j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

1/13

The mechanism of action of retigabine (ezogabine),

a first-in-class K+ channel opener for the treatment of epilepsy*Martin J. Gunthorpe, yCharles H. Large, and zRaman Sankar

*NewFrontiers SciencePark, GlaxoSmithKline plc, Harlow, Essex, United Kingdom; yMedicines Research Center,

GlaxoSmithKline S.p.A., Verona, Italy; and zDepartments of Pediatricsand Neurology, David Geffen

School of Medicine at UCLA, Los Angeles, California, U.S.A.

SUMMARY

The pharmacologic profile of retigabine [RTG (interna-

tional nonproprietary name); ezogabine, EZG (U.S.

adopted name)], is different from all currently approved

antiepileptic drugs (AEDs). Its primary mechanism ofaction (MoA) as a positive allosteric modulator of

KCNQ25 (Kv7.27.5) ion channels defines RTG/EZG as

the first neuronal potassium (K+) channel opener for the

treatment of epilepsy. KCNQ25 channels are predomi-

nantly expressed in neurons and are important determi-

nants of cellular excitability, as indicated by the

occurrence of human genetic mutations in KCNQ chan-

nels that underlie inheritable disorders including, in the

case of KCNQ2/3, the syndrome of benign familial neona-

tal convulsions. In vitro pharmacologic studies demon-

strate that the most potent action of RTG/EZG is at

KCNQ25 channels, particularly heteromeric KCNQ2/3.

Furthermore, mutagenesis and modeling studies have

pinpointed the RTG/EZG binding site to a hydrophobic

pocket near the channel gate, indicating how RTG/EZG

can stabilize the open form of KCNQ25 channels; the

absence of this site in KCNQ1 also provides a clear expla-

nation for the inbuilt selectivity RTG/EZG has for potas-

sium channels other than the KCNQ cardiac channel.

KCNQ channels are active at the normal cell resting

membrane potential (RMP) and contribute a continual

hyperpolarizing influence that stabilizes cellular excitabil-

ity. The MoAof RTG/EZG increases the number of KCNQ

channels that are open at rest and also primes the cell to

retort with a larger, more rapid, and more prolongedresponse to membrane depolarization or increased neu-

ronal excitability. In this way, RTG/EZG amplifies this nat-

ural inhibitory force in the brain, acting like a brake to

prevent the high levels of neuronal action potential burst

firing (epileptiform activity) that may accompany sus-

tained depolarizations associated with the initiation and

propagation of seizures. This action to restore physio-

logic levels of neuronal activity is thought to underlie the

efficacy of RTG/EZG as an anticonvulsant in a broad

spectrum of preclinical seizure models and in placebo-

controlled trials in patients with partial epilepsy. In this

article, we consider the pharmacologic characteristics of

RTG/EZG at the receptor, cellular, and network levels as

a means of understanding the novel and efficacious MoA

of this new AED as defined in both preclinical and clinical

research.

KEY WORDS:KCNQ, M-current, Antiepileptic drugs,

Anticonvulsant, Antiepileptic.

The introduction of retigabine [RTG (international non-

proprietary name), Marketing Authorization in Europe; ez-

ogabine (EZG; U.S. adopted name)] as a novel antiepilepticdrug (AED) for the adjunctive treatment of partial-onset sei-

zures represents the culmination of >20 years of preclinical

and clinical research. RTG/EZG, first named D-23129 (free

base) or D-20443 (HCl salt version), was developed in the

1980s by the East German company Arzneimittelwerk

Dresden, which became part of the ASTA Medica group

after German reunification. Initial interest in its potential as

a novel treatment for epilepsy stemmed from researchefforts on flupirtine, a congener compound that was

successfully developed as a nonopioid centrally acting

analgesic in Europe (marketed as Katadolon). Flupirtine

demonstrated weak evidence of anticonvulsant efficacy fol-

lowing its submission to the National Institutes of Health

Anticonvulsant Drug Development program in the 1980s

(Rostock et al., 1996) and a small exploratory clinical trial

(Porter et al., 2007). These early efficacy findings were

serendipitous and devoid of an appreciation of the

mechanism of action (MoA) or pharmacologic properties

that delivered them. Rather, similar to the majority of AEDs

available today (Meldrum & Rogawski, 2007), in vivo

Accepted November 10, 2011; Early View publication January5, 2012.Address correspondence to Martin J. Gunthorpe, Neurosciences CEDD,

GlaxoSmithKline, New Frontiers Science Park North, Third Avenue,HarlowCM19 5AW, U.K. E-mail: [email protected]

Wiley Periodicals, Inc. 2012 International LeagueAgainst Epilepsy

Epilepsia, 53(3):412424, 2012

doi: 10.1111/j.1528-1167.2011.03365.x

CRITICAL REVIEW AND INVITED COMMENTARY

412

8/12/2019 j.1528-1167.2011.03365.x

2/13

efficacy from preclinical models was the key factor support-

ing development of both flupirtine and then retigabine as a

novel epilepsy treatment; the understanding of the MoA

would come later. In the case of RTG/EZG, quantitative

structural optimization efforts led to an improved efficacy

[10-fold based on characterization of ED50 (median effec-tive dose) values in the rat maximal electroshock seizure

(MES) test] and pharmaceutical profile relative to flupirtine

that were key considerations in the selection of this agent

for further development as a novel anticonvulsant (Rostock

et al., 1996).

A large body of novel research has now yielded a detailed

appreciation of the pharmacologic MoA of RTG/EZG.

Early efforts demonstrated a lack of potent action at pre-

viously established AED targets such as sodium (Nav) and

calcium channels (Cav) (Fig. 1A). In addition, action on

the c-aminobutyric acid (GABA)ergic system including

GABAA receptors was restricted to relatively high RTG/EZG concentrations. These studies fueled initial sugges-

tions that RTG/EZG possessed a new and unique MoA

(Rundfeldt, 1997, 1999; Rundfeldt & Netzer et al., 2000a).

The first data linking RTG/EZG to the modulation of potas-

sium (K+) channelsa previously untapped superfamily of

ion channels that could conceptually deliver an inhibitory

effect on neurotransmission, and hence provide a new

means to treat epilepsywas published in 1997 (Rundfeldt,

1997). However, the crucial breakthrough was the discovery

and cloning of a new family of K+ channels KCNQ (Kv7

by more recent accepted nomenclature) (Alexander et al.,

2008)that was found to be genetically linked to a form of

inherited human epilepsy known as benign familial neonatalconvulsions (BFNC) (Biervert et al., 1998; Charlier et al.,

1998; Singh et al., 1998). The potent and selective pharma-

cologic action of RTG/EZG as a positive allosteric modula-

tor (opener) of key members of this family, such as

KCNQ2/3, was published by multiple groups shortly there-

after (Main et al., 2000; Rundfeldt & Netzer, 2000b; Wic-

kenden et al., 2000). Furthermore, the fact that these

channels were also demonstrated to underlie the native M-

current, a K+ conductance negatively regulated by musca-

rinic ligands, led to a newfound appreciation for the MoA of

RTG/EZG at the cellular level (Tatulian et al., 2001).

