Typical migraine is a common debilitating disorder affecting approximately 12% of the Western population [1]. A large US epidemiological study indicated that the disease affects around 18% of women and 12 % of men in the overall population [2]. The disease has two main subtypes, migraine without aura (MO) affecting ~ 73% and migraine with aura (MA) affecting about 33%. Migraine with aura is defined as the more severe type and both subtypes can occur within the same family. The number of genes involved in common migraine is unknown; although genes for a rare severe sub-type of migraine Familial Hemplegic Migraine (FHM) have been identified on chromosome 19 [3] and recently FHM type 2 has been identified on chromosome 1 [4].

Mutations in the neuronal calcium channel gene (CACNA1A) on 19p13.1-13.2 have been shown to cause FHM [3]. Calcium channel genes may also be involved in typical migraine (MA and MO) as some studies have indicated that the FHM locus may contribute to migraine with and without aura [5]. We have previously reported linkage in one large typical migraine family [6] to the CACNA1A region on chr19. Also in a study by Terwindt et al. (1998) [7] the authors detected an FHM mutation in the CACNA1A gene in a typical migraine patient with MA, suggesting that FHM may be a rare and severe form of MA.

Typical migraine has been extensively studied on chromosome 19 showing positive linkage to causal markers in the C19p13 area [5,6,8-10], but only in FHM has the chromosome 19 gene and mutations within been identified [3]. The NOTCH3 gene on C19p13.2-p13.1 has previously been shown to be a gene involved in CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) [11]. CADASIL is an inherited stroke syndrome that leads to dementia. The key features of the disorder are migraine, recurrent subcortical events, and dementia, in association with diffuse white matter abnormalities on neuroimaging [12]. The gene encodes a large transmembrane receptor [13], which has been shown to be expressed in vascular smooth muscle cells in adult human tissues [14]. In a study by Joutel et al. 1997 [13] this large gene, containing 33 exons, was found to contain a cluster of mutations in exons 3 and 4 affecting 64% (32/50) of tested patients with the CADASIL disorder. Of the other 36% of CADASIL sufferers, 26% (13/50) have mutations in exons 2,5,11,14,18,19,22 or 23 and the other 10% (5/50) had no detectable mutation in the NOTCH3 gene [13]. Mutations identified to date in NOTCH3have all been missense mutations involving a loss or gain of a cysteine amino acid residue [13].

The NOTCH3 gene may also be implicated in migraine, as migraine shows some symptomatic overlap with CADASIL. This has been suggested from studies performed by Hutchinson et al. (1995) [15] and Verin et al. (1995) [16] where the authors have found some correlations between the two disorders. Hutchinson et al. (1995) [15] used MRI to study 15 members of an Irish family, 10 of whom had evidence of CADASIL. Five members of this family had hemiplegic migraine. They proposed that hemiplegic migraine may be an allelic disorder to CADASIL. Migraine with aura (MA) is also included as one of the clinical hallmarks of CADASIL, with many CADASIL patients showing an unusually high frequency of attacks of migraine with atypical aura [17]. In 2003, Oberstein et al. [18] found that among 6 individuals who carried a mutation in the NOTCH3 gene, there was an increase in white matter hyperintensities on brain MRI, compared to controls and migraine with aura was more common in these subjects than in their controls. In addition, a recent association study by Schwaag et al. identified a significant association of the synonymous rs1043994 polymorphism with migraine [19]. However, not all NOTCH3 mutations may be involved in migraine, as an association study by Borroni et al. [20] indicated that the that functional polymorphism T6746C of NOTCH3 did not have any association with migraine in the tested Italian population.

The TNFSF7 gene (Tumor Necrosis Factor Ligand Superfamily, member 7) localized on C19p13 is a surface antigen found on activated, but not resting, T and B lymphocytes [21]. It is a 19 amino acid protein containing a 20-amino acid hydrophilic N-terminal domain that lacks a signal sequence; an 18-amino acid hydrophobic region that presumably functions as a transmembrane anchor; and a C-terminal domain that contains 2 potential N-linked glycosylation sites is extracellular classifying TNFSF7 as a type II transmembrane protein. TNFSF7 is homologous to the ligands of the TNF receptor family, including TNF-alpha, TNF-beta and the CD40 ligand, showing 19 to 24% amino acid sequence identity in the extracellular region. [21]. TNF-alpha and TNF-beta have both been previously associated with migraine. Rainero et al. 2004 [22], recently found an association between the tumor necrosis factor-alpha -308 G/A gene polymorphism and migraine. In a group of 299 migraine patients and 306 control subjects, the association of this polymorphism with the occurrence and clinical characteristics of migraine was tested. They found that homozygosity for the G allele was associated with an increased risk of migraine (odds ratio [OR] = 2.85, p <0.001) [22]. Also Empl et al. 2003 [23] suggested that TNF-alpha could contribute to migraine pain generation and in their study of TNF-alpha and its soluble receptor sTNF-RI, they found that migraine patients tended to have less concentration levels of sTNF-RI (794 +/- 158 pg/ml) than controls (945 +/- 137 pg/ml) [23]. The authors suggested that if TNF-alpha plays a role in migraine physiopathology, migraine patients may lack sufficient antagonistic sTNF-RI to neutralize hyperalgesic TNF-alpha during a migraine attack, causing pain [23]. The TNF-beta gene has been associated with migraine without aura. Trabace et al. 2002 [24] found the frequency of the TNFB*2 allele of the TNF-beta gene was significantly increased in patients with migraine without aura as compared with the control group (78.72% versus 61.4%, Pc =.004) [24]. They stated that carriage of the TNFB*2 allele confers a high risk for the development of migraine without aura. The data supported the hypothesis that TNF-beta could be a susceptibility gene in migraine without aura [24]. Since TNFSF7 is homologous to the ligands of the TNF receptor family i.e. TNF-alpha and TNF-beta, (and is localized to a migraine susceptibility area at C19p13) we decided to investigate this gene for association with migraine.

