Prof. Nollaig Parfrey, UCC Department of Pathology, Cork University Hospital, Cork, Ireland.
Objective.— We studied a large Irish Caucasian pedigree with familial hemiplegic migraine (FHM) with the aim of finding the causative gene mutation.
Background.— FHM is a rare autosomal-dominant subtype of migraine with aura, which is linked to 4 loci on chromosomes 19p13, 1q23, 2q24, and 1q31. The mutations responsible for hemiplegic migraine have been described in the CACNA1A gene (chromosome 19p13), ATP1A2 gene (chromosome 1q23), and SCN1A gene (chromosome 2q24).
Methods.— We performed linkage analyses in this family for chromosome 1q23 and performed mutation analysis of the ATP1A2 gene.
Results.— Linkage to the FHM2 locus on chromosome 1 was demonstrated. Mutation screening of the ATP1A2 gene revealed a G to C substitution in exon 22 resulting in a novel protein variant, D999H, which co-segregates with FHM within this pedigree and is absent in 50 unaffected individuals. This residue is also highly conserved across species.
Conclusions.— We propose that D999H is a novel FHM ATP1A2 mutation.
Familial hemiplegic migraine (FHM) is a rare autosomal-dominant subtype of migraine with aura (MA), which is classified by the International Classification of Headache Disorders as a familial MA that in cludes motor weakness.1 The worldwide prevalence of hemiplegic migraine is unknown. However, a Danish epidemiological study revealed a prevalence of 0.01%, with both familial and sporadic forms of this condition occurring with equal frequency.2
There are 4 genetic variants of FHM described. FHM type 1 (FHM1; MIM 141500), which is responsible for approximately 50% of all FHM families,3 is linked to chromosome 19p134 and is caused by mutations in the CACNA1A gene that encodes for the α1A subunit of the P/Q-type neuronal calcium channel, Cav2.1.5 FHM type 2 (FHM2; MIM 602681) accounts for 20-30% of known FHM families,6 and is due to mutations in the ATP1A2 gene, which is located on chromosome 1q23.7 The ATP1A2 gene encodes the α2 subunit of Na+-, K+-ATPase, a plasma membrane enzyme that countertransports Na+ and K+ across cell membranes.8 A third locus of FHM (FHM type 3, FHM3) has recently been described on chromosome 2q24 (MIM609634).9 A missense mutation in the neuronal voltage-gated sodium channel gene, SCN1A, was found in 3 FHM families in which CACNA1A and ATP1A2 gene mutations were excluded.9 In addition to the above established FHM loci, a fourth locus (MIM 607516) is described in an American family (of German/native American origin) linked to chromosome 1q31.10 No responsible gene has yet been reported in this locus.
Clinical manifestations of FHM usually occurs in the first 2 decades of life.11 Presentation may include migraine or nonmigraine headaches with motor, visual, sensory auras, and speech dysphasia.1 In addition, fever,12 confusion, coma,11,13-17 seizures,3,11,13,15-19 transient or permanent cerebellar signs,11,16,17,19-21 and rarely, mental retardation can occur.12,15,17,19,22 Permanent cerebellar signs are more common in FHM1,11,21 while seizures are a more frequent feature in FHM2.3,11,13,15-19 In this study, we investigated an Irish family with clinical features of FHM, determined linkage to the FHM2 locus, and described a novel mutation in the ATP1A2 gene.
MATERIALS AND METHODS
FHM Family.— A large Irish family with autosomaldominant FHM was recruited (Fig. 1). A total of 37 family members from the latter 3 generations were interviewed and examined by one of the authors (DF). Initially, 2 FHM probands (Fig. 1: IV:6, IV:33) belonging to 2 large FHM families were separately diagnosed. It was subsequently discovered that these 2 families were in fact related, yielding one large kindred. There are a total of 30 family members affected with FHM, of which 21 are still alive.Of these, 19 participated in this study, 1 declined participation, and another was living abroad. The histories of the deceased family members affected with FHM were obtained from relatives' historical accounts. The majority of relatives recognized the symptoms of hemiplegic migraine, as it was common within the family. In addition to the 19 participants with FHM, 4 family members who had migraine without aura (MO; Fig. 1: IV:3, IV:19, IV:39, V:7), 3 family members who had typical MA but without motor weakness (Fig. 1: IV:10, V:38, V:53), and 12 unaffected family members also participated in this study. The diagnosis of migraine was based on the International Classification of Headache Disorders.1 All clinical diagnoses of FHM and other types of migraine in this family were made prior to genetic testing. The majority of the FHM subjects also had MA attacks without the motor weakness and some of them also suffered from MO at other times (Fig. 1). Brain imaging was performed only in subjects attending the hospital or the migraine clinic as patients and was not routinely prescribed as part of this research.
All 19 FHM subjects described sensory and motor auras during their hemiplegic migraine attacks. All except 3 of these subjects also had visual auras (Fig. 1; V:6, V:40, V:54) and all except 3 individuals reported speech dysphasia during their migraine attacks (Fig. 1; V:6, V:40, V:54). In addition, all FHM subjects except V:6 and V:56 (Fig. 1) had accompanying dysarthria, gait ataxia, and upper limb clumsiness during their hemiplegic migraine attacks.