After two decades of research and >100 publications onRTG/EZG, our understanding of RTG/EZG pharmacology

at the receptor, cellular, and network level has advanced sig-

nificantly. We now have a greater understanding of the role

of KCNQ channels in the control of neuronal excitability

(Gribkoff, 2003; Surti & Jan, 2005; Maljevic et al., 2008;

Brown & Passmore, 2009) that provides a rigorous founda-

tion from which to interpret the broad anticonvulsant effi-

cacy of RTG/EZG defined in preclinical seizure models

(Large et al., 2012). Crucially, during this timeframe, these

preclinical findings have also been shown to translate into

significant efficacy for patients in a full development pro-

gram that culminated in two large-scale phase 3 double-

blind placebo-controlled trials in patients with partial

(focal) epilepsy [RESTORE 1 (Study 301) and RESTORE 2

(Study 302) (Brodie et al., 2010; French et al., 2011)].

These data come full circle in defining KCNQ K+ channels

as a fully validated human target for the treatment of this

major disorder which, even with the availability of a largenumber of approved agents (Fig. 1B), still presents a high

unmet medical need. Indeed, due to underlying differences

in disease pathology and the acknowledged limitations

regarding effectiveness or tolerability of previously

approved agents, up to a third of patients are considered

refractory to treatment and a further third are in need of bet-

ter treatment options (Kwan & Brodie, 2000; Kwan et al.,

2010). Such findings have led to a constant call for new

AEDs with novel MoA. In this article, we will consider the

effects of RTG/EZG pharmacology at the receptor, cellular

and network level as a means to understand the efficacy and

utility of this new AED with a novel MoA for the treatmentof epilepsy.

KCNQChannels, a Family ofK+

Channels That ControlCellularExcitability in Humans

We now know that KCNQ channels comprise a family

of five related genes (Fig. 2A). Each gene encodes a

KCNQ subunit with a 6-transmembrane topography and a

pore loop that can assemble to form homotetrameric chan-

nels, and in some cases, heterotetrameric channels. The

key importance of these channels as modulators of cellu-lar excitability in the central and peripheral nervous sys-

tem and other organs in the body where they are

expressed (Supporting Information Table S1) is clear from

the occurrence of a range of human genetic mutations in

KCNQ (Kv7) K+ channels that underlie several inheritable

disorders. These include long QT syndrome that arises

from loss of function of KCNQ1, a predominantly cardiac

ion channel that contributes to the repolarization of the

cardiac action potential (Jentsch, 2000; Shieh et al., 2000)

and a rare form of dominant deafness that is thought to

arise from impaired function of KCNQ4 in the outer hair

cells of the cochlea organ in the inner ear that disruptsnormal K+ balance and excitability (Kharkovets et al.,

2000, 2006). In the case of KCNQ2/3, which are the pre-

dominant KCNQ channels expressed in the brain, a loss

of function due to a range of channel mutations or dele-

tions has been linked to BFNC, a form of epilepsy. The

key role that these channels play in maintaining neuronal

excitability is further highlighted by the fact that rela-

tively moderate (approximately 25%) reductions in KCNQ

channel activity due to impaired expression or function in

either subunit, can lead to a loss of control over neuronal

excitability in the brain resulting in a seizure

predisposition (Maljevic et al., 2008). This finding is in

413

Retigabine (Ezogabine) Mechanism of Action

Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

3/13

A

B

Figure 1.

The unique MoA of retigabine (RTG)/ezogabine (EZG) (A) Schematic representation of a typical neurone in the central nervous sys-

tem highlighting the primary site and mechanism of action (MoA) of currently approved antiepileptic drugs (AEDs) for the treatment

of partial epilepsy. (B) AEDs can be broadly divided into two classes: drugs that inhibit neuronal excitation or increase inhibition.

Briefly, the MoAs associated with these classes are as follows: Sodium (Na+

) channel blockers inhibit neuronal action potential firingand transmission by promoting inactivation and reducing contributions to electrical activity at the axon initial segment (AIS) as well as

on the axon itself. Calcium (Ca2+) channel modulators reduce excitatory transmission by reducing presynaptic neurotransmitter

release, a Ca2+-dependent process. Drugs that bind SV2A may also cause this effect whereas glutamatergic drugs will reduce the

effects of this neurotransmitter on AMPA or NMDA receptors on the postsynaptic membrane. Drugs may affect GABA receptors in a

number of ways: via direct positive allosteric modulation of GABAAreceptor activity (e.g., benzodiazepines), or indirectly, by increas-

ing levels of GABA via inhibition of GABA transaminase (e.g., vigabatrin) or GABA transporter-1 (GAT1, e.g., tiagabine). RTG/EZG is

unique in that it acts as a positive allosteric modulator of KCNQ potassium (K +) channels leading to an inhibition of high-frequency

action potential firing at the AIS due to increased hyperpolarization. KCNQ channels are also present on dendrites and the axon (not

shown), the increased activity of which may also contribute to the anticonvulsant efficacy of RTG/EZG. N.B. Some AEDs also have

multiple pharmacological activities that may contribute to efficacy. In the case of RTG/EZG, positive allosteric modulation of GABA Areceptors may also occur at high concentrations. AMPA, a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; Cl), chloride;

GABA,c-aminobutyric acid; NMDA, N-methyl-D-aspartate; SV2A, synaptic vesicle 2A.

Epilepsia ILAE

414

M. J. Gunthorpe et al.

Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

4/13

concordance with preclinical data in which transgenic

KCNQ2+/)mice, deficient in one copy of theKcnq2gene,

show increased sensitivity to the chemo-convulsive agent

pentylenetetrazole (Watanabe et al., 2000). Consequently,

it has been widely suggested that strategies to increase

activity of KCNQ2/3 represents a novel targeted means to

restore the control of neuronal excitability in patients withepilepsy.

RTG/EZG Is a PositiveAllosteric Modulator(Opener) of KCNQ25

In vitro pharmacological approaches combined with

structural modeling and site-directed mutagenesis studies

have now determined the pharmacological action of RTG/

EZG to be at the KCNQ25 K+ channels. Initial insight into

the novel action of RTG/EZG was derived from studies on

native neuronal K+ channels. Electrophysiological studies

on NG108-15 neuroblastoma, hNT cells (a cell line derived

from human neuronal cells) and mouse cortical neurons in

culture identified a pronounced effect of low concentrations

of RTG/EZG (0.1 lM) on the enhancement of an unidenti-

fied K+ current in these cells (Rundfeldt, 1997). These same

studies also provided first data on the selective and more

potent action of RTG/EZG on K

+

channels than on other tar-gets associated with the pharmacology or MoA of AEDs

(Fig. 1B and see below).

