OUP user menu

Mutations of sodium channel α subunit type 1 (SCN1A) in intractable childhood epilepsies with frequent generalized tonic–clonic seizures

Tateki Fujiwara, Takashi Sugawara, Emi Mazaki‐Miyazaki, Yukitoshi Takahashi, Katsuyuki Fukushima, Masako Watanabe, Keita Hara, Tateki Morikawa, Kazuichi Yagi, Kazuhiro Yamakawa, Yushi Inoue
DOI: http://dx.doi.org/10.1093/brain/awg053 531-546 First published online: 1 March 2003


A group of infant onset epilepsies manifest very frequent generalized tonic–clonic seizures (GTC) intractable to medical therapy, which may or may not be accompanied by minor seizures such as myoclonic seizures, absences and partial seizures. They include severe myoclonic epilepsy in infancy (SMEI) and intractable childhood epilepsy with GTC (ICEGTC). They are commonly associated with fever‐sensitivity, family history of seizure disorders and developmental decline after seizure onset. Mutations of the neuronal voltage‐gated sodium channel α subunit type 1 gene (SCN1A) were recently reported in SMEI patients. To clarify the genotypic differences in this group of epilepsies, we searched for SCN1A abnormalities in 25 patients with SMEI and 10 with ICEGTC, together with the family members of 15 patients. Frameshift mutations in SCN1A were observed in four patients, nonsense mutations in five patients, missense mutations in 21 patients, other mutations in two patients and no mutation in five patients. SMEI patients showed nonsense mutations, frameshifts, or missense mutations, while ICEGTC patients showed only missense mutations. Study of both parents of 11 patients revealed that the mutations in these patients were de novo. However, two mothers had the same missense mutations as their ICEGTC children, and they had generalized epilepsy with febrile seizures plus. Here we suggest that SMEI and ICEGTC represent a continuum with minor phenotypic and genotypic differences.

  • Keywords: intractable childhood epilepsies; severe myoclonic epilepsy in infancy; generalized tonic–clonic seizures; febrile seizures; voltage‐gated sodium channel
  • Abbreviations: FS = febrile seizures; GEFS+ = generalized epilepsy with febrile seizures plus; GTC = generalized tonic–clonic seizures; ICEGTC = intractable childhood epilepsy with generalized tonic–clonic seizures; PCR = polymerase chain reaction; SCN1A = neuronal voltage‐gated sodium channel α subunit type 1 gene; SMEI = severe myoclonic epilepsy in infancy


Childhood epilepsies with frequent generalized tonic–clonic seizures (GTC)

A group of patients manifest very frequent and intractable convulsions from infancy. The seizures are often induced by fever, and may be accompanied later by additional seizures such as myoclonic seizures, absence seizures or complex partial seizures. The clinical entity of this epilepsy type was first described by Dravet (1978) and termed ‘severe myoclonic epilepsy in infancy’ (SMEI) or Dravet syndrome. This epileptic syndrome is characterized by fever‐sensitive and refractory generalized clonic, GTC or unilateral seizures beginning in the first year of life (Dravet, 1978; Dravet et al., 1982, 1992; Commission on Classification and Terminology of the International League against Epilepsy, 1989). The seizures often culminate in status epilepticus. Development before seizure onset is normal, usually with no prior injury to the brain. Appearance of myoclonic seizures during the course of the disease confirms the diagnosis of SMEI. Some patients have additional complex partial seizures or absence seizures, and the psychomotor development declines in almost all patients.

Seino and Higashi (1979), on the other hand, reported a group of 24 patients with refractory epilepsy in childhood characterized by frequently occurring GTC. The seizures began before 1 year of age, often induced by fever, and tended to recur in cluster or status. All patients showed mental decline, and some showed ataxia or hypotonia. The EEG was characterized by diffuse theta waves, and spikes or sharp waves were rarely observed. This group of patients was later designated ‘high voltage slow wave–grand mal syndrome’ due to the characteristic EEG features (Wada et al., 1983; Kanazawa, 1992), or ‘intractable childhood epilepsy with generalized tonic–clonic seizures’ (ICEGTC) (Watanabe et al., 1989a, b; Fujiwara et al., 1992).

