Autosomal Dominant Lateral Temporal Epilepsy: Two Families with Novel Mutations in the LGI1 Gene


Address correspondence and reprint requests to Dr. P. Hedera at Department of Neurology, Vanderbilt University, 465 21st Avenue South, 6140 MRB III, Nashville, TN 37232-8552, U.S.A. E-mail:


Summary: Purpose: Mutations in the leucine rich, glioma inactivated gene (LGI1) were recently described in a small number of families with autosomal dominant lateral temporal epilepsy (ADLTE). ADLTE is characterized by partial seizures with symptoms suggestive of a lateral temporal onset, including frequent auditory aura. Here we report the results of clinical and genetic analyses of two newly identified families with ADTLE.

Methods: We identified two families whose seizure semiology was suggestive of ADLTE. Evaluation included a detailed history and neurologic examination, as well as collection of DNA. The coding sequence of the LGI1 gene from affected subjects from both families was analyzed for evidence of mutation.

Results: Each patient had a history of partial seizures, often with secondary generalization earlier in the course. Auditory aura was reported by approximately two thirds of affected patients in each pedigree. Novel mutations in LGI1 were detected in both families. A heterozygous single-nucleotide deletion at position 329 (del 329C) was detected in affected individuals from one family, whereas patients from the second family had a nonsynonymous variation, corresponding to C435G.

Conclusions: We identified two novel mutations in the LGI1 gene. The phenotype of these two families was similar to that of other kindreds with ADLTE, as auditory aura was absent in one third of affected individuals. Our results further support that LGI1 mutations should be considered in patients with a history of partial seizures if the semiology of seizures is consistent with the onset in the lateral temporal lobe.

Autosomal dominant temporal lobe epilepsy (ADLTE) is an inherited temporal lobe epilepsy characterized by focal seizures with symptoms suggestive of the lateral temporal lobe origin, such as auditory auras and a relatively benign clinical course (1,2). This type of epilepsy was described by Ottman et al. (3), who designated it as autosomal dominant partial epilepsy with auditory features (ADPEAF). They identified a family in which affected individuals had partial seizures with frequent auditory hallucinations, even though other sensory symptoms also were present; auditory aura, together with the absence of autonomic and motor symptoms, suggested a lateral temporal lobe localization. The same group mapped ADLTE (ADPEAF) to a locus on chromosome 10q24 and subsequently identified mutations in the leucine-rich, glioma-inactivated gene (LGI1) (3,4). LGI1, isolated from glioma tissue and hypothesized to play a role in tumor progression, is expressed predominantly in neurons, with the highest levels of expression found in the temporal lobes (4,5). Although the pathogenesis of epilepsy due to LGI1 mutations remains unclear, its role in ADLTE has been further supported by identification of additional disease-causing mutations in a small number of families with this type of epilepsy (6–10). Moreover, it has become apparent that many subjects with LGI1 mutations did not experience an auditory aura, and the same gene was mutated in kindreds with aphasic seizures (6,10). The term ADLTE was proposed to reflect the anatomic origin of seizures in this syndrome (2,5,10). Here we report clinical and genetic analysis of two kindreds with ADLTE due to novel mutations in the LGI1 gene.


Clinical and laboratory studies

We identified two probands (III-4/A from family A and III-1/B from family B; Fig. 1), from our epilepsy patient database, who had a partial epilepsy with prominent auditory features and a family history of epilepsy. Six affected, five unaffected, and two married-in individuals from family A and three affected, one unaffected, and one married-in individual from family B consented to participate in this study, approved by the Institutional Review Board at Vanderbilt University. A comprehensive medical history was obtained through a personal interview of every participating subject, and information about the presence of seizures, age at onset, frequency of seizures, the presence and nature of aura and semiology, and possible risk factors was collected. All subjects, including those who did not report seizures, were specifically asked about auditory phenomena that may precede seizures or occur in isolation. Seizures were classified based on clinical description. A detailed neurologic examination was performed on every subject. Interictal 18-channel electroencephalograms (EEG) were obtained on subjects III-4/A and III-13/A. We used a digital electroencephalograph (XLTEK; Toronto, ON, Canada) and the 10-20 electrode system; the recording included hyperventilation and intermittent photic stimulation. Subject III-1/B previously had one 4-h outpatient and one 4-day inpatient video-EEG monitoring study at our institution. A board-certified clinical electrophysiologist (B.A-K.) interpreted the results of all neurophysiology studies.

