Progress in brain research, volume 213

1 Corresponding author: Tel.: +4989-5160-4468; Fax: +4989-5160-4470,e-mail address: ortrud.steinlein@med.uni-muenchen.de Abstract Autosomal dominant nocturnal frontal lobe epilepsy ADNFL

New Haven, Connecticut

USA

Donald G SteinAsa G Candler Professor

Department of Emergency Medicine

Emory UniversityAtlanta, GeorgiaUSA

Dick F SwaabProfessor of Neurobiology

Medical Faculty, University of Amsterdam;Leader Research team Neuropsychiatric DisordersNetherlands Institute for Neuroscience

AmsterdamThe Netherlands

Howard L Fields

Professor of NeurologyEndowed Chair in Pharmacology of AddictionDirector, Wheeler Center for the Neurobiology of Addiction

University of CaliforniaSan Francisco, California

USA

First edition 2014

Copyright# 2014 Elsevier B.V All rights reserved

No part of this publication may be reproduced or transmitted in any form or by any means,electronic or mechanical, including photocopying, recording, or any information storage andretrieval system, without permission in writing from the publisher Details on how to seekpermission, further information about the Publisher’s permissions policies and our

arrangements with organizations such as the Copyright Clearance Center and the CopyrightLicensing Agency, can be found at our website:www.elsevier.com/permissions

This book and the individual contributions contained in it are protected under copyright by thePublisher (other than as may be noted herein)

Notices

Knowledge and best practice in this field are constantly changing As new research andexperience broaden our understanding, changes in research methods, professional practices, ormedical treatment may become necessary

Practitioners and researchers must always rely on their own experience and knowledge inevaluating and using any information, methods, compounds, or experiments described herein

In using such information or methods they should be mindful of their own safety and the safety

of others, including parties for whom they have a professional responsibility

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,assume any liability for any injury and/or damage to persons or property as a matter of productsliability, negligence or otherwise, or from any use or operation of any methods, products,instructions, or ideas contained in the material herein

ISBN: 978-0-444-63326-2

ISSN: 0079-6123

For information on all Elsevier publications

visit our website atstore.elsevier.com

Printed and bound in Great Britain

Ste´phanie Baulac

Sorbonne Universite´s, UPMC Univ Paris 06, UM 75; INSERM, U1127; CNRS,

UMR 7225, and Institut du Cerveau et de la Moelle e´pinie`re, ICM, Paris, France

Department of Paediatrics, and Department of Clinical Neurosciences, Faculty

of Medicine and Alberta Children’s Hospital Research Institute, University of

Calgary, Calgary, Alberta, Canada

Antonio Gambardella

Institute of Neurology, Department of Medical Sciences, University Magna

Graecia, Catanzaro, Italy

David A Greenberg

Battelle Center for Mathematical Medicine, Nationwide Children’s Hospital and

Pediatrics Department, Wexner Medical Center, Ohio State University, Columbus,

OH, USA

Ingo Helbig

Division of Neurology, The Children’s Hospital of Philadelphia, Philadelphia, USA

Shinichi Hirose

Department of Pediatrics, School of Medicine, and Central Research Institute for

the Molecular Pathomechanisms of Epilepsy, Fukuoka University, Fukuoka,

Institute of Neurology, Department of Medical Sciences, University Magna

Graecia, Catanzaro, Italy

Holger Lerche

Department of Neurology and Epileptology, Hertie Institute for Clinical Brain

Research, University of Tu¨bingen, Tu¨bingen, Germany

Atul Maheshwari

Department of Neurology, Developmental Neurogenetics Laboratory, Baylor

College of Medicine Houston, TX, USA

v

Carlo Nobile

CNR-Neuroscience Institute, Section of Padua, Viale G, Colombo, Padova, ItalyJeffrey L Noebels

Department of Neurology, Developmental Neurogenetics Laboratory; Department

of Neuroscience, and Department of Molecular and Human Genetics, BaylorCollege of Medicine, Houston, TX, USA

Harvey B Sarnat

Department of Paediatrics; Department of Pathology (Neuropathology), andDepartment of Clinical Neurosciences, Faculty of Medicine and Alberta Children’sHospital Research Institute, University of Calgary, Calgary, Alberta, CanadaOrtrud K Steinlein

Institute of Human Genetics, University Hospital, Ludwig-Maximilians-University,Munich, Germany

There has never been a time before in the history of epilepsy research when scientists

reached such a high level of knowledge The amazing amount of data collected

within the last two decades greatly facilitated our understanding of basic concepts

of epileptogenesis On the other hand, this progress came with the insight that the

mechanisms underlying seizure generation are far more complex than previously

thought Twenty years ago, the introduction of the concept of epilepsies as

channe-lopathies seemed to offer a plausible pathogenetic concept Since then, it has become

obvious that ion channels are only part of the story, and that even abona fide

mu-tation within an ion channel cannot be taken as a proof that disturbed channel

func-tion directly translates into neuronal hyperexcitability More complex mechanisms

have to be considered, and some of them might even precede the first clinically

vis-ible seizure by many years or even decades It also has become obvious that a, most

likely large, number of genes exist that are directly associated with symptomatic or

genetic epilepsies but are neither coding for an ion channel subunit nor for a protein

that has any detectable interactions with such an ion channel Apparently, new

path-ogenetic concepts are needed to guide researchers through the ever-increasing

com-plexity of a field that less than half a century ago had still been dominated by the

hypothesis that a single “epilepsy gene” exists Nowadays, it is clear that a large

number of epilepsy genes hide in our genome, and that these genes are able to cause

seizures by many different mechanisms, both directly and indirectly The selection of

topics presented by the chapters in this book reflects this pathogenetic heterogeneity

as far as this is even possible in a single volume These chapters are not aiming to

simply present a summary of facts but rather try to offer the reader a broad view of the

scientific concepts, theories, and approaches that presently dominate the different

fields in epilepsy research The group of authors that contributed to this book is

as heterogeneous as the epilepsies themselves, including geneticists,

electrophysiol-ogists, and clinical researchers This makes for a lively and sometimes refreshingly

controversial discussion, providing the readers with a wealth of different views,

hy-potheses, and ideas that hopefully create a fertile ground for the development of

suc-cessful future research strategies

Ortrud K Steinlein

vii

1 Corresponding author: Tel.: (+49)89-5160-4468; Fax: (+49)89-5160-4470,

e-mail address: ortrud.steinlein@med.uni-muenchen.de

Abstract

Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) was the first epilepsy in

humans that could be linked to specific mutations It had been initially described as a

chan-nelopathy due to the fact that for nearly two decades mutations were exclusively found in

sub-units of the nicotinic acetylcholine receptor However, newer findings demonstrate that the

molecular pathology of ADNFLE is much more complex insofar as this rare epilepsy can also

be caused by genes coding for non-ion channel proteins It is becoming obvious that the

dif-ferent subtypes of focal epilepsies are not strictly genetically separate entities but that

muta-tions within the same gene might cause a clinical spectrum of different types of focal

Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) was first described

as a distinct familial partial epilepsy in 1994 (Scheffer et al., 1995) Although rare, it

is often referred to not least because of its status as the very first idiopathic epilepsy in

humans for which the underlying genetic cause had been identified (Steinlein et al.,

1995) This was achieved at a time when molecular genetics was still a rather new

field, 300,000-marker genome-wide association studies unheard of, and

high-throughput sequencing a vision rather than daily routine Genotyping of only about

200 polymorphic markers led to the identification of a strong candidate locus for

ADNFLE on the tip of the long arm of chromosome 20 in a large Australian family

Progress in Brain Research, Volume 213, ISSN 0079-6123, http://dx.doi.org/10.1016/B978-0-444-63326-2.00001-6

that included more than 25 affected individuals (Phillips et al., 1995) At that time,this chromosomal region was already in the process of being characterized due to thefact that some years previously it had been identified as a candidate region for an-other type of rare monogenic idiopathic epilepsy, named benign familial neonatalconvulsions (BFNCs) (Leppert et al., 1989) It turned out that the region on chromo-some 20q contains two different ion channel subunit genes,CHRNA4 encoding thea4-subunit of the neuronal nicotinic acetylcholine receptor and the voltage-gated po-tassium channel geneKCNQ2 (Steinlein et al., 1994) The latter one was proven to bethe major gene for BFNC, whileCHRNA4 (and some years later CHRNB2) was iden-tified as one of the main genes that cause ADNFLE (Biervert et al., 1998; De Fusco

et al., 2000; Singh et al., 1998; Steinlein et al., 1995) The identification of these firsttwo seizure-related genes introduced the concept of epilepsies as channelopathies, aconcept that has by now gotten firmly established by the discovery of several addi-tional epilepsy-causing ion channel genes Today, nearly 20 years later, ADNFLE isagain attracting attention by teaching us that one and the same disorder can be both achannelopathy and a non-ion channel disorder (Dibbens et al., 2013; Ishida et al.,2013; Ishii et al., 2013; Martin et al., 2013) (Fig 1;Table 1)

The nAChR subunit genesCHRNA4 and CHRNB2 are responsible for the clinicalphenotype in about 12–15% of ADNFLE patients with a strong family history(Steinlein et al., 2012) Both genes are expressed throughout the brain and the pro-teins they encode ensemble to build one of the most widely expressed nAChRs(3a4/2b2 or 2a4/3b2) in mammalian brain The ubiquitous expression pattern of thisnAChR subtype is surprising given that mutations in these genes cause a seizure

pital lobe

Occi-Frontal lobe Parietal lobe

Temporal lobe

CHRNA4 CHRNB2 CHRNA2 DEPDC5 KCNT1

phenotype that originates from the frontal lobe and rarely shows secondary

general-ization So far, it can only be speculated about the pathomechanisms that prevent

CHRNA4 and CHRNB2 mutations from having a more widespread effect

A possible explanation for this phenomenon could be that in most parts of the brain

the effect the mutations have on neuronal excitability can be compensated by other

nAChR subunits Another possibility would be that genes from other ion channel

families or even non-ion channel genes are involved in this restricted seizure activity

So far, nearly all of the ADNFLE mutations identified within CHRNA4 or

CHRNB2 are missense mutations that cause amino acid exchanges within the second,

less often the first, transmembrane domain (Bertrand et al., 2005; Cho et al., 2003; De

Fusco et al., 2000; Hirose et al., 1999; Magnusson et al., 2003; Phillips et al., 1995;

Steinlein et al., 1995) The nAChR genes encode receptor subunits with four

trans-membrane domains These are either directly or indirectly contributing to the

struc-ture that forms the walls of the ion channel and to the governing of the channels

opening and closing mechanism The second transmembrane domain, consisting

of helical segments forming an inner ring (TM2) that shapes the pore, can be

regarded as a hot spot for ADNFLE mutations Several of these mutations have been

identified more than once in unrelated families from different countries or even

con-tinents This includes the neighboring mutationsSer280Phe and

CHRNA4-Ser284Leu that are so far the most commonly detected ADNFLE mutations (Cho

et al., 2003; Hirose et al., 1999; Ito et al., 2000; McLellan et al., 2003; Phillips

et al., 2000; Rozycka et al., 2003; Steinlein et al., 1995, 2000) These two mutations

are only separated by a few amino acids, but nevertheless differ markedly with

respect to both their biopharmacological characteristics and the severity of the

clin-ical phenotype they are associated with Most of the patients carrying

CHRNA4-Ser280Phe present with an “epilepsy-only” phenotype, while many of those

withCHRNA4-Ser284Leu have additional neurological symptoms such as

mild-to-moderate mental retardation Furthermore, the latter group of patients tend to have

an unusually early age of onset, while carriers of theCHRNA4-Ser280Phe mutation

Table 1 Clinical phenotypes associated with ADNFLE genes

Genes Function Clinical phenotypes

CHRNA4/

CHRNB2

Ion channel ADNFLE CHRNA2 Ion channel NFLE (ADNFLE?)