The relation of these two epilepsy groups, or the borderland of SMEI (Ernst et al., 1988; Ogino et al., 1988; Oguni et al., 1994, 2001; Kanazawa, 2001), has been intensively discussed. Watanabe et al. (1989a, b) investigated 82 patients who had intractable and frequent generalized clonic, tonic–clonic, or unilateral clonic seizures since the first year of life without preceding pathological events. They distinguished three groups: group 1 (31 patients) without myoclonic or absence seizures (ICEGTC); group 2 (30 patients) with myoclonic seizures and, in some cases, with absence or partial seizures (SMEI); and group 3 (21 patients) with myoclonic seizures long after infancy or frequent absences instead of myoclonias. There were no differences between the three groups in onset age, sex, family history, neurological abnormalities and EEG findings. Nine patients in group 1, 10 in group 2 and 10 in group 3 were followed for >5 years (Fujiwara et al., 1992). Follow‐up results showed that, in all the three groups, the convulsive seizures persisted whereas the other types of seizures tended to disappear as the patients grew, and the prognosis of mental development was poor. Other authors supported this finding. Ohki and colleagues reported that the most consistent core seizure type was tonic–clonic convulsion and the background EEG activity changed to a pattern of centro‐parietal dominant theta waves after 1 year of age (Ohki et al., 1997). These characteristic EEG activities persisted, although the seizure symptoms other than GTC varied considerably among the patients. The persistence of GTC in the long‐term follow‐up was also noted by Giovanardi Rossi et al. (1991), even though they indicated the increased importance of partial seizures.

Thus, the major feature that differentiates between SMEI and ICEGTC is the presence or absence of myoclonic seizures, although minor differential features have been suggested, such as the predominance of typical generalized clonic–tonic–clonic convulsions in SMEI (Fujiwara et al., 1992). It was also suggested that the incidence of severe mental dysfunction and abnormal brain imaging was slightly lower and the prognosis was slightly better in ICEGTC (Kanazawa, 1992).

Ogino et al. (1989) investigated 10 patients without myoclonia or absences, and found that these patients had many characteristics in common with SMEI except the absence of myoclonic seizures, and speculated that common pathophysiological factors may be involved in these cases and SMEI. Kanazawa (1992, 2001) considered that the two groups represent portions of a continuum and speculated that GTC in combination with other seizure types complicate the pathophysiological conditions; especially, the combination with myoclonic seizures increases the severity of epilepsy.

Aicardi (1994) reported that 18 out of 50 patients (36%) in his series had few or no myoclonic seizures. Aicardi (1991, 1994) therefore insisted that the cases without myoclonia also belong to the same epileptic syndrome as typical SMEI cases, and that transition exists ranging from full‐fledged cases with intense myoclonic activity to non‐myoclonic cases. He observed that, in one sibship, one child manifested SMEI whilst the other had convulsions only. He therefore considered the term SMEI misleading and that ‘severe polymorphic epilepsy of infants’ might be more appropriate.

Doose et al. (1998) investigated patients with epilepsy of infancy with frequent and often prolonged febrile and afebrile GTC as the only seizure type. The most common triggering factor was fever or immersion in a hot bath. With advancing age, the symptomatology became increasingly polymorphic due to additional seizure symptoms. EEG was normal during the first weeks or even months, and subsequently all patients almost regularly exhibited diffuse 4–7 Hz rhythms. In the subsequent course, the EEG became increasingly polymorphic. Doose and colleagues therefore speculated that epilepsy of infancy with GTC represents different parts of a broad spectrum of related conditions, overlapping with other forms of early childhood epilepsy such as SMEI, a severe type of myoclonic–astatic epilepsy, and early childhood absence epilepsy with GTC (Doose et al., 1998).

In summary, accumulating evidence has gradually brought about a consensus that intractable childhood epilepsies with frequent GTC consist of a continuum with phenotypes such as cases with minor seizures mainly myoclonic seizures (SMEI), and cases without minor seizures (ICEGTC).

Clinical genetic background

SMEI has often been associated with a family history of convulsive disorders (Scheffer et al., 2001). Benlounis et al. (2001) found febrile convulsions in 22 relatives and epilepsy in nine relatives of 65 SMEI patients, while Singh et al. (2001) also reported 39 family members of 12 SMEI probands to be affected by seizure disorders. Several studies have reported occurrence of SMEI in siblings (Dravet et al., 1992; Singh et al., 2001; Veggiotti et al., 2001).

Likewise, a family history of seizures was reported in 24% of patients with intractable GTC by Doose et al. (1998), in four of 10 cases by Ogino et al. (1989), and in four of 10 intractable GTC cases other than SMEI by Kanazawa (1992). Aicardi (1994) and Ogino et al. (1989) described pairs of siblings in whom one manifested the complete picture of SMEI, while the other had only GTC. Doose et al. (1998) found characteristic EEG traits of diffuse 4–7 Hz rhythms also in the relatives, indicating the role of genetic factors in the pathogenesis of these GTC epilepsies.

These data thus suggested that these epilepsies might share a common genetic background.

Mutations of SCN1A

Recently, mutations of the neuronal voltage‐gated sodium channel α subunit type 1 gene (SCN1A) were found in SMEI (Claes et al., 2001). We also demonstrated the mutation to be frameshift or nonsense type, although there were some cases showing no mutation in SCN1A (Sugawara et al., 2002).

