Seizure 23 (2014) 69–73

A newly discovered LGI1 mutation in Korean family with autosomaldominant lateral temporal lobe epilepsy

Moon Kyu Lee a,1,2, So Won Kim b,c,1, Ji Hyun Lee b,e, Yang-Je Cho a, Doh-Eui Kim a,Byung In Lee a, Ho Min Kim d, Min Goo Lee b,c, Kyoung Heo a,*a Department of Neurology, Epilepsy Research Institute, Yonsei University College of Medicine, Seoul, Republic of Koreab Department of Pharmacology, Pharmacogenomic Research Center for Membrane Transporters, Yonsei University College of Medicine, Seoul,

Republic of Koreac Brain Korea 21 Project for Medical Science, Yonsei University, Seoul, Republic of Koread Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Koreae Research Center for Human Natural Defense System, Yonsei University College of Medicine, Seoul, Republic of Korea

A R T I C L E I N F O

Article history:

Received 6 August 2013

Received in revised form 3 October 2013

Accepted 4 October 2013

Keywords:

LGI1

Mutation

Korean ADLTE family

Genetic analysis

Clinical features

A B S T R A C T

Purpose: A new leucine-rich glioma-inactivated 1 gene (LGI1) mutation inducing an amino acid sequence

substitution was found in a Korean family with autosomal dominant lateral temporal lobe epilepsy

(ADLTE). We report the clinical features and characteristics of this newly identified LGI1 mutation.

Methods: Clinical data were collected from a large ADLTE family. All exons and flanking regions of the

LGI1 gene were directly sequenced. 243 healthy controls were screened for the putative mutation. The

‘Sorting Tolerant From Intolerant’ algorithm was employed for the prediction of mutated LGI1 protein

stability. LGI1 protein secretion was confirmed in vitro by immunoblotting assay.

Results: The main clinical characteristics included a young age at onset (mean, 12.4 years), diverse

phenotypic manifestations, the occurrence of generalized tonic–clonic seizures, and a favorable

prognosis. The genetic analysis detected a nonsynonymous single nucleotide polymorphism of

c.137G > T coding for p.C46F in the five affected family members. This variant was not found in the

normal control population and one unaffected family member. All the amino acids substituted for

cysteine at position 46 of the LGI1 protein were predicted to damage protein stability in in silico analysis.

Mutated C46F protein was retained within the cell at the immunoblotting assay.

Conclusion: We identified a new LGI1 mutation in a large Korean ADLTE family which appeared to be

involved in the development of epilepsy through suppressing LGI1 protein secretion.

� 2013 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Seizure

jou r nal h o mep age: w ww.els evier . co m/lo c ate /ys eiz

1. Introduction

One of the most widely investigated gene in partial epilepsy isthe leucine-rich glioma-inactivated 1 (LGI1) gene. Partial epilepsywith autosomal dominant inheritance and semiological featuresconsistent with lateral temporal lobe seizure has been reported asautosomal dominant lateral temporal lobe epilepsy (ADLTE) orautosomal dominant partial epilepsy with auditory features. The

* Corresponding author at: Department of Neurology, Yonsei University College

of Medicine, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-752, Republic of Korea.

Tel.: +82 2 2228 1607; fax: +82 2 393 0705.

E-mail address: kheo@yuhs.ac (K. Heo).1 These authors are joint first authors.2 Present address: Department of Neurology, Gangneung Asan Hospital,

University of Ulsan College of Medicine, Gangneung 210-711, Republic of Korea.

Tel.: +82 33 610 3195; fax: +82 33 610 4960.

1059-1311/$ – see front matter � 2013 British Epilepsy Association. Published by Else

http://dx.doi.org/10.1016/j.seizure.2013.10.001

gene of ADLTE was localized to the long arm of chromosome 10 in1995, and epilepsy related LGI1 mutation was discovered in2002.1,2 In relatively small series, LGI1 mutations have been foundin about 50% of ADLTE families.3–5 On the other hand, acollaborative study of the Commission of Genetics of the ItalianLeague Against Epilepsy reported that LGI1 mutations were foundonly in 10 (30%) of 33 ADLTE families. Also, they demonstrated asignificantly lower penetrance rate and relative disease risk in non-LGI1-mutated families compared with LGI1-mutated pedigrees,suggesting that a complex inheritance pattern may underlie aproportion of these families.6 Sporadic lateral temporal lobeepilepsy cases with ictal auditory manifestations have beendescribed as idiopathic partial epilepsy with auditory features(IPEAF). De novo LGI1 mutations have also been identified in IPEAF,and account for about 2% of sporadic cases with IPEAF.7,8 Twenty-six LGI1 point mutations related to epilepsy have been reported todate.9 Microdeletion of LGI1 was found in 2012.10

vier Ltd. All rights reserved.