Figure 1.

Pedigrees of the kindred A and B. *Individuals in whom DNA was collected. Roman numerals denote the generation and Arabic numerals in the text indicate the position in the pedigree, counted from the left.

Genetic linkage analysis

Phenotype was determined as definitely affected or unaffected before genetic analysis. DNA was available from all subjects, with the exception of the subject III-3/B. Family A was tested for linkage to the locus on chromosome 10q24 that was previously identified in ADLTE by using microsatellite markers D10S200, D10S510, and D10S571 (2,3). We did not test linkage in family B because of very limited statistical power for a meaningful analysis. Two-point linkage analysis was performed with the FASTLINK program by using an AD mode of disease inheritance and disease allele frequency of 0.001. We assigned penetrance 0.90 for LOD score calculations. Allele frequencies were determined by genotyping of 90 healthy, unrelated individuals with similar ethnic background (180 chromosomes).

Analysis of the LGI1 gene

All eight exons were amplified by using the polymerase chain reaction (PCR) with published primers and conditions (6). PCR products were purified through Sephadex G-50 columns and sequenced by using an ABI PRISM dRhodamine Terminator Cycle Sequencing Ready Reaction and the ABI PRISM 3100 Genetic Analyzer (PE Applied Biosystems, Foster City, CA, U.S.A.). Each exon was sequenced in both directions in two affected and one unaffected member from each family. Observed sequence changes were further analyzed for segregation with disease in every affected and unaffected subject within a given family. We also sequenced 200 normal chromosomes by using 100 healthy unrelated control subjects to permit detection of benign polymorphisms for each detected mutation.


Clinical description and laboratory evaluation

Phenotypic features of 11 affected subjects from families A and B are summarized in Table 1. Each patient had a history of partial seizures; simple partial seizures (SPSs) were the most common, but most also had complex partial seizures (CPSs). However, the majority of patients had nocturnal secondarily generalized tonic–clonic seizures (SGTCSs) as a presenting type of seizure with an average age at onset at 15 years (range, 11–34 years) and developed daytime CPSs and SPSs later in the course of the disease. Auditory hallucinations varied from a simple sound to more complex melodies. Affected individuals with an auditory aura from family A also described a loss of hearing during seizures. Auditory features were absent in three of seven affected subjects in family A and in two of five in family B. Three subjects (II-8/A, III-1/A, and III-11/A) reported precipitation of seizure by unexpected high-pitched sounds and the subject III-2/B by a low-pitched sound; furthermore, flashing lights also were associated with a reflex seizure precipitation (subjects III-11/A and III-2/B). Only four patients reported a seizure-free interval >1 year, but the majority of patients had relatively infrequent seizures; history of status epilepticus was reported by one individual (III-8/A), who later achieved good control of seizures. Every affected subject had normal intelligence and neurologic examination did not show any significant abnormalities in affected subjects. Interictal EEG was within normal limits in all studied subjects; no seizures were recorded during EEG monitoring of the subject III-1/B and the studies also were within normal limits.

Table 1. Clinical features in the two families



Age at



Auditory aura
and natural
history of aura
Seizure types, age
at appearance (if
known), listed in
order of frequency


