Autosomal Dominant Nocturnal Frontal Lobe Epilepsy in a Spanish Family With a Ser252Phe Mutation in the CHRNA4 Gene FREE

AUTOSOMAL dominant nocturnal frontal lobe epilepsy (ADNFLE) has been recently defined in an Australian kindred of English and French-Canadian origin.^{1 - 2} This epileptic syndrome has long been misdiagnosed as nocturnal paroxysmal dystonia,^{3 - 4} parasomnias, or psychiatric disturbances.⁵ It is characterized by clusters of brief nocturnal motor seizures, which originate in the frontal cortex.⁶ The clinical onset is in childhood. The results of interictal electroencephalogram (EEG) are usually normal, as are computed tomographic scans and magnetic resonance images. Seizures are usually well controlled by carbamazepine; however, the recurrence risk after drug withdrawal often persists for life. This syndrome is inherited as an autosomal dominant disorder with a penetrance of 70% to 80%.

A similar but not necessarily identical syndrome has been described in about 35 families of different ethnic origins (English, French-Canadian, Norwegian, and Spanish), including 28 Italian families who were identified in a specialized sleep disorder clinic.^{7 - 12}

Linkage to chromosome 20q13.3 was found in an Australian family,¹³ and a C→T transition was identified, causing a replacement of a serine with phenylalanine in codon 252 (homologous to Ser 248 of the Torpedo a subunit) in the α4-subunit (CHRNA4) of the neuronal nicotinic acetylcholine receptor.¹⁴ A second mutation in the same exon (an insertion of 3 nucleotides in codon 263) has been reported in a Norwegian family. Expression of mutated complementary DNA in oocytes confirmed the functional role of this insertion and suggested 2 possible mechanisms for the disorder: either a decrease or an increase in nicotinic receptor activity.^{9 ,15}

However, with the exception of these 2 families, no mutations were detected in the other families in the vicinity of codon 252, suggesting either the existence of mutations in other parts of exon 5, or in other exons of this gene, or that this disorder could be clinically and/or genetically heterogeneous.^{7 ,12} Recently, a second locus at 15q24 has been described, confirming the genetic heterogeneity of this disorder.¹⁶

We report a family from the south of Spain that fulfills the clinical criteria for ADNFLE and that carries the same mutation as that reported in the Australian family, confirming the high degree of clinical homogeneity in the phenotypic expression of this syndrome and the causative role of the Ser252Phe mutation.

Twenty-eight members of a Spanish family from Seville, including 11 affected individuals and 2 obligate carriers, underwent clinical, electrophysiologic, and genetic analysis. The pedigree is shown in Figure 1. Five of the patients were followed up for 10 years in the Outpatient Consultation of the Hospital Valme of Seville. All patients were reinterviewed by one of us (J.G.), using a semistructured questionnaire for epileptic disorders.¹⁷

Interictal EEG, with 19 electrodes positioned according to the international 10-20 system, using bipolar and referential montages, was performed in 9 patients (Table 1). Hyperventilation and intermittent photic stimulation were used as activators. In 4 patients, an EEG was obtained during diurnal sleep, without video monitoring.

Computed tomography was performed in 3 patients, and magnetic resonance images were obtained from 5 patients with the use of a 0.5-T field strength unit (Table 1).

After informed consent was obtained, venous blood samples were collected from 16 subjects (11 affected, 4 asymptomatic, and 1 relative). DNA was extracted from whole blood by standard procedures.

Four fluorescent markers (loci D20S120, D20S102, D20S171, and D20S173) from the Généthon map, spaced at an average distance of 15 centimorgans from each other,¹⁸ flanking the locus of the CHRNA4 gene on chromosome 20q13.3, were typed on the DNA from the 16 individuals sampled. Microsatellite primers were labeled with either 6-Fam, Hex, or Tet phosphoramides.¹⁹ Polymerase chain reactions were performed in 192-well microtiter plates in a final volume of 15 mL containing 30 ng of genomic DNA, 0.16-mmol/L dNTPs and NBL:, 1× NBL buffer, 0.33-mmol/L concentrations of each primer, and 0.03 U of Taq polymerase. After a hot-start procedure (enzyme is added at 94°C after denaturation for 5 minutes at 96°C), 26 polymerase chain reaction cycles were performed (40 seconds at 94°C, 30 seconds at 55°C). The rest of the procedure was similar to that described by Reed et al.¹⁹ Individuals 134702 and 88415 from the Centre d'Etude du Polymorphisme Humain (Paris, France) (CEPH) families¹⁸ were genotyped in each set of reactions as allele size references.

A simulation by the SLINK program²⁰ was performed with the use of 4 isofrequent alleles at each marker locus, assuming that the disorder is an autosomal dominant condition, with a frequency of the susceptibility allele of 0.001, a penetrance of 70%, and no phenocopies. Lod score computations were performed with the MLINK subroutine of the LINKAGE package (version 5.10).²¹ Allele frequencies used were based on unrelated individuals from 8 CEPH families. The analysis was also performed with the use of allele isofrequencies and/or different penetrance values, which produced no drastic modification of the results (data not shown).

Polymerase chain reaction fragments between nucleotides 535 and 825 of the fifth exon of the CHRNA4 gene were amplified with primers 5′-GGCGAGTGGGTCATCGTGG-3′ and 5′-GATGACCAGTGAGGTGGACG-3′, according to the previously published sequences.²² Polymerase chain reaction products were purified by means of P100 chromatography and sequenced directly by the AmpliTaq FS Dye Terminator cycle sequencing kit and ABI Prism 377 DNA sequencer.