Independently from the elucidation of the precise K+

channel entities targeted by RTG/EZG, other research led to

the cloning and characterization of the KCNQ channels

(Jentsch, 2000). The clear potential for these targets to rep-

resent a novel approach to the treatment of epilepsy cata-

lyzed the formation of an immediate hypothesis regarding

whether the drugs anticonvulsant properties were mediated

via a direct pharmacological action on these channels. Inde-

pendent results from a number of separate groups provided

clear and consistent confirmation that RTG/EZG exhibited

Figure 2.

Pharmacological action and selectivity of retigabine (RTG)/ezogabine (EZG) at KCNQ25 channels. ( A) The KCNQ potassium (K+)

ion channel family consists of five genes with the indicated phylogeny. Each gene encodes a KCNQ subunit with a 6-transmembrane

structure and a pore loop that can assemble to form homo-, and in some cases, heterotetrameric channels. (B) A summary of the

pharmacological potency of RTG/EZG on known homo- and heteromeric KCNQ channels defining its selective action at KCNQ25

(see text for further details). (C) Illustrates the architecture of a KCNQ channel assembled from four subunits (IIV, blue) where two

transmembrane domains (S5, S6) from each are highlighted in magenta and line the aqueous integral ion channel pore that is selective

for K+ ions (adapted with kind permission from Springer Science + Business Media: Meldrum & Rogawski (2007, p. 59), Figure 3).

Detailed modeling and mutagenesis studies have shown that RTG/EZG binds to KCNQ25 channels at this location near the channel

gate leading to a stabilization of the channel open state. Key amino acid residues glycine (G)301 and tryptophan (W)236 for RTG/EZGbinding are indicated. The lack of W236 in KCNQ1 explains the lack of activity of RTG/EZG at this cardiac channel. EC50, half maximal

effective concentration; ECM, extracellular matrix; IC50, half maximal inhibitory concentration.

Epilepsia ILAE

415


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

5/13

potent action as a positive allosteric modulator (i.e., opener

of KCNQ channels) that would serve to reduce neuronal

excitability (Main et al., 2000; Rundfeldt & Netzer, 2000b;

Wickenden et al., 2000; Otto et al., 2002). Detailed

pharmacological studies have demonstrated the drugs

selectivity of action on KCNQ channels (Fig. 2B). RTG/EZG is effectively devoid of activity at KCNQ1, the cardiac

KCNQ channel, and exhibits a higher potency at channels

assembled from KCNQ2/3, which are linked to human epi-

lepsy, over those assembled from KCNQ4 or KCNQ5 (Main

et al., 2000; Rundfeldt & Netzer, 2000b; Wickenden et al.,

2000, 2001; Tatulian et al., 2001; Yeung et al., 2008). One

could therefore cite the overall rank order of potency of

RTG/EZG based on determinations of the half maximal

effective concentration (EC50) as KCNQ3 > KCNQ2/

3 = KCNQ3/5 > KCNQ2 > KCNQ4 = KCNQ5, covering

the concentration range from 0.6 to 6.4 lM(Fig. 2B). How-

ever, in interpreting the relevance of this pharmacology forthe effects in vivo, one must also be cognizant of the fact

that the KCNQ heteromers KCNQ2/3 and KCNQ3/5 are

assembled more efficiently in cells than their homomeric

counterparts and can contribute an order of magnitude more

K+ current than homomeric channels (Lerche et al., 2000;

Main et al., 2000; Wickenden et al., 2000). These hetero-

meric channels will therefore likely predominate amongst

KCNQ channels in delivering the cellular effects and effi-

cacy of RTG/EZG. As a consequence, RTG/EZG is perhaps

best considered as a KCNQ25 active drug that delivers its

predominant anticonvulsant MoA through the heteromeric

KCNQ2/3 and KCNQ3/5 channels that are the main constit-

uents of the neuronal M-current.The in vitro functional characterizations of RTG/EZG

KCNQ pharmacology are now wholly supported by the

results of elegant modeling and structure-function studies

using site-directed mutagenesis. These approaches, based

on the recent advances in our structural knowledge of K+

channel function (Swartz, 2004) have identified the precise

location and key amino acids such as glycine (G)301 and

tryptophan (W)236 contributing to the RTG/EZG binding

site, a hydrophobic pocket near the channel gate (Fig. 2C)

(Schenzer et al., 2005; Wuttke et al., 2005; Lange et al.,

2009). Crucially, these key amino acids involved in RTG/

EZG binding are conserved in KCNQ25 and their locationnear the channel gate provides an explanation for why RTG/

EZG binding can increase KCNQ channel function: namely

a stabilization of the ion channel in the open K+ conducting

form (Schenzer et al., 2005; Wuttke et al., 2005). These

data are also informative regarding the lack of RTG/EZG

opener action at KCNQ1; the lack of key glycine and trypto-

phan residues means that the RTG/EZG binding site is

absent in this family member (Fig. 2C). Rather, a weak

inhibitory effect of RTG/EZG at high concentrations (half

maximal inhibitory concentration [IC50] values approxi-

mately 100lM) is reported (Fig. 2B) that most likely

reflects a direct occlusion of the KCNQ1 channel pore, simi-

lar to the action of chromanol or some benzodiazepines

(Seebohm et al., 2003; Lerche et al., 2007).

RTG/EZG Selectivityof Action:PrimaryMoA Is at KCNQ

Channels

A large number of studies have been conducted to explore

the pharmacology of RTG/EZG, initially as a search for an

underlying MoA to explain its anticonvulsant efficacy, and

subsequently to further elucidate the broader selectivity of

this agent for consideration as a potential medicinal agent.

The results of these studies are summarized below and com-

pared in Table 1, with their likely relevance for therapeutic

effect in epilepsy patients where, at the maximum daily dose

of 1,200 mg (achieved using divided doses of 400 mg three

times daily [t.i.d.]), mean free average plasma concentra-

tions (Cave) of RTG/EZG were 0.83lM and maximummeanfree plasma concentrations (Cmax) were predicted to

be approximately 1 lM(based on a population pharmacoki-

netic analysis of data obtained from the RESTORE trials

(Tompson et al., in preparation). Since RTG/EZG is highly

permeable and not subject to P-glycoproteinactive transport

across the bloodbrain barrier, free plasma concentrations

provide an accurate indication of the free brain concentra-

tions achieved.

Effectson other K+ channels

To ascertain whether the effects of RTG/EZG were selec-

tive for KCNQ channels over other K+ channels, a range of

functional electrophysiological studies have been conductedon exemplar channels from other families that make up the

K+ channel superfamily (Alexander et al., 2008). RTG/

EZG was shown to be without significant activity at mem-

bers of the 2TM K+ family of channels (e.g., inward rectifi-

ers KIR 2.1 [IRK1] or KIR 3.1) as well as 4TM family and

the 2-pore domain channels, such as K2P1.1 (TWIK1) at

concentrations 100 lM. Members of the 6TM family e.g.,

Kv1.5, hERG, which includes voltage-gated [Kv] channels

and EAG and Ca2+ activated channels (in addition to KCNQ

channels), were also relatively unaffected, with IC50values

of >50 lM and 59 lM, respectively (GSK/Valeant data on

file, personal communications). These data emphasize theremarkably selective action of RTG/EZGon a subset of K+

channels, namely KCNQ25 over the broader superfamily

of K+ channels that occur in humans.

