CHAPTER 10 218 10 Overlap between migraine and cluster headache Comorbidities between disorders can be the result of either sharing of genetic and/or environmental factors, or a causal relationship. GWAS data can be used to investigate a possible shared genetic architecture in correlation studies where the proportion of variance two trait share due to genetics is measured (the genetic correlation rG). In Chapter 6 and Chapter 7 susceptibility loci for cluster headache were identified and in Chapter 8 risk loci for migraine were identified. Chapter 6 and a parallel GWAS paper on cluster headache independently of each other showed genetic overlap between migraine and cluster headache, as one (UFL1/FHL5) of the four genes associated with cluster headache was the same as in migraine.105 This is not that surprising, as there was already a presumed link between cluster headache and migraine as they both share key clinical characteristics (unilateral headache and cranial autonomic symptoms), pathophysiological mechanisms (involvement of the trigeminal vascular system) and response to therapy (CGRP monoclonal antibodies).106 This overlapping locus was reason to further investigate the genetic overlap with a correlation analysis, this is a parameter that describes the proportion of variance two traits share due to genetic causes. In Chapter 6 we found a positive correlation of 33% between migraine and cluster headache. However, due to the relatively small sample size of the cluster headache GWAS this was not significant, therefore, it was important to perform this analysis in a larger cohort. In a meta-analysis of these cluster headache GWAS (Chapter 7), there was possibility to further investigate the possibility of an overlapping genetic architecture between both disorders. In the European meta-analysis a total of three (FHL5, PLCE1, LRP1) of all seven cluster headache risk loci were similar to migraine risk loci. Apart from these three overlapping loci, none of the other 119 known migraine risk loci showed an association with cluster headache. Genetic correlation analysis did show that there was an overall genetic correlation between cluster headache and migraine, suggesting a shared genetic architecture. It could be argued that this overlap would be due to “contamination” of migraine cases in our cluster headache dataset. However, there are findings that would contradict this. In case the signal in our cluster headache meta-analysis came from migraine patients, one would expect that the exact same loci would come up. LRP1 is associated with the lowest p-value in migraine, but the second most prominent locus in migraine is PRDM16. This locus was already reported in 2011 in the second migraine GWAS, which only found three risk loci (LRP1, PRDM16 and TRPM8).107 This locus shows no signal in the cluster headache meta-analysis. Hence, the absence of PRMD16, argues that there is no contamination of migraine patients in the cluster headache sample. The third locus associated gene in migraine is FHL5, but all other top-associated genes up-until PLEC1, which is the 11th associated gene in migraine, show no signal in cluster headache. Thus, it seems that there is some sharing in genetics architecture between cluster headache and migraine, but that their exact genetic signature differs. In fact, investigation of the top 200 ranked gene-based p-values between both disorders found an overlap of only 10 genes (data not shown), which confirms the notion that the two disorders have an overall different genetic signature. The advantage of gene-based p-values is that it is only based on SNPs that are located within a gene, therefore, you have more power. Furthermore, we show
RkJQdWJsaXNoZXIy MTk4NDMw