Mutations of SCN1A were also reported in patients with generalized epilepsy with febrile seizures plus (GEFS+) in the form of missense type (Escayg et al., 2000, 2001; Wallace et al., 2001; Sugawara et al., 2001; Ito et al., 2002). GEFS+ is characterized by a wide spectrum of phenotypes including febrile seizures (FS), FS plus (either FS continuing beyond the age of 6 years and/or afebrile GTC) and FS plus with other minor seizure types (Scheffer et al., 1997; Singh et al., 1999; Ito et al., 2002). Severe form of epilepsy such as myoclonic–astatic epilepsy was also included in this familial epilepsy syndrome (Singh et al., 1999). Because the family history of seizures in SMEI is in keeping with the spectrum of seizure phenotypes seen in GEFS+, Singh et al. (2001) suggested that SMEI might also be included in GEFS+ and that it represents the most severe phenotype in the GEFS+ spectrum.

This study aimed: (i) to confirm our previous finding regarding SMEI; (ii) to extend the molecular genetic investigation to ICEGTC; and (iii) to clarify the relation between these GTC epilepsies and GEFS+ syndrome. We hypothesized that: (i) the genetic abnormalities of intractable GTC epilepsies with and without minor seizures can be found in the Na channel gene but with slight genotypic differences corresponding to the varied phenotypes, and (ii) these mutations are also different from those of the conventional GEFS+ syndrome because the GTC epilepsies are far more severe and intractable than conventional GEFS+ syndrome.

Patients and methods


We studied one pair of monozygotic twin, the phenotype of whom had been reported in detail elsewhere (Fujiwara et al., 1990), and 33 unrelated Japanese patients with intractable epilepsy characterized by frequent GTC and an early onset of seizures (before 1 year of age). All patients were highly sensitive to an increase of body temperature caused by fever or taking a bath.

The 35 patients consisted of 20 males and 15 females (see Table 1) ranging in age from two to 31 years (mean 15.7 years) at the time of study. They were followed for 0–27 years (mean 12.2 years). Seizure first started between 1 and 11 months of age (with a majority between the 3rd and 10th month) as unilateral clonic, generalized clonic or GTC. Personal antecedent was noted in Patient 12 who had mild asphyxia at delivery, and in Patients 17 and 34 who were small at delivery (without asphyxia). All patients including Patients 12, 17 and 34 developed normally before the seizure onset. In all patients, the convulsive seizures tended to occur in status epilepticus or in cluster, especially during the early course of the epilepsies.