Effectson GABA receptors and GABAergic

neurotransmission

The effects of RTG/EZG on GABAAreceptors and other

aspects of GABAergic neurotransmission have been exam-

ined in a range of in vitro studies, initially based on the

exploration of the compounds MoA. Electrophysiological

studies on GABAA-receptor-mediated currents in cultured

cortical neurons identified a potentiating effect (Rundfeldt

416


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

6/13

& Netzer, 2000a; Otto et al., 2002). RTG/EZG can therefore

additionally act as a positive allosteric modulator of the

GABAA receptor; however, higher concentrations of RTG/

EZG than are active at KCNQ channels are typically

required to produce significant potentiation (i.e., 10 lM).

More detailed assessment of the effects of RTG/EZG has

also been undertaken on a range of different GABAArecep-tor subtypes expressed in Xenopus oocytes (GSK/Valeant

data on file, personal communication). The results from

these studies showed that concentrations of RTG/

EZG10 lM were required to cause significant augmenta-

tion of the GABAAreceptor response. The potency of RTG/

EZG differed somewhat depending on the GABAAreceptor

subunit combination, with the following rank order:

a1b3c2 = a1b2c2 > a3b2c2 = a2b2c2 > a5b2c2 = a1b

2(N265S)c2 = a1b1c2 (GSK/Valeant data on file, per-

sonal communication). RTG/EZG can therefore act at the

major GABAAreceptor isoforms in the brain (e.g., a1b3c2

and a1b2c2 together comprise approximately 50% of theCNS receptors (Wafford, 2005). However, in all cases, the

potentiating effects were not inhibited by the benzodiaze-

pine site antagonist flumazenil, indicating that the action

of RTG/EZG is not through the benzodiazepine site on

the receptor, consistent with results from the radioligand

binding studies using a range of GABA receptor ligands

(see below) and published studies (Rundfeldt & Netzer,

2000a).

In a separate line of investigation, Kapetanovic et al.

(1995) uncovered an apparent effect of RTG/EZG on

GABA metabolism whereby levels of newly synthesized

GABA in rat hippocampal slices were increased following

exposure to 20 lMRTG/EZG. This effect was greater than

seen for other AEDs and different to flupirtine, a compound

that is also active at the KCNQ and GABAAchannels in the

range of 1030 lM(Popovici et al., 2008); however, no fur-

ther studies have shed light on the particular mechanism

underlying this effect and in vivo microdialysis studies did

not provide any evidence of increases in GABA occurringin vivo (Rundfeldt C, personal communication October,

2010; see also Straub et al., 2001).

Further studies regarding the MoA of RTG/EZG at the

GABAA receptor have concluded that RTG/EZG interacts

with a novel site on the GABAA receptor complex that is

allosterically coupled with the binding sites for the agonist

GABA and Org 20549 (a neuroactive steroid); EC50values

for the displacement of these GABAAreceptor ligands were

23 lM (van Rijn & Willems-van Bree, 2003; van Rijn &

Willems-van Bree, 2004). Studies by Otto et al. (2002) have

also demonstrated that RTG/EZG can potentiate inhibitory

postsynaptic currents mediated by direct activation ofGABAAreceptors in mouse cortical neurons. Again, signifi-

cant effects on peak current or charge transfer required

concentrations of RTG/EZG 10 lM, and were markedly

larger at 50lM indicating that this was the start of the

concentration-response curve (Otto et al., 2002). Overall,

these data demonstrate that RTG/EZG can modulate

GABAergic neurotransmission. Although in vitro concen-

trations do not necessarily translate directly to in vivo con-

centrations, the concentrations required to drive GABA

effects are thought to be higher than those that are effective

at KCNQ channels or attained in patients with epilepsy

(Table 1).

Table 1. KCNQ channels: the primary site for RTG/EZG MoA

Pharmacological action Effect

Level of activity or EC50 orIC50where determineda

Ratio of activity : FreeCmaxor Cave at 1,200 mg/day in patients

with epilepsyb

KCNQ Positive allosteric modulator EC50 = 1.6 lM at KCNQ2/3 1GABA Positive a llosteric m odulator a t GABAA

receptors (nonbenzodiazepinesite)

Significant effects at 10lM inthe

majority of studies

10-fold

Effects on GABAmetabolism Significant effectsat20lM 20-fold

Calcium channels Weak inhibitor IC50> 100 lMat neuronal Cav channels

(29%inhibition at 100lM)

>100-fold

Sodium channels Weak inhibitor IC50> 100 lMat neuronal Nav channels

(25% at 100lM)

>100-fold

Glutamate receptors Noeffectat NMDA, AMPA,

or kainate receptors

Noeffectup to10 lM >10-fold

Other: Broad selectivity

profile

Noadditionalactivities detected Nosignificant interactionsin 62assays

of ion channels, transporters, enzymes,

and 2nd-messenger systems at 10lMc

>10-fold

AMPA,a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; Cave, average plasma concentration; Cmax, maximum plasma concentration at steady state; Cmin,minimum plasma concentration at steady state; EC50, half maximal effective concentration; GABA, c-aminobutyric acid; IC50, half maximal inhibitory concentration;

NMDA,N-methyl-D-aspartate.aSee text and the following references for further information: (Kapetanovic et al., 1995; Otto et al., 2002; Rundfeldt & Netzer, 2000a; Rundfeldt & Netzer,

2000b; Tatulian et al., 2001; van Rijn & Willems-van Bree, 2003; van Rijn & Willems-van Bree, 2004).bEstimated as 1.0 lMbased on mean Cmaxat 1,200 mg = 1,520 ng/mL from population pharmacokinetic analysis of the RESTORE trial data. Mean C aveconcen-

trations at this dose were 1,261 ng/mLor 0.83lM (GSK/Valeant data on file; Plasma protein binding = 7980%; RTG/EZG molecular weight = 303.3 Da).cSee Table S2 for supporting information.

417


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

7/13

Effects on Na+ and Ca2+ channels

The effects of RTG/EZG on Nav channels has been

assessed using whole-cell patch clamp electrophysiology on

differentiated NG108-15 neuroblastoma cells. Only weak

inhibition was detected with 9.1 3.4%, 20.1 2.9%, and

24.9 5.9% reduction in Nav current measured followingincubation with either 1, 10, or 100 lMRTG/EZG, respec-

tively. Similar studies on mixed Cavcurrents studied in dif-

ferentiated NG108-15 neuroblastoma cells also detected

weak activity of RTG/EZG, with significant effects only

observed at concentrations above 10 lM. The greatest inhi-

bition of mixed N-, T- and L-type currents was determined

to be 15.7 3.0%, 47.3 10%, and 60.1 7.3% reduction

following incubation with 10, 30, and 100 lM RTG/EZG,

respectively (GSK/Valeant data on file, personal communi-

cation). In conclusion, RTG/EZG exhibited only weak

inhibitory effects at voltage-gated Navand Cavchannel cur-

rents at predominantly supratherapeutic concentrations.

