ion channel drug discovery high throughput assay ... · the radiolabel ligand-binding assay was...

ASSAY and Drug Development TechnologiesVolume 2 Number 5 2004copy Mary Ann Liebert Inc

Technology Review

High Throughput Assay Technologies for Ion Channel Drug Discovery

Wei Zheng1 Robert H Spencer2 and Laszlo Kiss2

Abstract Ion channels represent a class of membrane spanning protein pores that mediate the fluxof ions in a variety of cell types To date 400 ion channels have been cloned and characterizedand some of these channels have emerged as attractive drug targets Several existing medicationselicit their therapeutic effect through the modulation of ion channels underscoring the importance ofion channels as a target class for modern drug discovery To meet the increasing demand for high-throughput screening of ion channels assay technologies have evolved rapidly over the past 5ndash10years In this article the authors review the technologies that are currently used for the screening ofion channels The technologies discussed are binding assays ion flux assays fluorescence-basedassays and automated patch-clamp instrumentation

543

Introduction

ION CHANNELS ARE membrane-spanning proteins thatform pores through which inorganic ions such as Na

K Ca2 and Cl can rapidly traverse the cell mem-brane down their electrochemical gradient (Table 1) Itis well established that ion channels play a vital role inneuronal signal transduction neurotransmitter releasemuscle contraction cell secretion enzyme activation andgene transcription To date upwards of 400 different hu-man ion channel genes have been identified Mutationsthat disrupt or alter channel function have been associ-ated with many diseases (ldquochannelopathiesrdquo) includinghypertension cardiac arrhythmia diabetes cystic fibro-sis and a variety of neuronal disorders1ndash5 Thus ion chan-nels represent an important class of molecular targets fordrug development

The role of ion channels in drug safety has alsoemerged as an important issue in the last several yearsSince 1985 five drugs have been withdrawn from themarket because they prolong the cardiac QT interval and

in several cases produce a lethal ventricular arrhythmiaknown as torsade de pointes6ndash8 The mechanism under-lying this toxic effect involves inhibition of one or moreof the cardiac ion channels (1) the hERG potassiumchannel (Ikr) (2) the KCNQ1KCNE1 potassium chan-nel (Iks) and (3) the SCN5A sodium channel910

Ion channels can be broadly grouped into two majorclasses1112 ligand-gated and voltage-gated ion channels(Table 2) The primary distinguishing feature is that volt-age-gated ion channels do not have endogenous li-gands but are activated by changes in the membrane potential Ion channels can exist in multiple states suchas the closed open and inactivated states Voltage-gatedion channels transition (gate) between these states in re-sponse to changes in membrane potential In contrast li-gand-gated channels transition between these states inresponse to the binding and unbinding of a ligand In theopen state ions can flow through a single ion channelpore at prodigious rates of over 107 ionss Cell-basedfunctional assays are an essential requirement for thescreening of ion channels at both the primary and sec-

Departments of 1Automated Biotechnology and 2Molecular Neurology Merck Research Laboratories West Point PA

ABBREVIATIONS AAS atomic absorbance spectrometry CCD charge-coupled device FLIPR fluorescent-imaging plate reader FRETfluorescence resonance energy transfer hERG human ether-go-go-related gene HTS high throughput screening IC50 50 inhibitory concen-tration SPA scintillation proximity assay

bullbull

5315_16_p543-552 11504 1235 PM Page 543

ondary levels Traditional methods developed for HTSof ion channels such as binding ion flux and fluores-cent probes measure ion channel activity indirectlyPatch-clamp electrophysiology is regarded as the ldquogoldstandardrdquo for measuring ion channel activity and phar-macology Patch-clamp allows for the direct real-timemeasurement of ion channel activity but in its traditionalformat is low-throughput and requires a high degree ofoperator skill Hence drug screening assays for ion chan-nels in comparison to those for enzyme and receptor tar-gets have compromised data quality for throughput Re-cently a number of new screening technologies have beendeveloped and improved for ion channel assays and arepoised to change this13ndash15 A summary of the currentlyavailable screening technologies is listed in Table 3

Demands on Ion Channel Screening Assays

The current demands for ion channel screening in thedrug discovery process can be grouped into three mainareas primary screening assays secondary screening as-says and ion channel safety assessment

Primary screening assays (HTS)