KCNT1 Ion channel (signaling

function?) Malignant migrating partialseizures

Early infantile epileptic encephalopathy Severe ADNFLE DEPDC5 Non-ion channel Focal epilepsy with variable foci

ADNFLE

The question mark indicates that the clinical phenotype overlaps with that previously described in other

ADNFLE families but might not be identical

develop their seizures at an average age that is typical for most nAChR-caused turnal frontal lobe epilepsies (Bertrand et al., 2002; Cho et al., 2003; Hirose et al.,1999; Ito et al., 2000; McLellan et al., 2003; Phillips et al., 2000; Rozycka et al.,2003; Steinlein et al., 1995, 2000) On a molecular level, the two mutations differedsignificantly with respect to their carbamazepine sensitivity, an antiepileptic drugthat in vivo was shown to be highly effective on CHRNA4-Ser280Phe carryingnAChRs but not on those with the mutationCHRNA4-Ser284Leu (Bertrand et al.,

noc-2002) These results, gained from the analysis of nAChRs expressed in Xenopusoocytes, fit in with the observation that patients with the mutation CHRNA4-Ser280Phe usually benefit from carbamazepine treatment, while sufficient seizurereduction is rarely achieved by carbamazepine monotherapy in patients carryingCHRNA4-Ser284Leu (Bertrand et al., 2002; Cho et al., 2003; Hirose et al., 1999;Ito et al., 2000; McLellan et al., 2003; Phillips et al., 2000; Rozycka et al., 2003;Steinlein et al., 1995, 2000)

The term nocturnal frontal lobe epilepsy describes a large group of partial epilepsiesthat are heterogeneous in origin ADNFLE as a rare monogenic disorder only ac-counts for a small proportion of these epilepsies that are mostly either symptomatic

or multifactorial Patients with sporadic as well as familial nocturnal frontal lobe ilepsy mostly show hypermotoric seizures with movements and vocalizations Due totheir often bizarre nature, the seizures might be misdiagnosed, for example, as a kind

ep-of nonepileptic movement disorder, night terrors, or pseudoseizures alograms (EEGs) are not always helpful to establish the diagnosis because, as com-monly found in frontal lobe epilepsies, they tend to be normal both interictally andictal Consciousness is usually not impaired during seizures and postictal confusion

Electroenceph-is not observed Diurnal seizures might happen; however, most of these seizures cur during daytime naps, while seizures during wakefulness are a rare and infrequentevent (Scheffer et al., 1995; Vigevano and Fusco, 1993) Seizure onset most oftenhappens during the second decade of life; however, it can vary considerably evenwithin the same family (mean age of onset is 14 years (1410 years)) In many in-dividuals, seizures become milder and less frequent once they reach middle age

oc-A possible explanation for this phenomenon could be the normally occurring subtledecline in the number of expressed nAChRs with age

The very first reports described ADNFLE as a rather benign type of epilepsy thataffects otherwise healthy individuals and is readily controlled by carbamazepine.However, follow-up reports put a question mark behind this initial assessment Thiswas mainly due to the frequency with which additional major neurological symptomswere found in patients affected by this “benign” epilepsy A considerable degree ofinterindividual variation is observed with respect to the neuropsychological develop-ment of the patients It can range from normal intelligence to selected cognitive def-icits to different degrees of mental retardation Cognitive impairment seems to be

frequently associated with certain ADNFLE mutations while being rather rare with

other mutations (Bertrand et al., 2005; Hirose et al., 1999; Steinlein et al., 2012) The

same applies to psychiatric symptoms such as schizophrenia-like psychosis that was

present in most patients from a Norwegian ADNFLE family but is not usually seen in

other patients with the same disorder (Magnusson et al., 2003) The results of a

meta-analysis including 19 families with 10 different mutations in either CHRNA4 or

CHRNB2 suggest that some of these mutations are frequently associated not only

with epilepsy but also with additional major cognitive or psychiatric symptoms,

while other ADNFLE mutations are preferentially found in “epilepsy-only” families

Another feature in which patients with ADNFLE demonstrate considerable

inter-individual variability is the sensitivity with which their seizures respond to

antiepi-leptic drug treatment In many families, especially those with mutations such as

CHRNA4-Ser280Phe (previous name Ser248Phe), seizures are sufficiently

con-trolled by the antiepileptic drug carbamazepine Seizures in patients with other

ADNFLE mutations (for example, CHRNA4-Ser284Leu or CHRNA4-Thr293Ile

(previously named Ser252Leu or Thr265Ile)) do not respond easily to

carbamaze-pine or other antiepileptic drugs and might require a multidrug treatment strategy

Quite often, the latter type of ADNFLE mutation is associated with an increased risk

for major comorbidities such as mental retardation or psychiatric symptoms (Cho

et al., 2003; Hirose et al., 1999; Steinlein et al., 1995) The reservation must be made,

however, that for most nAChR mutations the number of known ADNFLE families is

still too small to derive reliable genotype–phenotype relations from them

So far, only a single mutation (Ile279Asn) has been described inCHRNA2, a gene

that encodes one of the majora-subunits of the nAChR (Aridon et al., 2006; Combi

et al., 2009; Gu et al., 2007) The mutation was found in a family of Italian origin in

which 10 members were affected by nocturnal epilepsy The seizure phenotype was

characterized by arousal from sleep, followed by prominent fear sensation and

ton-gue movements Compared to other ADNFLE families, a rather high rate of

noctur-nal wanderings was reported It is therefore not entirely clear yet if the phenotype in

this family is indeed ADNFLE or if is better classified as a separate entity of

noc-turnal frontal lobe epilepsy (Hoda et al., 2009) Analyses of CHRNA2-Ile279Asn

on a molecular basis showed that expression of nAChRs carrying this mutation in

Xenopus oocytes significantly increases the number of receptors expressed at the

membrane surface The mutated receptors also yielded higher ACh-evoked currents

and showed a markedly increased sensitivity toward their natural agonist

acetylcho-line Taken together, it can be concluded that, comparable to the impact ADNFLE

mutations within theCHRNA4 and CHRNB2 genes have, the CHRNA2 mutation

results in a gain-of-function effect, at least in theXenopus oocyte model system This

effect was even stronger whenCHRNA2 was coexpressed with CHRNB2 instead of

CHRNB4 (Aridon et al., 2006; Hoda et al., 2009)

5 BIOPHARMACOLOGICAL PROFILES OF nAChR MUTATIONS

The CHRNA2 mutation Ile279Asn displayed, similar to previously describedepilepsy-causing mutations in nAChRs, its own distinct biopharmacological profile.Unlike the wild-type receptor, which already responds to the open-channel blockercarbamazepine at low doses, blocking of theCHRNA2-Ile279Asn-carrying nAChRswas only achieved at carbamazepine doses above 40mM (Hoda et al., 2009) It istherefore possible that carbamazepine would be ineffective if used as an antiepilepticdrug in patients withCHRNA2-Ile279Asn-caused nocturnal seizures It has not beenentirely understood yet why some ADNFLE mutations are good responders with re-gard to carbamazepine and some are not It had been speculated that these differencesmight be due to the specific position of the mutated amino acids in relation to the ionchannel lumen (Hoda et al., 2009) However, there are no clear differences betweenthe mutations of carbamazepine responders versus nonresponders with respect totheir localization within the protein

It is also not understood yet why only some nAChR mutations exhibit the samegain-of-function effect after application of either nicotine or acetylcholine, whileother mutations differ with respect to these two agonists For reasons unknown, some

of the nAChR mutations are less, while others are more sensitive toward nicotinewhen compared to the wild type (Hoda et al., 2008) It is therefore possible thatsmoking or other means of nicotine consumption might exacerbate seizure activity

in carriers of certain mutations, while others might benefit from it The latter effecthas already been shown for patients from two Norwegian families with ADNFLEmutations CHRNA4-776ins3 and CHRNA4-Ser248Phe (Brodtkorb and Picard,

2006) However, it is possible that experiments involving nicotine application arenot safe enough to be conducted with carriers of mutations showing high affinity to-ward nicotine For example, theCHRNA2 mutation Ile279Asn, which belongs to thelatter group, displays a rather pronounced gain-of-function effect after nicotine ap-plication in theXenopus model (Hoda et al., 2009) It is therefore possible that nic-otine consumption would have a marked negative effect on seizure control in patientswith this mutation

Recently, genomic mapping in a family with genetic epilepsy led to the discovery ofKCNT1 as a new causative gene for both ADNFLE and epileptic encephalopathy.Extension of the study subsequently identified several additional familial and spo-radic cases ofKCNT1-caused ADNFLE (Heron et al., 2012) The clinical phenotypediffered from the above described nAChR-caused ADNFLE insofar as the nocturnalseizures were frequently associated with major comorbidities Intellectual disabilitywas more common in KCNT1 mutation carriers than in those with CHRNA4 orCHRNB2 mutations, and the same was true for psychiatric symptoms or behavioralproblems There was also a significantly lower age of onset (mean 6 years, 4 years

lower than in nAChR-caused ADNFLE), together with complete penetrance

(com-pared to an average 60–80% in nAChR-caused ADNFLE) Interestingly, another

report was published in the same journal issue also describing patients with

muta-tions inKCNT1, but those patients did not present with ADNFLE but with the much

more severe phenotype of malignant migrating partial seizures of infancy (Barcia

et al., 2012) This disorder is part of the group of early onset epileptic

encephalop-athies that includes different catastrophic childhood epilepsies with poor prognosis

In most of these epilepsies, the outcome is characterized by psychomotoric disability

that can be severe Patients with malignant migrating partial seizures of infancy

experience pharmacoresistant polymorphic focal seizures and psychomotor

develop-ment arrests after the first months of life EEGs demonstrate that the seizures are not

always concentrating on the same region (as they do from the frontal lobe in

ADN-FLE) but can arise from various areas of the brain Furthermore, during a seizure,

epileptic activity can be seen to migrate from one part of the brain to another one

(hence the name of the syndrome)

TheKCNT1 gene (also known as SLACK or ENFL5) encodes an outwardly

rec-tifying sodium-activated potassium channel (KCa4.1) that (although calcium does not

act as its main activator) due to sequence homologies belongs to the subfamily of

calcium-activated potassium channels KCNT1 is expressed in two alternatively

spliced isoforms of different length Both isoforms contain six putative

membrane-spanning regions as well as an extended COOH terminus The open

prob-ability of KCa4.1 channels increases with depolarization, implying intrinsic voltage

dependence In brain, the KCa4.1 potassium channel has been shown to be widely

expressed (Joiner et al., 1998) Furthermore, the KCNT1 gene shows a high level

of sequence conservation, a fact that strongly suggests an important functional role,

most likely in neuronal excitability Interestingly, the major areas ofKCNT1

expres-sion include the substantia nigra, frontal cortex, deep cerebellar nuclei, trigeminal

system, subthalamic nuclei, rubrospinal tract, reticular formation, and

vestibulo-ocular tract Several of these brain structures participate in the regulation of

move-ment and posture, and it has therefore been postulated that KCa4.1 might be important

for motor control (Bhattacharjee et al., 2002) Such a role would fit well with the

finding that mutations inKCNT1 are one of the causes of ADNFLE, a disorder

char-acterized mainly by motor seizures

So far, it is not entirely clear how mutations in theKCNT1 gene are able to cause

two different seizure phenotypes Obviously, the mutations found in malignant

mi-grating partial seizures of infancy all occurredde novo, as it is mostly the case in

intellectually debilitating genetic conditions with an age of onset before adulthood

However,de novo mutations are also found in some ADNFLE patients Both the

KCNT1 mutations causing severe ADNFLE and the ones found in malignant

migrat-ing partial seizures of infancy are missense mutations (Barcia et al., 2012; Heron

et al., 2012) In the latter condition, there seem to be hotspots withinKCNT1 for this

mutational mechanism because both the mutation Gly288Ser and Arg428Gln

oc-curred more than once in unrelated patients (Barcia et al., 2012; Ishii et al.,

2013) Several of the so far published mutations occurred at different CpG sites

within theKCNT1 gene These sites are known to be prone to mutational events, mostlikely because the methylated cytosines are vulnerable to spontaneous deaminationinto thymine.

Most of the mutations in both severe ADNFLE and malignant migrating partialseizures of infancy are affecting conserved amino acid residues that are locatedwithin the large COOH-terminal region ofKCNT1, a region of so far unknown func-tion It is therefore not clearly evident if KCNT1-caused epilepsy belongs to thegroup of channelopathies, or if it is caused by different mechanisms The COOH-terminal region is located within the cytoplasma and contains several conservedsequence motifs These are believed to facilitate the interaction with a network ofproteins that regulate channel activity One of the genes involved in this network

isFMR1, encoding the FMRP protein that is involved in rapid, activity-regulatedtransport of mRNAs and has an important role in synaptogenesis and neuronal plas-ticity (Deng et al., 2013) Loss of FMRP is known to cause one of the most commoninherited mental retardation disorders, the fragile X syndrome (Kremer et al., 1991).Under physiological conditions, FMRP binds selectively to sequences at the KCa4.1COOH terminus and subsequently activates the potassium channel (Brown et al.,

2010) Mutations within the COOH terminus might introduce conformational ations that could interfere with the binding of FMRP, causing changes in the firingpattern of neurons expressing the KCa4.1 potassium channel It has also been spec-ulated that the binding to KCa4.1 might in reverse modulate functions of FMRP such

alter-as the regulation of the transport of its cargo mRNAs or the activity-dependent creases in the translation of these mRNAs (Li et al., 2009; Schutt et al., 2009).The direct functional association between the potassium channel KCa4.1 andFMRP, a protein whose loss of function is known to inflict profound intellectual dis-ability, might explain the observation thatKCNT1 mutations not only cause epilepsybut are responsible for a much broader neurological phenotype that can include bothcognitive and psychiatric features (Barcia et al., 2012) However, it cannot be ex-cluded that other, so far unknown, mechanisms are responsible for the clinical spec-trum associated with KCNT1 mutations The COOH-terminal region of KCa4.1contains sequence motifs such as phosphorylation sites, tandem regulators of potas-sium conductance domains, and nicotinamide adenine dinucleotide binding sites.One of the mutations discovered in ADNFLE patients affects an amino acid withinsuch a nicotinamide adenine dinucleotide binding site, while another one targets anamino acid directly adjacent to this site Both mutations were associated with anADNFLE phenotype that was more pronounced than that observed with mutations

in-in other parts of the C termin-inus (Heron et al., 2012) It is therefore possible that aclose relationship exists between the position of theKCNT1 mutation and the clinicalphenotype