Effects on glutamate receptors

The potential effects of RTG/EZG on glutamatergic

neurotransmission were evaluated in a range of functional

electrophysiological studies on recombinant receptors

expressed inXenopus oocytes or HEK293 cells and native

receptors in cortical neurones in culture. No effect of RTG/

EZG was observed on N-methyl-D-aspartate (NMDA)-

induced currents in cortical neurons at 10 lM. RTG/EZG

was also inactive versus currents induced by 300 lM kainate

at a concentration of 10 lM, but did reduce currents by 17%

at 100 lM, suggesting a small inhibitory effect at high con-

centrations. Although a significant block of kainate(10 lM)-induced currents was observed in these studies,

more detailed follow-up investigations examining the effect

of RTG/EZG on different cloned alpha-amino-3-hydroxy-

5-methyl-4-isoxazolepropionic acid (AMPA) and kainate

receptors, combined with additional studies on kainate-

induced currents in neurons did not detect any significant

inhibitory effects of RTG/EZG (GSK/Valeant data on file,

personal communication). These data, in conjunction with

the radioligand binding assay data summarized below, sup-

port the conclusion that RTG/EZG does not interact signifi-

cantly with glutamate receptors.

Cross-screening selectivity profile: radioligandbinding

assaysand additional electrophysiological studies

Radioligand binding studies with RTG/EZG at concen-

trations of 10)9, 10)7 or 10)5 M conducted at NovaScreen

Biosciences (Hanover, MD, U.S.A.) according to standard

methods (GSK/Valeant data on file, personal communica-

tion) did not show significant interaction with the known

modulator binding sites on 62 receptors, ion channels, trans-

porters, enzymes, and second messengers (percent inhibi-

tion

8/12/2019 j.1528-1167.2011.03365.x

8/13

Quantification of this effect (the EC50to shift this curve to

50% of its maximum [V1/2] is 1.6 lM; Fig. 3A[iv]) can be

used to deduce the potent action of RTG/EZG that explains

why more KCNQ channels are open at a given membrane

potential in the presence of the drug, particularly following

membrane depolarization (e.g., compare the magnitude ofresponse obtained at )60 mV, close to the normal cell

RMP, or at a more depolarized level such as )30 mV that

may occur following neuronal activity, in the presence or

absence of RTG/EZG (Fig. 3A[iii]). This effect translates to

larger K+ outward currents and an increased hyperpolarizing

effect on the cell (Main et al., 2000; Rundfeldt & Netzer,

2000b; Wickenden et al., 2000; Tatulian et al., 2001). It is

worth noting that the increase in K+ current is achievedsolely by this means of increasing the probability of a

KCNQ channel being open, as RTG/EZG does not alter the

Figure 3.

Retigabine (RTG)/ezogabine (EZG) enhances magnitude and duration of KCNQ channel action to resist depolarization, and shifts

voltage-dependence (VD) of KCNQ activation, leading to more rapid, prolonged, and increased levels of KCNQ channel opening in

response to depolarizing stimuli. (A) (i) Whole-cell potassium currents evoked by voltage steps between )100 and +30 mV in

KCNQ2/3-expressing cells are significantly larger in the presence of (ii) 10 lMRTG/EZG and these also exhibit faster kinetics of acti-vation and slowed deactivation (blue arrows; see also B). (iii) Increasing concentrations of RTG/EZG from 0.1 to 10 lM(iv) reversibly

shifted VD of KCNQ2/3 activation by approximately 30 mV to more hyperpolarized potentials, with an EC50of 1.6 lM. (Reproduced

with permission from the American Society for Pharmacology and Experimental Therapeutics, Wickenden et al. (2000), p. 597) ( B)

Expanded portions of KCNQ current responses clearly highlight the effect of RTG/EZG on the rate of channel activation/deactiva-

tion. (Reproduced with permission from the American Society for Pharmacology and Experimental Therapeutics, Main et al. (2000))

(C) RTG/EZG action on KCNQ channel gating: 10 lMincreased the rate of KCNQ activation (sact) and reduced the rate and contri-

bution of KCNQ slow deactivation (sslow). When combined with VD changes, RTG/EZG leads to an overall increase in KCNQ chan-

nel open probability (Popen) i.e., more KCNQ channels will be open at a given voltage in the presence of RTG/EZG. ( D) This effect

underlies the pronounced hyperpolarizing effect of RTG/EZG on the cell resting membrane potential (RMP). In the example shown,

1 lMRTG/EZG was able to reduce RMP from approximately)63 to)70 mV, which would dramatically reduce excitability (Repro-

duced with permission from the American Society for Pharmacology and Experimental Therapeutics, Main et al. (2000, p. 259)).

Epilepsia ILAE

419


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

9/13

single-channel conductance of individual KCNQ2/3 chan-

nels (Tatulian & Brown, 2003).

More careful examination of the KCNQ current traces in

response to the step-depolarizations illustrated also indi-

cates that RTG/EZG alters the rate of KCNQ activation and

deactivation (blue arrows in Fig. 3A[ii]). This effect hasbeen studied in detail and is clear when viewed on an

expanded timeframe (Fig. 3B) (Main et al., 2000; Wicken-

den et al., 2000; Tatulian et al., 2001). These studies have

demonstrated that RTG/EZG increases the rate of KCNQ

activation (sact) and reduces the rate of the slower compo-

nent of a two-step deactivation process (sslow) (Fig. 3C).

These kinetic effects yield more KCNQ channels in the

open rather than closed positions at a given membrane

potential underlying the marked increase in open-channel

probability (Popen) and indicate that RTG/EZG also

enhances the timeframe over which KCNQ channels impact

cellular excitability.

The hyperpolarizing influence of RTG/EZG stabilizes

the RMPof neurons

The effects of RTG/EZG on KCNQ channels are consis-

tent with studies of the agent on native M-currents in neurons

(Tatulian et al., 2001; Yue & Yaari, 2004), providing confi-

dence in the translation of the measured in vitro pharmacol-

ogy at recombinant receptors to native systems. Studies on

neurons indicate that KCNQ K+ channels are active at the

cell RMP and, unlike the majority of voltage-gated channels,

do not inactivate such that they contribute a continual hyper-

polarizing influence on cells (Tatulian et al., 2001; Otto

et al., 2002; Yue & Yaari, 2004). This is evident from studieswith the KCNQ antagonists XE-991 or linopirdine that typi-

cally cause a depolarization of the RMP and increase in neu-

ronal excitability and action potential firing. The ability of

RTG/EZG to achieve the converse through recruitment of

additional KCNQ channels resulting in a concentration-

dependent and marked hyperpolarizing effect on cell RMP

has also been revealed by similar studies (Fig. 3D).