HTS has seen a tremendous advance during the last10 years and remains the first critical step in the dis-covery of lead chemical structures for novel drug tar-gets To increase the probability of success for findingnew leads from HTS many companies have investedheavily in expanding both the diversity and quality oftheir compound libraries For most mid- and large-sizedcompanies the library collection has grown to 400000ndash1000000 (or more) compounds The standard para-digms used to screen these libraries have evolved to au-tomated 384-well or higher-density single-compoundtest formats Minimal throughput of 30000 (ideally100000) compoundsday has become the requisite An important consideration in screening compound li-braries of this size is the materials used in assay devel-opment (including dead volume during robotic screen-ing positive and negative controls etc) which cantypically account for 30ndash60 of the total reagent andconsumable requirements for completion of a single tar-get screen

Zheng et al544

TABLE 1 ION CONCENTRATIONS INSIDE

AND OUTSIDE OF A CELL

Concentration

Ion Extracellular Intracellular

Na 145 mM 12 mMK 45 mM 140 mMCa2 18 mM 01ndash02 M (cytosol resting cell)

100 M (cytosol stimulated cell)Mg2 15 mM 08 mMCl 116 mM 4 mMpH 74 71

TABLE 2 CLASSIFICATION OF ION CHANNELS

Channel type Activator Ion permeability TM domains a

Ligand-gated ion channelsIP3R IP3 Ca2 6-TMCNG cAMP Na K Ca2 6-TMnAChR ACh nicotine Na K Ca2 4-TM5-HT3 5-HT Na K Ca2 4-TMGABAAC GABA Cl 4-TMGlycine GABA Cl 4-TMNMDA Glutamate NMDA Na K Ca2 3-TMAMPA Glutamate AMPA Na K Ca2 3-TMKainate Glutamate Na K Ca2 3-TMP2X P2Z ATP Na 2-TM

Voltage-gated ion channelsK channels membrane potential K 6-TMNa channels membrane potential Na 24-TMCa2 channels membrane potential Ca2 24-TMCl channels membrane potential Cl 12-TM

TM transmembrane IP3 inositol trisphophate IP3R IP3 receptor CNG cyclic nucle-otide-gated cAMP cyclic adenosine monophosphate ACh acetylcholine nAChR nico-tinic ACh receptor 5-HT 5-hydroxytryptamine (serotonin) GABA -aminobutyric acidNMDA N-methyl-D-aspartate AMPA -amino-3-hydroxy-5-methylisoxazole-4-propi-onic acid

aThe pore-forming subunit of these channels

5315_16_p543-552 11504 1235 PM Page 544

Assay Technologies for Ion Channels 545

TA

BL

E3

SC

RE

EN

ING

TE

CH

NO

LO

GIE

SF

OR

ION

CH

AN

NE

LT

AR

GE

TS

Ass

ay t

ype

Thr

ough

put

For

mat

Cos

t pe

r w

ella

Com

men

ts

Mem

bran

e bi

ndin

gM

ediu

mh

igh

96-w

ellb

Med

ian-

high

Lim

ited

by

liga

nd a

vail

abil

ity

and

stru

ctur

e N

ot f

unct

iona

l38

4-w

ell

(SP

A)

Med

ian

Ele

ctro

phys

iolo

gyV

olta

ge-c

lam

ppa

tch-

clam

pV

ery

low

Sin

gle

cell

bC

lass

ic a

nd g

old

stan

dard

IonW

orks

HT

Low

-med

ium

384-

wel

lbH

igh

Low

sea

l re

sist

ance

(G

)

Pat

chX

pres

s 70

00L

ow16

-wel

lbV

ery

high

G

seal

but

low

thr

ough

put

Ion

flux

ass

ayR

adio

isot

opes

(45

Ca2

22

Na

86

Rb

)L

ow96

-wel

lbM

edia

nL

ow s

igna

l-to

-noi

se r

atio

no

t fo

r H

TS

Rb

flux

ass

ay (

atom

ic a

bsor

banc

e)M

ediu

m96

384

-wel

lbL

owF

or K

ch

anne

lsF

luor

esce

nce

dye

Ca2

dye

Hig

h96

384

-wel

lbL

owndashm

edia

nL

imit

ed f

or C

a2pe

rmea

ble

chan

nels

Mem

bran

e po

tent

ial

Mem

bran

e po

tent

ial

kit

Hig

h96

384

-wel

lbM

edia

nndashhi

ghIC

50sh

ifte

d to

the

rig

ht i

n ce

rtai

n ty

pes

of c

hann

els

FR

ET

-bas

edH

igh

963

84-w

ellb

Low

ndashmed

ian

a Cos

t pe

r w

ell

is c

alcu

late

d on

ly f

or t

he s

peci

al c

onsu

mab

le r

eage

nt (

eg

S

PA

bea

ds

dye

or

spec

ial

plat

ech

ip)

base

d on

the

hig

hest

pla

te d

ensi

ty a

vail

able

b H

eter

ogen

eous

ass

ay (

cell

was

h is

req

uire

d)