View this table:
Table 1

Clinical features of 35 patients and the results of mutation analysis

Patient no.DiagnosisSexPresent age (years)Age of seizure onsetFamily historyPast historySeizure typesSeizure frequency (present)Neurological findingsEEG findings (other than HVS)NeuroradiologyMutation
1SMEIMale1911 monthsMonozygotic twin: SMEI (case 2); maternal uncle: febrile convulsionNoneClonic seizure: 11 months–; GTC: 1 year–; myoclonic seizure: 2–3 years; cps: 2 years 9 monthsGTC: monthlySevere MDRare spikes, rare multifocal spikes, sp–wCT: normalc.3637C→T, R1213X
2SMEIMale1910 monthsMonozygotic twin: SMEI (case 1); maternal uncle: febrile convulsionNoneClonic seizure: 10 months–; GTC: 1 year–; myoclonic seizure: 2 years 4 months–3 years 3 months; cps: 2 years 11 monthsGTC: weeklySevere MDRare spikes, rare multifocal spikes, sp–wCT: normalc.3637C→T, R1213X
3SMEIFemale103 monthsPaternal cousin: febrile convulsionNoneMyoclonic seizure: 3 months–3 years; cps: 5 months–; GTC: 8 months–; absence: 1year–GTC: weekly; cps: weekly; absence: dailyAtaxia, severe MD, slight hypotonusF‐dominant sp–w, poly sp–w; MEG, B–F dipole clusterCT: lateral ventricle enlargement, L>Rc.4223G→A, W1408X
4SMEIMale172 monthsNoneNoneGTC: 9 months–; myoclonic seizure: 11 months–4 years; cps: 2 months–GTC: daily; cps: weeklyAtaxia, severe MDRare bilateral diffuse sp–wCT: normalc.5419delATGTTCTATGAG, 1807delMFYE
5SMEIFemale24 monthsBrother: partial epilepsyNoneGTC: 4 months–; myoclonic seizure: 11 months–GTC: weekly; myoclonia: dailyModerate MDNo epileptic paroxysmsCT: slight frontotemporal atrophyexon 4 splice site GT→AT
6SMEIFemale113 monthsMother, aunt: febrile convulsionNoneClonic seizure: 3 months; hemiclonic‐ or GTC: 5 months–; cps: 9 months–2 years; myoclonic seizure: 5 months–2 years 6 months; absence: 2 yearsGTC: weeklySevere MDSpikes inF, OCT, MRI: normalc.2855G→A, W952X
7SMEIMale254 monthsNoneNoneHemiclonic seizure: 4 months–; GTC: 6 months–; myoclonic seizure: 1–3 yearsGTC: weeklyAtaxia, severe MDsp–w, spikes, multifocal spikesCT: moderate atrophyc.3852G→A, W1284X
8SMEIMale27 monthsUncle: febrile convulsionNoneClonic seizure: 7 months–; GTC: 1 year–; myoclonic seizure: 1 year–GTC: monthly; myoclonia: dailyModerate MDSpikes in F, TMRI: normalc.2835delC, 946RfsX953
9SMEIFemale210 monthsMaternal cousin, paternal cousin: epilepspyNoneGTC: 4 months–; myoclonic seizure: 4 months–; cps: 1 year 3 months–GTC: monthly; myoclonia: daily; cps: monthlyNDNo epileptic paroxysmsCT: normalc.530delG, G177 fs X180, c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
10SMEIMale112 monthsFather, maternal cousin: febrile convulsion; second cousin: clonic seizures at age 2 yearsNoneHemiclonic seizure: 2 months–; GTC: 5 months–; myoclonic seizure: 6 months–?; absence: 6 months–?; cps: 6 months–9 yearsGTC: weeklyAtaxia, severe MD, slight pyramidal signsp–w, poly sp–w, multifocal spikes, photo sensitivityCT, MRI: normalc.2196insCACCCTGT, Q732fsX749
11SMEIMale246 monthsFather: febrile convulsionNoneHemiclonic seizure: 6 months–4 years; GTC: 1 year–; myoclonic seizure: 2–21 years; cps: 13 months–4 years; absence: 2–7 yearsGTC: weeklyAtaxia, severe MDsp–w, poly sp–w, multifocal spikesCT: slight atrophyc.5292insT, F1765fsX1794, c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
12SMEIFemale283 monthsMaternal aunt: febrile convulsionAsphyxiaGTC: 6 months–; myoclonic seizure: 3–5 months; 2 years 6 months–4 years; cps: 7 yearsGTC: monthlyAtaxia, severe MDRare sp–wCT: normalc.307A→G, S103G (0/109), c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
13SMEIMale196 monthsMaternal uncle: febrile convulsionNoneGTC: 6 months–; myoclonic seizure: 2–4 yearsGTC: weeklyAtaxia, severe MDPoly sp–w, rare spikes in FCT: normalc.5054C→A, A1685D (0/109)
14SMEIFemale117 monthsNoneNoneHemiclonic seizure: 7 months–; GTC: 7 months–; myoclonic seizure: 11–13 months, 7 years; absence: 11–13 months; cps: 5 yearsGTC: monthlyAtaxia, severe MDRare sp–w, rare multifocal spikesCT, MRI: normalc.335C→T,T112I (0/109)
15SMEIMale77 monthsNoneNoneMyoclonic seizure: 4–6 years; hemiclonic seizure: 7 months–; GTC: 10 months–; cps: 4 yearsGTC: yearlyModerate MDPoly sp–w, sp–w, multiple spikes; MEG: diffuse dipolesMRI: normalc.3693T→A, S1231R (0/94)