So far, the number of known families withKCNT1 mutations is too small to duce reliable genotype–phenotype correlations that could be useful for the geneticcounseling of affected families However, strong genotype–phenotype relationshipsseem to exist, a hypothesis that is supported by the fact that the few recurrentKCNT1mutations that have been reported are causing a roughly uniform phenotype (Barcia

de-et al., 2012; Heron de-et al., 2012; Ishii de-et al., 2013) So far, there is no evidence that a

mutation might be able to cause both severe ADNFLE and malignant migrating

par-tial seizures of infancy There is, however, some indication that the phenotypic

spec-trum associated withKCNT1 mutations might not be restricted to either of these

clinical phenotypes A case report described a single patient with ade novo KCNT1

mutation and a severe clinical phenotype that includes profound psychomotor

retar-dation, microcephaly, deficient neuronal myelination, and therapy-resistant

myo-clonic seizures (Vanderver et al., 2013) The EEG did not show any signs

compatible with the migrating partial seizures seen in the above described patients,

and it is therefore possible thatKCNT1 mutations are associated with a broader

spec-trum on early infantile epileptic encephalopathies

Mutations within theDEPDC5 gene (alternative name KIAA0645) have been

re-cently found in patients with familial focal epilepsies (Dibbens et al., 2013; Ishida

et al., 2013; Martin et al., 2013) The phenotypic spectrum in these families included

the subtypes ADNFLE, familial temporal lobe epilepsy, and familial focal epilepsy

with variable foci The majority of these mutations are nonsense mutations that can

be expected to introduce premature stop codons resulting in nonsense-mediated

mRNA degradation, thus causing a loss-of-function effect The frequency of

DEPDC5 mutations in patients with familial focal epilepsy was estimated to be about

12–27%, renderingDEPDC5 one of the most frequent causes detected so far in

ge-netic epilepsy The penetrance seems to be lower when compared to ADNFLE

caused by nAChR mutations; however, this might be a bias due to the so far low

num-ber of known families (Ishida et al., 2013) The clinical phenotype in patients with

DEPDC5 mutations is rather benign insofar as most of them are of normal intellect

without any detectable structural brain lesions However, autism spectrum disorder

or intellectual disability has been described in some affected individuals (Dibbens

et al., 2013; Ishida et al., 2013; Martin et al., 2013)

DEPDC5 encodes the DEP domain-containing protein 5 that is ubiquitously

expressed in human tissues The DEP domain was named from the initials of three

proteins, disheveled (Dsh), Egl-10, and pleckstrin (Klingensmith et al., 1994; Koelle

and Horvitz, 1996) So far, not much is known about the structure and function of

eitherDEPDC5 or its DEP domain There are a few reports that discuss a possible

role ofDEPDC5 in the pathogenesis of different malignancies; however, the

evi-dence for a role ofDEPDC5 in carcinogenesis is far from conclusive The gene is

located on chromosome 22 in a region that was found to harbor a homozygous

de-letion common to two cases of glioblastoma (Seng et al., 2005) However, structural

aberrations of chromosome 22 are a frequent occurrence in astrocytic tumors, and

this observation could have therefore been caused by coincidental occurrence

In another report, an intronic single nucleotide polymorphism within the

DEPDC5 gene was described as a risk factor significantly associated with the

likelihood of progression to hepatocellular carcinoma in patients with chronic viralhepatitis (Miki et al., 2011) The polymorphism rs1012068 was detected in a genome-wide association study, remained significant after Bonferroni correction for multipletesting, and was confirmed in a follow-up replication study (Miki et al., 2011) Theassociation between rs1012068 and the risk for hepatocellular carcinoma can therefore

be regarded as robust; however, with an odds ratio of about 2 in males (lower in males), the risk conferred by this polymorphism is rather small Furthermore, it is un-known if rs1012068 itself is functional or if it only acts as a placeholder for anotherlinked polymorphism within or outsideDEPDC5 Additional albeit indirect evidencefor a role ofDEPDC5 in tumorigenesis is provided by the observation that DEPDC1, agene containing a DEP domain similar to that present inDEPDC5, has been linked tobladder carcinogenesis (Kanehira et al., 2007) Further studies are needed to clarify ifDEPDC5 indeed participates in the molecular pathology of malignancies and, if so,how this relates to its established role in epileptogenesis

fe-It is nevertheless interesting that with DEPDC5 yet another epilepsy gene hasbeen identified that was discussed as a possible cancer gene before being discovered

to cause a monogenic type of epilepsy The same happened several years previouslywith a different gene,LGI1 on chromosome 10q, which is responsible for autosomaldominant temporal lobe epilepsy (also named autosomal dominant partial epilepsywith auditory features) (Chernova et al., 1998; Gu et al., 2002; Kalachikov et al.,

2002) It is therefore tempting to speculate that at least some of the pathways leading

to cancer might have parts in common with those that cause epilepsy One such nection could be provided by the putative role of DEPDC5 within the mTORpathway

con-The mTOR complex 1 is known as one of the most important regulators of cellgrowth and has been frequently found to be deregulated in different common mul-tifactorial disorders including malignancies and diabetes mellitus (Laplante andSabatini, 2012) The mTOR complex 1 is able to sense amino acid levels by inter-acting with a complex signaling machine (Zoncu et al., 2011) Part of this machine isthe GATOR complex, a multiprotein Rag-interacting complex that contains theDEPDC5 protein as one of its components (Bar-Peled et al., 2013) Experimentallyinduced loss of function in GATOR resulted in hyperactive mTOR complex 1 sig-naling (Bar-Peled et al., 2013) This observation is of interest with respect to epilep-togenesis because aberrant mTOR complex 1 signaling is known to causedisturbances in neuronal migration and cortical lamination This has been demon-strated in different neuronal migration disorders, including tuberous sclerosis Themolecular mechanisms leading to tuber formation during brain development in pa-tients with tuberous sclerosis include loss-of-function mutations in either theTSC1 ortheTSC2 gene that are part of the mTOR signaling cascade This results in consti-tutive mTOR activation which in turn interferes with the development of the cerebralcortex (Prabowo et al., 2013; Tsai et al., 2012) A possibility would be that mutations

inDEPDC5 have a less dramatic effect on mTOR complex 1 signaling but disturb itenough to introduce microscopic changes in brain cytoarchitecture or synaptic

connectivity that are able to promote focal epilepsy without causing macroscopically

visible structural malformations

ADNFLE was the first epilepsy in humans for which mutations have been identified

By now, it has also become the prototype of a neurological disorder that can be

caused by genes coding for either ion channels or non-ion channels The ADNFLE

patients belonging to either one of these genetic subgroups are not easily

distinguish-able from each other on the basis of their clinical characteristics alone One possible

indication pointing toward a non-ion channel origin can be the observation of family

members with other types of focal epilepsies, but this does not apply to all families

concerned Non-ion channel ADNFLE patients tend to have a more severe

pheno-type, both with respect to an earlier age of seizure onset and a higher frequency

of additional major symptoms such as mental retardation However, seriously

af-fected individuals or even whole families with a severe course of the disorder are

also found in the group of ion channel ADNFLE patients, rendering the clinical

course a not very reliable criterion to differentiate between both groups As in many

other rare types of epilepsy, genetic testing has therefore become a routine

instru-ment in the classification of ADNFLE patients Nevertheless, mutations are still only

detectable in far less than half of the ADNFLE families This implicates that

addi-tional genes exist that are able to cause this clinical phenotype Given the rapid

pro-gress in sequencing technologies, it can be expected that at least some of these genes

will be identified within the next few years It will be most interesting to see to which

functional classes these genes belong, and if these new genes are able to further shed

light on the obviously complex pathomechanisms that underlie nocturnal frontal lobe

epilepsy Already, the clinical similarities between the two groups of patients pose

the question whether the proteins encoded by the ion channel and non-ion channel

ADNFLE genes are involved in some of the same, so far unknown functional

pathway(s) Uncovering such common pathways not only will greatly facilitate

our understanding of the molecular basis of epileptogenesis but hopefully will also

be able to reveal new therapeutic targets

REFERENCES

Aridon, P., Marini, C., Di Resta, C., Brilli, E., De Fusco, M., Politi, F., Parrini, E., Manfredi, I.,

Pisano, T., Pruna, D., Curia, G., Cianchetti, C., Pasqualetti, M., Becchetti, A., Guerrini, R.,

Casari, G., 2006 Increased sensitivity of the neuronal nicotinic receptor alpha 2 subunit causes

familial epilepsy with nocturnal wandering and ictal fear Am J Hum Genet 79, 342–350

Barcia, G., Fleming, M.R., Deligniere, A., Gazula, V.R., Brown, M.R., Langouet, M.,

Chen, H., Kronengold, J., Abhyankar, A., Cilio, R., Nitschke, P., Kaminska, A.,

Boddaert, N., Casanova, J.L., Desguerre, I., Munnich, A., Dulac, O., Kaczmarek, L.K.,

Colleaux, L., Nabbout, R., 2012 De novo gain-of-function KCNT1 channel mutationscause malignant migrating partial seizures of infancy Nat Genet 44, 1255–1259.Bar-Peled, L., Chantranupong, L., Cherniack, A.D., Chen, W.W., Ottina, K.A., Grabiner, B.C.,Spear, E.D., Carter, S.L., Meyerson, M., Sabatini, D.M., 2013 A tumor suppressor com-plex with GAP activity for the Rag GTPases that signal amino acid sufficiency tomTORC1 Science 340, 1100–1106.

Bertrand, D., Picard, F., Le Hellard, S., Weiland, S., Favre, I., Phillips, H., Bertrand, S.,Berkovic, S.F., Malafosse, A., Mulley, J., 2002 How mutations in the nAChRs can causeADNFLE epilepsy Epilepsia 43, 112–122

Bertrand, D., Elmslie, F., Hughes, E., Trounce, J., Sander, T., Bertrand, S., Steinlein, O.K.,

2005 The CHRNB2 mutation I312M is associated with epilepsy and distinct memory icits Neurobiol Dis 20, 799–804

def-Bhattacharjee, A., Gan, L., Kaczmarek, L.K., 2002 Localization of the Slack potassium nel in the rat central nervous system J Comp Neurol 454, 241–254

chan-Biervert, C., Schroeder, B.C., Kubisch, C., Berkovic, S.F., Propping, P., Jentsch, T.J.,Steinlein, O.K., 1998 A potassium channel mutation in neonatal human epilepsy.Science 279, 403–406

Brodtkorb, E., Picard, F., 2006 Tobacco habits modulate autosomal dominant nocturnal tal lobe epilepsy Epilepsy Behav 9, 515–520

fron-Brown, M.R., Kronengold, J., Gazula, V.R., Chen, Y., Strumbos, J.G., Sigworth, F.J.,Navaratnam, D., Kaczmarek, L.K., 2010 Fragile X mental retardation protein controlsgating of the sodium-activated potassium channel Slack Nat Neurosci 13, 819–821.Chernova, O.B., Somerville, R.P., Cowell, J.K., 1998 A novel gene, LGI1, from 10q24 is rear-ranged and downregulated in malignant brain tumors Oncogene 17, 2873–2881.Cho, Y.W., Motamedi, G.K., Laufenberg, I., Sohn, S.I., Lim, J.G., Lee, H., Yi, S.D., Lee, J.H.,Kim, D.K., Reba, R., Gaillard, W.D., Theodore, W.H., Lesser, R.P., Steinlein, O.K., 2003

A Korean kindred with autosomal dominant nocturnal frontal lobe epilepsy and mentalretardation Arch Neurol 60, 1625–1632

Combi, R., Ferini-Strambi, L., Tenchini, M.L., 2009 CHRNA2 mutations are rare in theNFLE population: evaluation of a large cohort of Italian patients Sleep Med

10, 139–142

De Fusco, M., Becchetti, A., Patrignani, A., Annesi, G., Gambardella, A., Quattrone, A.,Ballabio, A., Wanke, E., Casari, G., 2000 The nicotinic receptor beta 2 subunit is mutant

in nocturnal frontal lobe epilepsy Nat Genet 26, 275–276

Deng, P.Y., Rotman, Z., Blundon, J.A., Cho, Y., Cui, J., Cavalli, V., Zakharenko, S.S.,Klyachko, V.A., 2013 FMRP regulates neurotransmitter release and synaptic informationtransmission by modulating action potential duration via BK channels Neuron