5315_16_p543-552 11504 1235 PM Page 545

Secondary screening assays (HTS hit confirmation and lead optimization)

The throughput requirement for these types of assaysis much lower than HTS but the demands on data qual-ity are higher Unlike primary screening compound titra-tion with eight to 10 different concentrations in duplicateis usually needed to determine the IC50 value for eachcompound tested in a secondary screen Often one ormore different types of assays are performed to confirmthe activity of a compound The required screeningthroughput for secondary assays is on the order of tensto hundreds of compounds per day

Ion channel safety assessment

Cardiac ion channel safety has received a lot of atten-tion in the past 5ndash7 years Identifying compounds thathave the potential to produce QT prolongation early inthe development process is of industry-wide interest In-hibition of cardiac hERG channels has been identified asthe mechanism underlying the cardiac toxicity of severaltherapeutic agents Considerable efforts have been de-voted to develop a reliable high-throughput assay forthese channels Because of public health safety concernsthe US Food and Drug Administration currently requiresthat the activity of all novel agents for which an investi-gational new drug application is filed be evaluated againstthe hERG channel Therefore hERG assays must be ofthe utmost quality reproducibility and reliability Prac-tically the throughput requirement for these assays mustbe similar to secondary screening assays

Current Ion Channel Screening Technologies

Radioligand binding assay

The radiolabel ligand-binding assay was developed inthe 1960s It has been extensively used for drug screen-ing of many targets including ion channels Binding as-says were utilized most extensively in the 1980s and 1990sbefore cell-based functional assays were made availablefor HTS Binding assays incorporate the use of a ligandthat is labeled with a radioactive tracer such as 3H or 125IBinding of the labeled ligand to a specific site on a chan-nel protein can be displaced by an unlabeled compoundif it binds to the same site on the protein The activity ofthe unlabeled compound can be quantified by its ability(IC50) to compete with the labeled ligand Filtration bind-ing assays utilize a glass fiber filter-mounted 96-well plateto separate free ligands with the ligandndashchannel proteincomplex This assay requires a plate wash step which lim-its the screening throughput The SPA uses solid scintil-lant-containing beads to capture cell membranes The la-beled ligands bind to these membrane-coated beadswhich enables homogeneous detection due to the trans-

fer of energy from labeled ligands to SPA beads in prox-imity This SPA binding assay can be miniaturized into384- and 1536-well formats with a throughput of 50000ndash100000 compounds per day at moderate cost

Binding assays provide no information about the effectof novel agents on ion channel function For example anagonist cannot be distinguished from an antagonist in abinding assay Additionally if a compound interacts withthe channel protein at a site distinct from the labeled li-gand it will not be detected in a binding assay An ex-ample of this is the hERG 3H-MK-499 binding assay Ithas been reported that the potency of certain compoundsin the 3H-MK-499 binding assay can differ by more than100-fold when compared to the potency of the same com-pounds measured in a functional voltage-clamp hERG as-say Voltage-gated ion channels do not have endogenousligands and hence exogenous toxins or compounds areused as the labeled ligands The structure diversity of hitsidentified by binding assay-based primary screens for ionchannel targets is often limited

Fluorescent dye probes calcium-sensing dye

Fluo-3 and Fluo-4 are the most commonly used fluo-rescent dyes for the measurement of changes in intracel-lular calcium Cells are loaded with the dye and an ex-ogenous stimulus is applied to elicit the influx of calciumions The dyes are excited at a wavelength of 480 nm andemit a strong signal at 525 nm (Fig 1) The fluorescenceintensity of these dyes increases proportionally with theelevation of intracellular free calcium concentration In anon-stimulated cell the intracellular free calcium con-centration (01ndash02 M) is four orders of magnitude lessthan the extracellular calcium concentration (2 mM) Incontrast the intracellular and extracellular concentrationdifference between K Na and Cl ions is muchsmaller (20-fold) and does not change radically whencells expressing channels that conduct one or more ofthese ions are stimulated Hence fluorescent dyes that are sensitive to these ions are not widely utilized In cellsthat express channels that conduct Ca2 influx of Ca2

through open channels in the cell membrane can producelarge changes in intracellular Ca2 concentrations Thesechanges in intracellular Ca2 concentration can be de-tected with the Fluo dyes The FLIPR system (Molecu-lar Devices Sunnyvale CA) developed in the early1990s utilizes a CCD-camera-based system to detect and capture the fluorescent signal emitted by the dyesThroughput in this screening platform has been maxi-mized with the addition of a fluorescence quenching sub-stance to the assay buffer to suppress extracellular dyefluorescence thereby eliminating the need for a cell wash step Throughput of up to 80000 data pointsday in 384-well format with relatively low reagent cost can be achieved A similar platform the Functional DrugScreening System (FDSS Hamamatsu Hamamatsu City