As shown in the pedigree (Figure 1), there were 11 affected individuals and 2 obligate carriers. The disorder was transmitted in an autosomal dominant mode with incomplete penetrance. The age at seizure onset ranged from 3 to 12 years (mean, 7.6 years). Only 2 patients had antecedents of febrile seizures. No history of head trauma, meningitis, encephalitis, or any other neurologic condition that was likely to have caused epilepsy was found. The pregnancy, delivery, and postnatal period had been uneventful in all patients. Results of clinical examination were normal in all of the family members examined (11 affected patients, 3 asymptomatic carriers, and 2 healthy individuals). None of the patients had evidence of intellectual disability or psychiatric conditions.

Seizures always occurred during sleep periods, usually after the beginning of sleep. None of the patients manifested diurnal attacks. Seizures generally started with a sudden awakening with malaise and suffocation, followed by motor manifestations, with dystonic posturing of limbs and myoclonic jerks. Patients remained conscious but were frequently unable to speak during the episode. Secondary generalization was rare, and there were no postictal phenomena. Seizures were brief, shorter than 1 minute, and appeared in clusters with up to 20 seizures per night, alternating with periods with sporadic attacks or complete absence of symptoms. Two children, subjects III-5 and III-7, aged 3 and 5 years, respectively, had seizures consisting of brief episodes (approximately 1-2 minutes) of crying with absence of response that occurred in clusters, several times every night during several days. One child (subject III-6) occasionally had isolated episodes consisting of sitting in the bed, saying a few words, and continuing to sleep, without crying or showing terror. These episodes suggested somniloquia.

The interictal EEG was normal except in 2 patients who showed epileptic activity over frontal areas. The diurnal sleep EEG performed in 4 patients was normal. Neuroimaging was normal in all of the patients studied.

Carbamazepine was the most effective antiepileptic drug used, although 2 patients responded well to valproate sodium. Some patients developed tolerance, and after a period with a good response to drugs, there was poor control of seizures, even with high doses of antiepileptic drugs or polytherapy with several drugs.

Clinical information for the patients is summarized in Table 1.

Simulation of linkage by the SLINK program gave a lod score of 2.43 at θ=0. From the 4 fluorescent markers from the Généthon map, a suggestive linkage was found with locus D20S171 (lod score, 1.47 at θ=0) (data not shown). Sequencing of the portion of exon 5 of the CHRNA4 gene that contains the 2 previously described mutations was undertaken. A missense mutation C→T that causes replacement of a serine by phenylalanine at amino acid position 252 was detected (Figure 2). This mutation segregated with the disorder in all 11 affected members, 2 obligate carriers (subjects I-3 and II-2), and 1 asymptomatic child (subject III-2). Therefore, the infrequent episodes seen in the child III-6 cannot be associated with the disease, since she does not carry the mutation present in the other patients of the family. The spouse II-1 does not carry the mutation.

Two polymorphisms described previously²² and associated with the mutation Ser252Phe in the same exon at position 555 and 594 (consisting in both cases of a C→T transition in exon 5) were also analyzed. These 2 polymorphisms were not found.

The family we describe shows a homogeneous clinical picture that is consistent from one patient to another and is in accordance with that previously described in other families with ADNFLE.^{1 - 2 ,8 ,12}

Although this syndrome is rare, it is probably more common than previously suspected. Our confirmation of the mutation in the CHRNA4 gene in a third family shows that this condition is not merely anecdotal, as has been suggested, but is a well-defined clinical entity. This knowledge is important because it can help in the diagnosis in children whose parents are also affected. Diagnosis of this disorder in young children is difficult because the symptoms are often dismissed as nightmares.

The presence of the same mutation in 3 large families, with an almost identical clinical syndrome but with different ethnic origins, confirms the involvement of the CHRNA4 gene in ADNFLE. Two polymorphisms described previously in the Australian kindred²² and associated with the mutation Ser252Phe in the same exon at position 555 and 594 (consisting in both cases of a C→T transition in exon 5) were not found, excluding a common ancestral haplotype shared by the Australian and Spanish families. Quite surprisingly, no mutations in exon 5 were found in 15 families by Oldani et al.¹² Moreover, there is no linkage to markers in the 20q13.2-13.3 region in the 5 largest families. This could be because of the small number of markers in this region, their low degree of informativity, the large distance between the available markers leading to recombination events, or the sequence analysis of only a part of the fifth exon of the gene.

Nevertheless, as reported by others,^{7 ,16} genetic heterogeneity exists in ADNFLE, reinforced by the absence of strict inclusion criteria and the difficulty of ascertaining the clinical features when no EEG monitoring data are available.

Clinical characteristics of families with ADNFLE caused by mutations in the CHRNA4 gene are slightly different from those in families not linked to chromosome 20q. In both groups, onset is in childhood or adolescence, and the disorder is characterized by clusters of brief nocturnal motor seizures with hyperkinetic or dystonic manifestations, which persist through adult life. The seizures frequently respond to antiepileptic drugs, especially carbamazepine, although recurrence on withdrawal of medication is frequent.^{2 ,12} However, the presence of enuresis, diurnal attacks, or other subjective complaints, such as tiredness on awakening or excessive diurnal somnolence, has only been reported in families without linkage to chromosome 20q. Therefore, ADNFLE is a clinically and genetically heterogeneous disorder, although some genotype-phenotype correlations exist.

Accepted for publication February 16, 1999.

This work was supported by the Association Française contre les Myopathies, Paris, France, and the Health Department of the Basque Government, Vitoria, Basque Country, Spain (1996-1997).

Reprints: Amets Sáenz, BSc, Experimental Unit, Hospital Ntra. Sra. de Aránzazu, P/ Dr Beguiristain s/n, Apdo Correos 477, 20080 San Sebastián (Basque Country), Spain (e-mail: uniexpe2@chdo.osakidetza.net).