77, 696–711

Dibbens, L.M., de Vries, B., Donatello, S., Heron, S.E., Hodgson, B.L., Chintawar, S.,Crompton, D.E., Hughes, J.N., Bellows, S.T., Klein, K.M., Callenbach, P.M., Corbett, M.A.,Gardner, A.E., Kivity, S., Iona, X., Regan, B.M., Weller, C.M., Crimmins, D., O’Brien, T.J.,Guerrero-Lopez, R., Mulley, J.C., Dubeau, F., Licchetta, L., Bisulli, F., Cossette, P.,Thomas, P.Q., Gecz, J., Serratosa, J., Brouwer, O.F., Andermann, F., Andermann, E.,van den Maagdenberg, A.M., Pandolfo, M., Berkovic, S.F., Scheffer, I.E., 2013 Mutations

in DEPDC5 cause familial focal epilepsy with variable foci Nat Genet 45, 546–551

Gu, W.L., Brodtkorb, E., Steinlein, O.K., 2002 LGI1 is mutated in familial temporal lobe ilepsy characterized by aphasic seizures Ann Neurol 52, 364–367

ep-Gu, W., Bertrand, D., Steinlein, O.K., 2007 A major role of the nicotinic acetylcholine

recep-tor gene CHRNA2 in autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) is

unlikely Neurosci Lett 422, 74–76

Heron, S.E., Smith, K.R., Bahlo, M., Nobili, L., Kahana, E., Licchetta, L., Oliver, K.L.,

Mazarib, A., Afawi, Z., Korczyn, A., Plazzi, G., Petrou, S., Berkovic, S.F., Scheffer, I.E.,

Dibbens, L.M., 2012 Missense mutations in the sodium-gated potassium channel gene

KCNT1 cause severe autosomal dominant nocturnal frontal lobe epilepsy Nat Genet

44, 1188–1190

Hirose, S., Iwata, H., Akiyoshi, H., Kobayashi, K., Ito, M., Wada, K., Kaneko, S.,

Mitsudome, A., 1999 A novel mutation of CHRNA4 responsible for autosomal dominant

nocturnal frontal lobe epilepsy Neurology 53, 1749–1753

Hoda, J.C., Gu, W., Friedli, M., Phillips, H.A., Bertrand, S., Antonarakis, S.E., Goudie, D.,

Roberts, R., Scheffer, I.E., Marini, C., Patel, J., Berkovic, S.F., Mulley, J.C., Steinlein, O.K.,

Bertrand, D., 2008 Human nocturnal frontal lobe epilepsy: pharmocogenomic profiles of

pathogenic nicotinic acetylcholine receptor beta-subunit mutations outside the ion channel

pore Mol Pharmacol 74, 379–391

Hoda, J.C., Wanischeck, M., Bertrand, D., Steinlein, O.K., 2009 Pleiotropic functional effects

of the first epilepsy-associated mutation in the human CHRNA2 gene FEBS Lett

583 (10), 1599–1604

Ishida, S., Picard, F., Rudolf, G., Noe, E., Achaz, G., Thomas, P., Genton, P., Mundwiller, E.,

Wolff, M., Marescaux, C., Miles, R., Baulac, M., Hirsch, E., Leguern, E., Baulac, S., 2013

Mutations of DEPDC5 cause autosomal dominant focal epilepsies Nat Genet

45, 552–555

Ishii, A., Shioda, M., Okumura, A., Kidokoro, H., Sakauchi, M., Shimada, S., Shimizu, T.,

Osawa, M., Hirose, S., Yamamoto, T., 2013 A recurrent KCNT1 mutation in two sporadic

cases with malignant migrating partial seizures in infancy Gene 531, 467–471

Ito, M., Kobayashi, K., Fujii, T., Okuno, T., Hirose, S., Iwata, H., Mitsudome, A., Kaneko, S.,

2000 Electroclinical picture of autosomal dominant nocturnal frontal lobe epilepsy in a

Japanese family Epilepsia 41, 52–58

Joiner, W.J., Tang, M.D., Wang, L.Y., Dworetzky, S.I., Boissard, C.G., Gan, L.,

Gribkoff, V.K., Kaczmarek, L.K., 1998 Formation of intermediate-conductance

calcium-activated potassium channels by interaction of Slack and Slo subunits Nat

Neurosci 1, 462–469

Kalachikov, S., Evgrafov, O., Ross, B., Winawer, M., Barker-Cummings, C., Martinelli

Bone-schi, F., Choi, C., Morozov, P., Das, K., Teplitskaya, E., Yu, A., Cayanis, E.,

Penchaszadeh, G., Kottmann, A.H., Pedley, T.A., Hauser, W.A., Ottman, R., Gilliam, T.C.,

2002 Mutations in LGI1 cause autosomal-dominant partial epilepsy with auditory

fea-tures Nat Genet 30, 335–341

Kanehira, M., Harada, Y., Takata, R., Shuin, T., Miki, T., Fujioka, T., Nakamura, Y.,

Katagiri, T., 2007 Involvement of upregulation of DEPDC1 (DEP domain containing 1)

in bladder carcinogenesis Oncogene 26, 6448–6455

Klingensmith, J., Nusse, R., Perrimon, N., 1994 The Drosophila segment polarity gene

di-shevelled encodes a novel protein required for response to the wingless signal Genes

Dev 8 (1), 118–130

Koelle, M.R., Horvitz, H.R., 1996 EGL-10 regulates G protein signaling in the C elegans

nervous system and shares a conserved domain with many mammalian proteins Cell

84, 115–125

Kremer, E.J., Pritchard, M., Lynch, M., Yu, S., Holman, K., Baker, E., Warren, S.T.,Schlessinger, D., Sutherland, G.R., Richards, R.I., 1991 Mapping of DNA instability atthe fragile-X to a trinucleotide repeat sequence P(Ccg)N Science 252, 1711–1714.Laplante, M., Sabatini, D.M., 2012 mTOR signaling in growth control and disease Cell

149, 274–293

Leppert, M., Anderson, V.E., Quattlebaum, T., Stauffer, D., O’Connell, P., Nakamura, Y.,Lalouel, J.M., White, R., 1989 Benign familial neonatal convulsions linked to geneticmarkers on chromosome 20 Nature 337, 647–648

Li, C., Bassell, G.J., Sasaki, Y., 2009 Fragile X mental retardation protein is involved in tein synthesis-dependent collapse of growth cones induced by semaphorin-3A Front Neu-ral Circuits 3, 11

pro-Magnusson, A., Stordal, E., Brodtkorb, E., Steinlein, O., 2003 Schizophrenia, psychotic ness and other psychiatric symptoms in families with autosomal dominant nocturnal fron-tal lobe epilepsy caused by different mutations Psychiatr Genet 13, 91–95

ill-Martin, C., Meloche, C., Rioux, M.F., Nguyen, D., Carmant, L., Andermann, E., Gravel, M.,Cossette, P., 2013 A recurrent mutation in DEPDC5 predisposes to focal epilepsies in theFrench-Canadian population Clin Genet.http://dx.doi.org/10.1111/cge.12311 [ahead ofprint]

McLellan, A., Phillips, H.A., Rittey, C., Kirkpatrick, M., Mulley, J.C., Goudie, D.,Stephenson, J.B.P., Tolmie, J., Scheffer, I.E., Berkovic, S.F., Zuberi, S.M., 2003 Pheno-typic comparison of two Scottish families with mutations in different genes causing auto-somal dominant nocturnal frontal lobe epilepsy Epilepsia 44, 613–617

Miki, D., Ochi, H., Hayes, C.N., Abe, H., Yoshima, T., Aikata, H., Ikeda, K., Kumada, H.,Toyota, J., Morizono, T., Tsunoda, T., Kubo, M., Nakamura, Y., Kamatani, N.,Chayama, K., 2011 Variation in the DEPDC5 locus is associated with progression to he-patocellular carcinoma in chronic hepatitis C virus carriers Nat Genet 43, 797–800.Phillips, H.A., Scheffer, I.E., Berkovic, S.F., Hollway, G.E., Sutherland, G.R., Mulley, J.C.,

1995 Localization of a gene for autosomal dominant nocturnal frontal lobe epilepsy tochromosome 20q 13.2 Nat Genet 10, 117–118

Phillips, H.A., Marini, C., Scheffer, I.E., Sutherland, G.R., Mulley, J.C., Berkovic, S.F., 2000

A de novo mutation in sporadic nocturnal frontal lobe epilepsy Ann Neurol 48, 264–267.Prabowo, A.S., Anink, J.J., Lammens, M., Nellist, M., van den Ouweland, A.M.W., Adle-Biassette, H., Sarnat, H.B., Flores-Sarnat, L., Crino, P.B., Aronica, E., 2013 Fetal brainlesions in tuberous sclerosis complex: TORC1 activation and inflammation Brain Pathol

23, 45–59

Rozycka, A., Skorupska, E., Kostyrko, A., Trzeciak, W.H., 2003 Evidence for S284L tion of the CHRNA4 in a white family with autosomal dominant nocturnal frontal lobeepilepsy Epilepsia 44, 1113–1117

muta-Scheffer, I.E., Bhatia, K.P., Lopes-Cendes, I., Fish, D.R., Marsden, C.D., Andermann, E.,Andermann, F., Desbiens, R., Keene, D., Cendes, F., et al., 1995 Autosomal dominantnocturnal frontal lobe epilepsy A distinctive clinical disorder Brain 118, 61–73.Schutt, J., Falley, K., Richter, D., Kreienkamp, H.J., Kindler, S., 2009 Fragile X mental re-tardation protein regulates the levels of scaffold proteins and glutamate receptors in post-synaptic densities J Biol Chem 284, 25479–25487

Seng, T.J., Ichimura, K., Liu, L., Tingby, O., Pearson, D.M., Collins, V.P., 2005 Complexchromosome 22 rearrangements in astrocytic tumors identified using microsatellite andchromosome 22 tile path array analysis Genes Chromosomes Cancer 43, 181–193

Singh, N.A., Charlier, C., Stauffer, D., DuPont, B.R., Leach, R.J., Melis, R., Ronen, G.M.,

Bjerre, I., Quattlebaum, T., Murphy, J.V., McHarg, M.L., Gagnon, D., Rosales, T.O.,

Peiffer, A., Anderson, V.E., Leppert, M., 1998 A novel potassium channel gene, KCNQ2,

is mutated in an inherited epilepsy of newborns Nat Genet 18, 25–29

Steinlein, O., Smigrodzki, R., Lindstrom, J., Anand, R., Kohler, M., Tocharoentanaphol, C.,

Vogel, F., 1994 Refinement of the localization of the gene for neuronal nicotinic

acetylcholine-receptor alpha(4) subunit (Chrna4) to human-chromosome

20q13.2-Q13.3 Genomics 22, 493–495

Steinlein, O.K., Mulley, J.C., Propping, P., Wallace, R.H., Phillips, H.A., Sutherland, G.R.,

Scheffer, I.E., Berkovic, S.F., 1995 A missense mutation in the neuronal nicotinic

acetylcholine-receptor alpha-4 subunit is associated with autosomal-dominant nocturnal

frontal-lobe epilepsy Nat Genet 11, 201–203

Steinlein, O.K., Stoodt, J., Mulley, J., Berkovic, S., Scheffer, I.E., Brodtkorb, E., 2000

Inde-pendent occurrence of the CHRNA4 Ser248Phe mutation in a Norwegian family with

noc-turnal frontal lobe epilepsy Epilepsia 41, 529–535

Steinlein, O.K., Hoda, J.C., Bertrand, S., Bertrand, D., 2012 Mutations in familial nocturnal

frontal lobe epilepsy might be associated with distinct neurological phenotypes Seizure

21, 118–123

Tsai, V., Parker, W.E., Orlova, K.A., Baybis, M., Chi, A.W., Berg, B.D., Birnbaum, J.F.,

Estevez, J., Okochi, K., Sarnat, H.B., Flores-Sarnat, L., Aronica, E., Crino, P.B., 2012

Fetal brain mTOR signaling activation in tuberous sclerosis complex Cereb Cortex

24, 315–327 http://dx.doi.org/10.1093/cercor/bhs310

Vanderver, A., Simons, C., Schmidt, J.L., Pearl, P.L., Bloom, M., Lavenstein, B., Miller, D.,

Grimmond, S.M., Taft, R.J., 2013 Identification of a novel de novo p.Phe932Ile KCNT1

mutation in a patient with leukoencephalopathy and severe epilepsy Pediatr Neurol

50, 112–114

Vigevano, F., Fusco, L., 1993 Hypnic tonic postural seizures in healthy children provide

evidence for a partial epileptic syndrome of frontal lobe origin Epilepsia 34, 110–119

Zoncu, R., Bar-Peled, L., Efeyan, A., Wang, S.Y., Sancak, Y., Sabatini, D.M., 2011 mTORC1

senses lysosomal amino acids through an inside-out mechanism that requires the vacuolar

H+-ATPase Science 334, 678–683

Potassium channel genes

and benign familial neonatal

Snezana Maljevic1, Holger Lerche

Department of Neurology and Epileptology, Hertie Institute for Clinical Brain Research, University

of Tu¨bingen, Tu¨bingen, Germany

1 Corresponding author: Tel.: +49-7071-29-81922; Fax: +49-7071-29-4698,

e-mail address: snezana.maljevic@uni-tuebingen.de

Abstract

Several potassium channel genes have been implicated in different neurological disorders

in-cluding genetic and acquired epilepsy Among them,KCNQ2 and KCNQ3, coding for KV7.2

and KV7.3 voltage-gated potassium channels, present an example how genetic dissection of an

epileptic disorder can lead not only to a better understanding of disease mechanisms but also

broaden our knowledge about the physiological function of the affected proteins and enable

novel approaches in the antiepileptic therapy design In this chapter, we focus on the neuronal