RTG/EZG reduces spike-frequency adaptation,

preferentially impacting high-frequency firing

Studies on the firing properties of rodent cortical and hip-

pocampal neurons indicate an important role of M-currentsin setting the electroresponsive properties of neurons and

controlling their intrinsic firing frequencies and patterns in

response to depolarization (Hetka et al., 1999; Otto et al.,

2002; Yue & Yaari, 2004). Such studies have highlighted

that in addition to enhancement of the neuronal M-current

promoting maintenance of the RMP, RTG/EZG can also

reduce subthreshold excitability (i.e., it influences the range

of underlying electrical activity in the cell that, if suffi-

ciently depolarized, can reach the threshold to trigger the

firing of one or more action potentials). It appears that

KCNQ channels are particularly important in this regard

since they are highly expressed at the axon initial segment

(AIS; see Fig. 1A) (Devaux et al., 2004; Chung et al., 2006;

Pan et al., 2006), the part of the neuron that integrates the

net depolarizing and hyperpolarizing influences on the cell

and from which action potentials originate; KCNQ channels

are also present on dendrites and the axon (Devaux et al.,

2004; Chung et al., 2006; Pan et al., 2006) meaning thatthey can also impact dendritic integration and neuronal

transmissioneffects that could also meaningfully contrib-

ute to the anticonvulsant efficacy of RTG/EZG.

In studies on rat CA1 pyramidal (hippocampal) neuro-

nes, RTG/EZG showed pronounced concentration-depen-

dent (110 lM) attenuation of higher frequency or burst

firing evoked by prolonged depolarizing stimuli (Fig. 4A)

(Yue & Yaari, 2004). In contrast, single action potentials

recorded in response to short duration depolarizing stimuli

were unaffected by the same concentrations of RTG/EZG.

This sparing effect on lower frequency activity by RTG/

EZG is in line with the primary contribution of KCNQchannels to the medium afterhyperpolarization (mAHP),

an observed conductance event that serves to control the

excitability of neurones over a timescale in the order of

tens to hundreds of ms (Storm, 1989; Gu et al., 2005).

Such an effect reflects the slow kinetics of KCNQ activa-

tion that, even in the agents presence, are too slow to

inhibit the onset of a solitary action potential, but are well

placed to impact subsequent excitability. Therefore, the

enhanced KCNQ recruitment following the initial depolar-

ization of an action potential in the presence of RTG/EZG

does lead to a reduced spike afterdepolarization that is

responsible for the attenuation of subsequent action poten-

tials leading to the phenomenon of spike frequency adap-tation. Such differential behavior is likely to underlie the

tolerability or therapeutic index achieved with RTG/EZG

preclinically or in patients with epilepsy since, from a ris-

ing dose or concentration-dependence perspective, the

higher levels of neuronal activity or burst firing that are

likely to accompany seizures will be inhibited prior to

more physiological levels of firing.

RTG/EZG attenuates epileptiform activity in rat and

human brain tissue

RTG/EZG has been shown to attenuate epileptiform

activity recorded ex vivo in a number of brain slicepreparations including rat hippocampal slices treated

with proconvulsant agents such as the K+ channel blocker

4-aminopiridine (4-AP), bicuculline, low Mg2+, or NMDA

(Armand et al., 1999, 2000; Dost & Rundfeldt, 2000).

RTG/EZG also reduced in vitro spontaneous bursting in

the entorhinal cortex of rats that had been treated with kai-

nate to induce status epilepticus (Smith et al., 2007) and

attenuated recurrent epileptiform discharges evoked by 4-

AP in combination with bicuculline (Armand et al., 1999).

Although no clinical data are available, these findings sug-

gest the potential of K+ channel openersfor the treatment

of status epilepticus, in particular drug-resistant status

420


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

10/13

(Large et al., 2012 for additional comment). Importantly,

from a human translational perspective, similar studies

have also been carried out in human brain tissue. Straub

et al. (2001) demonstrated antiepileptic effects of RTG/

EZG in neocortical slice preparations from 17 patients

who underwent surgery for the treatment of intractable

(i.e., pharmacoresistant) epilepsy. RTG/EZG decreased the

occurrence of spontaneous rhythmic sharp waves that were

recorded in these brain slices ex vivo (Fig. 4B) and also

suppressed epileptiform field potentials evoked by use of

low Mg2+ (Straub et al., 2001). These data provide evi-

dence for the presence of functional KCNQ channels in

these brain regions, hence supporting the potential for

RTG/EZG to provide therapeutic benefit for patients suffer-

ing from such intractable focal seizures. In addition, they

emphasize that RTG/EZG can deliver sufficient KCNQ-

mediated inhibition to effectively restore physiological

levels of neuronal excitability at the network level, an

B

A

Figure 4.

Retigabine (RTG)/ezogabine (EZG) action on native KCNQ channels can reduce hyperexcitability, high frequency action potential fir-

ing and epileptiform activity (EA) in the brain. (A) RTG showed pronounced attenuation of higher frequency or burst firing in studies

on rat CA1 pyramidal (hippocampal) neurones. Short-duration depolarizing stimuli (a1a4) evoked single action potentials that were

unaffected by 110 lMRTG/EZG. In contrast, prolonged depolarizing stimuli evoked burst firing of neurones and this activity was

preferentially reduced by RTG/EZG in a concentration-dependent manner, due to the kinetics of KCNQ activation that are too

sloweven in the presence of RTG/EZGto inhibit solitary action potentials. However, enhanced KCNQ recruitment by the initial

depolarization in the presence of RTG/EZG leads to a reduced afterdepolarization that is responsible for the attenuation of sub-

sequent action potentials (spike frequency adaptation). (Adapted with permission from the Society for Neuroscience, Yue & Yaari

(2004)) (B) RTG/EZG displayed inhibitory properties on EA recorded in human cortical brain tissue resected from patients undergo-

ing surgery for intractable partial epilepsy. The recordings show spontaneous sharp wave activity ex vivo (note the higher temporal

resolution of the upper trace that shows individual events occurring in the longer time-base traces below) that were sensitive to

RTG/EZG, highlighting the presence of functional KCNQ channels in these brain regions and hence the potential for RTG/EZG to

provide therapeutic benefit for patients suffering from these intractable focal seizures (Adapted from Straub et al. (2001), 2001,

with permission from Elsevier BV).Epilepsia ILAE

421


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

11/13

effect that probably underlies its anticonvulsant activity in

a broad range of seizure types in the intact brain (Rostock

et al., 1996; Tober et al., 1996; Mazarati et al., 2008) (see

Large et al., 2012 for review).

Discussion

In this review we have considered the novel MoA of

RTG/EZG, originally discovered as a compound with anti-

convulsant activity but unknown MoA, and now defined as

a selective positive allosteric modulator (opener) of

KCNQ25 channels. Significant research efforts have con-

tributed to the two main aspects of this pairing: (1) the con-

tinued development of RTG/EZG as a potential treatment

for epilepsy through its clinical development over the last

decade and (2) the fundamental research into the physiolog-

ical role of KCNQ channels and their potential contribution

to diseases such as epilepsy. With the completion of pivotalclinical studies defining the dose-dependent efficacy and

tolerability of RTG/EZG in patients with partial epilepsy,

both of these avenues of research validate the KCNQ chan-

nels as a useful therapeutic target for the treatment of epi-

lepsy (Fig. 5).

