Cover Page
The handle http://hdl.handle.net/1887/47855 holds various files of this Leiden University dissertation
Author: Eising, E.
Title: Exploring genes and pathways involved in migraine
Issue Date: 2017-03-15
&
Summary
suMMARy
The research in this thesis was aimed at identifying genes and molecular pathways involved in the pathogenesis of migraine. To this end, research of animal models was combined with research of headache patients. Two gene expression analyses were performed in brain tissue obtained from transgenic mice that express human pathogenic mutations in voltage-gated CaV2.1 Ca2+ channels known to cause familial hemiplegic migraine type 1 (FHM1), a monogenic subtype of migraine with aura. Next, a study was conducted aimed at analyzing epigenetic marks at migraine candidate genes in FHM1 mice, as a first exploration of a possible role of epigenetics in migraine pathophysiology. In addition to the studies in migraine-relevant mouse models, the genetic architecture of migraine was explored by performing various gene-based pathway analyses to mine the extensive data sets that were collected by the International Headache Genetics Consortium (IHGC) and used for recent genome wide association studies (GWAS) for the common forms of migraine. With these data, the overlap in genetic architecture of migraine with aura (MA) and migraine without aura (MO) was explored, as well as the cell types and brain regions that are likely involved in the pathophysiology of migraine and the two migraine subtypes. Finally, a gene expression profiling study was performed in blood of patients suffering from primary headache, in this case cluster headache. Combining the various approaches let to the identification of migraine-relevant genes, pathways and cell types that helps to unravel pathophysiological mechanisms of migraine and other primary headache disorders.
Chapters 2 and 3 describe gene expression studies of brain tissue from FHM1 transgenic migraine mouse models to identify genes and pathways involved in migraine pathophysiology.
In Chapter 2, the effects of the FHM1 R192Q and S218L missense mutations on gene expression were explored under naïve, basal conditions using microarray technology. The FHM1 R192Q mutation is associated with ‘pure’ FHM, while the FHM1 S218L mutation is associated with hemiplegic migraine, cerebellar ataxia, seizures, and mild head trauma-induced brain edema.
Identified changes in expression profiles were anticipated to shed light on why the R192Q and S218L mutations are associated with different clinical severity and symptoms in patients. Caudal cortical and cerebellar tissue was investigated because of their relevance to migraine aura and ataxia, respectively. Expression differences in cerebellum were most pronounced in FHM1 S218L mice, as expected from the cerebellar ataxia phenotype in mice (and patients) with this mutation. In contrast, there were only very modest changes in cortical RNA expression profiles in either of the FHM1 mouse models; more precisely, only 23 genes were differentially expressed of which nine overlapped between both mutant strains. This study showed that the pronounced consequences of the FHM1 mutations on patient phenotypes and cortical neurobiology are not reflected in cortical gene expression profiles.
In Chapter 3 experimentally induced cortical spreading depression (CSD), the underlying mechanism of the migraine aura, was used as a migraine-relevant trigger to augment the effects of the FHM1 R192Q mutation on cortical gene expression profiles. By using KCl application onto the dura above the caudal cortex, multiple CSD events were induced that were recorded by
217
&
Summary
an electrode in the frontal cortex. For sham experiments, NaCl applications that do not elicit CSD events were used. Cortical gene expression profiles, assessed by a technique named deep Serial Analysis of Gene Expression sequencing from brain tissue in between the application and recording sites, showed that CSD induced inflammatory responses in the cortex of both FHM1 R192Q and wild-type mice. Notably, a gene set could be identified that is up-regulated after CSD, specifically in FHM1 R192Q mutant mice. Genes from this gene set are associated with interferon-related inflammatory signaling and are up-regulated under immune-stimulated conditions while they are not co-expressed in the unstimulated brain. The inflammatory signature upon induction of CSD in FHM1 R192Q mice suggests that the FHM1 mouse brain is much more susceptible to inflammatory triggers. Furthermore, our findings indicate a role for interferon-mediated signaling pathways in migraine.
Genetic and environmental factors play an equally large role in conferring risk for developing migraines. Environmental influences are envisaged to affect epigenetic mechanisms, e.g. DNA methylation and post-translational modifications (PTMs) of histone proteins, and can thereby result in gene expression changes relevant for migraine pathophysiology. In Chapter 4 we lay the ground for our epigenetic study by reviewing the basics of epigenetic mechanisms relevant to migraine. Besides playing a role in establishing cellular and developmental stage-specific regulation of gene expression, epigenetic processes are important for programming lasting responses to environmental signals. Epigenetic mechanisms may explain how non-genetic endogenous and exogenous factors, such as female sex hormones, stress hormones and inflammation may modulate migraine characteristics like attack frequency. The potential for developing drugs that specifically target epigenetic mechanisms, which may open up exciting new avenues for the prophylactic treatment of migraine, is discussed as well.
