Genetics of Idiopathic Epilepsies


Address correspondence and reprints requests to Dr. S. Hirose at Department of Pediatrics, School of Medicine, Fukuoka University, 45-1, 7-chome Nanakuma, Jonan-ku, Fukuoka 814-0180, Japan. E-mail:


Summary: Purpose: To search for clues to molecular genetics of common idiopathic epilepsy syndromes. Genetic defects have been identified recently in certain inherited epilepsy syndromes in which the phenotypes are similar to those of common idiopathic epilepsies.

Methods: Mutations identified as the causes of inherited idiopathic epilepsies were reviewed.

Results: Mutations of the genes encoding two subunits of the neuronal nicotinic acetylcholine receptor were found in autosomal dominant nocturnal frontal lobe epilepsy. Mutations of two K+-channel genes were identified in benign familial neonatal convulsions. Mutations of the genes encoding several subunits of the voltage-gated Na+-channel and γ-aminobutyric acid (GABA)A receptor also were identified as the underlying causes of various epilepsy syndromes, such as autosomal dominant epilepsy with febrile seizures plus, benign familial neonatal infantile seizures, and autosomal dominant juvenile myoclonic epilepsy. Mutations within the same gene may result in different epilepsy phenotypes. Thus, the Na+ channel, GABAA receptor, and their auxiliaries may be involved in the pathogenesis of various types of epilepsy. Some forms of juvenile myoclonic epilepsy, idiopathic generalized epilepsy, and absence epilepsy may result from mutations of Ca2+ channels. Mutations of the Cl channel have been recently found to be associated with a certain type of epilepsy. The recent discovery that mutations of LGI1, a gene encoding a nonchannel molecule, are associated with autosomal partial epilepsy with auditory features may provide a new insight into our understanding of the genetics of idiopathic epilepsy.

Conclusions: These findings suggest the involvement of brain channelopathies in the pathogenesis of certain types of idiopathic epilepsy.

Idiopathic epilepsy is a group of epilepsies with no apparent etiology other than a genetic predisposition. In general, idiopathic epilepsy is not complicated with neurodegenerative processes such as mental retardation and cerebellar ataxia. Genetic defects have been identified in certain familial epilepsy syndromes in which the phenotypes are similar to those of common idiopathic epilepsies. However, the familial epilepsy syndromes show dominant inheritance with high penetrance, whereas such significant inheritance is not observed in common idiopathic epilepsies. Nevertheless, the similarities in symptoms between such familial epilepsies may provide us with clues to the genetics of common idiopathic epilepsies. This article reviews such genetic identifiers among familial epilepsies to address the genetic abnormalities underlying common idiopathic epilepsies (1–3). Table 1 summarizes familial idiopathic epilepsies and responsible genes.

Table 1. Familial idiopathic epilepsies and responsible genes
Disorder (McKusick)ChromosomeGenes (products)Reference
ADJME (MIM 600669)5q34GABRA1 (GABAA receptor)52
ADNFLE1 (MIM 6000513)20q13.2-q13.3CHRNA4 (ACh receptor)4–7
ADNFLE3 (MIM 605375)1p21CHRNB2 (ACh receptor)8,9
ADPEAF (MIM 600512)10q24LGI1 (non-channel)44,45
BFNC1 (MIM 125370)20q13.3KCNQ2 (K+-channel)13–19
BFNC2 (MIM 121201)8q24KCNQ3 (K+-channel)20,21
BFNIS2q23-q24.3SCN2A (Na+-channel)34
EA1 (MIM 160120) with partial epilepsy12p13KCNA1 (K+-channel)49
EA2 (MIM 108500) with epilepsy19qCACNA1A (Ca2+-channel)51
FS & afebrile seizures (MIM 604233)2q23-q24.3SCN2A (Na+-channel)32
FS+ & absence5q34GABRG2(GABAA receptor)36
GEFS+1 (MIM 604233)19q13.1SCN1B (Na+-channel)28,29
GEFS+2 (MIM 604233)2q24SCN1A (Na+-channel)28–31
GEFS+3 (MIM 604233)5q34GABRG2 (GABAA receptor)35
IGE (MIM 600669)3q26CLCN2 (Cl-channel)53
IGE (MIM 600669)2q22-23CACNB4 (Ca2+-channel)50
JME (MIM 254770)2q22-23CACNB4 (Ca2+-channel)50


Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) is characterized by clusters of brief seizures during light sleep and is often misdiagnosed as nightmare or parasomnia. Linkage analyses mapped three genetic loci for ADNFLE to chromosomes 20q13.3, 15q24 and 1p21. ADNFLE with genetic locus on chromosomes 20q13.2, 15q24 and 1p21 are now referred to as ADNFLE type 1 (MIM 600513), 2 (MIM 603204) and 3 (MIM 605375), respectively. The phenotypes of the three types of ADNFLE are indistinguishable.

For ADNFLE type 1, two missense (c.839C>T; S280F and c.851C>T; S284L) and an insertional mutations in the neuronal nicotinic acetylcholine receptor α4 subunit (CHRNA4) gene, CHRNA4, have been found in families with ADNFLE (4–7). For ADNLE type 3, two missense mutations (c.859G>C; V287L and c.859G>A; V287M) of the CHRNB2 gene have been reported (8,9). All mutations found were heterozygous.

CHRNA4 and CHRNB2 are components of neuronal acetylcholine receptor (nAChR), a ligand-gated ion channel in the brain. They are assembled into a hetero-pentamer that functions as a nAChR, and (α4)2(β2)3 are considered to be the dominant subtype of nAChR in the brain. CHRNA4 and CHRNB2 span the plasma membrane 4 times. The second membrane-spanning domain, known as M2, lines the ion pore of nAChR. Interestingly, all mutations identified so far in both CHRNA4 and CHRNB2 are located in the M2 region.

Electrophysiologic characteristics of nAChR bearing these mutations have been examined. In short, the CHRNA4 mutations seem to lead to “loss of function” of nAChR, whereas the CHRNB2 mutations seem to lead to “gain of function,” although the phenotypes of ADNFLE resulting from both mutations are indistinguishable (8–11). Thus, the molecular pathogenesis of epilepsy of ADNFLE due to abnormalities of nAChR requires further investigation.

A sporadic case of NFLE was found to be caused by c.851C>T; S284L, one of the CHRNA4 mutations found in ADNFLE type 1. This discovery implies that some cases of idiopathic epilepsy syndromes may be caused by a single genetic abnormality rather than multifactorial causes, the accepted etiology for idiopathic epilepsies (12).


BFNC is characterized by clusters of generalized and partial seizures afflicting exclusively neonates and is known to remit spontaneously. However, the incidence of subsequent epilepsy later in life also is high in individuals with BFNC during childhood. BFNC is classified into two groups; BFNC1 (MIM 121200) is linked to chromosome 20q13.3 and thought to be the major phenotype, whereas BFNC2 (MIM 121201) is linked to chromosome 8q24. Mutations of the KCNQ2 and KCNQ3 genes have been identified as the underlying abnormalities of BFNC1 and 2, respectively. To date, >10 different mutations of the KCNQ2 gene have been discovered (13–19) whereas only two mutations including one found in a Japanese pedigree with BFNC2 were identified in the KCNQ3 gene (20,21). All mutations found were heterozygous.

Abnormalities of both KCNQ2 and KCNQ3 lead to dysfunction of M-current, which is a slowly activating and inactivating low-threshold K+ current that controls the subthreshold of neuronal excitation (22). KCNQ2 and KCNQ3 synergistically contribute to the formation of the native M current. Both are believed to assemble into heteromeric tetramers to function as active K+ channels. Because M-current is believed to regulate the subthreshold electroexcitability of neurons, reduced K+ currents resulting from mutations so far identified in either KCNQ2 or KCNQ3 are considered to lead to relative hyperexcitability of neurons. Dysfunction of either KCNQ2 or KCNQ3 can thus result in indistinguishable convulsions in BFNC 1 and 2.