KV7 channels and associated genetic disorders—channelopathies, in particular benign familial

neonatal seizures, epileptic encephalopathy, and peripheral nerve hyperexcitability

(neuro-myotonia, myokymia) caused byKCNQ2 or KCNQ3 mutations Furthermore, strategies using

KV7 channels as targets or tools for the treatment of epileptic diseases caused by neuronal

hyperexcitability are being addressed

Keywords

KCNQ2, KCNQ3, M-current, retigabine, heterologous expression, dominant-negative effect,

haploinsufficiency, developmental expression

Each of approximately 85 billion neurons in the human brain greatly relies in its

function on the specific expression of relatively small proteins—ion channels—in

its membrane These proteins provide a unique milieu in which information can

be generated and transmitted to control both movement of the little toe and creation

of a space shuttle or The Fifth Symphony In other words, ion channels form selective

pores for different ions, which can open and close in a regulated manner and thus

determine the ion flux over membrane, presenting the basis of the electrical

excit-ability Essentially, changes in membrane potential allow opening of voltage-gated

Progress in Brain Research, Volume 213, ISSN 0079-6123, http://dx.doi.org/10.1016/B978-0-444-63326-2.00002-8

ion channels, whereas binding of specific chemical messengers (neurotransmitters)evokes ion passage through ligand-gated ion channels (Lerche et al., 2005).Voltage-gated ion channels are responsible for the generation of action potentialsand their conduction along the axons, as well as for establishing and revokingmembrane potential at rest When action potentials arrive on the presynapticmembrane, they induce Ca2+influx and the release of neurotransmitters, which bind

to the ligand-gated postsynaptic channels and provide the information transmissionbetween cells Neurons can be distinguished by the chemical messengers theyrelease: excitatory neurons communicate via glutamate or acetylcholine, whereasneurotransmitters produced by inhibitory neurons are g-aminobutyric acid (GABA)and glycine Ion channels are further characterized by specific temporal and spatialdistribution and can inhabit different neuronal compartments In excitatory pyrami-dal cells, the specific voltage-gated sodium channels, such as NaV1.2 and NaV1.6(Liao et al., 2010), or potassium channels, such as KV7.2 and KV7.3 (Maljevic

et al., 2008), are expressed in the axon initial segments (AISs), the origin site ofaction potentials In contrast, the NaV1.1 sodium channel is found at the AISs

of the inhibitory neurons (Ogiwara et al., 2007) Ligand-gated channels occupypostsynaptic membranes in dendrites, but some voltage-gated ion channels are alsofound at these sites (Vacher et al., 2008)

The majority of genetic defects detected thus far in idiopathic epilepsies affection channels Genetic alterations can affect channel function and thereby alterthe electrical impulse, modifying neuronal excitability and driving networks ofneurons into synchronous activity, which can finally lead to an epileptic seizure.Moreover, mutations within the postsynaptic receptors can affect the conductionbetween cells and thus present an epilepsy-causing defect

Within a healthy brain, ion channels, ingrained in membranes of excitatory andinhibitory neurons, are providing a neuronal balance Epileptic seizures can beelicited by disruption of this balance caused by ion channel defects and treated byanticonvulsants that are mainly affecting ion channels It is a challenge to ourunderstanding how and why genetic alterations resulting in epileptic seizures donot cause disease phenotype interictally Furthermore, genetic epilepsy disordersoccur at certain age and can in some cases remit spontaneously, indicating that spe-cific patterns of ion channel function or expression may be responsible for the seizureprecipitation

Potassium channel genes cover a number of important physiological functionsand have, therefore, been under a detailed investigation in relation to genetic epilep-sies Indeed, in the past 20 years, several potassium channels have been associatedwith epilepsy Especially one potassium channel family, the KCNQ channels,drew much attention since mutations in the KCNQ genes have been linked todifferent human inherited diseases Mutations inKCNQ2 and KCNQ3 genes werethe first potassium channel mutations associated with an epileptic phenotype inbenign familial neonatal seizures (BFNS) (Biervert et al., 1998; Charlier et al.,1998; Schroeder et al., 1998) In the meantime, the phenotype spectrum related tothese channels extended, including among others severe epileptic encephalopathy

(EE) (Weckhuysen et al., 2012) In parallel, development of the newly approved drug

retigabine, which is acting as an opener of these channels, has started a new era in the

development of antiepileptic drugs

Subunits of potassium (K+) channels are encoded by approximately 80 genes (KCN)

in mammals, and present the most divergent of all ion channel families They are

widely expressed throughout the body having various physiological functions

(Coetzee et al., 1999) The specificity of these channels for K+ over other cations

is defined by a highly conserved amino acid sequence, the so-called GYG signature

sequence, which enables selective transmission of K+ by replacing the six water

molecules that surround these ions The K+channel from aStreptomyces lividans

bacterium KscA was the first crystallized ion channel (Doyle et al., 1998)

Subse-quently determined crystal structures of mammalian channels revealed that

confor-mational changes, which open and close the pore, take place within its inner part

in response to membrane depolarization, binding of Ca2+ or other regulatory

mechanisms (Long et al., 2005)

Based on the number of transmembrane (TM)-spanning regions in each subunit

and their physiological and pharmacological characteristics, K+ channels are

grouped into 2TM, 4TM, and 6TM or 7TM families (Gutman et al., 2005) All

potassium channel genes are thought to emerge by gene duplication from a single

ancestor gene (Jegla et al., 2009) having 2TM segments This structure is

character-istic for the inward-rectifier K+channel family (KIR), including ATP-sensitive K+

channels which associate with sulfonylurea subunits to regulate cellular metabolism

and G-protein-coupled KIR channels As in the majority of K+ channel families,

functional pore is formed by four subunits (Hibino et al., 2010) As a matter of fact,

the 4TM K+channel family is the only one in which the functional pore is formed by

two subunits These channels are unique because they contain two instead of one

pore-forming loop The 4TM, responsible for the leak currents in neuronal cells,

are active at rest and have constitutively open channel gate (Plant et al., 2013)

K+ channels, which are voltage-insensitive and activated by low concentrations

of internal Ca2+, comprise the 6TM family of “small-conductance” (SK) and

“intermediate-conductance” (IK) KCachannels Ca2+does not bind directly on these

channels but is instead bound to calmodulin (CaM), which induces conformational

changes resulting in pore opening (Wei et al., 2005) In the 7TM KCa1.1, so-called

big-conductance (BK) channels, the N-terminus makes a seventh pass through

the membrane to the extracellular side These channels are expressed in a broad

variety of cells and binding of Ca2+is not dependent on its association with CaM

(Shieh et al., 2000)

The largest family of K+channels is encoded by about 40 genes and encompasses

voltage-gated (KV) channels KVchannels consist of four a-subunits, each

contain-ing 6TM regions, which form a scontain-ingle pore (Fig 1) (Gutman et al., 2005) A short

amino acid sequence, containing positively charged arginine residues, forms thefourth TM segment S4 responsible for the channel regulation by voltage and there-fore named the voltage sensor In response to changes in membrane potential, con-formational changes within this region will affect the movement of the channel gate

in the intracellular side of the pore-forming S5–S6 loop Amino and C-terminaldomain are located inside the cell and can vary in their length in different subunits.The subunit assembly domain is usually found at the N-terminus, except in KCNQ(KV7) and hERG (KV11) channels, where it is located in the C-terminus These parts

of the channel also contain binding domains for auxiliary subunits or other regulatoryproteins (Gutman et al., 2005)

Functional diversity of potassium channels is further increased by theirheteromerization into dimers (KIR) or tetramers (KV) and interaction with a number

of auxiliary subunits mRNA splicing and posttranslational modifications alsocontribute to the K+channel diversity (Gutman et al., 2005; Shieh et al., 2000)

2.1 HOW POTASSIUM CHANNELS REGULATE NEURONAL

is determined by the inactivation of NaVchannels, as well as by efflux of K+ions due

to the concentration gradient across the membrane upon the opening of KVchannels

FIGURE 1

Structure and function of KVchannels Functional voltage-gated potassium channels (KV)are made of four subunits A typical structure of a single KVchannel subunit is shown on theleft Due to their different characteristics, KVchannels play diverse physiological roles.Whereas several KVchannels are directly implicated in the membrane repolarization during

an action potential, KV7.2/KV7.3 are active in the subthreshold range and important forthe regulation of resting membrane potential and prevention of repetitive firing (middle).Examples for specific somatodendritic and axonal localization of KVchannels are presented

on the right

(Lehmann-Horn and Jurkat-Rott, 1999) Slower than the NaVchannels, some KV

channel subunits generate fast K+ currents across the membrane, which can also

inactivate and are recognized as A-type potassium currents (Shieh et al., 2000)

Inactivation is a state of the channel protein in which, although still in the open

conformation, the channel pore is not permeable due to occlusion by an amino

terminal sequence (fast, N-type or ball-and-chain sequence inactivation) or a

con-formational change within a pore (slow, P- or C-type inactivation) An important

potassium current in neurons is the so-called M-current, a noninactivating slow

current which is activated at subthreshold voltages and can be regulated by

musca-rinic agonists, which is where the name comes from (Brown and Adams, 1980)

Physiologically, the A-currents will have a larger impact on the initial action

poten-tials within a spike train whereas M-current will determine the response to multiple

spikes, when A-current is inactivated (Bean, 2007; Brown and Adams, 1980)

Typical A-type KV channels are found in KV1–KV4 subfamilies, while KV7

(KCNQ) and KV11 (hERG) produce the M-currents (Fig 1) (Shieh et al., 2000)

Within the central and peripheral nervous systems, the a subunits of KVchannel

family are expressed in both neurons and glial cells and besides excitability also

affects Ca2+signaling, secretion, volume regulation, proliferation, and migration

Within a single neuron, they can occupy different subdomains indicating their

spe-cialized physiological roles (Jensen et al., 2011) For instance, KV2 and KV4 present

somatodendritic channels, KV1 subunits are found on axons and nerve terminals,

KV7 reside mainly at AISs and nodes of Ranvier, and KV3 are expressed in dendritic

or axonal domains, depending on the neuronal cell type or a splice variant (Fig 1)

(Vacher et al., 2008) A variety of molecular mechanisms, including interactions

with other neuronal proteins, determine specific distribution of KV channels in

neuronal membrane subdomains, which is also dependent on and regulated by

neuronal activity (Jensen et al., 2011; Misonou and Trimmer, 2004)

2.2 POTASSIUM CHANNELS IN EPILEPSY AND RELATED DISORDERS

The major physiological roles that potassium channels play in the nervous system

indicate they may be involved in a number of neuronal disorders characterized by

increased excitability, such as epilepsy, migraine, naturopathic pain, ataxia, and others

Diseases caused by dysfunction of ion channels are called “channelopathies.” Before

we concentrate on the neonatal seizures and the associated neuronalKCNQ2/3

chan-nelopathies, we will shortly address the involvement of other potassium channels in

epilepsy and pertinent diseases

2.2.1 Mutations in KV1.1 Cause Episodic Ataxia

KCNA1 gene encodes KV1.1 channel, which is the human homolog of theShaker

potassium channel of the fruit flyDrosophila melanogaster Mutations causing a loss

of function of the Shaker channel in fruit flies are related to the leg-shaking

pheno-type occurring episodically or upon ether anesthesia As mentioned before, KV1.1

channels mediate the fast-inactivating A-currents known to regulate the repolarizing

phase of an action potential (Shieh et al., 2000) Mutations inKCNA1 have beenassociated with episodic ataxia type 1 (EA-1), a human equivalent of the Shakermutation phenotype, characterized by seconds-to-minutes-long ataxia and repetitivedischarges in distal musculature (myokymia) occurring interictally (Browne et al.,

1994) In some cases, complex partial or tonic–clonic seizures have been reported

As a matter of fact, compared to healthy individuals EA-1 patients are about 10 timesmore likely to develop seizures (Rajakulendran et al., 2007) Interestingly, the KV1.1knock-out mouse model also exhibits an epilepsy phenotype and reveals alteredaxonal conduction of action potentials (Smart et al., 1998) Among the loss-of-functionmechanisms caused by distinctiveKCNA1 mutations are altered kinetics, reduced cur-rent amplitudes, or trafficking defects of the KV1.1 channel (Rajakulendran et al.,

2007) Expression of KV1.1 mutations in neurons suggested their major effect wasincreased neurotransmitter release (Heeroma et al., 2009)

2.2.2 KCa1.1 Mutation Linked to Paroxysmal Dyskinesia and Epilepsy

BK channels are ubiquitously expressed and open in response to both Ca2+increaseand voltage A KCa1.1 (geneKCNMA1) mutation has been detected in a large familywith generalized epilepsy and paroxysmal dyskinesia (Du et al., 2005) Paroxysmaldyskinesias present a heterogeneous group of rare neurological disorders featuringsudden, unpredictable, disabling attacks of involuntary movement (hyperkinesias),which may require life-long treatment Functional studies revealed an increased cal-cium sensitivity predicting a gain of function and neuronal hyperexcitability by apresumably faster action potential repolarization (Du et al., 2005)