Additional clinical features found in the FHM subjects include epilepsy and mental retardation in subject V:40 (Fig. 1), single episodes of childhood febrile convulsions in 2 subjects (Fig. 1; IV:6, IV:12), and childhood seizures in 3 subjects (Fig. 1; III:2, V:2, V:6). Individual V:40 had complex partial seizures with secondary generalization. Subject III:2 was 77 years old and could only give a retrospective account of generalized tonic-clonic seizures in childhood. No previous medical records were available. Both V:2 and V:6 had absence seizures in childhood and historical accounts were obtained from their parents.
Mild cerebellar signs including subtle intention tremor and nystagmus were found in 5 subjects (Fig. 1; IV:2, IV:20, IV:30, IV:33, V:2) in between migraine attacks. These findings were noted incidentally on clinical examination. Apart from their migraine attacks, these subjects were otherwise asymptomatic. With the exception of IV:33 (Fig. 1), who had an MRI brain scan because she was a migraine clinic patient, none of the other subjects had brain imaging. The brain MRI was normal in IV:33. In particular, there were no cerebellar white matter changes or atrophy. Additional functional studies of cerebellar dysfunction on these FHM subjects were not performed.
Our study methods are in accordance with the Declaration of Helsinki and were approved by the Clinical Research Ethics Committee. Blood samples from affected and unaffected family members were taken after informed consent was obtained.
In addition, 50 Irish Caucasian control subjects were recruited in whom migraine and epilepsy phenotypes were excluded. Peripheral blood samples were obtained from these subjects after informed consent was obtained.
Linkage Analyses.— Genomic deoxyribonucleic acid (DNA) was extracted from peripheral blood lymphocytes using phenol-chloroform method or commercially available DNA extraction kits (Qiagen, Crawley, UK). DNA was then amplified by polymerase chain reaction (PCR) using fluorescently labeled primer pairs for microsatellite markers flanking the chromosomal regions of interest. The markers studied were D1S2635, D1S2707, and D1S2705 for the FHM2 locus on chromosome 1q23. All primer sequences are available on the University of California (UCSC)'s human genome browser database. Individuals were genotyped by analyzing the amplified PCR products using the GENESCAN software on the ABI 310 automated DNA analyzer (Applied Biosystems, Foster City, CA, USA). Linkage analysis was then performed using the FASTLINK software program. Information for allele frequencies was obtained from The Centre d'Etudes du Polymorphisme Humaine (CEPH) database. The disease model used was autosomal-dominant inheritance with 80% penetrance13 and the disease gene frequency was taken to be to be 0.0001.2
Mutation Screening.— For the FHM family, all exons of the ATP1A2 gene were screened by PCR amplification of genomic DNA from 2 affected subjects (Fig. 1: III:2, IV:6) and 2 unaffected subjects (Fig. 1: III:3, IV:21).The genomic structure was available from the UCSC human genome browser database. Details of the intron-exon structures and primer sequences are available on request. Amplified PCR products of all exons were purified (Jetquick Spin Columns, Genomed, Löhne, Germany) and cycle-sequenced using a dye terminator kit (Applied Biosystems). Sequence analysis was then performed using the ABI 310 automated sequencer (Applied Biosystems).
Once the segregating variant was identified, DNA from 50 Irish control subjects was screened for the putative mutation in exon 22 using restriction enzyme analysis. PCR amplification was carried out using the primers 5'AGAAGAGGCTGTTGGAAGA AGACA and 5'GCAGGAACCAGTAGTGGGAG TGGA. Amplified PCR products were incubated overnight at 37°C with the restriction enzyme NlaIII. The presence of the variant adds a restriction site in the 583 bp fragment; the digestion produces five DNA fragments for the wild-type allele (18, 26, 74, 112, and 353 bp) and six DNA fragments for the mutant alleles (18, 26, 74, 112, 135, and 218 bp).
Linkage Analyses.— Linkage to chromosome 1q23 was established with all three microsatellite markers achieving significant logarithm of the odds (LOD) scores (Table). The most significant LOD score was +5.6 for D1S2705 at θ = 0.
Table Table.—. Scores of Microsatellite Markers Showing Linkage to the FHM2 Locus (Chromosome 1q23)
Mutation Screening.— The nucleotide 2995 G→ C substitution in exon 22 of the ATP1A2 gene was found to be disease segregating within this FHM family (Fig. 2) and is present in all subjects affected with hemiplegic migraine. This point mutation results in the replacement of the acidic aspartic acid residue with a basic histidine at position 999 in the Na+-, K+-ATPase protein (D999H). This mutation was not found in the 100 control chromosomes that were screened.
