Mutations in LGI1 gene in Japanese families with autosomal dominant lateral temporal lobe epilepsy: The first report from Asian families


Address correspondence to Akio Ikeda, M.D., Ph.D., Department of Neurology, Kyoto University Graduate School of Medicine, 54 Shogoin-Kawaharacho, Sakyo-ku, Kyoto 606-8507, Japan. E-mail:


Autosomal dominant lateral temporal lobe epilepsy (ADLTE) caused by LGI1 (leucine-rich gene, glioma-inactivated-1) mutations is a rare familial epileptic syndrome characterized by the auditory ictal manifestation and rare nocturnal generalized seizures. We have examined the sequence of the LGI1 gene in four Japanese families with lateral temporal lobe epilepsy having characteristic auditory features, and identified one novel (1421G>A), and one reported (1418C>T) point mutation each in two families. These two mutations were 3 bp apart in the LGI1 gene and caused adjoining amino acid substitutions. The two families presented different clinical phenotypes and seizure control to drug treatment. These findings suggest that LGI1 mutations in Japanese ADLTE families may not be uncommon, and that diverse clinical phenotypes make adequate diagnosis of ADLTE difficult when only based on clinical information.

Autosomal dominant lateral temporal lobe epilepsy (ADLTE, MIM 600512) is a rare form of lateral temporal lobe epilepsy with characteristic manifestations such as auditory and visual auras, seizures triggered by auditory stimuli, and rare nocturnal generalized seizures (Gu et al., 2005). The majority of hereditary idiopathic epileptic disorders are caused by mutations of genes encoding ion channel components. The only exception in human so far is ADLTE, which is caused by mutations in LGI1, leucine-rich glioma inactivated-1, a gene encoding a secreted neuronal protein (Fukata et al., 2006).

In 1995, the linkage analysis of a large family with partial epilepsy showed strong evidence for localization of a gene for partial epilepsy on 10q22-24 (Ottman et al., 1995). In 2002 the LGI1 was identified as a causative gene, and its mutations in five families with ADLTE were reported (Kalachikov et al., 2002). To date more than 13 point mutations with amino acid substitutions and one insertion, five deletions, three point mutations resulting in premature truncations, and one point mutation resulting in skipping two exons have been reported (Morante-Redolat et al., 2002; Fertig et al., 2003; Gu et al., 2005; Chabrol et al., 2007). No pathologic polymorphism was found in the promoter region of LGI1 (Bovo et al., 2008). Clinical semiology seems to vary among families, but no analyses with regard to the different ethnics were done because of absence of reports of ADLTE from Asian countries. Herein we report the first, non-Caucasian ADLTE families with LGI1 mutations. The content of this article has not appeared elsewhere except for in abstract form (Ikeda et al., 2007).

Subjects and Methods


In 2005 and 2006, we had four families who were clinically consistent with a diagnosis of ADLTE. They were treated for their seizure disorders at Kyoto University Hospital and resided in Kyoto province. Two families of four carried different heterozygous mutations in the LGI1 gene, the family “O” and the family “U” (Figs. 1 and 2). Two patients in the family “O” and three patients in the family “U” were treated for their seizure disorders and psychiatric symptoms at Kyoto University Hospital. The detailed clinical manifestation of the two families is described in the Results.

Figure 1.

Pedigree of a five-generation family “O” with autosomal dominant lateral temporal lobe epilepsy (ADLTE). The proband case is indicated by the arrow. The DNA samples from IV-5 and IV-7 carry 1418C>T mutation in the LGI1 gene. The clinical significance of this family is the diversity of the symptoms, for example, panic attack–like symptoms and seizures with auditory features. It seems that the affected phenotype of seizure skips generations.

Figure 2.

Pedigree of a two-generation family “U” with autosomal dominant lateral temporal lobe epilepsy (ADLTE). The proband case is indicated by the arrow. The DNA sample from I-1 (the father of the proband) carries 1421G>A mutation in the LGI1 gene (cited from Ikeda et al., 2000, with permission).