M.K. Lee et al. / Seizure 23 (2014) 69–7370

Recently we have encountered a Korean ADLTE family, andidentified a new LGI1 mutation. We report the phenotypes, geneticfeatures, and laboratory findings.

2. Methods

2.1. Subjects and data collection

A large Korean family comprising 52 members (excluding theirspouses) came to our attention by simultaneously encounteringaffected family members in the epilepsy clinic of SeveranceHospital. Detailed history taking revealed autosomal dominantinheritance of epilepsy within this family. After the approval of theInstitutional Review Board of our institution, six family members[five affected members (III-15, IV-8, IV-9, IV-14, and IV-16), andone healthy member (IV-18)] who signed informed consents wereinterviewed and underwent genetic analysis. Data collection wasfocused on seizure semiology, age at seizure onset, and treatmentoutcomes. The routine scalp EEG recordings were available in sixpatients. Brain magnetic resonance imagings (MRIs) werereviewed.

2.2. Genetic analysis

We performed exon scanning of all LGI1 (NG_011832.1 fromNCBI database) coding exons and splice junction sites by directsequencing to identify variants in six family members. Addition-ally, any putative mutation was screened in 243 Korean populationdata from our database. Genomic DNA was extracted from theperipheral whole blood using a nucleic acid isolation device,QuickGene-mini80 (Fujifilm, Japan). The PCR protocol has beenoptimized for amplifying LGI1 based on UCSC In-Silico PCR (http://genome.ucsc.edu/cgi-bin/hgPcr?command=start). The PCR pro-ducts were purified using a MultiScreen384-PCR Filter Plate(Milipore, USA). The purified products were then sequenced usinga BigDye Terminator Cycle Sequencing Kit and an ABI 3730xlautomated sequencer (Applied Biosystems, USA). Mutation anal-yses were performed using Phred, Phrap, Consed, Polyphred 5.04software (http://droog.mbt.washington.edu/PolyPhred.html).

2.3. Gene cloning

Coding sequence of the LGI1 gene with Xho I and EcoR I enzymesites was obtained from the LGI1 gene in the pBluescriptR vector(Cat no. IRATp970G0723D, Image ID 4811956, Imagenes GmbH,Germany) using PCR method (forward primer: 50-CCGCTCGAGGC-CACCATGGAATCAGAAAGAAGC-30, reverse primer: 50-GGAATTCT-CATGCGCTTAAGTCAAC-30), and inserted into the pcDNA3.1(-)vector. After that, site-directed mutagenesis PCR was performed tomake the C46R and C46F mutagenized LGI1 clones (forward primerfor C46R: 50-CAAAATGCCCTGCCGTGCGTACTTGTACCAAAGAT-30,reverse primer for C46R: 50-ATCTTTGGTACAAGTACGCACGG-CAGGGCATTTTG-30, forward primer for C46F: 50-CCAAAATGCC-CTGCCGTGTTTACTTGTACCAAAGATAAT-30, reverse primer forC46F: 50- ATTATCTTTGGTATAAGTAAACACGGCAGGGCATTTTGG-30, and these primers were PAGE purified).

2.4. Immunoblotting

Wild type, C46R, and C46F of LGI1 clone and pcDNA3.1(-) weretransfected into the Human Embryonic Kidney (HEK) 293T cells at6-well plate using LipofectamineTM and PlusTM reagents (Cat no.18324-020 and 10964-021, respectively, Invitrogen, USA), and thetransfecting solution was replaced with serum free media forterminating transfection. 40 h after transfection, serum free mediawas separately collected and centrifuged. The supernatant was