In addition, examination of all the coding sequence of the ATP1A2 gene did not detect other mutations including those that have been previously described in the literature (Fig. 2).7,14,16-18,23-26
We describe a novel missense mutation, D999H, in the ATP1A2 gene in an Irish FHM pedigree. This is the first ATP1A2 gene mutation to be described in an Irish FHM family. There are 20 known ATP1A2 gene mutations causing FHM (Fig. 2), of which 18 are missense mutations.7,14,16-18,23-26
The ATP1A2 gene encodes for the α2 subunit of the Na+-, K+-ATPase plasma membrane enzyme, which couples ATP hydrolysis to countertransport Na+ and K+ across cell membranes. The Na+-, K+- ATPase protein is composed of three heteromeric subunits, the α catalytic subunit, the β regulatory subunit, and the γ subunit (function unknown).8 Of the α subunits, the α1 subunit is ubiquitously expressed while the α2 subunit is expressed in muscle, adipose tissue, and central nervous tissue, particularly astrocytes.8,27 This α2 subunit is composed of an N-terminal region containing 4 membrane-spanning domains (M1-4) and a C-terminal region containing 6 membrane-spanning domains (M5-10), which are linked by a large intracellular loop. This large M4-5 loop contains critical functional sites and undergoes major conformational changes during the enzymatic cycle.28 Eleven of the 20 known ATP1A2 gene mutations causing FHM2 are located within this loop (T345A, T376M, R593W, A606T, G615R, R689Q, E700K, D718N, M731T, L764P, and V628M; see Fig. 2).7,14,16-18,24,25,29 The D999H variant we have identified is located in the M10 trans-membrane domain, which is situated in the C-terminus cation transporting ATPase domain (Fig. 2).17 It is adjacent to the P979L mutation, which was associated with interictal mild cerebellar signs (mild dysarthria and nystagmus) in affected individuals.17 Interestingly, only the P979L and G301R ATP1A2 gene mutations were reported to cause interictal cerebellar signs in FHM.16,17 No functional studies were published on these mutations, although the G301R mutation is also located on a trans-membrane protein (Fig. 2; M3). With the exception of the W887R mutation, no functional studies were published for the other FHM2 causing ATP1A2 gene mutations located within the C-terminal region. Functional studies of the W887R mutation have shown a loss of function effect of the Na+-, K+-ATPase protein and have demonstrated that mutated alleles caused functional haplo-insufficiency of the β subunit binding site within this domain.7 As the Na+-, K+-ATPase exchanges intracellular Na+ for extracellular K+, loss of the Na+-, K+-ATPase function results in raised extracellular K+, which facilitates cortical spreading depression,30 a mechanism postulated to cause migraine aura.31 This ATPase dysfunction also results in raised intracellular Na+, which will increase intracellular Ca2+ levels as a result of a decrease in Na+-Ca2+ exchange. This raised intracellular Ca2+ also facilitates cortical spreading depression.32 Functional studies on CACNA1A gene mutations responsible for FHM also supports the effect of raised intracellular calcium in the pathogenesis of migraine aura.33 Other ATP1A2 gene mutations that have also demonstrated a loss of function effect of the Na+-, K+- ATPase protein include R593W, G615R, L764P, and D718N.7,17,25,26
Of the 19 FHM subjects within this family, 4 have histories of seizures and 2 have had histories of febrile convulsions. Seizures have been described more frequently in FHM2 families (9 of 15) than FHM1 families (5 individuals in 35 families). Also of interest is that 5 FHM subjects within this family demonstrated subtle interictal cerebellar signs. These findings were noted incidentally on routine clinical examination and confirmed by a senior neurologist. However, further functional assessment of cerebellar dysfunction and MRI brain imaging were not routinely performed. Only one subject, who attended the migraine clinic as a patient, had a MRI brain scan, which was normal. Cerebellar signs occurring during hemiplegic migraine attacks have been reported in FHM2 families.16 However, permanent cerebellar signs persisting between migraine attacks are uncommon in FHM2.11,21 Only 9 individuals in all of the FHM2 families reported to date exhibited interictal cerebellar signs.16,17 Another family with possible linkage to chromosome 1q21-23 also exhibited interictal cerebellar signs in the form of gazeevoked nystagmus and scanning speech in 3 family members.19
Within our pedigree, only one family member with the ATP1A2 D999H mutation did not have the clinical phenotype of FHM (Fig. 1: IV:37). This finding is consistent with incomplete penetrance as previously described for FHM2.13
It has been suggested that typical MA and FHM are part of a spectrum of migraine disorders associated with auras.11 One study found 2 novel ATP1A2 mutations in MA and MO,34 while another study failed to find any significant variants.35 We did not detect any ATP1A2 mutation in family members who do not have FHM but who suffer from either MA or MO. Our findings do not support ATP1A2 mutations as the pathogenesis of MA or MO.
The D999H mutation co-segregates with FHM within this family and was present in all the family members affected with hemiplegic migraine. This mutation was also excluded in 50 control subjects and is located in a residue that is highly conserved across different species (Fig. 3).These findings are highly suggestive that D999H is the novel causative ATP1A2 mutation for FHM in this Irish pedigree.
Acknowledgments: We thank the family members for their participation and Mr.David Shilling for his technical support.We also thank the Migraine Association of Ireland for their support in patient recruitment.