At the receptor, cellular and network level, the importance

of KCNQ channels as key regulators of neuronal excitability

is clear: they are able to work as a powerful inhibitory force

in the brain by providing a continual hyperpolarizing influ-

ence to maintain or control the cell RMP and reduce sub-

threshold excitability. This key function provides a clear

underlying explanation for the MoA of RTG/EZG that stabi-

lizes KCNQ25 channels in the open position, increasingthe recruitment of such channels at rest, and particularly fol-

lowing depolarization, that act to exert a hyperpolarizing

effect on the cell. RTG/EZG therefore effectively primes the

cell to resist firing bursts of action potentials that occur dur-

ing the sustained depolarizations associated with the initia-

tion and generalization of seizures. These mechanistic

effects are consistent with the anticonvulsant properties of

RTG/EZG demonstrated in in-vitro and in-vivo models of

epilepsy and in patients with epilepsy (Fig. 5) (see Large

et al., 2012).

The pharmacological profile of RTG/EZG is therefore

different from all currently approved AEDs and may offeran additional treatment option for patients. Initial clinical

trials were focused on the adjunctive use of RTG/EZG in

patients with partial epilepsy; however, based on the broad-

spectrum antiepileptic potential for this agent defined in a

wide range of epilepsy models of partial (focal), general-

ized, idiopathic, and refractory epilepsy (see Large et al.,

2012), further clinical investigation is warranted to explore

the utility of RTG/EZG for use in other seizure types.

A further avenue for investigation will be the use of RTG/

EZG in combination with other AEDs, as is common in

clinical practice. An agent with a fundamentally different

MoA such as RTG/EZG may offer good potential for use in

combination with AEDs that work in other ways (Fig. 1). In

this regard, combination studies conducted in animal mod-

els are encouraging, demonstrating additive or even syner-

gistic activity with commonly used AEDs such as valproateand lamotrigine (Luszczki et al., 2009). Further preclinical

and clinical studies are now warranted to address this spe-

cific aspect of AED therapy, and to understand which AEDs

may best complement RTG/EZG to maximize efficacy and

reduce the potential for side-effects in patients.

Acknowledgments

The authors thank David Gibson PhD, CMPP of Caudex Medical, NewYork, NY, USA (supported by GlaxoSmithKline and Valeant PharmaceuticalsInternational) for providing editorial assistance in the preparation of themanuscript, Chris Crean (Valeant Pharmaceuticals International) for expert

Figure 5.

Concentrations of retigabine (RTG)/ezogabine (EZG) associ-

ated with anticonvulsant efficacy in patients with epilepsy and

the pharmacological effects measured at KCNQ2/3 in vitro.

Bar height represents the median reduction in seizure fre-

quency (right hand y-axis) endpoint achieved in the two Phase 3

RESTORE trials of RTG/EZG administered t.i.d. (Brodie et al.,

2010; French et al., 2011) that assessed 600, 900, or 1,200 mg/

day (similar data were achieved with the alternative endpoint

assessing the proportion of patients achieving 50% reductions

in seizures). Bar width indicates the range of concentrations

based on the range of mean free Cmin to mean Cmax for each

dose calculated by population pharmacokinetic analysis. The

mean Cave values achieved at each dose were 0.40, 0.60, and

0.83 lMat 600, 900, and 1,200 mg, respectively, and are indi-

cated by the white dotted line superimposed on each bar. Thein vitro RTG/EZG concentrationresponse profile (black

squares and associated curve fit) for the shift in V1/2of KCNQ

channel activation (left hand y-axis) is also overlaid for direct

comparison. Cave, average plasma concentration; Cmax, maxi-

mum plasma concentration at steady state; Cmin, minimum

plasma concentration at steady state.

Epilepsia ILAE

422


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

12/13

input regarding the clinical pharmacokinetic data cited, and Steve WhitePhD (University of Utah, Salt Lake City, UT, USA) for initial discussionsregarding content and a review of the manuscript.

Disclosure

We confirm that we have read the Journal s position on issues involvedin ethical publication and affirm that this article is consistent with those

guidelines. Martin J. Gunthorpe and Charles H. Large are former employ-ees of GlaxoSmithKline. Raman Sankarhas received basic research supportfrom Valeant Pharmaceuticals and NTP, has participated in clinical trialssponsored by Pfizer, and has served as a speaker and/or paid consultant forGSK, UCB, Lundbeck, NTP, and Sunovion.

References

Alexander SP, Mathie A, Peters JA. (2008) Guide to Receptors and Chan-nels (GRAC), 3rd edition.Br J Pharmacol 153(Suppl. 2):S1S209.

Armand V, Rundfeldt C, Heinemann U. (1999) Effects of retigabine(D-23129) on different patterns of epileptiform activity induced by 4-

aminopyridine in rat entorhinal cortex hippocampal slices. NaunynSchmiedebergs Arch Pharmacol 359:3339.Armand V, Rundfeldt C, Heinemann U. (2000) Effects of retigabine

(D-23129) on different patterns of epileptiform activity induced by lowmagnesium in rat entorhinal cortex hippocampal slices. Epilepsia41:2833.

Biervert C, Schroeder BC, Kubisch C, Berkovic SF, Propping P, JentschTJ, Steinlein OK. (1998) A potassium channel mutation in neonatalhuman epilepsy. Science 279:403406.

Brodie MJ, Lerche H, Gil-Nagel A, Elger C, Hall S, Shin P, Nohria V,Mansbach H; RESTORE 2 Study Group. (2010) Efficacy and safety of

ezogabine (retigabine) in refractory partial epilepsy. Neurology

75:18171824.Brown DA, Passmore GM. (2009) Neural KCNQ (Kv7) channels. Br J

Pharmacol 156:11851195.Charlier C, Singh NA, Ryan SG, Lewis TB, Reus BE, Leach RJ, Leppert

M. (1998) A poremutation in a novel KQT-like potassium channelgene

in an idiopathic epilepsy family.Nat Genet18:5355.Chung HJ, Jan YN, Jan LY. (2006) Polarized axonal surface expression of

neuronal KCNQ channels is mediated by multiple signals in theKCNQ2 and KCNQ3 C-terminal domains. Proc Natl Acad Sci USA

103:88708875.Devaux JJ, Kleopa KA, Cooper EC, Scherer SS. (2004) KCNQ2 is a nodal

K+ channel.J Neurosci 24:12361244.Dost R, Rundfeldt C. (2000) The anticonvulsant retigabine potently sup-

presses epileptiform discharges in the low Ca ++ and low Mg++ modelin the hippocampal slice preparation. Epilepsy Res 38:5366.