The exact pathogenesis of age-dependent characteristics and propensity for subsequent epilepsies in BFNC have not been elucidated. We recently showed that KCNQ channels serve as the predominant inhibitory system in the CNS in neonates; GABAergic transmission governs the inhibitory system afterward. Deficient KCNQ-channels thus cause convulsions in neonates. Our finding also suggests that in the mature CNS, dysfunction of KCNQ channels itself cannot affect seizure threshold during the resting stage but can contribute to seizure activity under the condition of a neuronal hyperexcitable stage. Accordingly, deficient KCNQ channels are probably involved in the age-dependent characteristics (occurrence and spontaneous remission) and propensity for subsequent epilepsies in BFNC (23,24).


A small proportion of patients with febrile seizures (FSs) subsequently have afebrile seizures later in life. This clinical subset of FSs includes a clinical entity referred to as generalized epilepsy with febrile seizures plus (GEFS+); the affected individuals have not only FSs but also various types of afebrile seizures (25,26). GEFS+ is now subdivided into four groups: GEFS+1, resulting from mutations of SCN1B, GEFS+2 resulting from mutations of SCN1A, GEFS+3 resulting from defects of GABAA receptor, and FSs associated with afebrile seizures resulting from mutation of SCN2A (see later). However, the term GEFS+ may be ill defined, because it is not clear whether “plus” means FSs after age 6 years or afebrile seizures preceded by frequent FSs. In addition, GEFS+ may be inappropriate nomenclature, because seizures in individuals who are thought to have GEFS+ apparently encompass partial-seizure phenotypes. We propose that this phenotype be called “autosomal dominant epilepsy with febrile seizure plus” (ADEFS+) instead (27).

The cause of ADEFS+ is mutations within two genes encoding the α subunit of neuronal voltage-gated Na+ channels. Neuronal voltage-gated Na+ channels, the main generators of action potentials of neurons, are composed of three subunits; an α subunit and two auxiliary β subunits, β1 and β2. The α subunit is a large pore-forming molecule and sufficient to function as a Na+ channel on its own. At least three α subunits exist in the central nervous system: α1, α2, and α3, and the corresponding genes are SCN1A, SCN2A, and SCN3A, respectively.

Two missense mutations have been found in the gene encoding the Na+ channel β1 subunit, SCN1B, compared with ∼10 missense mutations of SCN1A (28–31). Recently, we found an interesting nucleotide substitution of SCN2A (c.562C>T, R188W) in a patient with frequent FSs and partial epilepsy (32). This is the first evidence that genetic abnormality of SCN2A is associated with a disease in human and mammals. Thus, it seems that various Na+-channel subunits and their modulators can be involved in the pathogenesis of the GEFS+ phenotype.

Missense mutation in SCN1B (c.363C>G; C121W) is a loss-of-function mutant and may allow a persistent inward Na+ current in neurons (i.e., neuronal hyperexcitability) (29). Recent electrophysiologic characterization of the SCN1A mutations demonstrated augmentation of Na+ influx in neurons. Consequently, similar to mutation of SCN1B, increased neuronal hyperexcitability of the deficient channel due to SCN1A mutations is expected (33). Electrophysiologic studies of the Na+ channel bearing the mutation revealed that the mutation leads to fast desensitization. Because other electrophysiologic properties of the mutant channel, such as conductance and recovery, are comparable with wild type, R188W should augment Na+ influx. Thus, the Na+ channel harboring R188W is associated with hyperexcitability of neurons, which in turn induces seizure activity (32). Intriguingly, two missense mutations of SCN2A were recently found in individuals with benign familial neonatal-infantile seizures (BFNIS) in two families (34).