Zheng et al546

5315_16_p543-552 11504 1235 PM Page 546

Japan) is also available for this type of assay Calcium-sensing dyes have been used extensively for voltage-gatedchannels and ligand-gated receptors that conduct Ca2

ions Since the membrane potential is not controlled influorescence-based assays the potency of compounds thatdisplay ldquostate-dependentrdquo or ldquovoltage-dependentrdquo antag-onism can be significantly weaker when compared to po-tencies obtained in the patch-clamp assay

Fluorescent dye probes voltage-sensing dyes

Voltage-sensing dyes are used to track changes inmembrane potential Oxonol derivative voltage-sensingdyes are negatively charged and associate with the out-side layer of a hyperpolarized cell membrane When thecell membrane is depolarized (ie the inner layer of thecell membrane becomes more positively charged) thedye moves into the inner layer of cell membrane Oxonolvoltage-sensing dyes were discovered in the 1960s butwere not commercialized for use in large-scale screeningof ion channels until the mid-1990s when plate readersbecame available1617

FRET-based voltage-sensing dye FRET incorporatesthe use of a pair of dyes to monitor changes in membranepotential1819 The FRET donor is a coumarin dye linkedto a phospholipid that inserts into the outer leaflet of thecell membrane and the FRET acceptor is an oxonol de-rivative In a hyperpolarized cell 400 nm excitation of thecoumarin produces FRET and excites the oxonol deriva-tive also associated with the outer cell membrane whichthen emits a fluorescent signal at 580 nm When the cellmembrane is depolarized the oxonol derivative movesinto the inner layer of the cell membrane thereby creat-ing a greater physical distance between it and the coumarin

dye FRET is disrupted because of the physical distance(100 nm) between the two dyes Under these conditionsthe emission from coumarin (460 nm) is enhanced whilethe emission from the oxonol is reduced (Fig 2) Theevents are quantified as a ratio of emission detected fromthe FRET donor and FRET acceptor FRET-based volt-age dyes can provide a relatively rapid temporal resolu-tion (approximately seconds) in comparison to calcium-sensing dye (approximately minutes) The radiometricmeasurement of change in membrane potential helps to reduce assay artifacts The Voltage Ion Probe Reader(VIPR Aurora Discovery San Diego CA) was specifi-cally designed for FRET-based assays and has a through-put of 35000ndash50000 compounds per day (384-well for-mat) Drawbacks to this approach include (1) special andcostly instrumentation is required (2) the dyes are onlycapable of monitoring slow membrane potential changesand thus their use is limited to a select group of ion chan-nels and (3) it requires two cell wash steps which placesa limit on throughput A multiwavelength FLIPR instru-ment that can collect data using radiometric dyes was re-cently introduced (FLIPRTetra Molecular Devices) Thisinstrument collects images at different wavelengths inrapid succession (1 s) rather than simultaneously

A proprietary membrane-potential dye kit (MolecularDevices) This kit has been used for the homogeneousmeasurement of changes in membrane potential with sev-eral potassium channels2021 It utilizes a voltage-sensingdye mixed with proprietary fluorescent quenchers Thetemporal resolution of this dye is in the range of minutesslower than the FRET-based voltage-sensing dye com-bination Throughput is maximized by the use of aquencher which enables homogeneous assay format The


FIG 1 Structure (left panel) and spectrum (right panel) of Fluo-3 a calcium-sensing dye Fluo-3 is added as the membrane-permeable non-fluorescent acetoxymethyl (AM) ester form in the loading buffer In the cytosol endogenous esterases hydrolyzeit to form a free acid (salt form) which is not membrane-permeant and becomes fluorescent in the presence of calcium with ex-citation at 488 nm and emission at 525 nm