Gene analysis

Genetic analysis of ADLTE was approved by the Kyoto University Graduate School of Medicine Ethical Committee, and written informed consent was obtained from all the patients and the control participants investigated in the present study. DNA from the five patients in the four families were analyzed in this study. The control sample consisted of DNA from 50 healthy Japanese volunteers who gave informed consent. Using primers previously reported, eight exons of the LGI1 gene were amplified and subsequently sequenced on an ABI310 sequencer (Applied Biosystems, Foster City, CA, U.S.A.) (Michelucci et al., 2003). Mutations in the LGI1 gene were confirmed by NruI restriction enzyme digestion.


Clinical manifestation

Family “O” had a history of epilepsy in nine members in five generations (Fig. 1). A proband, IV-5, a 27-year-old-woman, had the first generalized tonic–clonic seizure (GTCS) at the age of 10 years while asleep. Currently she has GTCSs every 1–2 months, almost daily auras, and weekly complex partial seizure (CPS), despite high-dose polytherapy such as carbamazepine, zonisamide, and clonazepam. The patient had several types of auras as follows. Auditory auras consisted of high-pitched sound, becoming gradually louder, resulting in inability to understand the spoken language (aphasic state). Visual auras were elementary, flashing light, and metamorphopsia. The patient also had déjà vu, jamais vu, and autonomic auras (i.e., palpitation, cold sweat, and pale face). All of the auras occurred in various combinations and also occurred independently. Those auras, especially the latter ones, were frequently followed by panic attack–like symptoms (i.e., sudden onset of fear and anxiety). The patient also confessed that she had auditory hallucination that she should throw her child out of the window. This occurred independent of the ordinary auras described above. Electroencephalography (EEG) showed no clear epileptiform discharges, despite repetitive sleep EEGs, and magnetic resonance imaging (MRI) showed the small volume of the left superior temporal gyrus. Fluorodeoxyglucose–positron emission tomography (FDG-PET) by statistical parametric mapping (SPM) analysis revealed decreased glucose metabolism in the left lateral temporal area.

The younger sister of the proband, IV-7, a 26-year-old-woman, also had the first GTCS while asleep at the age of 11 years, and currently has GTCSs once a year, and auditory auras (high-pitched sound) and déjà vu, each several times per week. The patient had no visual auras, jamais vu, autonomic auras, or psychiatric symptoms such as anxiety attacks. It was reported that EEG showed no clear epileptiform discharges.

Among their family members, four had apparent psychiatric symptoms such as an outrage of emotion and extraordinary explosive violent behaviors. One of them (II-12) was admitted to the psychiatric hospital for several months, and two of them (II-9 and II-12) had sudden death. Five had both psychiatric symptoms and seizures like the proband. For seizure disorders, at least five (II-12, III-2, III-3, IV-5, and IV-7) had both GTCS and partial seizures. Auditory auras were observed in at least two members (III-3 and IV-1).

The phenotype of the family “U” was already reported in detail before the LGI1 gene was identified as a causative gene of ADLTE in 2002 (Ikeda et al., 2000), and was relatively uniform (Fig. 2). Briefly, I-1, the father of the proband (II-3) was a 51-year-old, full-time employee with a past history of febrile convulsion. At the age of 16 years, his first nocturnal GTCS occurred. His habitual seizures also started with a high-pitched buzzing sound, followed by loss of awareness. The patient often noticed that his seizures were triggered by the telephone bell. His seizures were completely controlled after adding carbamazepine to valproic acid. Two children (II-3 was the proband in Ikeda et al., 2000) also had first GTCS and partial seizures since the second decade, and the seizures were completely controlled by carbamazepine. All of them had normal MRI, and EEG showed sharp transients in the left frontotemporal areas only in I-1, but no clear epileptiform discharges were recorded.