French JA, Abou-Khalil BW, Leroy RF, Yacubian EM, Shin P, Hall S,Mansbach H, Nohria V; On behalf of the RESTORE 1/Study 301 Inves-tigators. (2011) Randomized, double-blind, placebo-controlled trial of

ezogabine (retigabine) in partial epilepsy.Neurology 76:15551563.Gribkoff VK. (2003) The therapeutic potential of neuronal KCNQ channel

modulators. Expert Opin Ther Targets 7:737748.

Gu N, Vervaeke K, Hu H, Storm JF. (2005) Kv7/KCNQ/M and HCN/h, butnot KCa2/SK channels, contribute to the somatic medium after-hyper-polarization and excitability control in CA1 hippocampal pyramidalcells.J Physiol 566:689715.

Hetka R, Rundfeldt C, Heinemann U, Schmitz D. (1999) Retigabinestrongly reducesrepetitive firingin rat entorhinal cortex. Eur J Pharma-col 386:165171.

Jentsch TJ. (2000) Neuronal KCNQ potassium channels: physiology androle in disease.Nat Rev Neurosci 1:2130.

Kapetanovic IM, Yonekawa WD, Kupferberg HJ. (1995) The effects of

D-23129, a new experimental anticonvulsant drug, on neurotransmitteramino acids in the rat hippocampus in vitro. Epilepsy Res 22:167173.

Kharkovets T, Hardelin JP, Safieddine S, Schweizer M, El-Amraoui A,Petit C, Jentsch TJ. (2000) KCNQ4, a K+ channel mutated in a form ofdominant deafness, is expressed in the inner ear and the central auditorypathway. Proc Natl Acad SciUSA97:43334338.

Kharkovets T, Dedek K, Maier H, Schweizer M, Khimich D, Nouvian R,Vardanyan V, Leuwer R, Moser T, Jentsch TJ. (2006) Mice with alteredKCNQ4 K+ channels implicate sensory outer hair cells in human pro-gressive deafness. EMBOJ25:642652.

Kwan P, Brodie MJ. (2000) Early identification of refractory epilepsy. NEnglJ Med342:314319.

Kwan P, Arzimanoglou A, Berg AT, Brodie MJ, Allen HW, Mathern G,

Moshe SL, Perucca E, Wiebe S, French J. (2010) Definition of drugresistant epilepsy: consensus proposal by the ad hoc Task Force of theILAE Commission on Therapeutic Strategies. Epilepsia 51:10691077.

Lange W, Geissendorfer J, Schenzer A, Grotzinger J, SeebohmG, FriedrichT, Schwake M. (2009) Refinement of the binding site and mode ofaction of the anticonvulsant Retigabine on KCNQ K+ channels. Mo lPharmacol 75:272280.

Large CH, Sokal DM, Nehlig A, Gunthorpe MJ, Sankar R, Crean CS,VanLandingham KE, White S. (2012) The spectrum of anticonvulsantefficacy of retigabine (ezogabine) in animal models: Implications forclinical use. Epilepsia 53:425436.

Lerche C, Scherer CR, Seebohm G, Derst C, Wei AD, Busch AE, Stein-meyer K. (2000) Molecular cloning and functional expression ofKCNQ5, a potassium channel subunit that may contribute to neuronalM-current diversity.J Biol Chem 275:2239522400.

Lerche C, Bruhova I, Lerche H, Steinmeyer K, Wei AD, Strutz-Seebohm

N, Lang F, Busch AE, Zhorov BS, Seebohm G. (2007) Chromanol293B binding in KCNQ1 (Kv7.1) channels involves electrostatic inter-actions with a potassium ion in the selectivity filter. Mol Pharmacol

71:15031511.Luszczki JJ, Wu JZ, Raszewski G, Czuczwar SJ. (2009) Isobolographic

characterization of interactions of retigabine with carbamazepine,lamotrigine, and valproate in the mouse maximal electroshock-inducedseizuremodel.Naunyn Schmiedebergs Arch Pharmacol 379:163179.

Main MJ, Cryan JE, Dupere JR, Cox B, Clare JJ, Burbidge SA. (2000)Modulation of KCNQ2/3 potassium channels by the novel anticonvul-sant retigabine.Mol Pharmacol 58:253262.

Maljevic S, Wuttke TV, Lerche H. (2008) Nervous system KV7 disorders:breakdown of a subthreshold brake.J Physiol 586:17911801.

Mazarati A, Wu J, Shin D, Kwon YS, Sankar R. (2008) Antiepileptogenicand antiictogenic effects of retigabine under conditions of rapid kind-ling: an ontogenic study. Epilepsia 49:17771786.

Meldrum BS, Rogawski MA. (2007) Molecular targets for antiepileptic

drug development.Neurotherapeutics 4:1861.Otto JF, Kimball MM,Wilcox KS.(2002)Effects of the anticonvulsant reti-

gabine on cultured cortical neurons: changes in electroresponsive prop-erties and synaptic transmission.Mol Pharmacol 61:921927.

Pan Z, Kao T, Horvath Z, Lemos J, Sul JY, Cranstoun SD, Bennett V,Scherer SS, Cooper EC. (2006) A common ankyrin-G-based mecha-nism retains KCNQ and NaV channels at electrically active domains ofthe axon.J Neurosci 26:25992613.

Popovici F, Dorostkar M, Boehm S. (2008) The non-opioid analgesic flu-pirtine is a modulator of GABAA receptors involved in pain sensation.

BMC Pharmacol 8(Suppl. 1):A14.Porter RJ, Nohria V, Rundfeldt C. (2007) Retigabine. Neurotherapeutics

4:149154.Rostock A, Tober C, Rundfeldt C, Bartsch R, Engel J, Polymeropoulos EE,

Kutscher B, Loscher W, Honack D, White HS, Wolf HH. (1996) D-23129: a new anticonvulsant with a broad spectrum activity in animal

models of epileptic seizures. Epilepsy Res 23:211223.Rundfeldt C. (1997) The new anticonvulsant retigabine (D-23129) actsas an opener of K+ channels in neuronal cells. Eur J Pharmacol336:243249.

Rundfeldt C. (1999) Characterization of the K+ channel opening effect ofthe anticonvulsant retigabine in PC12 cells. Epilepsy Res 35:99107.

Rundfeldt C, Netzer R. (2000a) Investigations into the mechanism of actionof the new anticonvulsant retigabine. Interaction with GABAergic andglutamatergic neurotransmission and with voltage gated ion channels.

Arzneimittelforschung 50:10631070.Rundfeldt C, Netzer R. (2000b) The novel anticonvulsant retigabine acti-

vates M-currents in Chinese hamster ovary-cells tranfected with humanKCNQ2/3 subunits.Neurosci Lett282:7376.

Sankar R, Wu J, Shin D, Mazarati A. (2008) Retigabine inhibits rapidkindling epileptogenesis in rats through the activation of Kv7.2/7.3channels. 8th European Congress on Epileptology, Berlin, Germany,September 2125, 2008. Abstract T153.

423


Epilepsia, 53(3):412424, 2012doi: 10.1111/j.1528-1167.2011.03365.x

8/12/2019 j.1528-1167.2011.03365.x

13/13

Documents

j.1528-1167.2011.03365.x