16SMEIFemale254 monthsNoneNoneHemiclonic seizure: 4 months–; GTC: 4 months–; myoclonic seizure: 3–6 years; cps: 3 yearsGTC: monthlySevere MDsp–w, spikes in B–F, L–FCT: normalc.2954A→T, N985I (0/93)
17SMEIFemale148 monthsNoneSFD (2270 g)Myoclonic seizure: 14 months–13 years; hemiclonic‐ or GTC: 8 months–; cps: 2 years; absence: 14 months–13 yearsGTC: weeklyAtaxia, severe MDsp–w, poly sp–w, photosensitivityCT: slight atrophyc.5492T→C, F1831S (0/94)
18SMEIMale35 monthsPaternal cousin: febrile convulsionNoneGTC: 5 months–; cps: 6 months; myoclonic seizure: 11 months–GTC: yearly; myoclonia: dailySlight MDMultifocal spikes, sp–w, poly sp–w, photosensitivityCT: normalc.793G→T, G265W (0/95)
19SMEIFemale286 monthsUncle:, febrile convulsionNoneHemiclonic‐ or GTC: 6 months–; myoclonic seizure: 6–11 years; absence: rareGTC: monthlySevere MDsp–w, rare focal spikesCT: slight atrophyc.2878A→G, M960V (0/93)
20SMEIMale275 monthsNoneNoneHemiclonic‐ or GTC: 5 months–; myoclonicseizure: 1 year 6 months–4 years 3 months; cps: 1 year 6 months–2 years 11 monthsGTC: yearlyAtaxia, severe MDsp–w, multifocal spikesCT: moderate atrophyc.5434T→G, W1812G(0/94), c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
21SMEIFemale1910 monthsPaternal uncle: febrile convulsionNoneGTC: 10 months–; myoclonic seizure: 3 years 1 month–4 years; cps: 13 yearsGTC: weeklyModerate MD Rare spikes in F; MEG: B–CP dipole clusterMRI: mild atrophyc.3199A→G, T1067A (A/G 20/103, G/G 2/103)
22SMEIMale313 monthsPaternalcousin, maternal uncle: febrile convulsionNoneGTC: 3 months–; absence: 4 years; myoclonic seizure: 1 year 6 months–5 yearsGTC: yearlyAtaxia, severe MDRare sp–w, spike in R–FCT: slight atrophy, L>RND
23SMEIFemale91 monthsBrother: febrile convulsionNoneHemiclonic seizure: 1 month–; GTC: 4 months–; myoclonic seizure: 3 years–GTC: monthly; myoclonia: dailySlight MDPoly sp–w, rare spikes in L>R–O; photosensitivityMRI: normalND
24Atypical SMEIFemale2610 monthsPaternal aunt:isolated convulsionNoneAbsence with myoclonia: 2 years 6 months–4 years; clonic seizure or GTC: 10 months–21 yearsFree for 5 yearsSevere MDRare sp–w, poly sp–wCT: normalc.3789C→G, F1263L (0/95)
25Atypical SMEIMale254 monthsNoneNoneHemiclonic‐ or GTC: 4 months–; absence: 1 year–GTC: weekly; absence: rareAtaxia, severe MDsp–w, poly sp–w, rare multifocal spikes; photo‐ and pattern sensitivityCT: slight atrophyc.1028G→A, G343E (0/111)
26ICEGTCMale194 monthsSister: Down’s syndrome; maternal cousin: epilepsyNoneHemiclonic seizure: 4 months–; GTC: 7 months–; cps: 7–17 yearsGTC: monthlySevere MDB–F sharp wave, spike‐slowCT: normalc.2935G→A, G979R (0/93)
27ICEGTCMale259 monthsMother: epilepsy; sister: congenital heart diseaseNoneGTC: 9 months–GTC: monthlyAtaxia, moderate MDSpike in R–F, sp–w, multifocal spikesMRI: normalc.5126C→T, T1709I (0/109), c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
28ICEGTCMale183 monthsNoneNoneGTC: 3 months–; cps: 1–17 yearsGTC: weeklySevere MDNo epileptic paroxysmsNot donec.4894C→T, P1632S (0/109)
29ICEGTCFemale135 monthsNoneNoneGTC: 5 months–; cps: 1 year 2 months–7 yearsGTC: weeklyAtaxia, severe MD, autisticRare sp‐slow, spikes in L‐, R–T, B–OCT, MRI: normalND
30ICEGTCFemale54 monthsMother: febrile convulsionNoneClonic–, hemiclonic‐, GTC: 4 months–GTC: weeklyModerate MDsp–w, multifocal spikesCT, MRI: normalND
31ICEGTCMale213 monthsMaternal cousin: febrile convulsionNoneHemiclonic seizure: 3 months–; GTC: 1 year–; cps: 2 yearsGTC: dailyAtaxia, severe MD, pyramidal signMultifocal spikesCT: normalc.2948T→C, V983A (0/93), c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
32ICEGTCFemale2511monthsNoneNoneGTC: 11 months– GTC: yearlyAtaxia, moderate MD, L‐hemiparesisSpikes in L–FT, spike‐slow CT: normalc.5422T→C, F1808L (0/94)
33ICEGTCMale39 monthsNoneNoneHemiclonic seizure: 9 months–; GTC: 1 year–GTC: monthlyModerate MD, hyperkineticRare sharp waves in L–TCT: normalc.2422A→T, T808S (0/95), c.3032A→T, N1011I (0/103), c.3199A→G, T1067A (A/G 20/103, G/G 2/103)
34ICEGTCMale38 monthsMother: febrile convulsionsSFDGTC: 8 months–GTC: yearlyNDNo epileptic paroxysmsCT: normalc.4831G→T, V1611F (0/93)
35ICEGTCMale37 monthsNoneNoneHemiclonic seizure: 7 months–; GTC: 11 months–GTC: weeklyModerate MDRare spikesMRI: normalND