This study investigated the migraine susceptibility locus C19p13 by testing two genes, from this genomic region, that may possibly be involved in the disorder. The NOTCH3 gene was analysed by sequencing all exons with known CADASIL mutations through a family previously linked to C19p13 [6]. The TNFSF7 gene was also investigated using SNP case-control association analysis in a population of migraineurs and matched controls.

The procedure for genetic diagnostic testing for CADASIL mutations has recently been revised. We undertook a similar approach to this revision in testing for these mutations in common migraine. A study in Britain [28], found 15 different point mutations in the NOTCH3 gene in 48 families, 73% of which were in exon 4, 8% in exon 3, and 6% in each of exons 5 and 6. The authors suggested that on the basis of this spectrum the suggested protocol for genetic diagnostic testing for CADASIL would be to screen exon 4 and proceed to mutational screening of exons 3, 5, and 6 where indicated [28]. A similar approach can be undertaken in diagnostic testing for other diseases with known mutations in genes causing a disorder, including FHM whereby two genes and corresponding mutations have been identified in causing this rare subtype of migraine.

The NOTCH3 gene has been localised to C19p13, a region showing linkage to MF1 a typical (not FHM) migraine pedigree. Our sequencing results for this gene, specifically testing exons with known CADASIL mutations proved to be negative. These results indicate that common migraine, at least in this pedigree, is not caused by mutations in the NOTCH3 gene.

TNF-alpha and TNF-beta have both been previously associated with migraine [22, 24]. Since TNFSF7 is localized to the same migraine susceptibility area at C19p13 and shows homology to the ligands of the TNF-alpha and TNF-beta genes (both localized at C6p21.3), we decided to also investigate this gene for involvement in migraine. A suitable informative SNP (rs1862511) was selected for an association analysis study involving a matched case-control population. Overall chi-square results comparing migraine (total) and MA sub-populations with sex, age and ethnicity matched controls proved to be non-significant, with P values of P = 0.136, P = 0.589 respectively obtained (Table 4). These values were both above the threshold of P<0.05 for significance. However, the MO subgroup displayed signs of possible involvement of the TNFSF7 SNP with this sub-population compared to controls. P values for genotype and allele analysis were P = 0.036 and P = 0.017, respectively (Table 4). These results indicate that this genetic variant may play a role in migraine without aura and warrant further investigation.

Overall the results presented here do not support a relationship between NOTCH3 and common migraine. Although CADASIL patients often display migraine symptoms there was no evidence that CADASIL mutations from the NOTCH3 gene are involved in typical migraine in our C19p13 linked [6] migraine family samples. However, this does not discount the possibility of a role for mutations in the as yet unscreened exons in NOTCH3 in migraine susceptibility. The TNFSF7 gene also localized to the C19p13 region did not show any association with typical migraine or the MA subtype, however a weak association was found with MO affected individuals. Since this gene shows homology to the TNF receptor family and both TNF-alpha and beta have previously shown association with typical migraine [22, 24], the TNFSF7 gene (localizedon C19p13, a migraine linked region, see [5,6,8-10]) was investigated in migraine. Our association study showed signs of possible involvement of this gene with the MO subtype only, but clearly more work needs to be done in increasing sample size numbers and also confirmation from other research groups to give an indication of a relationship of this gene to migraine without aura.

It is clear that typical migraine is a complex disorder that may involve several genes on various chromosomes. Since the C19p13 region has been implicated in a number of studies it is considered a ‘hotspot’ for the disorder. More extensive sequencing of the NOTCH 3 gene that may identify novel mutations that relate to migraine should be undertaken. Further studies need to be performed with TNFSF7 and MO affected individuals and also other possible TNF receptor family homologues on C19p13, such as TNFSF9 (C19p13.3) and TNFSF14 (C19p13.3) should be tested for migraine involvement. It may be useful to explore the TNF receptor-migraine interaction further by studying these genes localized to the C19p13 region.