5315_16_p543-552 11504 1235 PM Page 547

FIG 2 Schematic illustrationof the assay mechanism forFRET-based voltage sensing (A)Structure of the phospholipid-linked FRET donor (coumarin)and FRET acceptor (oxonol) Theexcitation and emission frequen-cies of the FRET donor are 400and 460 nm respectively andemission from the FRET acceptoris 580 nm The emission fromboth the FRET donor and accep-tor are measured simultaneously

using the VoltageIon Probe Reader (Aurora Discovery Inc) (B) At the resting membrane potential (approximately 60 mV)fluorescence energy is transferred from the coumarin donor to the oxonol acceptor resulting in fluorescence at 580 nm Duringdepolarization the FRET acceptor (oxonol) translocates to the inner membrane leaflet resulting in a decrease in FRET emissionat 580 nm and a concomitant increase in emission from the FRET acceptor at 460 nm The relative intensity at 460 and 580 nmprovides a readout of changes in the membrane potential of the cell

quencher(s) absorb the emission of the voltage-sensitivedye when it is positioned in the outer layer of the cellmembrane When the cell membrane is depolarized thedye moves to the inner layer of cell membrane and uponexcitation emits a detectable signal (Fig 3) The through-put of this assay (in 384-well format) is 60000ndash80000compounds per day and the screening cost is relativelyhigh (because of the price of dye kit) A similar mem-brane potential dye kit ACTOne is also available fromBD Biosciences (Franklin Lake NJ)

The homogeneous nature of this assay combined withthe use of a CCD-based imaging instrument that mea-sures the entire plate at once helps to minimize well-to-well variation This is important since the signal-to-noiseratio in this assay is low (15ndash25) We have utilized thistype of an assay for an HTS on a chloride channel in

which the signal-to-noise ratio was only 14 Despite thissmall signal-to-noise window the hit confirmation ratefor the HTS screen was 56

A direct comparison of this assay with a FRET-basedvoltage-sensing dye assay on the same channel revealedthat the activities of small molecule compounds were lesspotent in the no-wash assay (W Zheng et al unpub-lished data) We have also observed that the activities ofpeptides and peptideprotein-based toxins were greatlyreduced in the no-wash dye assay but could be detectedin the FRET-based dye assay The noted reduction in ac-tivity may be a result of interference from the quencher(s)or other unidentified components in this no-wash dye kitor that the dyes themselves associate not only with theinner layer of cell membrane but also the membranes ofsubcellular organelles

Zheng et al548

A

B

5315_16_p543-552 11504 1235 PM Page 548


FIG 3 FLIPR membrane potential assay A voltage-sensitive dye together with fluorescence quenchers is added to cells Thedistribution of dye across the membrane is related to the membrane potential At the resting membrane potential (approximately60 mV) voltage-sensitive dye molecules are associated with the extracellular leaflet of the cell membrane where nonndashmem-brane-permeable quenchers in the assay buffer diminish their fluorescence Upon depolarization dye molecules rapidly translo-cate into the inner membrane resulting in an increase in fluorescence intensity The membrane potential dye kit is available fromMolecular Devices

FIG 4 Schematic illus-trations of a non-radioac-tive ion flux assay usingAAS (A) Diagram of thegeneral procedure for load-ing cells with tracer ionand subsequent activationeither by membrane depo-larization (high K) or ad-dition of ligand to initiatethe flux of ions across thecell membrane (B) Tracer

ion concentrations are sensitively quantified using an atomic absorption spectrometer Vaporized atoms are converted to theirground state within a flame where they absorb light of specific wavelengths PMT photomultiplier tube An automated platereader for 96384-well plates is available from Aurora Biomed Inc (wwwaurorabiomedcom)

A

B

Ion flux assay

Radioactive ion flux assay Radiotracers can be usedto measure the flux of ions moving in or out of a cell viaion channels expressed in the cell membrane Radiotrac-ers are available for every class of ion channels 86Rb

for potassium channels 22Na for sodium channels45Ca2 for calcium channels and 36Cl for chloride chan-nels Although these radiotracers have been used for over20 years in ion channel assays their application for HTSdrug screening has been limited Tracer assays are het-erogeneous and slow requiring both a tracer loading andwash steps Only the steady-state function of ion chan-nels can be measured with radiotracers Signal-to-noise

ratio is low because of incomplete removal of extracel-lular tracer after loading or the continuous leak of tracerout of cells Concerns over excessive radioactive wasteand safety also limit radioactive flux assay use in HTS

AAS Ion flux assays using non-radioactive ion tracersanalyzed in AAS have been utilized since the 1950s Com-mercial instrumentation for high-throughput AAS screen-ing has only become available within the last few years(Fig 4) AAS assays have been developed for a variety ofvoltage-gated22 and ligand-gated channels23 For this assaynon-radioactive tracer ions are loaded into cells expressingthe channel of interest Ion flux is then initiated when chan-nels are activated by a ligand or by depolarizing the cell