collected into the Amicon1 Ultra-4 3K Centrifugal Filters (Cat. no.UFC800324, Millipore, USA) and centrifuged. The remainingconcentrated serum free media was added with 2� lysis buffer(40 mM Tris, 300 mM NaCl, 20% Glycerol, 2% NP-40, and proteaseinhibitor cocktail). Cell was washed twice with 1� phosphatebuffered saline and collected with rubber scraper at the 1� lysisbuffer (20 mM Tris, 150 mM NaCl, 10% Glycerol, 1% NP-40, andprotease inhibitor cocktail), then sonicated for 15 s. Cell lysate wascentrifuged and the supernatant was collected. 2� Tricine samplebuffer (Cat. no. KTR020, Koma Biotech Inc, Korea) was added tothe prepared media and cell lysate. EzWay PAG System Precast Gel(4–12%) and Tricine running buffer (Cat. no. KG50105 and KTR030,respectively, Koma Biotech Inc.) were used for gel electrophoresis.Anti LGI1 (C-19) and Actin (I-19) antibodies (Cat. no. sc-9583 andsc-1616, respectively, Santa Cruz, USA) were employed as primaryantibodies, and ImmunopureR antibody Rabbit anti-Goat IgG(H + L) (Cat. no. 31402, Thermo Scientific, USA) was used as asecondary antibody. Actin beta was used as a control for the celllysate.

2.5. In silico analysis

Model for leucine-rich repeat (LRR) domain of LGI1 wascalculated using automated modeling mode in the SWISS modelserver (http://swissmodel.expasy.org/). Two crystal structures inthe LRR family of VLRB61 and Slit with four LRRs as templatemodels were adopted for modeling process. Model figures wereprepared using programs PyMol (DeLano Scientific). The aminoacid substituting effect for LGI1 protein stability was predictedusing the ‘Sorting Tolerant From Intolerant’ (SIFT) algorithm at theonline tool of http://sift.jcvi.org.11

3. Results

3.1. Clinical description

None of family members had the information for seizure historyof the first generation. Among all family members, elevenmembers experienced seizures: two of four in the secondgeneration, two of 17 in the third generation, six of 11 in thefourth generation, and one of 19 in the fifth generation (Fig. 1A).None of them had antecedents for seizure. Three of the familymembers were deceased (II-2, II-3, and III-12), and out of the eightwho were still alive, six (III-15, IV-8, IV-9, IV-14, IV-16, and IV-19)were taking antiepileptic drugs (AEDs). The mean age at seizureonset for the eight living members was 13.6 years (range, 10 to 22years). The current mean age of unaffected (n = 8) offspring of theaffected fourth generation (IV-11, IV-14, and IV-16) is 13.1 years(range, 7–18 years). All affected family members were right-handed. The three deceased members had generalized tonic–clonicseizures (GTCSs), and no living family member remembered thecharacteristics of deceased members’ auras. All living affectedmembers had GTCSs. Patients IV-11 and V-9 had only one GTCS intheir lifetime. GTCSs occurred predominantly during sleep in fourmembers (III-15, IV-8, IV-16, and IV-19). Four members (III-15, IV-9, IV-14, and IV-16) had complex partial seizures (CPSs). Twomembers (III-15 and IV-14) described specific seizure precipitantssuch as psychic stress, speech, and visual stimulus. The auras ofseven living members were as follows: simple or complex auditoryhallucinations (IV-8 and IV-9), vertiginous auras (IV-8, IV-16 andIV-19), complex visual hallucinations (III-15), and psychic auras(IV-14 and V-9). Six patients (III-15, IV-8, IV-9, IV-14, IV-16, and IV-19) underwent routine scalp EEG recording with nasopharyngealelectrodes. One patient (IV-16) showed synchronous bilateraltemporal sharp waves. Brain MRIs of five patients (III-15, IV-8,IV-14, IV-16, and IV-19) revealed no significant findings. Seizure

Fig. 1. Family pedigree and mutation. (A) Korean ADLTE family pedigree having LGI1 mutation of C46F. Circles denote females; squares denote males; deceased family

members are indicated with diagonal line; affected subjects are marked by black symbols; subjects with unknown information for seizure history are marked by gray

symbols. (B) Direct sequencing results of p.46 (c.137) are compared between wild type and mutant subject. (C) Immunoblotting assay of LGI1 wild type, C46R and F at the

HEK293T cell lysate and media. Beta actin was confirmed for endogenous control. (D) LGI1 protein structures of wild type (WT) and C46F with disulfide bonds of cysteine rich

and LRR domains. The cysteine rich domain N-terminal to LRR, LRR, and cysteine-rich domain C-terminal to LRR are colored pink, yellow, and blue, respectively. Disulfide

bonds are indicated as orange lines.

