Abstract
It has been suggested that folate metabolism could be involved in migraine pathogenesis. We analysed the 5′, 10′ -methylenetetrahydrofolate reductase (MTHFR) genotypic distribution in a large migraine sample. We genotyped 230 migraine patients (152 migraine without aura (MO) and 78 migraine with aura (MA)) and 204 nonheadache controls. The incidence of TT homozygosis for migraine in general (12%), MO (9%) and MA (18%) did not significantly differ from that found in healthy controls (13%). Differences were significant when the frequency of TT homozygosis between MA and MO (P = 0.03, OR = 2.34, 95% CI = 1.04-5.26) was compared. There was a tendency for a higher frequency of the MTHFR T allele in the MA group (42%) as compared to MO (29%) and controls (36%). These differences were significant only in the case of MA vs. MO (P = 0.006, OR = 1.75, 95% CI = 1.15-2.65). These results could indicate that the MTHFR C677T polymorphism, causing mild hyperhomocystinaemia, might be a genetic risk factor for experiencing aura among migraineurs. Overall, however, there was no association between migraine and the C677T MTHFR polymorphism.
Introduction
Migraine is a genetically heterogeneous disorder. The genetic basis of migraine has been discussed over time regarding aspects of inheritability, mode of transmission and candidate genes. Except for the rare variant of familial hemiplegic migraine, caused by point mutations in CACNA1 (1) and ATP1A2 (2) genes, no other specific genes for nonhemiplegic migraine have ever been found. Recently, it has been suggested that folate metabolism could be involved in the pathogenesis of migraine with (MA) and without (MO) aura. In two independent association studies, Kowa et al. (3) and Kara et al. (4) proposed the C677T variant of the 5′,10′-methylenetetrahydrofolate reductase (MTHFR; OMIM♯236250) as a risk modifier factor for both migraine in general and, mainly, for MA in a Japanese and in a Turkish population. Here we tested the role of the C677T allelic variant in migraine in a larger sample of European patients.
Patients and methods
Subjects
We recruited 230 consecutive migraine patients (mean age 40.8 ± 15.3 years) in our general neurological clinic. Migraine patients with a stroke history were excluded. The control group comprised 204 nonheadache healthy volunteers (mean age 39.9 ± 13.8 years). All migraine patients and healthy controls were interviewed by an experienced neurologist. The migraine patients were diagnosed as having IHS MO (152 cases; mean age 43.5 ± 16.0 years) or MA (78 cases; mean age 35.7 ± 12.5 years) (5). Informed consent was obtained from all patients and controls. All were Caucasians living in the same region.
Genetic analysis
Genotyping of the C677T-MTHFR polymorphism was performed on leucocyte genomic DNA samples by polymerase chain reaction (PCR) fragment-length polymorphism analysis, by using primers, PCR conditions, and identification of HinFI restriction fragments, as already described (6).
Statistical analysis
For statistical analysis MA and MO patients were grouped together and termed ‘migraine in general’. Genotypic and allele frequencies were computed using the SPSS program, version 10.0. Hardy–Weinberg equilibrium adjustment and differences in frequency of C677T-MTHFR alleles and genotypes were calculated by the χ2 method. The significance level was set at P < 0.05. The relative risk for the different genotype carriers was estimated calculating the odds ratios (OR) and 95% confidence intervals (CI).
Results
Both genotypic and allelic frequencies fit the Hardy–Weinberg equilibrium (Table 1). The incidence of TT homozygosis for migraine in general (12%), MO (9%) and MA (18%) did not significantly differ from that found in healthy controls (13%). Differences reached significance, however, when the frequency of TT homozygosis between MA and MO (P = 0.03, OR = 2.34, CI = 1.04–5.26) was compared. There was a tendency for a higher frequency of MTHFR T677 allele in the MA group (42%) as compared to MO (29%) and controls (36%). Again these differences were significant only in the case of MA vs. MO (P = 0.006, OR = 1.17, CI = 1.15–2.65) (Table 2).
Genotype distributions and allele frequencies of the MTHFR gene
P < 0.05 vs. MO.