2.2.3 KV4.2 and Acquired Epilepsy

KV4 channels control somatodendritic excitability by generating subthreshold A-typecurrents In the experimental model of temporal lobe epilepsy, pilocarpine-inducedstatus epilepticus decreased protein levels and increased posttranslational modi-fications of KV4.2 (encoded byKCND2) enhancing dendritic excitability and neuronalactivity and thus promoting the seizure initiation and/or propagation (Bernard

et al., 2004)

In the following, we will introduce the major features of KV7 voltage-gated sium channels and focus on the physiological role, clinical pictures, genetics, andpathophysiology ofKCNQ2/3-associated channelopathies

potas-3.1 MEET THE KCNQS

Five voltage-gated delayed rectifier K+channels (KV7.1–5), encoded by theKCNQgene family and often referred to as KCNQ1–5 channels, gained a VIP (very impor-tant potassium) channels status very soon after their discovery The excitement was

not only because it was recognized that four genes from this small gene family

associate with hereditary human diseases, but also that among them first two

potas-sium channel genes related to genetic epilepsy were found Even more, the products

of these two epilepsy genes were shown to be mainly responsible for the generation

of the slow potassium M-current, previously known for almost two decades as one of

the important regulators of neuronal excitability Lastly, a drug synthesized during

the clinical evaluation of flupirtine was proven to be specifically binding to the pore

sequence of KCNQ channels expressed in brain and has recently been introduced

into the market as a first-in-class anticonvulsive exploiting opening of KVchannels

as a mechanism If we have your attention now, it is time to meet the key players

3.1.1 KCNQ1

KCNQ1 gene, coding for KV7.1 channel proteins, was the first cloned channel of this

family and has been identified using a positional cloning approach on chromosome

11p15.5 in families with long QT syndrome type 1 (Wang et al., 1996) Like in all

the other KV channels, KV7.1 subunits assemble into tetramers, but present the

only KV7 subunit that cannot form heterotetramers with other KV7 family members

Instead, KV7.1 a-subunits coassemble with auxiliary KCNE1 b-subunits, also known

as minK or IsK, to create channels that generate the slow delayed rectifier K+current,

IKs, which plays a key role in cardiac late-phase action potential repolarization

(Barhanin et al., 1996; Sanguinetti et al., 1996) Besides in the heart, KV7.1/KCNE1

channels are expressed in the inner ear, thyroid gland, lung, gastrointestinal tract, the

small intestine, pancreas, forebrain neuronal networks and brainstem nuclei, and in the

ovaries (Goldman et al., 2009; Jespersen et al., 2005) These channels are also found in

the proximal and distal tubule of the nephron (Vallon et al., 2001), which together with

theKcne1 ( / ) mice phenotype, including hypokalemia, urinary and fecal salt

wast-ing, and volume depletion, suggests the importance of these channels for the kidney

function (Arrighi et al., 2001; Vallon et al., 2001; Warth and Barhanin, 2002)

Long QT syndrome (LQTS) presents a disorder of cardiac repolarization, which

predisposes affected individuals to ventriculartorsade de pointes tachyarrhythmias

and cardiac sudden death In fact, two syndromes characterized by LQTS have been

associated withKCNQ1 loss-of-function mutations: autosomal dominant Romano–

Ward and the recessive Jervell and Lange-Nielsen syndrome In the latter, long

QT is combined with congenital deafness (Wang et al., 1996) The KV7.1 mutations

often cause a strong suppression of the remaining WT currents, i.e., the

dominant-negative effect (Maljevic et al., 2010; Schmitt et al., 2000) Since mouse models

carrying LQTS mutations develop spontaneous seizures, a possible role of KV7.1

in epileptogenesis has also been suggested (Goldman et al., 2009)

3.1.2 KCNQ2 and KCNQ3

Two different approaches were used to identifyKCNQ2 and KCNQ3 genes:

screen-ing of a human brain cDNA library usscreen-ing aKCNQ1-derived sequence (Yang et al.,

1998) and positional cloning in families with BFNS (Biervert et al., 1998; Charlier

et al., 1998; Schroeder et al., 1998) The corresponding protein subunits KV7.2 and

KV7.3 are found expressed throughout different brain regions and can form and heterotetrameric channels, which conduct slowly activating and deactivatingcurrent elicited at subthreshold membrane potentials, the so-called M-current(Wang et al., 1998) A number of mutations associated with neonatal seizures,Rolandic epilepsy (Neubauer et al., 2008), or more severe EE (Weckhuysen et al.,

homo-2012) phenotypes have been detected so far, with the majority of them affectingthe KCNQ2 channels One rare single-nucleotide polymorphism in KCNQ3 hasbeen linked with autism spectrum disorders (Gilling et al., 2013)

3.1.3 KCNQ4

TheKCNQ4 gene has been cloned from human retina cDNA library using a KCNQ3partial cDNA probe In parallel, a missense dominant-negative mutation was iden-tified in this gene, which cosegregated with an inherited autosomal dominant form ofnonsyndromic progressive hearing loss (DFNA2) (Jentsch, 2000; Kubisch et al.,

1999) KCNQ4 mRNA is expressed in outer hair cells of the inner ear and lowexpression is found in the brain (Kubisch et al., 1999), restricted to the structures

of the brainstem, predominantly within the nuclei contributing to the central auditorypathway (Kharkovets et al., 2000) The cochlear nerve appeared not to be KV7.4immunoreactive (Kharkovets et al., 2000)

KV7.4 subunits can form homo- and heterotetrameric channels with KV7.3,yielding M-like currents (Kubisch et al., 1999) Detected DFNA2 mutations show

a loss of function either by a haploinsufficiency mechanism or by a negative effect (Maljevic et al., 2010) It has been proposed that dominant-negativemutations preferentially cause all-frequency hearing loss with younger onset, whilemutations following a haploinsufficiency mechanism are related to a late-onsethearing impairment affecting only high frequencies (Topsakal et al., 2005).Two generated mouse models, aKcnq4 / knock-out and a mouse carrying adominant-negative DFNA2 mutation (KCNQ4dn/+), exhibited a hearing loss overseveral weeks with KCNQ4dn/+ mice showing a slower progression (Kharkovets

dominant-et al., 2006) The analysis revealed depolarization and degeneration of outer haircells, indicating that a disrupted potassium efflux due to the absence of KV7.4 cur-rents can lead to the potassium overload of cells and their progressive devolution(Kharkovets et al., 2006)

3.1.4 KCNQ5

The last cloned member of theKCNQ gene family, encoding the KV7.5 subunit, is theKCNQ5 It was cloned from a human brain cDNA library by homology screeningwithKCNQ3 (Schroeder et al., 2000) and use of theKCNQ5 gene sequence identifiedfrom a GenBank search (Lerche et al., 2000) KV7.5 can form heteromers with KV7.3and its distribution is similar to KV7.2 and KV7.3: the splice variant I is found in thebrain and splice variant II and III in skeletal muscles (Lerche et al., 2000; Schroeder

et al., 2000) No mutations related to epilepsy or other hereditary human disordershave been identified so far (Kananura et al., 2000; Maljevic et al., 2010)

3.2 STRUCTURAL AND FUNCTIONAL HALLMARKS OF KV7.2/3

CHANNELS

All KV7 channel share a typical structure of other voltage-gated potassium channels,

meaning that each subunit comprises six transmembrane (TM/S1–6) regions, a

voltage-sensing, arginine-rich S4 segment and a pore formed by loops between

S5 and S6 harboring the GYG sequence With no crystal structure available so

far, the length and borders of the six TM segments are based on hydrophobicity

prediction or use of homology modeling based on the known structure of other

potassium channels (Doyle et al., 1998; Long et al., 2005) The amino (N-) and

carboxy (C-) terminal domains are positioned intracellularly As in other KV

chan-nels, four subunits interact to form a functional channel pore In contrast to the

majority of KVchannels with a tetramerization (T1) domain at their N-terminus,

the assembly of KV7 channel subunits occurs via a domain localized at the

C-terminus (Maljevic et al., 2003; Schmitt et al., 2000; Schwake et al., 2003)

Furthermore, the KV7 C-terminus is exceptionally long and contains many

regula-tory domains (see below)

KV7.2–KV7.5 homotetramers, as well as their heteromeric combinations with

KV7.3 channels, produce the M-current, a slow subthreshold potassium current

which can be abrogated by the activation of muscarinic acetylcholine receptors

(Brown and Adams, 1980; Wang et al., 1998) As previously described, the

M-current is important for the control of the membrane potential and can impede

repetitive neuronal firing

In heterologous systems, the homomeric KV7.3 currents are not greater than

background potassium currents (Schroeder et al., 1998; Wang et al., 1998) On

the other hand, coexpression of KV7.3 and KV7.2 in an equimolar ratio generates

at least 10-fold larger currents in Xenopus oocytes than KV7.2 alone, suggesting

the formation of heteromers Differential sensitivity to TEA, a common KVchannel

blocker, with KV7.2 being more sensitive than KV7.3, was used to confirm the

for-mation of heteromers Expression of tandem KV7.3/7.2 constructs in a nonneuronal

cell line revealed an intermediate TEA sensitivity, which was indistinguishable from

the one obtained for the M-current recorded from the cervical superior ganglion SCG

in adult rats (Hadley et al., 2003; Wang et al., 1998) Thus, the suggested

stoichiom-etry of KV7.2 and KV7.3 subunits in the SCG is 1:1 Moreover, a particular amino

acid residue, localized in the proximity of the GYG sequence in the KV7.3 pore

do-main, was shown as responsible for the detainment of the KV7.3 homotetramers in

the endoplasmic reticulum in neurons In contrast, when combined with KV7.2,

KV7.3 subunits are able to reach the surface membrane as part of the heterotetrameric

complex (Gomez-Posada et al., 2010)

One possible explanation for the current augmentation of KV7.2/KV7.3

heterote-tramers is that compared to their surface expression in the homomeric constellation,

the number of KV7.2 subunits reaching the plasma membrane when they are part of

this heteromeric channel complex with KV7.3 is significantly increased (Schwake

et al., 2000) Apart from the effects on trafficking to the surface, other molecular

mechanisms involving regulatory actions of certain parts of channel proteins orindividual amino acids have been proposed (Etxeberria et al., 2004; Maljevic

et al., 2003) Role of regulating proteins, such as CaM, will be discussed below

3.2.1 What Happens at the C-terminus?

KV7 channels’ particularly long C-terminus is involved in the assembly, trafficking,and gating of these channels (Haitin and Attali, 2008) The predicted secondarystructure shows four helical regions (helices A–D), which are conserved in all familymembers (Yus-Najera et al., 2002) The proximally located helices A and B areresponsible for the interaction with CaM and involved in channel trafficking andgating Distally located helices C and D are thought to form coiled-coil assemblies(Haitin and Attali, 2008)

3.2.1.1 Assembly of KV7 Channels

The study ofSchmitt et al (2000)was the first to reveal that the C-terminal part mayplay a role in the assembly of the KV7 subunits They defined a short amino acidstretch in the C-terminus essential for the functional expression of these channelsand could further show that one of the mutations associated with the Jervell andLange-Nielsen syndrome affected the subunit assembly via the C-terminus Thisfinding prompted research on other KV7 channels Using a chimeric approach, parts

of KV7.1, which does not interact with any other KV7 subunit, were exchanged with

KV7.2 and KV7.3, to demonstrate that the assembly of KV7.2 and KV7.3 channelsalso happens at the C-terminus via a so-called A-domain (Maljevic et al., 2003;Schwake et al., 2003) Ensuing biochemical and structural dissection of the subunitinteraction domain (Schwake et al., 2006; Wehling et al., 2007), including itscrystallization in the KV7.4 (Howard et al., 2007), revealed two coil-coiled stretches,corresponding to helices C and D Whereas helix C shows high conservation among

KV7 channels, the more divergent helix D is probably the one determining subunitassembly specificity

3.2.1.2 Regulation of the M-current

Several regulatory molecules interact with the KV7 channel via their C-terminus.Among them are CaM, A-kinase anchoring proteins (AKAPs), protein kinase-C(PKC), phosphatidylinositol 4,5-bisphosphate (PIP2), and syntaxin 1A (syx)(Haitin and Attali, 2008; Regev et al., 2009) CaM, whose binding site is created

by helices A and B in the KV7.2 C-terminus, is promoting folding and trafficking

of the channel to the plasma membrane (Etxeberria et al., 2008; Haitin and Attali,

2008) Syntaxin is binding to a partially overlapping region with CaM, but exertsopposite effects on the channel function (Regev et al., 2009) Furthermore, a trimericcomplex formed by the AKAP79/150 protein and PKC (Hoshi et al., 2003) is alsobinding at the C-terminus Activated PKC inhibits the channel by phosphorylatingserine residues in helix B and may, therefore, have a significant contribution inthe transmitter-mediated inhibition of KV7 channels (Delmas and Brown, 2005).PIP2is suggested to stabilize the open state of neuronal KV7 channels (Li et al.,

2005), probably by binding to the proximal part of the C-terminus (Haitin and Attali,