*The number of amino acids is as descriped by Escayg et al. (2000). The numerals in parentheses indicate the number of subjects with mutation per investigated control population. Abbreviations: B = bilateral; C = central; cps = complex partial seizure; F = frontal; HVS = high voltage slow waves; L = left; MD = mental decline; MEG = magnetoencephalography; ND = not detected; O = occipital; P = parietal; R = right; SFD = small for date; sp–w = spike–wave complex; T = temporal.

Neuroimaging findings were often normal or mildly atrophic. Psychomotor development slowed down after the second year, and ataxia was observed in 16 patients. No gross neurological deficits such as hemiplegia were observed except in Patient 32.

Family history showed febrile convulsions in the relatives of 14 patients, isolated convulsion in the relative of one patient, epilepsy in the relatives of four patients, and both epilepsy and febrile convulsion in the relatives of one patient and a pair of twins. In sum, the family history for seizures was positive in 22 patients (62.8%).

EEG findings in the 35 patients are shown in Table 1. In some cases, epileptic discharges were not observed at the initial examination but became evident thereafter. High voltage 4–7 Hz diffuse slow background activities were observed in the majority of cases. Photosensitivity or pattern sensitivity was noted in five patients.

During the course of the epilepsies, myoclonic seizures were observed in 23 patients, while atypical absences were observed in 10 patients and complex partial seizures in 20 patients. The GTC were refractory to all kinds of antiepileptic medication at maximal doses. At present, GTC persist in all but one patient (Patient 24). In most cases, frequency of GTC was between weekly and monthly occurrence.

The 35 patients were divided into two groups according to the presence or absence of ‘combined’ generalized seizures other than GTC throughout the course of epilepsy. Twenty‐three patients (Patients 1–23 in Table 1), who had myoclonic seizures met the criteria of SMEI according to the proposal of the Commission on Classification and Terminology of the International League against Epilepsy (1989). The history of perinatal problem in Patient 12 was exceptional, but all other features of this patient converged to a diagnosis of SMEI. Two patients (Patients 24 and 25) had frequent absence seizures (Patient 25 had very frequent self‐induced photosensitive absence seizures with myoclonia) and refractory GTC, but no overt myoclonic seizures. We included these patients in the SMEI group designating as atypical SMEI. Ten patients (Patients 26–35 in Table 1) were classified as ICEGTC according to the definition mentioned above. Although three young patients (Patients 33–35 in Table 1), aged 3 years, might have minor seizures such as myoclonic seizures in later life, they were included in the ICEGTC group.

The blood samples of both parents of 12 patients (Patients 1, 2, 8, 13, 15, 18, 20, 24, 29, 31, 33 and 34), the mothers of two patients (Patients 12 and 17) and a father of one patient (Patient 26) were also studied, in addition to those of 111 normal unrelated volunteers whose results served as control.

Statistical analysis was performed using the Kruskal–Wallis test and one‐way factorial ANOVA (analysis of variance).

Written informed consent was obtained from the parents or responsible adults where necessary, and the study protocol was approved by the Ethical Committees of Shizuoka Medical Institute of Neurological Disorders and of the Institutional Review Board of RIKEN‐BSI.

Mutational analysis

Genomic DNA was extracted from heparinized blood samples of affected and unaffected individuals using a commercially available kit (QIAamp DNA Blood Midi Kit, Qiagen Research Genetics, Valencia, CA, USA). This DNA was amplified by polymerase chain reaction (PCR) with Pyrobest (TAKARA Shuzo Co. Ltd, Kyoto, Japan) and analysed by sequencing. PCR primers were designed to amplify all 26 coding exons of SCN1A defined by a comparison of the cDNA (GenBank accession No. AY043484) and genome sequences (Nos. AC010127 and AC021673). PCR products were sequenced directly with a dideoxy terminator kit and analysed by an automated sequencer (model ABI3700, Applied Biosystems, NJ, USA).

In order to confirm the DNA sequencing results, either of two additional methods was used. In the first method, mutant alleles were detected by the addition or loss of restriction enzyme sites within the PCR products. In the second method, the PCR products were subcloned. A single deoxyadenosine nucleotide was transferred to the 3′ ends of PCR products. The A‐tailed PCR products were then ligated into pCR®2.1 TOPO® plasmid vector (Invitrogen, Carlsbad, CA, USA), heat‐shocked into competent E.coli DH5α‐T1® (Invitrogen, Carlsbad, CA, USA), and grown on LB agar plates containing 100 µg/ml ampicillin (Nacalai tesque, Kyoto, Japan), 30 µM isopropyl β‐d‐thiogalactoside (IPTG) (Nacalai tesque) and 40 µg/ml X‐gal (Nacalai tesque). Ten or more distinct single colonies were picked and sequenced.