5315_16_p543-552 11504 1235 PM Page 549

membrane with a high K buffer (50ndash80 mM) The con-centration of the tracer ion in the supernatant andor withinthe cells is then measured and the percentage of efflux (orinflux) is calculated The signal-to-noise ratio of this assayis typically 6ndash10 and the reagent cost for screening is verylow Both single-channel and multichannel instruments arecurrently available for 96-well and 384-well screens withmoderate throughput (Aurora Biomed San Diego) Reportsdescribing the application of this technology for screeningassays have been primarily focused on potassium channelssuch as hERG KCNQ2 and Ca2-activated potassiumchannels where Rb is used as the tracer ion24ndash26

Electrophysiology

Patch-clamp Patch-clamp electrophysiology methodsare regarded as the gold standard for measurement ofcompound activity on ion channels in vitro Through anelectrode attached to the cell membrane the current gen-erated by the ions flowing through ion channels expressedin the cell membrane can be measured while the mem-brane potential is voltage-clamped (Fig 5) The activityof ion channels is measured directly and in real-time De-spite the high-quality data generated by this method inits traditional format patch-clamping has limited use indrug screening for ion channel targets because of very

low throughput However in the past 2 years several au-tomated patch-clamp instruments have been developedand are now commercially available28

Automated patch-clamp electrophysiology Traditionalpatch-clamp electrophysiology incorporates the use of aglass micropipette electrode microfabricated from glasscapillary tubes for controlling the membrane potentialwhile measuring ionic current flow The breakthrough thatmade automating this process for truly higher throughputfeasible was the development of the planar patch-clampelectrode The traditional single micropipette electrode wasreplaced with a planar substrate with an array of micro-apertures2728 The first commercially available instrument(IonWorks Molecular Devices) uses a planar 384-welldisposable polyimide plastic plate In order to performpatch-clamp recordings cells (in suspension) are firstadded to electrically isolated wells on the PatchPlatetradeEach well on the PatchPlate contains a single aperture forthe patch-clamping of a cell A slight negative pressure isused to pull the cell membrane into the aperture andachieve a 50ndash600 m seal Electrical access is achievedvia a perforating agent applied at the bottom surface of thePatchPlate Recent reports show that the pharmacology onseveral voltage-gated ion channels accurately reflects tra-ditional patch-clamp data The success rate of patch is be-tween 60 to 90 and the assay throughput for process-ing each 384-well plate is 1 h From 2000 to 3000 cellscan be patch-clamped in a day and 50ndash100 dose responsescan be acquired with this system representing a 100-foldincrease in throughput over the traditional patch-clamptechnique2930 The low seal resistance limits the use of thistechnology to cell lines with robust and homogeneous ex-pression of voltage-gated ion channels

Another automated planar patch-clamp instrument thatrecently became commercially available uses a 16-welldisposable glass chip (PatchXpress 7000A AxonMolec-ular Devices) As in the traditional patch-clamp a G sealresistance can be achieved on these chips with a successrate of 20ndash70 In this system electrical access is achievedby rupturing the cell membrane underneath the apertureAsynchronous operation and the integrated fluidics allowligand-gated as well as voltage-gated channels to be as-sayed with this instrument Pharmacological studies withthis instrument demonstrate a good correlation with tradi-tional patch-clamp data The two- to 10-fold increase inthroughput this system offers is attractive for detailed stud-ies of ion channel pharmacology as well as directed screen-ing of small sets of compounds31 Ideally this system fitswell alongside the previously discussed 384-well instru-ment Whereas the 384-well instrument is utilized to filterthrough hundreds to thousands of compounds the 16-wellsystem is used to perform detailed electrophysiologicalstudies on the identified leads The high cost of consum-ables will greatly impact the use of these instruments The

Zheng et al550

FIG 5 Schematic illustration of patch-clamp configurations(A) standard micropipette patch-clamp configuration and (B)planar chip patch-clamp configuration27

A

B

5315_16_p543-552 11504 1236 PM Page 550

evolution of automated patch-clamp electrophysiology isonly in its first stage Several other automated planar patch-clamp and oocyte clamp systems32ndash36 are available or inlate-stage development and the next few years promise tobe an exciting time for this technology

Perspectives for Ion Channel Screening Technologies

Automated patch-clamp electrophysiology

The quality and throughput of automated patch-clampinstrumentation will continue to improve in the next 3ndash10 years with advances in microfabrication and micro-machining technologies for existingnovel planar elec-trode substrates This should drive down the cost of consumables and make the technology more readily ac-cessible and widely used for the screening of ion chan-nels The immediate impact of this technology will be forsecondary screening and ion channel safety assessment

Non-invasive detection of ion channel activity

Current microelectrode based patch-clamp methods areinvasive and can disrupt intracellular physiology Micro-electrode array technology is a new approach for non-in-vasive extracellular recording of ion channel activity37

