812. Zwart JA, Dyb G, Hagen K, Odegard KJ, Dahl AA, Bovim G, Stovner LJ (2003) Depression and anxiety disorders associated with headache frequency. The Nord-Trondelag Health Study. Eur J Neurol 10(2):147-152. Zwart JA, Dyb G, Holmen TL, Stovner LJ, Sand T (2004) The prevalence of migraine and tension-type headaches among adolescents in Norway. The Nord- Trondelag Health Study (Head-HUNT-Youth), a large population-based epidemiological study. Cephalalgia 24(5):373-379.
249