Odds ratios for TT genotype in migraine vs. controls and in MO vs. MA cases
ns, not significant.
Discussion
In contrast to the two recent positive papers, our study showed a lack of association between the MTHFR C667T polymorphism and migraine. Our data suggest, however, that folate metabolism might play a role in the puzzling pathophysiology of MA. In fact, the MTHFR C677T polymorphism increases by over two-fold the risk for experiencing aura among migraineurs. The two previous studies found a much higher OR (6.5 and 3.5), for the risk of MA vs. controls in individuals carrying the TT genotype (3, 4). These authors also found an increased risk for migraine in general with this genotype. Both the somewhat smaller migraine sample and ethnicity could explain the overweight of the MTHFR TT genotype in MA and also in migraine in general as well as the broad confidence intervals given in these two previous studies. Nearly half of their MA patients in the Japanese study, in fact, shared the TT genotype. Such a high genotypic frequency has never been found even in patients with thromboembolic disease and would be unusual in a polygenic disorder as common as migraine is. As an example of possible ethnicity influence, high frequency of the mitochondrial 11084 mutation was reported in Japanese migraineurs (7), but this mutation has not been found in Caucasians with migraine (8, 9). In the study from Turkey there was just one case with a TT genotype within only 23 patients diagnosed as MA (4.3%) as compared to 2 TT cases in the control group (1.5%). Even though the numbers in our study should still be increased to optimize the statistical power, clearly the results of the Turkish study with the very small number of MA cases must be interpreted with great caution. In any case, both our results and those of Kowa et al. (3) coincide in that sharing the T677 allele is associated with a higher risk of experiencing aura among migraineurs.
Even though the concept of MO and MA as same or different entities is still debated, we agree with the position that migraine aura is a parallel track process to the pain, being facilitated or triggered by the same mechanism which is responsible for migraine pain and other symptoms (10). This hypothesis would also predict that migraine aura and headache would be dissociated genetically. In a polygenic and clinically heterogeneous condition such as migraine, it seems attractive to propose that there may be quite common gene polymorphisms which can predispose to experience aura within migraineurs. This could explain the very usual clinical observations of patients suffering both MA and MO episodes, the coexistence of MA and MO members in the same family or even why many migraine patients never experience a typical migraine aura.
What could the possible explanations for this apparent association between the C677T polymorphism and MA be? The C677T mutation in the MTHFR enzyme is the most common cause of moderate hyperhomocysteinaemia (11). This mutation produces a thermolabile enzyme that predisposes to hyperhomocysteinaemia, especially in folate- or vitamin B12-deficient patients. MTHFR also contains a flavin adenine dinucleotide (FAD), which raises the possibility that vitamin B2 status is a determinant of fasting homocysteine levels. There are several trials suggesting a beneficial effect of folate (11, 12) and vitamins B12 (13, 14) and B2 (15, 16) in migraine patients. Interestingly, the antimigraine effects of vitamin B2 were more patent in MA patients (15). Storer and Goadsby (17) developed a model of cranio-vascular pain using cats and demonstrated that spontaneous trigeminal cell firing was accelerated by DL-homocysteic acid, the oxidized derivative that mimics the effect of homocysteine on arteries, this suggesting that hyperhomocysteinaemia could sensitize the dura matter and/or cerebral arteries and predispose to migraine. The impact of MTHFR thermolabile variant on plasma homocysteine levels is still unclear (11). The few studies which have measured homocysteine levels in migraine individuals have offered contradictory results (18, 19). Evers et al. (18) however, found elevated serum levels of homocysteine only in MA patients, and not in MO or in tension-type headache, which would also concur with our results. Finally, both increased homocysteine levels and migraine, especially MA, are independent risk factors for brain infarction (11, 20–22). It would be interesting to test the MTHFR genotypes in patients with migrainous infarction as defined by the IHS.
Footnotes
Acknowledgements
This work was supported by Grants A28/02 of ‘Fundación Marqués deV aldecilla’ and by the ‘Centro Investigación de Enfermedades Neurológicas’, Nodo HUMV/UC, ISCIII, Spain.