Currently this technology has been used to record ionchannel activity in tissue slices and cultured cardiac cellsin single-wellchamber format Currently throughput islow but information content is high With the miniatur-ization of microelectrode arrays and the development ofmultiwell detection its application for ion channel screen-ing will be further explored Other label-free detectiontechnologies such as resonant acoustic profiling micro-plate differential calorimetry atomic force microscopyand microwave spectroscopy may also be developed fornon-invasive and high-throughput detection of ion chan-nel functions in the future

Next generation true high-throughput instrumentation

The current version of automated patch-clamp tech-nology cannot meet the demands of HTS for large com-pound collections The fluorescent dye-based and ion fluxassays only measure steady-state ion channel activitywhich may not reflect the physiological condition of ionchannels Additionally agonists or toxins are usually re-quired to activate channels in these screens The next gen-eration of patch-clamp-based screening technologies willhave to incorporate even higher throughput without com-promising data quality

Ion channel biology

The pore forming subunit of an ion channel is madeup of multiple subunits Each class of ion channel also


has multiple auxiliary subunits that can modulate the ac-tivity of the subunit Voltage-gated calcium channelsfor example consist of 1 2ndash and subunits andeach of these subunits has multiple isoforms and splicevariants The strategy for cell line generation of ion chan-nels is complicated by the variety of and auxiliary chan-nel subunits It is difficult not only to stably transfect allthe subunits into a cell line but also to select the ldquorightrdquocombination of these subunits Validation of biologicallyrelevant ion channel targets including auxiliary subunitcomponents will continue to be an important area for ionchannel drug discovery In addition the use of transfectedcell lines versus native cell lines for ion channel screen-ing warrants further investigation

Conclusions

Currently fluorescence-based assays remain the mostfrequently used method for the primary screening of largecompound collections in ion channel drug discovery Ionflux and automated patch-clamp assays are the choice forsecondary screening and lead optimization Although thescreening throughput and quality have been greatly im-proved compared to those 10 years ago ion channelscreening technologies need further innovation refine-ment and optimization Ion channel assays for future HTSwill have to be miniaturized into 1536-well or higherdensity formats to accommodate the increasing capacityfor the screening of multimillion compound librariesNew screening technologies are especially needed for theion channel targets that cannot be screened with existingtechnologies because of low channel expression in cellsIn addition more reliable and cost-effective methods areneeded for ion channel safety assessment

References

1 Davies NP Hanna MG The skeletal muscle chan-nelopathies distinct entities and overlapping syndromesCurr Opin Neurol 200316559ndash568

2 Mulley JC Scheffer IE Petrou S Berkovic SF Chan-nelopathies as a genetic cause of epilepsy Curr Opin Neu-rol 200316171ndash176

3 Pietrobon D Calcium channels and channelopathies of thecentral nervous system Mol Neurobiol 20022531ndash50

4 Kullmann DM The neuronal channelopathies Brain 20021251177ndash1195

5 Hubner CA Jentsch TJ Ion channel diseases Hum MolGenet 2002112435ndash2445

6 Ben-David J Zipes DP Torsades de pointes and proar-rhythmia Lancet 19933411578ndash1582

7 Belardinelli L Antzelevitch C Vos MA Assessing pre-dictors of drug-induced torsade de pointes Trends Phar-macol Sci 200324619ndash625

8 Fermini B Fossa AA The impact of drug-induced QT in-terval prolongatin on drug discovery and development NatRev Drug Discov 20032439ndash447

5315_16_p543-552 11504 1236 PM Page 551

26 Gill S Gill R Lee SS Hesketh JC Fedida D RezazadehS Stankovich L Liang D Flux assays in high throughputscreening of ion channels in drug discovery Assay DrugDev Technol 20031709ndash717

27 Wang X Li M Automated electrophysiology high through-put of art Assay Drug Dev Technol 20031695ndash708

28 Willumsen NJ Bech M Olesen SP Jensen BS KorsgaardMP Christophersen P High throughput electrophysiologynew perspectives for ion channel drug discovery Recep-tors Channels 200393ndash12

29 Laszlo K Bennett PB Uebele VN Koblan KS Kane SANeagle B Schroeder K High throughput ion-channel phar-macology planar-array-based voltage clamp Assay DrugDev Technol 20031127ndash135

30 Schroeder K Neagle B Trezise DJ Worley J IonworksHT a new high-throughput electrophysiology measure-ment platform J Biomol Screen 2003850ndash64

31 Xu J Guia A Rothwarf D Huang M Sithiphong K OuangJ Tao G Wang X Wu L A benchmark study withSealChiptrade planar patch-clamp technology Assay DrugDev Technol 20031675ndash684