M.K. Lee et al. / Seizure 23 (2014) 69–73 71

outcomes were favorable. Six of the eight family members (III-15,IV-9, IV-11, IV-16, IV-19, and V-9) had been seizure-free for �1year before the time of the last follow-up. Patient IV-11 was nevertreated with AED. Patient V-9 was treated with AEDs for threeyears but was able to discontinue medication. Two patients (IV-8and IV-14) receiving AEDs were still experiencing seizures. Clinicalprofiles are summarized in Table 1.

3.2. Gene analysis

One nucleotide variant, c.137G > T was shared within the fiveaffected members, and this variant was not detected at the

unaffected member and healthy controls. Each affected subject hadonly one c.137G > T variant allele (Fig. 1B). All tested affectedmembers had heterozygote amino acid of p.C46F encoded by thec.137G > T. which is located next to the c.136T > C coding p.C46Rknown as rs104894166. C46R variant was already reported at themembers of Norwegian family.12

3.3. Protein secretion

Wild type LGI1 protein was detected in both cell lysate andmedia. C46R and C46F mutation was only detected in the celllysate (Fig. 1C).

Table 1Clinical profiles.

Subject Sex Age

(year)

Onset

(year)

F/U

(year)

Seizure MRI EEG AED (mg/day) Outcome

III-15 F 62 13 3 Complex visual hallucinations, CPS, GTCS N N PB 90 SF

Nocturnal predominance of GTCS

Precipitated by emotional stress, conversation, and staring water

IV-8 F 50 12 3 Simple and complex auditory hallucinations, GTCS N N VPA 500, LTG 200,

TPM 250

1 GTCS/month

Nocturnal predominance of GTCS 1 aura/week

IV-9 M 49 10 3 Simple auditory hallucinations, CPS, GTCS N/A N CBZ 200, PHT 200,

PB 60

SF

IV-11 M 46 22 GTCS N/A N/A Never treated SF

IV-14 F 43 14 3 Fear, CPS, GTCS N N VPA 1000, LTG 200,

TPM 50

1 CPS/month

Precipitated by concentration 1 aura/week

IV-16 F 41 13 3 Dizziness, CPS, GTCS N Bi-T SWs LTG 250 SF

Nocturnal predominance of GTCS

IV-19 M 37 15 3 Dizziness, GTCS N N CBZ 400 SF

Nocturnal predominance of GTCS

V-9 M 23 10 Fear, GTCS N/A N/A Previously treated SF

F, female; M, male; (year), years; F/U, follow-up periods; CPS, complex partial seizure; GTCS, generalized tonic–clonic seizure; N, normal; N/A, not available; Bi-T, bilateral

temporal; SWs, sharp waves; AED, antiepileptic drug; PB, phenobarbital; VPA, valproic acid; LTG, lamotrigine; TPM, topiramate; CBZ, carbamazepine; PHT, phenytoin; SF,

seizure-free for �1 year before the time of the last follow-up.

M.K. Lee et al. / Seizure 23 (2014) 69–7372

3.4. Protein structure and stability

There is a LRR domain at the center of the two cysteine richdomains (Fig. 1D). Each cysteine rich domain has two disulfidebonds composed of four cysteine residues. Of the two disulfidebonds existed in the cysteine rich domain N-terminal to LRR, one islocated between the C42 and C48, and the other is between the C46and C55. In the case of the C46F mutation, two disulfide bonds inthe cysteine rich domain C-terminal to LRR are declined to one.LGI1 protein stability was predicted as ‘Tolerated’ or ‘Damaging’according to the amino acid substitution for C46. All amino acidssubstituted for cysteine were predicted as ‘Damaging’ status usingSIFT algorithm. Results of in silico analysis are shown in the on-linesupplementary data.