Family A 
 II-8Died at 6815SGTCFrequent SPS up to age 66Buzzing in the head, then loss of hearingSPS (30s), SGTC (15)Unexpected noise, especially high-pitched soundUnknown
 III-15412SGTCEarly seizures nocturnal, later diurnal, seizures controlled for >1 yrDid not appear till late 20s: “little shock” in head, brief buzzing sound, then loss of hearingSPS (late 20s), CPS (late 20s), SGTC (12)Startle, interruption of work by high-pitched voiceSevere headache, word-finding difficulties
 III-64410SGTCEarly seizures in sleep, since 20s, 3 CPSs per month, seizure free for 3 yrDid not appear until 14 yr, funny feeling, no auditory aura, no loss of hearingCPS (14), SGTC (10)NoneConfusion, severe headache
 III-84213SGTCEarly seizures in waking, history of status epilepticus; seizures controlled for >1 yrNo auditory aura; experienced a possible aura of feeling droopy at 41 yrCPS (late 20s), SGTC (13)Stress; menstrual periodFatigue; no postictal aphasia
 III-114012SGTCSeizures poorly controlled with 7–10 CPSs/mo, but rare SGTCsA feeling like she is going to pass out, then loss of hearing. No longer has an auraSPS (12), CPS (31), SGTC (12)Flashing lightsPostictal aphasia
 IV-52314CPSInfrequent CPS until 19 yr, then SGTCs (4–5/yr)No aura until 19, loss of hearing or familiar melodies (varies)CPS (14), SGTCS (19)NoneSevere headache
 IV-62115SGTCMostly SGTCs (every 3 mos), infrequent CPSsNo auditory aura, has an experience that includes déjà vuSGTC (15), CPS (17)NoneSevere headache, Word-finding difficulties
Family B 
 II-1Died at 5735SGTCHad only 2 nocturnal SGTC seizures that were 4 yr apart. Seizure free for 18 yr before his death. Came off AEDs early 50sNo aura reportedSGTC (35)NoneConfused, tired
 II-34411SGTCFrequent daily seizures (CPSs and up to 12/day SPSs), SGTC 3–4/moBuzzing, occasionally complex melodies (Beethoven's 5th symphony)SGTC (11), SPS (12), CPS (13)NoneWord-finding difficulties
 III-13613SGTCFirst 4 SGTCs were in sleep then had 2 in waking. She then developed SPSs and only occasional CPSs (could be only that she was distracted by the seizure). Only rare SPSs persistBrief ringing or feeling of change in pressure in the ear, sounding like a cicada sound; auditory aura was followed by intense fear/anxiety; aura was present from onsetSPS (28), SGTC (13), CPS (28)Possibly certain sounds; the distinction between extrinsic and intrinsic sounds is not clearConfusion, distraction, anxiety; no language difficulty
 III-23517SGTCFirst seizure nocturnal; second seizure was also SGTCs, precipitated by photic stimulation with the EEG. Had 6 total SGTC seizures, 2 of which diurnal. Later had infrequent auditory SPSsAura only with diurnal seizures; hears a low-pitched sound like a swinging rope or ringing; with bigger seizures, the sound gets louder in her head before generalizationSPS (17), SGTC (17)Low-pitched unexpected sounds, flashing lightsVery emotional, tired, sore all over after SGTC
 III-32815SGTCOnly nocturnal SGTCsNo aura reportedSGTCs (15)NoneConfusion

Genetic analysis

Pedigree analysis and linkage study

Both pedigrees demonstrated vertical transmission of epilepsy with two examples of male-to-male transmission in family A (data not shown) consistent with AD mode of inheritance. Whereas male-to-male transmission was absent in family B, the presence of epilepsy in three consecutive generations suggested an AD mode of inheritance. We identified one instance of probable incomplete penetrance (II-4/B to III-4/B); the subject III-4/B declined participation in the genetic analysis, but he did not have any additional risk factors for epilepsy, and his unaffected parent inherited a missense mutation in the LGI1 gene (see later). Positive two-point lod scores were obtained for all three markers analyzed (2.02 for both D10S200 and D10S571, and 1.79 for D10S520) for pedigree A with no recombination detected within the region analyzed.

Analysis of the LGI1 gene

Analysis of the LGI1 coding sequence in family A demonstrated a heterozygous single-nucleotide deletion at position 329 (del 329C, counted from the initiating methionine) resulting in a frameshift predicted to introduce a stop codon at the end of the third exon (Fig. 2A). This deletion was present in all affected individuals in the family, and analysis of 200 chromosomes from healthy controls did not identify the same change (Fig. 2B). All affected subjects from family B had a single-nucleotide variant at position 435 (C435G) in the exon 5 on one allele, predicted to result in a serine-to-arginine substitution at codon 145 (Fig. 2C). This missense mutation segregated with the disease and also was present in an unaffected parent (II-4/B) who had a child with seizures (III-4/B). Analysis of 200 normal chromosomes revealed that a cysteine residue was encoded at this position in all samples (Fig. 2D).