Mutational analysis was performed on all coding exons and splice sites of SCN1A. We detected a total of 30 heterozygous mutations in SCN1A, consisting of four frameshift mutations in four patients, four nonsense mutations in five patients, 20 missense mutations in 21 patients, and two other mutations in two patients. The presence of these mutations was confirmed by either of two additional methods: the restriction digestion of created or deleted restriction enzyme sites of the PCR products, or the sequencing of the ‘subcloned’ PCR products. The location of each mutation in SCN1A is illustrated in Fig. 1. The four frameshift mutations, a one‐base insertion (c.5292insT), an eight‐base insertion (c.2196–2203insCACCCTGT) and two one‐base deletions (c.2835delC and c.530delG) resulted in intragenic stop codons of F1765fsX1794, Q732fsX749, 946RfsX953 and G177fsX180. The four nonsense mutations (c.3637C→T, R1213X in two patients; c.4223G→A, W1408X; c.2855G→A, W952X; and c.3825G→A, W1284X) also resulted in truncated channels. Exon 4 splice site (GT→AT) mutation may lead to cryptic splicing or exon skipping resulted in generation of affected channel. A 12 nucleotide deletion (c.5419delATGTTCTATGAG, 1807delMFYE) was found at the C‐terminus region. The 19 missense mutations (c.307A→G, S103G; c.5054C→A, A1685D; c.335C→T, T112I; c.3693T→A, S1231R; c.2954A→T, N985I; c.5492T→C, F1831S; c.793G→T, G265W; c.2878A→G, M960V; c.5434T→G, W1812G; c.3789C→G, F1263L; c.1028G→A, G343E; c.2935G→A, G979R; c.5126C→T, T1709I; c.4894C→T, P1632S; c.2948T→C, V983A; c.5422T→C, F1808L; c.2422A→T, T808S; c.3032A→T, N1011I; c.4831G→T, V1611F) resulted in amino acid sequence not found in the control population (n = 93–111). However, a missense mutation (c.3199A→G, T1067A) detected in Patients 9, 11, 12, 20, 21, 27, 31 and 33 resulted in an amino acid sequence also found to be heterozygous in 20 individuals and homozygous in two individuals out of 103 control subjects. In the missense mutations, the amino acids were fairly well conserved as shown in Fig. 2. No mutation was detected in the coding region of SCN1A in five patients (Patients 22, 23, 29, 30 and 35).

Fig. 1 Mutations of voltage‐gated Na+ channel Nav1.1 gene in intractable childhood epilepsies with frequent GTC. (A) Diagram of the Nav1.1 protein showing four domains, each with six transmembrane helices. Filled square: nonsense mutations. Filled triangle: frameshift mutations leading to premature stop codons. Filled circle: missense mutations found in SMEI cases. Open circle: missense mutations found in the ICEGTC cases. Filled star: other mutation. A missense mutation (c.3199A→G, T1067A) considered to be a benign variant is excluded. Mutations of the SMEI cases include the results of our previous study (Sugawara et al., 2002); these are marked with an asterisk). The numbers of the patients are indicated in parentheses. (B) The c.3789C→G mutation in Patient 24 (arrow) creates an extra HinfI site within the PCR products. HinfI digestion resulted in additional two small fragments.

Fig. 2 Evolutionary conservations of SCN1A. Partial amino acid sequence alignments of SCN1A and other α‐subunit family members (GenBank accession nos. AY043484, M22253, AH010232, NM_012647, AJ251507, Y00766, M81758, M77235, AB027567.1, X82835, AF117907.1, AF188679, D37977, M22252). The alignments were performed using MacVector version 7.0 (Symantec Corp., Cupentino, CA, USA). Arrows above the sequences indicate the positions of the missense mutations identified in the present study. Fugu = Fugu rubripes; Electric eel = Electrophorus electricus.

Both parents of Patients 1, 2, 8, 13, 15, 18, 20, 24, 29, 31 and 33 were found to have normal alleles, indicating that the mutations in these patients were de novo. The mother of Patient 12 and the fathers of Patients 26 and 34 had also normal alleles. However, the mothers of Patients 27 and 34 had the same missense mutation as the patients. The mother of Patient 27 (c.5126C→T, T1709I) had a history of febrile convulsions until 6 years of age. She had febrile GTC since 14 years of age, initially several times a year, then gradually less on medication including Valproate. Although seizure freedom extended to several years, she then had occasional GTC when incompliant to medication. GTC occurred without a clear link to the awakening situation. The EEG showed diffuse bilateral spike waves, although rarely. CT was normal. The diagnosis of her epilepsy was idiopathic generalized epilepsy. The mother of Patient 34 (c.4831G→T, V1611F) had a history of 10 febrile convulsions until 10 years of age.

Of the 25 SMEI patients, five patients were found to have nonsense mutations, two patients had frameshift mutations, two patients had frameshift and missense (c.3199A→G, T1067A) mutations, 12 patients had missense mutations, two patients had other mutations (exon 4 splice site GT→AT and 1807delMFYE), and two patients had no mutation. Of the ten ICEGTC patients, seven patients were found to have missense mutations and three patients had no mutation. With respect to the correlation between phenotypes and genotypes, nonsense mutations and frameshift mutations were found only in SMEI patients, and the mutations in patients with ICEGTC were exclusively missense mutations.


SCN1A mutations in SMEI

We previously reported that nonsense and frameshift mutations are the major genotypes of SMEI (Sugawara et al., 2002). The present study revealed that the mutation could also occur as the missense or other types. Nevertheless, frameshift and nonsense mutations are the predominant abnormality seen in SMEI patients: together with the previous report, we had a total of seven patients with frameshift mutations and 12 patients with nonsense mutations out of 39 SMEI patients. The missense mutation found also in the control population (c.3199A→G, T1067A) was regarded as a benign variant unrelated to the disease (Escayg et al., 2001; Sugawara et al., 2001) and is excluded from the discussion below. The other 11 missense mutations not detected in the control subjects may be responsible for SMEI disease phenotype, although we need more control subjects and above all functional studies to confirm the pathogenicity, so that we may describe them as probable mutations. We failed to detect SCN1A mutations in three patients with SMEI in this study and four patients in our previous study. We have not ruled out the possibility of promoter mutations or mutations within intron sequences that might effect the splicing of RNA transcripts. Therefore, we have not fully excluded mutations within SCN1A. Alternatively, the responsible mutation may be present in genes other than SCN1A, such as in the GABA(A)‐receptor γ2 subunit gene as suggested by Harkin et al. (2002).