Gene analysis

Sequence analysis of eight exons of LGI1 in the two patients in family “O” (IV-5 and IV-7 in Fig. 1) revealed point mutation (1418C>T, numbering from the first nucleotide of the initiator codon) resulting in amino-acid substitution of serine to leucine at the 473rd residue. Sequence analysis of DNA from the patient in family “U” (I-1 in Fig. 2) also revealed point mutation (1421G>A) resulting in amino acid substitution of arginine to glutamine at the 474th residue. Each of the two mutations disrupts the restriction enzyme site of NruI. Digesting polymerase chain reaction (PCR) product of exon 8 with NruI using buffer provided by a manufacturer (New England BioLabos, MA, U.S.A.), the two mutations were screened in control participants. There was neither 1418C>T nor 1421G>A mutation in 50 control participants, suggesting that these base changes could be regarded as mutations, but not polymorphisms.

The serine at residue 473 and arginine at residue 474 of LGI1 were highly conserved between species and its homologs, LGI2, 3, and 4 (Fig. 3). The serine at the 473rd residue was considered to be phosphorylated, and the mutation might disrupt it.

Figure 3.

Multiple amino acid sequence alignment of human LGI1 and other species. The substituted amino acids, serine at the 473rd and arginine at the 474th, are highly conserved throughout other species as well as other subtypes of LGI families.

With regard to the clinical features, the reported symptoms in family “O” were consistent with previous reports of ADLTE, except that the degree and incidence of psychiatric symptoms were rather high. Symptoms in family “U” were almost common among pedigrees, except that only II-1 had déjà vu, suggestive of mesial temporal involvement. The two patients in family “O” also had déjà vu; however, other clinical features differed between them.


Because all reported ADLTE families with LGI1 mutations were Caucasians, we presumed that patients with LGI1-related epilepsy in Japan were rather rare, or that some genetic backgrounds may lower penetrance rates so that clinically they seem to be sporadic cases even though some families have LGI1 mutations. In the present study, however, we documented that two independent families in Kyoto province had different LGI1 mutations.

Clinical manifestations of family “O” were divided into two groups: one had seizures with auditory features and the other with psychiatric symptoms such as explosive violent behaviors and panic attack–like symptoms. The proband also had apparently nonepileptic auditory hallucination and one member was admitted to the psychiatric hospital. The psychiatric symptoms in ADLTE were also described in the previous reports, but they were much milder. When attention was paid only to seizures with auditory features in this family, epileptic symptoms appeared after skipping two consecutive generations. Therefore, it is most likely that psychiatric symptoms are one of constellations of LGI1-related ADLTE. The clinical manifestations of family “O” with S473L mutation were different from ones of the British decedent family in Australia, even though they had a mutation identical to our patients (Berkovic et al., 2004).

On the other hand, clinical symptoms of family “U” were nearly uniform, being consistent with previously reported, typical ADLTE. Mutation at the same codon was reported in Spanish family results in premature truncation, and their clinical symptoms were also mild and had the common features of ADLTE (Morante-Redolat et al., 2002).

Recently it was reported that ADAM22 served as a receptor for LGI1, the mutated form of LGI1 failed to bind to ADAM22, and LGI1 enhances α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor–mediated synaptic transmission in rat hippocampal slices. (Fukata et al., 2006). Schulte et al. (2006) reported that LGI1 protein assembled into the presynaptic Kv1 channels and inhibited inactivation by Kvbeta1. These findings may suggest that the underlying pathogenetic mechanism of ADLTE would be “loss of function” like other numerous familial epilepsies classified into “channelopathy.” However, the diversity of clinical phenotype probably depends on the mutated residue could be explained better by “gain of function” rather than “loss of function” mechanism, which would be resulting in homogenous phenotype. Further genotype–phenotype analyses in patients of different ethnic backgrounds may elucidate a part of the mechanisms.


This study was supported by the Research Grant for the Treatment of Intractable Epilepsy (19-1) from the Japan Ministry of Health, Labor and Welfare, and by the Scientific Research Grant (C2) (18590935) from the Japan Society for Promotion for Sciences (JSPS).

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Disclosure: None of the authors has any conflict of interest to disclose.