In Chapter 5 the effect of the FHM1 R192Q mutation on epigenetic mechanisms was investigated. Although the FHM1 R192Q mutation is not associated with large gene expression changes under basal conditions, as shown in Chapter 2, we envisaged that the FHM1 mutation may result in epigenetic changes already present under basal conditions that can prime the cells to respond to migraine-related triggers. Therefore, histone PTM levels were studied at regulatory regions of migraine candidate genes in the cortex of FHM1 R192Q mice using chromatin immunoprecipitation followed by qPCR (ChIP-qPCR). Ten different histone PTMs were investigated that can differentiate the activity state of promoters, enhancers and gene bodies. Genes were selected based on whether they surfaced from the gene expression profiling studies of Chapters 2 and 3 or from GWAS. Whereas histone PTM levels in the gene promoters correctly correlated with the expression levels of the respective genes, no epigenetic differences were identified that reflected differential expression of genes between the FHM1 R192Q and wild-type mice. In contrast, differential histone PTM levels could be identified at the promoter of GWAS hit Mef2d, and nominal significant differences were found at the promoters of the GWAS hits Tgfbr2 and Trpm8. These results indicate that the FHM1 mice might provide a suitable model to study the role of the GWAS hits in migraine pathophysiology.
218
&
Summary
In Chapters 6, 7 and 8 various gene-based pathway analyses were carried out of the extensive GWAS data sets for migraine that were obtained by efforts of the IHGC. Gene-based analyses are based on summing the SNP-based association signals per gene or per gene set and thus provide a powerful approach to identify additional associations in GWAS data, as they reduce the multiple testing burden that is a major obstacle in ordinary GWAS analyses that results from the testing of several hundred thousand to millions of SNPs.
In Chapter 6 the overlap in genetic architecture between MA and MO was examined on gene level. Using a novel gene-based (statistical) approach, individual genes and pathways were identified that were associated with both MA and MO. Gene-based tests were performed utilizing summary statistics results from 4,505 MA cases (with 34,813 controls) and 4,038 MO cases (with 40,294 controls) of the IHGC to examine the proportion of shared genes associated with MA and MO. A significant overlap was observed between both migraine types. Of the in total 1,514 genes with a nominally significant gene-based P-value (Pgene-based ≤ 0.05) in the MA subgroup, 107 also produced Pgene-based ≤ 0.05 in the MO subgroup. The proportion of overlapping genes is almost double the empirically derived null expectation, producing significant evidence of gene-based overlap (pleiotropy) (Pbinomial-test = 1.6×10-3). Combining results across MA and MO, six genes produced genome-wide significant gene-based P-values. Four of the genes, i.e.
TRPM8, UFL1, FHL5 and LRP1 were located in close proximity to previously reported genome- wide significant SNPs for migraine. TARBP2 and NPFF, separated by only 259 base pairs on chromosome 12q13.13, represent novel migraine risk genes. The genes that overlap in both migraine types were enriched for functions related to inflammation, the cardiovascular system and connective tissue.
In Chapter 7 GWAS data of 4,954 clinic-based patients with MA or MO (with 13,390 controls) of the IHGC was used to study the brain cell types that likely are involved in conferring the risk for migraine. To this end a gene set enrichment analysis was performed of glial cell- (astrocyte, microglia, and oligodendrocyte) and synapse-specific gene sets in the GWAS data set. The results revealed that astrocyte- and oligodendrocyte-specific gene sets are associated with migraine;
especially oligodendrocytic genes involved in protein modification and signal transduction and astrocytic genes involved in protein transport and cell proliferation. Observed differences for MA and MO indicated that both migraine types may have, at least in part, a different genetic background. In contrast to studies in FHM, in this study no support was found for a role of synaptic mechanisms.
Finally, in Chapter 8, GWAS data from 23,285 migraine cases and 95,425 population-matched controls of the IHGC was integrated with high-resolution spatial gene expression data of adult human brains from the Allen Human Brain Atlas to identify specific brain regions and molecular pathways involved in migraine pathophysiology. Two complementary approaches were used. For the first approach, modules of co-expressed genes were studied, which were calculated based on human brain expression data, for enrichment of genes that showed nominal association with migraine in the GWAS data. Enrichment of migraine-associated genes was observed for
219