We examined whether SMEI cases with different mutations or no mutations in the SCN1A may have different clinical features. The mean onset ages of minor seizures were 10.7 months, 16 months, 25 months and 17.9 months in patients with frameshift mutations, nonsense mutations, missense mutations, and no mutations, respectively. Minor seizures tended to occur earlier in patients with frameshift mutations than those with missense mutations (P = 0.0489). Frequency of convulsive seizures was slightly less in patients with missense mutations than those with nonsense mutations (P = 0.0683) (GTC occurred weekly or more in 57.1% of patients with frameshift mutations, 66.7% of those with nonsense mutations, 27.3% of those with missense mutations and 57.1% of those with no mutations). These clinical differences were minimal, but could become more evident with an increasing number of cases.

Mutations in ICEGTC

This study disclosed that not only SMEI but also ICEGTC showed mutations in SCN1A. The mutations in SMEI were frameshift, nonsense or missense, while the mutation in ICEGTC was exclusively missense so far. The loci of the missense mutations seen in ICEGTC were quite similar to those in SMEI, suggesting a genotypic continuity between these entities. Again, the pathogenicity of these missense mutations remains to be proved by functional analyses.

Mutations in relatives

Mutational analysis was performed in the relatives of 15 patients in this study and five patients described in our previous report (Sugawara et al., 2002). In the analysis, two mothers were revealed to harbour missense mutation of SCN1A identical to the probands. Phenotypes of these two mothers met the criteria of GEFS+. Both probands had ICEGTC. These important observations open the possibility that ICEGTC may be included in the GEFS+ syndrome, and give rise to the question: what is the factor that gives much severer phenotype only to the ICEGTC probands?

On the contrary, the nonsense mutation seen in the twin Patients 1 and 2 (c.3637C→T, R1213X), and that found in another twin patients and one unrelated patient reported in our previous report (c.4537C>A, S1516X) inevitably resulted in SMEI. Although familial investigation was limited in our sample, all tested nonsense or frameshift mutations were so far de novo, as was also reported by Claes et al. (2001). These de novo mutations in SCN1A are not consistent with the family history reported in series of studies (including ours) describing a high incidence of convulsive disorders in relatives (see Introduction). Other modifying factors might explain this discrepancy, but further studies should be performed to discuss it.

Nosology of childhood epilepsies with frequent GTC

The subjects of this study, SMEI and ICEGTC, had GTC as the core seizure symptom and additional minor seizures. There were no significant differences between the two groups in onset age, sex, family history, neurological abnormality, EEG findings and clinical course. They may represent a continuum with minor phenotypic variations.

Our previous and present studies showed that mutations in SCN1A were found at the high rate of 70–82% in both groups, with frameshift or nonsense mutations in SMEI and missense mutations in both groups. The mutations in SCN1A thus seem to play a significant, but not the exclusive role in the manifestation of these severe types of epilepsy and suggest a genotypic continuum of the two disorders. We should examine whether genotypic differences of frameshift, nonsense and missense mutations make different physiologic effects before phenotypic correlations are undertaken.

Relation to GEFS plus

A common feature of the epilepsies studied here is that the seizures are precipitated by fever, which is also a characteristic of GEFS+. Therefore, the present study also confirms a commonly accepted hypothesis that increased seizure susceptibility by fever indicates mutations in SCN1A.

Some authors speculated that SMEI might represent the most severe form of GEFS+ (Scheffer et al., 2001; Singh et al., 2001; Veggiotti et al., 2001). In the present series, SMEI mutations were de novo so far examined, and included frameshift and nonsense mutations. This is not consistent with the discussion of a GEFS+ spectrum (Ito et al., 2002).

However, we found two mothers with GEFS+ who had the same missense mutations in SCN1A as their children. These are the first reported cases associating ICEGTC with GEFS+ families, and suggest that this type of epilepsy may constitute the most severe form of GEFS+ spectrum. We should observe mutations of further patients and their families to examine whether this is also true for SMEI cases. Physiological studies may contribute to further understanding of these important intractable disorders of childhood and their relation to GEFS+.


We wish to thank all the patients, family members and control subjects who participated in this study and Drs Shigematsu, Shimomura, Tanaka, Kageyama and Uruno at the Shizuoka Medical Institute of Neurological Disorders for their support.


View Abstract