Genetic abnormalities that cause epilepsies associated with FSs have been identified recently in the gene encoding γ2 subunit of GABAA receptor, GABRG2. GABAA receptor, a ligand-gated Cl channel, functions as a tetramer consisting of several subunits. The main GABAA receptor in the CNS is composed of α1, β2, and γ2 subunits. GABAA receptor serves as the main inhibitory system in mature CNS but is rather excitatory in the developing CNS. One mutation recently found in a French family was identified in GABRG2 (a missense mutation c.983A>T; K328M) located in the linkage between transmembrane domains 3 and 4; the phenotype of the affected individuals was ADEFS+ (35).

GABAA receptors harboring K328M show reduced Cl currents in response to a physiologic ligand, γ-aminobutyric acid or GABA. Because GABAA receptors exert an inhibitory function, dysfunction of the GABAA receptor may lead to seizure activity (35). Another mutation of GABRG2 is a missense mutation (c.245G>A; R82Q) identified in a family in which the phenotype of affected individuals was FS+ followed by absences (36). The R82Q mutation resides within the first of two high-affinity benzodiazepine (BZD)-binding domains of the GABAA receptor. Receptors harboring the R82Q mutation show slow deactivation and rapid desensitization (37). Consequently, desensitized receptors may result in hypofunction of GABAergic inhibitory transmission. Interestingly, the mutation does not alter Cl current in response to GABA when expressed on oocytes but abolishes Cl current augmented by diazepam (36). A poor response of mutant receptors to BZDs, putative endogenous BZD-like substances, may be also involved in the pathogenesis of both febrile seizures and absences.

A great diversity of epilepsy phenotypes in the family with ADEFS+ has been confirmed as the responsible defect, which suggests two hypotheses. One is that monogenic epilepsy could display a wide variety of seizure phenotypes. In turn, some epilepsy syndromes that are currently classified as different epilepsy syndromes could be due to the same genetic defects or considered as allelic variants. Furthermore, because the epilepsy phenotypes of GEFS+1 encompass several common idiopathic epilepsy phenotypes such as absence and tonic–clonic seizures, common-type idiopathic epilepsies could result from a monogenic abnormality, although they are now considered as polygenic or multifactorial disorders.

Mutations of SCN1A were reported as a cause of severe myoclonic epilepsy in infancy (SMEI), although SMEI is neither familial nor idiopathic epilepsy (38–41). A nonsense mutation of GABRG2 also was identified in an individual with SMEI (42). These findings suggest that SMEI is a clinical spectrum of convulsive disorders that includes ADEFS+. However, the mutation of GABRG2 identified in SMEI causes not only channel dysfunction but also impairment of intracellular transport of the mutant GABAA receptors. Mutant receptors were precipitated in the endoplasmic reticulum (ER) and hence may cause ER stress followed by neuronal apoptosis (43). This suggests that the pathomechanism of SMEI is different from that of ADEFS+.


ADPEAF is a familial epilepsy syndrome characterized by lateral temporal epilepsy and auditory aura. Several mutations within the leucine-rich, glioma-inactivated 1 gene (LGI1) have been detected in individuals with ADPEAF (44,45). Mouse Lgi1 is expressed predominantly in neurons, particularly in the neocortex and limbic regions, where it is thought to be involved in temporal epilepsy. LGI1 is crowned as “glioma-inactivated” because loss of both copies of the gene promotes growth of glioma cells, whereas the physiologic function of the gene product has not been elucidated. The deduced amino acid sequence of the product suggests a secreted protein that has three leucine-rich repeats in the N-terminal portion and a novel repeat domain in the C-terminus (referred to as EPTP or EAR domain) (46,47). A similar domain is predicted in the product of the MASS1 gene, one nonsense mutation of which has been detected in a individuals with FSs (48). Thus, molecules with features of the domain may be candidates for a new gene responsible for epilepsy of FSs.


Other genetic abnormalities have been identified in individuals with epilepsy, albeit such epilepsies are often associated with a known syndrome or are found only in a few pedigrees. A missense mutation in KCNA1, a voltage-gated K+-channel gene, causes periodic ataxia type 1, EA1 (MIM 160120), and can sometime lead to partial seizures (49). This provides supporting evidence that not only KCNQ K+ channels, which are involved in M-current formation, but also other K+ channels expressed in the brain may be involved in the pathogenesis of certain types of epilepsy.