4. Discussion

The LGI1 protein has two main structural domains; four LRRs inthe N-terminal half, and seven epitempin (EPTP) repeats in the C-terminal portion.13 Three LGI1 mutations affecting conservedcysteine residue preceding LRR were reported to date.3,5,14

Disulfide bonds formed by cysteine residues play an importantrole in maintaining structural stability by protecting the LRRhydrophobic core from exposure.15 Although in silico analysisusing SIFT algorithm has a limitation for predicting the effect ofamino acid substitution that it does not use the protein structure,all amino acids substituted at the C46 position impaired theprotein stability on SIFT prediction.11 Most epilepsy-associatedmutant LGI1 proteins were retained within the cell except for theone encoded by the mutation (c.1219C > T).9 The C46F protein wasalso retained within the cell in our experiment. These findingsreemphasize that conserved cysteine residue is a critical region fornormal protein function and a single functioning copy of LGI1 geneis not enough for normal function, so that loss-of-functionmutations cause a dominant phenotype in this family.

The putative function of LGI1 protein is still unclear. The LGI1

protein has been found associated with different protein com-plexes and several molecular mechanisms possibly underlyingADLTE have been hypothesized. LGI1 forms membrane complexeswith Kv1.1 potassium channels. Both Kv1.1 and LGI1 are co-assembled with KV1.4 and Kvbeta1. LGI1 selectively preventsN-type inactivation mediated by the Kvbeta subunit. Changes ininactivation gating of presynaptic Kv1.1 potassium channels may

promote epileptic activity.16 Extracellularly secreted LGI1 linkstwo epilepsy-related receptors, ADAM22 and ADAM23, in the brainand organizes a transsynaptic protein complex that includespresynaptic potassium channels and postsynaptic AMPA receptorscaffolds. A lack of LGI1 disrupts this synaptic protein connectionand selectively reduces AMPA receptor-mediated synaptic trans-mission.17 Also, LGI1 binding to ADAM23 is necessary to correctlypattern neuronal morphology and altered anatomical patterningcontributes to ADLTE.18

Clinical findings were generally consistent with the character-istics of ADLTE. Focal seizures are clinically characterized byauditory auras in the majority of the cases. Other less frequentauras include aphasic, visual, psychic, autonomic, vertiginous, andother sensory symptoms.19 Two families with the C46R mutationhad prominent aphasic or auditory auras.12,14 The auras of ourpatients appeared to be more diverse compared to the aurasinduced by the C46R mutation. Five patients had psychic,dizziness, and visual auras. Only two patients had auditory auras.Ictal aphasia may not have been identified because of the lack ofprolonged EEG monitoring. Secondarily GTCSs are almost invari-ably present in patients with ADLTE which is consistent with ourresults.4 Seizures could be triggered by sounds, speech, or othersfactors in ADLTE. One patient with visual auras described visualstimuli as one of her seizure precipitants. The precipitating triggersmay reflect the natural function of epileptogenic focus in reflexseizure. The progression of the condition is usually benign. Six ofeight living family members achieved seizure freedom. In the fifthgeneration (n = 11) whose parent had seizures, only one member(V-9) experienced seizures. The paucity of seizure in the fifthgeneration may be explained by still low age for seizure onset orincomplete penetrance (67% and 61.3%, respectively, in twostudies).6,20

5. Conclusions

A new LGI1 mutation was identified in a large Korean ADLTEfamily. Sequencing of LGI1 revealed a heterozygous G to Ttransversion at position 137 of the cDNA resulting in a missensemutation in exon 1, which is predicted to substitute phenylalaninefor cysteine at position 46 of the LGI1 protein (C46F). Clinical andgenetic features were generally consistent with the previouslyreported cases with the exception of the diversity of auras amongthe affected family members. Further studies are required to

M.K. Lee et al. / Seizure 23 (2014) 69–73 73

evaluate the phenotypic diversities of LGI1 mutation relatedADLTE.

Conflict of interest statement

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

Acknowledgements

This study was supported by grant A111218-11-PG03 from theNational Project for Personalized Genomic Medicine, Korea Health21 R&D Project, Ministry for Health & Welfare, Republic of Korea.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the

online version, at http://dx.doi.org/10.1016/j.seizure.2013.10.001.

References

1. Ottman R, Risch N, Hauser WA, Pedley TA, Lee JH, Barker-Cummings C, et al.Localization of a gene for partial epilepsy to chromosome 10q. Nature Genetics1995;10:56–60.