2008) Interestingly, these molecules can act antagonistically or synergistically with

CaM, increasing the number of ways in which the function of these channels can be

regulated (Bal et al., 2010; Delmas and Brown, 2005; Etzioni et al., 2011)

3.2.1.3 Targeting and Localization of KV7.2/7.3 Channels

In neurons, KV7.2 and KV7.3 subunits are found in AISs and nodes of Ranvier

(Figs 1 and 2), but studies also indicate their expression in somatic and presynaptic

regions (Devaux et al., 2004; Hu et al., 2007; Maljevic et al., 2008; Martire et al.,

2004; Pan et al., 2006; Vacher et al., 2008) The localization of ion channel proteins

in the AIS is mediated by ankyrin G, large adaptor protein coupling membrane

proteins with actin–spectrin cytoskeleton In fact, the whole organization of the

AIS, site of generation of action potentials, is guided by ankyrin G (Rasband,

2010) The dense structure of AIS provides a necessary milieu for the detainment

and synergistic action of voltage-gated ion channels needed for the generation

and propagation of action potentials and also presents a diffusion barrier between

somatodendritic and axonal compartments of neurons

It was first shown for the NaV channels that a short conserved amino acid

sequence is crucial for the interaction with ankyrin G and thereby their targeting

to the AIS (Garrido et al., 2003; Lemaillet et al., 2003) In KV7.2 and KV7.3, the

ankyrin G interaction domain is found at their C-terminus (Pan et al., 2006), mapping

distally from the helix D Studies on neurons from the ankyrin G knock-out mice

show that lack of ankyrin G abolishes AIS targeting of both NaV and KV7.2/7.3

channels Furthermore, deletion of the KV7.3 ankyrin G binding domain had a

greater impact on the AIS targeting of the heteromeric KV7.2/7.3 complex than

the disruption of this domain in KV7.2 (Rasmussen et al., 2007) Interestingly,

ankyrin G binding motif only emerged in the vertebrate orthologues of Nav and

KV7 genes, coinciding with the development of myelination (Pan et al., 2006)

FIGURE 2

KV7.2 expresses in the axon initial segment (AIS) AIS is a neuronal compartment

with a high concentration of ion channels involved in action potential generation

Immunohistochemical staining of a mouse brain section reveals the colocalization of

KV7.2 and NaV1.2 channels in this region (Maljevic and Lerche, unpublished data)

Immunohistochemical analysis revealed that the density of ion channels out the AIS is not homogenous and can vary during development (Fig 2) Forinstance, NaV1.2 is found expressed earlier in the development at the AIS andsettles in its proximal part in adult neurons This process is paralleled by emergence

through-of NaV1.6, occupying the distal part of the axon (Liao et al., 2010) The gical meaning of the subdomain-specific localization is probably that low thresholddistal NaV1.6 is more important for the initiation of action potentials, whereas prox-imal high-threshold NaV1.2 plays a role in backpropagation to the soma (Hu et al.,

physiolo-2009) However, the developmental expression pattern is still unclear, although

it nicely explains the reminiscence of seizures caused by mutations in NaV1.2(see below)

Detailed analysis using confocal imaging and patch clamp recordings in the AISshowed that the density of KV7.2/7.3 channels is highest in the distal two-thirds ofthe AIS (Battefeld et al., 2014) The same study suggested that somatodendritic KV7channels could be robustly activated by the backpropagating action potentials, toabate afterdepolarization and repetitive firing On the other hand, axonal KV7channels may have a role in stabilizing the resting membrane potential, which in-creases the availability of NaVchannels and the action potential amplitude in nodes

of Ranvier (Battefeld et al., 2014)

3.3 EXPRESSION PATTERN OF NEURONAL KV7 CHANNELS

KV7.2 and KV7.3 channels are expressed together in different neurons of the centraland peripheral nervous system In the brain, they are found at different sites, includ-ing hippocampus, cortex, and thalamus, in both inhibitory and excitatory neurons(Cooper et al., 2000, 2001) However, in situ hybridization also indicates thatKCNQ2 and KCNQ3 mRNA is not always expressed in the same ratio (Schroeder

et al., 1998) This is supported by the observation that, in some neurons, only one

of the two subunits can be detected using immunohistochemistry (Cooper et al.,

2000) In rodents, both channels are found at low levels in the first postnatal days,and the expression increases within the first weeks of development (Geiger et al.,2006; Maljevic et al., 2008; Weber et al., 2006)

3.4 INSIGHTS FROM THE MOUSELAND

Several mouse models have been created to study the effects of either gene tion (knock-out models) or specific single amino acid exchanges found in patientsand inserted at the homologous site of the mouse gene (so-called humanizedknock-in mouse models) In the Kcnq2 knock-out line, / pups die right afterbirth due to pulmonary atelectasis In a hemizygous +/ constellation, animalshave no spontaneous seizures, but show increased sensitivity when pentylenetetra-zole is used to induce them (Watanabe et al., 2000) Removal of the Kcnq3does not produce any specific phenotype and these mice are viable (Tzingounisand Nicoll, 2008)

dele-To understand what happens in neurons lacking one or both alleles of KV7.2,

Robbins et al (2013)studied sympathetic neurons isolated from late Kcnq2 /

or +/ embryos Expectedly, quantitative PCR revealed lack of Kcnq2 mRNA

in the / and about 30% reduction in +/ neurons, translating into the absence

or reduction of the resulting M-current, respectively Interestingly, in both

geno-types, an increase in the expression of Kcnq3 and Kcnq5 mRNA was found

In neurons from the adultKcnq2 +/ mice, M-current level was same as in the

WT neurons, probably due to increased expression from the remaining allele as a

compensatory mechanism

To circumvent the early loss ofKcnq2 / mice,Peters et al (2005)designed a

conditionalKcnq2 knock-out model by introducing a dominant-negative mutant

un-der the antibiotic control so that it can be activated at different time points during

development Interestingly, induction of expression of this dominant-negative

Kcnq2 mutation in the right time window provoked spontaneous seizures,

accompa-nied with cognitive impairment and morphological changes in the hippocampus At

the time of generation of this mouse model, the severe phenotype seemed at

odds with the benign clinical pictures found in patients carryingKCNQ2 mutations,

but as it will turn out, corresponds well with the clinical picture ofKCNQ2-related

EE (Weckhuysen et al., 2012)

Two knock-in models, carrying either a KCNQ2 or a KCNQ3 BFNS-causing

mutation, have also been created (Singh et al., 2008) Homozygous mice revealed

reduced M-currents and showed spontaneous seizures throughout life, though not

limited to the early period of development, thus not faithfully reproducing the BFNS

phenotype The heterozygotes exhibited a reduced seizure threshold upon

appli-cation of convulsant drugs (Singh et al., 2008) The increased seizure susceptibility

also occurred in a sex-, mouse strain-, and seizure test-dependent manner (Otto

et al., 2009)

3.5 FUNCTIONAL ANALYSIS OF DISEASE-RELATED MUTATIONS

Ion channel defects can be examined in heterologous expression systems as well as

in neuronal cell lines and animal models The former implies expression of the

affected protein in a system free from endogenous channels with the same or

sim-ilar function Commonly used are different mammalian cell lines orXenopus laevis

oocytes The cRNA or cDNA encoding the WT and mutant channel is injected or

transfected in such cells and after providing enough time for production of

enco-ded proteins, analyzed in parallel using a combination of electrophysiological,

biochemical, or immunohistological techniques The obtained results show how

the mutant channel behaves or expresses compared to the WT Channels may

act differently in such expression systems in comparison to their native

environ-ment However, since the major question is whether a mutation significantly

affects channel function, data from heterologous systems are valuable initial step

in the functional analysis Furthermore, many of the obtained results could be

reproduced in animal models

Use of neuronal expression system has certain complications Namely, number

of available neuronal cell lines is limited, and rodent primary neuronal culturesare viable only for several weeks More importantly, neuronal cells express a wholerange of different channels, including the channel of interest, so it is a challenge torecognize particular effects of the analyzed protein As shown above, several mousemodels forKCNQ channels are available It is, however, unrealistic to expect thatgenetic murine models can be generated to study all detected mutations

3.6 KCNQ2 AND KCNQ3 CHANNELOPATHIES

The clinical phenotype first linked to mutations in the KCNQ2 and KCNQ3 was

a rare benign form of neonatal epilepsy (BFNS) Many reports of more complexphenotypes, including peripheral nerve hyperexcitability (PNH) and myokymia orRolandic epilepsy with centrotemporal spikes, emerged over time However, itwas only recently that a systemic analysis of a cohort of severely affected childrenwith refractory epilepsy and mental retardation introduced an EE as a clinicalphenotype related toKCNQ2 mutations This pattern with a spectrum of phenotypesassociated with mutations in a single gene has already been observed in other ionchannel genes associated with epilepsy, such asSCN1A or SCN2A, in which epilepticdisorders range from febrile seizures combined with heterogeneous generalizedepilepsy (GEFS +) to the severe myoclonic epilepsy of infancy (SMEI or Dravetsyndrome), and from benign familial neonatal–infantile seizures (BFNIS) to EE,respectively (Reid et al., 2009) As a rule, more severe phenotypes are usually linked

tode novo mutations affecting these genes

3.7 KCNQ2/3 MUTATIONS IN BFNS

Three epilepsy conditions beginning within the first year of life present mainly arily generalized focal seizures, have transient expression and generally benign out-come, and show autosomal dominant mode of inheritance Based on the exact onsettime and the genes involved we delineate among: BFNS starting typically beforethe fifth day of life (Rett and Teubel, 1964), BFNIS occurring between day twoand 6 months of age (Kaplan and Lacey, 1983), and benign familial infantile seizures(BFIS) emerging between 3 and 8 months of age (Vigevano et al., 1992) Interest-ingly, specific genes have been linked to each phenotype, withKCNQ2 and KCNQ3causing BFNS,SCN2A driving BFNIS, and PRRT2 being responsible for BFIS There-fore, genetic analysis can help define and differentiate among these syndromes

second-A recent study (Zara et al., 2013) in a group of patients with all three syndromesrevealed that a certain overlap exists between BFNS and BFNIS, since early occurringseizures, even if starting later than day 5 and therefore diagnosed as BFNIS, are likely

to be related toKCNQ2 mutations, although this is not always the case

3.7.1 Clinical Features and Genetics of BFNS

Starting in the first days of life, seizures in BFNS often occur in clusters and remitspontaneously after weeks to months If at all needed, treatment is only required for ashort period The onset of seizures is partial, and they are often accompanied with

hemi-tonic or -clonic symptoms, apnoeic spells, or clinically appear as generalized.

Electroencephalograms (EEGs) usually show normal interictal activity, whereas the

recorded ictal EEGs reveal a focal onset, and sometimes also bilateral synchrony

About 15% of patients may have recurring seizures later in life Inheritance is

autosomal dominant and a penetrance of 85% has been estimated Corresponding

to the benign outcome, the psychomotor development is in most cases normal

(Maljevic et al., 2008)

Sporadically, patients with mental retardation and difficult to treat epilepsies have

been described (Alfonso et al., 1997; Borgatti et al., 2004; Dedek et al., 2003; Schmitt

et al., 2005; Steinlein et al., 2007) A study of a cohort of BFNS families (Soldovieri

et al., 2014) reported that 5 out of the 17 families included one or two individuals with

more severe clinical picture, encompassing delayed psychomotor development,

intellectual disability, or other neurological features Other affected members in these

families had only benign neonatal seizures However, recent studies of cohorts of

severely affected patients introduced KCNQ2 mutations as a common cause of a

specific phenotype they described as KCNQ2-related EE (see below) Moreover,

one of the most common epilepsies in childhood, the so-called Rolandic or benign

epilepsy of childhood with centrotemporal spikes, has also been associated with

KCNQ2/3 mutations (Coppola et al., 2003; Neubauer et al., 2008) More than 50

mutations inKCNQ2 and 6 mutations in KCNQ3 have been described to cause BFNS

(Fig 3) Furthermore, deletions or duplications ofKCNQ2 gene are found in a

signif-icant proportion of BFNS families (Heron et al., 2007)

3.7.2 Pathogenic Mechanisms in BFNS

KV7.2 and KV7.3 mutations have been analyzed in heterologous systems, such as

X laevis oocytes and mammalian cell lines, and two mouse models carrying BFNS

mutations have also been generated (Jentsch, 2000; Maljevic et al., 2010) The

common feature of all studied mutations is a loss of function in both homomeric

and heteromeric channel conformations The mechanisms underlying loss of

func-tion include haploinsufficiency, gating alterafunc-tions and rarely a dominant-negative

effect (Fig 4) (Maljevic et al., 2008) Especially the cytoplasmic C-terminus of

KV7.2, the pore regions (S5–S6 segments) of both KV7.2 and KV7.3 channels,

and the voltage sensor S4 and the S1–S2 region of KV7.2 are affected by BFNS

mutations (Fig 3)

The common functional consequence of all mutations examined so far is a

reduc-tion of the resulting K+current (Fig 4) Even though a complete loss of function of

KV7.2 or KV7.3 is often observed (Jentsch, 2000; Lerche et al., 1999; Maljevic et al.,

2010), a coexpression of wild-type (WT) and mutant KV7.2 or KV7.3 with the WT of

the other subunit in a 1:1:2 ratio, translating the expected expression ratio in patients,

revealed a reduction in the current size of merely 20–25% compared with

coexpres-sion of both WTs This means that relatively small decline in the KV7.2/KV7.3

M-current appears to be sufficient to cause epileptic seizures in neonates (Bassi

et al., 2005; Jentsch, 2000; Lerche et al., 1999; Maljevic et al., 2008; Schroeder

et al., 1998; Singh et al., 2003; Soldovieri et al., 2014) Even in families with larger

phenotypic variability, including more severe neurological outcomes,in vitro studies

KV7.3, the recorded currents are strikingly increased for the WT and the majority of mutatedchannels In experiments mimicking the presumed ratio of KV7 subunits in a patientcarrying a KV7.2 mutation, expected decrease of recorded currents is 20–30% forhaploinsufficiency pathomechanism, or larger when the mutant exerts the dominant-negativeeffect BFNS mutations incline to the former and EE mutants to the latter mechanism.