Nonsense mutations of the Ca2+-channel β4 subunit gene, CACNB4, were found in a woman with juvenile myoclonic epilepsy (MIM 254770) and in members of a German family with generalized epilepsy (MIM 600669) and praxis-induced seizures. Although the majority of juvenile myoclonic epilepsies or generalized epilepsies may not be due to genetic abnormalities of CACNB4, these cases suggest the involvement of Ca2+ channel expressed in the brain in the pathogenesis of certain types of epilepsy (49). In fact, a nonsense mutation of the gene encoding the voltage-gated P/Q-type Ca2+ channel, CACNA1A, was reported in a patient with generalized tonic–clonic seizures and absence seizures associated with ataxia and learning difficulties. The mutation was heterozygous and de novo and impaired the channel function (51).

A heterozygous missense mutation of the gene encoding the α1 subunit of GABAA receptor, GABRA1, was recently detected in the autosomal dominant form of juvenile myoclonic epilepsy (ADJME) in a family (52). The GABAA receptor harboring the mutation impaired Cl influx through the channel in response to GABA, similar to the dysfunction of GABAA receptors harboring mutations of the γ2 subunit. Likewise, mutations of the CLCN2 gene encoding ClC-2, a voltage-gated Cl channel, were found to be associated with IGE (53). Impairment of the neuronal inhibitory system controlled by Cl influx can indeed result in epilepsy.


Acknowledgment:  Other members of the Epilepsy Genetic Study Group, Japan, (Chairperson, S.K.) are Goryu Fukuma, Hiromi Iwata, Hidetaka Akiyoshi, Minako Yonetani, Akiyo Hamachi, Atsushi Ishii, Tohru Kohashi, Takashi Sugawara, Kazuhiro Yamakawa, Tomoko Miyajima, Masatoshi Ito, Hiroshi Nagafuji, Hideki Muranaka, Syuji Wakai, Tamaki Miyamoto, Kiyoshi Chiba, Nobuo Koide, Kazuie Iinuma, Shinichi Niwa, Rumiko Kan, Tsuneo Ono, Eiichi Oguni, Noriko Horikawa, Masato Fukuda, Hiroyoshi Koide, Yuichi Goto, Teiichi Onuma, Hirohiko Hashimoto, Minoru Isomura, Yusuke Nakamura, Yasunori Oana, Miyako Oguni, Hirokazu Oguni, Makiko Osawa, Tasuku Moyajima, Junko Ohinata, Hiroshi Yamadera, Kiyoshi Hashimoto, Mariko Maesawa, Mana Kurihara, Minoru Hara, Keiichi Yamamoto, Komei Kumagai, Takashi Suzuki, Shoji Tsuji, Kotaro Endo, Manabu Wachi, Toru Konishi, Kyoko Maeda, Kozaburo Aso, Kazuyoshi Watanabe, Masayasu Tsuji, Yasuko Yamatogi, Takeshi Yasuda, Toshiaki Kugoh, Akira Sano, Kazumaru Wada, Naoki Akamatu, Kyoko Ito, Muneyuki Ito, Katumi Imai, Seiichiro Ueda, Takuya Ueno, Akihisa Okumura, Hiroshi Ono, Akiko Kamiishi, Ryutaro Kira, Ryozo Kumano, Hajime Tanaka, Yoshibumi Nakane, Fujito Nakatsu, Shinji Fushiki, Yoshiaki Mayanagi, Mikiya Morokiyo, Kotaro Morita, Takehiko Morimoto, Tsuyoshi Yanai, and Yoshiyasu Sugiura.

We thank Akiko Hayashi, Takako Umemoto, and Yumiko Oka for their excellent secretarial work. Our research was supported in part by grants from the Ministry of Education, Science, and Culture of Japan (S.K. and S.H.), the Epilepsy Research Foundation, Japan (S.K., A.M., and S.H.), and the Central Research Institute of Fukuoka University (A.M. and S.H.)