2. Kalachikov S, Evgrafov O, Ross B, Winawer M, Barker-Cummings C, MartinelliBoneschi F, et al. Mutations in LGI1 cause autosomal-dominant partial epilepsywith auditory features. Nature Genetics 2002;30:335–41.

3. Michelucci R, Poza JJ, Sofia V, de Feo MR, Binelli S, Bisulli F, et al. Autosomaldominant lateral temporal epilepsy: clinical spectrum, new epitempin muta-tions, and genetic heterogeneity in seven European families. Epilepsia2003;44:1289–97.

4. Berkovic SF, Izzillo P, McMahon JM, Harkin LA, McIntosh AM, Phillips HA, et al.LGI1 mutations in temporal lobe epilepsies. Neurology 2004;62:1115–9.

5. Ottman R, Winawer MR, Kalachikov S, Barker-Cummings C, Gilliam TC, PedleyTA, et al. LGI1 mutations in autosomal dominant partial epilepsy with auditoryfeatures. Neurology 2004;62:1120–6.

6. Michelucci R, Pasini E, Malacrida S, Striano P, Bonaventura CD, Pulitano P, et al.Low penetrance of autosomal dominant lateral temporal epilepsy in Italianfamilies without LGI1 mutations. Epilepsia 2013;54:1288–97.

7. Bisulli F, Tinuper P, Scudellaro E, Naldi I, Bagattin A, Avoni P, et al. A de novoLGI1 mutation in sporadic partial epilepsy with auditory features. Annals ofNeurology 2004;56:455–6.

8. Michelucci R, Mecarelli O, Bovo G, Bisulli F, Testoni S, Striano P, et al. A de novoLGI1 mutation causing idiopathic partial epilepsy with telephone-inducedseizures. Neurology 2007;68:2150–1.

9. Striano P, Busolin G, Santulli L, Leonardi E, Coppola A, Vitiello L, et al. Familialtemporal lobe epilepsy with psychic auras associated with a novel LGI1 muta-tion. Neurology 2011;76:1173–6.

10. Fanciulli M, Santulli L, Errichiello L, Barozzi C, Tomasi L, Rigon L, et al. LGI1microdeletion in autosomal dominant lateral temporal epilepsy. Neurology2012;78:1299–303.

11. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymousvariants on protein function using the SIFT algorithm. Nature Protocols2009;4:1073–81.

12. Gu W, Brodtkorb E, Steinlein OK. LGI1 is mutated in familial temporal lobeepilepsy characterized by aphasic seizures. Annals of Neurology 2002;52:364–7.

13. Nobile C, Michelucci R, Andreazza S, Pasini E, Tosatto SC, Striano P. LGI1mutations in autosomal dominant and sporadic lateral temporal epilepsy.Human Mutation 2009;30:530–6.

14. Pizzuti A, Flex E, Di Bonaventura C, Dottorini T, Egeo G, Manfredi M, et al.Epilepsy with auditory features: a LGI1 gene mutation suggests a loss-of-function mechanism. Annals of Neurology 2003;53:396–9.

15. Herrin BR, Cooper MD. Alternative adaptive immunity in jawless vertebrates.Journal of Immunology 2010;185:1367–74.

16. Schulte U, Thumfart JO, Klocker N, Sailer CA, Bildl W, Biniossek M, et al. Theepilepsy-linked Lgi1 protein assembles into presynaptic Kv1 channels andinhibits inactivation by Kvbeta1. Neuron 2006;49:697–706.

17. Fukata Y, Adesnik H, Iwanaga T, Bredt DS, Nicoll RA, Fukata M. Epilepsy-relatedligand/receptor complex LGI1 and ADAM22 regulate synaptic transmission.Science 2006;313:1792–5.

18. Owuor K, Harel NY, Englot DJ, Hisama F, Blumenfeld H, Strittmatter SM. LGI1-associated epilepsy through altered ADAM23-dependent neuronal morpholo-gy. Molecular and Cellular Neurosciences 2009;42:448–57.

19. Michelucci R, Pasini E, Nobile C. Lateral temporal lobe epilepsies: clinical andgenetic features. Epilepsia 2009;50(Suppl. 5):52–4.

20. Rosanoff MJ, Ottman R. Penetrance of LGI1 mutations in autosomal dominantpartial epilepsy with auditory features. Neurology 2008;71:567–71.