Figure 2.

A: Heterozygous deletion at the position 329 (del 329C) in exon 3, counted from the start codon (arrow), resulting in a frame shift (sequence of the sense strand is shown). B: Normal sequence of the same segment in unaffected individuals and healthy controls (sequence of the sense strand is shown). C: Heterozygous point mutation at the position 435 (C435G) in exon 5 (arrow) predicted to change serine to arginine at codon 145 (sequence of the antisense strand is shown). D: Normal sequence of the same segment in unaffected individuals and healthy controls (sequence of the antisense strand is shown).


We report two additional families with ADLTE caused by novel mutations in the LGI1 gene, further expanding our understanding of the mutational spectrum of this type of epilepsy. Every affected individual from family A had a single-nucleotide deletion in the third exon of the LGI1 gene, resulting in a truncated protein of only 79 amino acids. Of the 17 LGI1 mutations described to date (including this report), about half are predicted to result in a truncated protein, due to frameshift because of deletions, insertions or splice site mutations, or nonsense mutations (4,6–11). The presence of a truncated protein suggests that loss-of-function and haploinsufficiency are important mechanisms underlying disease pathogenesis. However, the occurrence of dominant-negative effects cannot be excluded in all cases. The missense mutation identified in family B is predicted to cause a change from serine to arginine within the third leucine-rich repeat in the LGI1 protein that is localized between amino acid residues 138 and 161. This domain is highly conserved and is proposed to define a new subfamily of proteins containing leucine-rich repeats. The serine corresponding to codon 145 is present in all known members of this subfamily (12). Leucine-rich repeats are important for protein–protein interaction and signal-transduction pathways (13). Even though the role of the leucine-rich repeats in the overall function of LGI1 is unknown, the presence of a disease-causing mutation in this highly conserved domain suggests its importance for a normal function and the pathogenesis of ADLTE.

We identified these two ADLTE families because of a typical history of auditory features in two probands, individuals III-4/A and III-1/B (Fig. 1). However, auditory aura was reported in only approximately two thirds of affected individuals in both families, and several patients who reported auditory hallucinations developed auditory features only later in the course of the disease. The frequency of reported auditory features varies among reported families, but in the majority of reported kindreds, it is in the same range or higher (2,3,10,11,14). It is important to ask probands with possible lateral temporal seizure onset who do not report auditory aura about other family members because an identification of auditory features in affected relatives can suggest involvement of the LGI1 gene.

Three affected individuals (II-8/A, III-1/A, and III-2/B) reported a frequent precipitation of their typical seizures by certain unexpected sounds. Furthermore, two subjects (III-11/A and III-2/B) reported seizures induced by flashing light. This is suggestive of sensory-evoked reflex seizure precipitation in association with the currently described mutations. The Frings audiogenic seizure-susceptibility mouse, a naturally occurring mouse mutant is considered a model for sensory-evoked reflex seizures. A homozygous single-nucleotide deletion in the gene Mass1, which codes for a fragment of the very large G protein–coupled receptor 1 (VLGR1), has been recently discovered to cause audiogenic seizures in these mice (15). A disease-causing mutation in the human ortholog MASS1 has been reported in one family with febrile and afebrile seizures (16). Although these patients did not have a history of reflex epilepsy, the role of the MASS1/VLGR1 gene in human epilepsy is intriguing. Sequence analysis of the MASS1/VLGR1 and LGI1 genes identified a new domain shared by these two proteins; it is presumed to play a role in the interaction of various proteins that may play an important role in epileptogenesis (17).

In summary, we report two new novel mutations in the LGI1gene, one predicted to cause premature termination of the protein and one missense mutation within the leucine-rich domain. Additional characterization of the mutational spectrum in this gene can further contribute to our understanding of the normal function of the LGI1 protein and its role in the epileptogenesis.


Acknowledgment:  We thank the members of the families described here, whose help and participation made this work possible. This work was supported in part by GCRC grant RR00095 to the VUMC General Clinical Research Center, Vanderbilt University Discovery Grant (J.S.S.), and NIH K08NS42743 (P.H.).