Modified after Maljevic et al (2011) and Orhan et al (2014)

revealed a comparably mild reduction of the maximal KV7.2/KV7.3 currents,

sug-gesting that additional genetic and environmental factors may contribute to

phenotypic variability in BFNS families (Steinlein et al., 2007)

Since the C-terminal domain contains many important regions, including the

tetramerization domain and binding sites for other regulatory proteins, impaired

tetramerization or reduced trafficking to the surface membrane could explain how

C-terminal mutations reduce KV7.2 currents For example, reduced surface

expres-sion was found for the KV7.2 mutant truncating the C-terminus (Schwake et al.,

2000) Two other C-terminal mutations were shown to disrupt the binding to

CaM (Richards et al., 2004) and for one of them impaired trafficking to the surface

membrane could be confirmed (Etxeberria et al., 2008) A mutation causing

frame-shift and prolongation of the channel protein was shown to decrease protein stability

(Soldovieri et al., 2006) For some mutations affecting the C-terminus impaired

regulation by syntaxin-1A, part of the presynaptic SNARE complex, known to

reduce KV7.2 currents, has been demonstrated (Soldovieri et al., 2014)

Interestingly, majority of reported mutations affecting the pore region in KV7.2 or

KV7.3 do not show a dominant-negative effect on the WT subunits, despite presence

of an intact C-terminal assembly domain (Charlier et al., 1998; Hirose et al., 2000;

Schroeder et al., 1998; Singh et al., 1998, 2003) These mutants probably affect ion

channel conductance and reduce K+ currents by a haploinsufficiency mechanism

However, for one trafficking-defective KV7.2 mutation located in the pore region,

a dominant-negative effect was reported (Maljevic et al., 2011) Remarkably, the

expression of this mutant channel in the surface membrane could be partially

restored by lowering the incubation temperature or by long exposure of cells to high

doses of retigabine, a neuronal KV7 channel opener

Interestingly, out of six mutations inKCNQ3 reported so far (Charlier et al., 1998;

Fister et al., 2013; Hahn and Neubauer, 2009; Hirose et al., 2000; Singh et al., 2003;

Soldovieri et al., 2014; Zara et al., 2013), five are found in the pore region and one in

the beginning of the S6 segment Functional analyses in heterologous systems

revealed a 20–40% reduction of KV7.2/KV7.3 currents (Jentsch, 2000; Singh

et al., 2003; Soldovieri et al., 2014), whereas one mutation was shown to cause a

dominant-negative effect (Sugiura et al., 2009; Uehara et al., 2008)

Changes in KV7.2 channel gating have been reported for the mutations perturbing

the S4 voltage sensor (Dedek et al., 2001; Miraglia del Giudice et al., 2000; Singh

et al., 2003; Soldovieri et al., 2007; Wuttke et al., 2007) Mutations affecting arginine

residues, thus altering positive charges within the S4 segment, cause rightward shift

of the activation curve accompanied with slowed activation and faster deactivation

kinetics (Miraglia del Giudice et al., 2000), together with decreased voltage

sensi-tivity (Castaldo et al., 2002), whereas mutations of noncharged residues produce

atypical gating, where rightward shift of the activation curve is accompanied with

a slowing of activation kinetics upon stronger depolarizing prepulses (Soldovieri

et al., 2007) Two KV7.2 mutations affecting the same positive charge (R207) and

exhibiting a pronounced dominant-negative effect have been associated with PNH

or BFNS and myokymia (see below)

Changes in channel gating have also been linked to the two mutations affectingthe S1–S2 extracellular loop (Hunter et al., 2006; Wuttke et al., 2008), both reveal-ing a significant reduction of the relative current amplitudes limited to subthresholdvoltages In reconstitution experiments supporting the presumedin vivo constella-tion of mutant and WT subunits, the observed changes were even smaller than the25% seen for other mutations, but sufficient to elicit prolonged bursts of actionpotentials and yield a lower threshold for infinitive firing in one compartmentneuronal model cell (Wuttke et al., 2008) An interaction of the mutated residueswith a positive charge within the voltage sensor S4 may be an explanation for thiseffect These findings emphasized in human disease model that the subthresholdvoltage range is most relevant for M channels to modulate neuronal firing.

3.7.2.1 Mechanisms of Spontaneous Seizure Remission in BFNS

Although our understanding of mechanisms underlining the occurrence of neonatalseizures has significantly increased in the last years, the question of their transientexpression, limited to the first days of life, remains puzzling

Could developmental changes in expression patterns of KV7.2 and KV7.3 channelsaccount for the neonatal seizure phenotype and spontaneous remission of seizures?The available data from rodent brains suggest a significant upregulation in expression

of both channels within the first three postnatal weeks (Geiger et al., 2006; Maljevic

et al., 2008; Weber et al., 2006), which means that only a small number of neuronal

KV7 channels are responsible for an adequate control of the subthreshold membranepotential in neonates’ brain If the critical amount of functional channels could not

be reached due to even a mild loss of function caused by mutations, the M-currentmay be on too low level, and this can lead to the generation of seizures By contrast,this happens only rarely in adulthood, when M channels are abundantly available, or anupregulation of other K+channels helps compensate for the M-channel deficit Fur-thermore, the proposed excitatory action of GABA in the immature brain could aggra-vate this effect (Okada et al., 2003) Namely, the intracellular concentration of Cl ions

in neurons is increased in the early postnatal period, and binding of GABA will elicitoutward Cl currents that cause membrane depolarization, just opposite to the hyper-polarizing effect of inward Cl currents in the mature brain Hence, with GABA acting

as a depolarizing signal, the M-current might have even more important role as an hibitor of neuronal firing and its importance would diminish parallel to the inhibitoryswitch of the GABAergic system In addition, the exclusive expression of a shorter,nonfunctional splice variant of KV7.2 in fetal brain which can attenuate KV7.2/

in-KV7.3-mediated currents (Smith et al., 2001) has been suggested to contribute tothe seizure occurrence and remission

patients performed byWeckhuysen et al (2012)that revealed thatKCNQ2 mutations

are responsible for this severe phenotype in about 10% of patients In their first

study, seven novel KCNQ2 mutations were detected, six of them occurring

de novo, and ensuing research established KCNQ2-related encephalopathy as

fre-quent early onset (neonatal) EE phenotype (Milh et al., 2013; Numis et al., 2014;

Weckhuysen et al., 2013)

3.8.1 Clinical and Genetic Features

Early onset EEs include a divergent group of syndromes characterized by early

occurrence of seizures correlated with impaired neurological development Patients

affected with KCNQ2-related encephalopathy present with pharmacoresistant

neonatal onset seizures with strong tonic component In contrast to BFNS patients,

the interictal EEG activity is characterized by burst suppression or multifocal spikes

and transient T1 and T2 hyperintensities of the basal ganglia have also been reported

(Weckhuysen et al., 2012) Whereas seizures generally remit by age of 3, profound

intellectual disability and motor impairment persist (Milh et al., 2013; Weckhuysen

et al., 2012, 2013) The level of impairment may vary as well as the ability to learn to

walk or speak by the age of 3 (Milh et al., 2013; Weckhuysen et al., 2013)

So far, three cohorts of patients presenting this phenotype have been screened

for mutations in theKCNQ2 gene Following initial study by Weckhuysen and

col-leagues, who sequenced bothKCNQ2 and KCNQ3 in the cohort of 80 patients, finding

sixde novo KCNQ2 mutations and a mosaic mutant in a patient with a milder

pheno-type, two other larger cohorts have been analyzed (Milh et al., 2013; Weckhuysen

et al., 2013) Percentage of patients carryingde novo mutations in these two studies

varied between 13% (11/84 patients; Weckhuysen et al., 2013) and 23% (16/71

patients;Milh et al., 2013), and the number of detectedde novo KCNQ2 mutations

rose to about two dozen Interestingly, using whole exome sequencing of 12 patients

with Ohtahara syndrome, presenting with similar features as theKCNQ2

encephalop-athy, threede novo KCNQ2 mutations could be detected (Saitsu et al., 2012)

Thorough phenotypic characterization in these studies has been accompanied

with the treatment response analysis (Numis et al., 2014; Weckhuysen et al.,

2013) Positive response to the KV7.2 channel opener retigabine was found in one

EE patient, whereas some responded well to carbamazepine But, seizure-free status

in these children did not seem to improve the severe psychomotor delay (Numis

et al., 2014)

3.8.2 Pathophysiologic Mechanisms of EE

Functional analysis of seven KCNQ2 encephalopathy mutations detected in the

initial report (Weckhuysen et al., 2012) revealed a loss of function of the mutant

KCNQ2 allele The study done in X laevis oocytes unveiled that five out of seven

analyzed mutants exhibited a strong dominant-negative effect on the WT subunits

This effect was found for only 4 out of more than 50 known BFNS mutations,

sug-gesting that it may present a prevailing mechanism behind the severe EE phenotype

(Orhan et al., 2014)

The EE mutations affect functionally important parts of the channel, including thevoltage-sensing S4, the pore, and the C-terminus domain (Fig 3) Not surprisingly,both mutations in the S4 segment cause large depolarizing shifts in voltage-dependentactivation, especially pronounced for the R213Q mutant This mutation affects argi-nine on the same position as one BFNS mutation (R213W).Miceli et al (2013)com-pared the effects of the two mutations and showed a more pronounced functionaldefect, mainly expressed as a dramatic decrease of voltage sensitivity, for theR213Q, which is possibly an explanation for the more severe EE phenotype Whenanalyzed inX laevis oocytes, the R213Q mutation had dramatically reduced currentamplitudes, although the surface expression seemed unaffected, suggesting thatmutated channels may fail to open in response to depolarization However, the mostimportant effect for both S4 mutations seems to be the prominent dominant-negativeeffect on coexpressed KV7.2 and KV7.3 WT subunits in a 1:1:2 ratio, mostlypronounced in the subthreshold range of an action potential (Orhan et al., 2014).

Of the three pore mutations, all showing dramatically abbreviated currents, twocannot be rescued by KV7.3 coexpression, which has not been observed for any other

KV7.2 mutation so far (Orhan et al., 2014) The current reduction in 1:1:2 coexpressionexperiments for these mutants is>50%, whereas a mere 20–30% reduction of the

KV7.2/KV7.3 current amplitude corresponding to a haploinsufficiency mechanism istypical for the majority of BNFS-causing mutations (Jentsch, 2000; Maljevic et al.,

2008) Moreover, for the only two BFNS mutations causing a dominant-negative effect

by an overall current amplitude reduction, the effect on the 1:1:2 currents was less nounced (Maljevic et al., 2011; Singh et al., 2003), marking a close genotype–phenotype correlation The third pore mutant yields small currents on its own and

pro-an almost 50% reduction in the 1:1:2 coexpression experiments (Orhan et al., 2014).Whereas S4 and pore mutations present with a clear dominant-negative effect,functional defects of the two C-terminal mutations are comparable to those of BFNSmutations Additionally, one of them shifts the activation curve to more depolarizedpotentials, which is completely reversed by its coexpression with the WT subunits.The reduced potassium currents may be related to a disrupted trafficking to the sur-face membrane However, the oocyte expression system is perhaps not best suited toanalyze channel trafficking and its interactions with other molecules present in a neu-ronal environment Since these mutations affect channel region comprising CaMbinding site, critical for the surface expression of the KV7 channels in neurons(Alaimo et al., 2009; Etxeberria et al., 2008), as well as many sequences for inter-action with other regulatory proteins, or posttranslational modifications (Delmas andBrown, 2005; Haitin and Attali, 2008; Hernandez et al., 2008), the real impact of thetwo C-terminal EE mutations can probably only be assessed in neurons

These functional data suggest that strikingly reduced M-current in the first days

of life not only leads to seizure generation but affects the normal neuromotor opment of affected children Notably, EE mutations seem to affect the critical func-tional parts of the channel and mutants significantly impair the function of the WTsubunits The arising question at this point is which kind of intervention could pos-sibly improve the outcome of the disease One of the interesting drug candidates to