32 Bruggemann A George M Klau M Beckler M Steindl JBehrends JC Fertig N High quality ion channel analysison a chip with the NPCcopy technology Assay Drug Dev Tech-nol 20031665ndash673

33 Kutchinsky J Friis S Asmild M Taboryski R Pedersen SVestergaard RK Jacobson RB Krzywkowski K SchroslashderRL Ljungstroslashm T Helix N Soslashrensen CB Bech M Willum-sen NJ Characterization of potassium channel modulatorswith QPatchtrade automated patch-clamp technology systemcharacteristics and performance Assay Drug Dev Technol20031685ndash693

34 Shieh C-C Trumbull JD Sarthy JF McKenna DG Pari-har AS Zhang XF Faltynek CR Gopalakrishnan M Au-tomated Parallel Oocyte Electrophysiology Test station(POETstrade) a screening platform for identification of li-gand-gated ion channel modulators Assay Drug Dev Tech-nol 20031655ndash663

35 Joshi PR Suryanarayanan A Schulte MK A vertical flow chamber for Xenopus oocyte electrophysiology andautomated drug screening J Neurosci Methods 200413269ndash79

36 Schnizler K Kuster M Methfessel C Fejtl M The robo-ocyte automated cDNAmRNA injection and subsequentTEVC recording on Xenopus oocytes in 96-well microtiterplates Receptors Channels 2003941ndash48

37 Stett A Egert U Guenther E Hofmann F Meyer T NischW Haemmerle H Biological application of microelectrodearrays in drug discovery and basic research Anal BioanalChem 2003377486ndash495

Address reprint requests toLaszlo Kiss PhD

Department of Molecular NeurologyMerck amp Co IncSumneytown Pike

West Point PA 19486

E-mail laszlo_kissmerckcom

9 Clancy CE Kurokawa J Tateyama M Wehrens XH KassRS K channel structure-activity relationships and mech-anisms of drug-induced QT prolongation Annu Rev Phar-macol Toxicol 200343441ndash461

10 Walker BD Krahn AD Klein GJ Skanes AC Yee R Druginduced QT prolongation lessons from congenital and ac-quired long QT syndromes Curr Drug Targets CardiovascHaematol Disord 20033327ndash335

11 Domene C Haider S Sansom MS Ion channel structuresa review of recent progress Curr Opin Drug Discov Dev20036611ndash619

12 Galligan JJ Ligand-gated ion channels in the enteric ner-vous system Neurogastroenterol Motil 200214611ndash623

13 Bennett PB Guthrie HR Trends in ion channel drug dis-covery advances in screening technologies Trends Bio-technol 200321563ndash569

14 Netzer R Bischoff U Ebneth A HTS techniques to in-vestigate the potential effects of compounds on cardiac ionchannels at early-stages of drug discovery Curr Opin DrugDiscov Dev 20036462ndash469

15 Worley JF Main MJ An industrial perspective on utiliz-ing functional ion channel assays for high throughputscreening Receptors Channels 20028269ndash282

16 Gonzalez JE Maher MP Cellular fluorescent indicatorsand voltageion probe reader (VIPR) tools for ion channeland receptor drug discovery Receptors Channels 20028283ndash295

17 Wolff C Fuks B Chatelain P Comparative study of mem-brane potential-sensitive fluorescent probes and their usein ion channel screening assays J Biomol Screen 20038533ndash543

18 Gonzalez JE Tsien RY Voltage sensing by fluorescenceresonance energy transfer in single cells Biophys J 1995691272ndash1280

19 Gonzalez JE Tsien RY Improved indicators of cell mem-brane potential that use fluorescence resonance energytransfer Chem Biol 19974269ndash277

20 Baxter DF Kirk M Garcia AF Raimondi A HolmqvistMH Flint KK Bojanic D Distefano PS Curtis R Xie YA novel membrane potential-sensitive fluorescent dye im-proves cell-based assays for ion channels J Biomol Screen2002779ndash85

21 Whiteaker KL Gopalakrishnan SM Groebe D Shieh CCWarrior U Burns DJ Coghlan MJ Scott VE Gopalakr-ishnan M Validation of FLIPR membrane potential dyefor high throughput screening of potassium channel mod-ulators J Biomol Screen 20016305ndash312

22 Tang W Kang J Wu X Rampe D Wang L Shen H LiZ Dunnington D Garyantes T Development and evalua-tion of high throughput functional assay methods for HERGpotassium channel J Biomol Screen 20016325ndash331

23 Terstappen GC Functional analysis of native and recom-binant ion channels using a high-capacity nonradioactiverubidium efflux assay Anal Biochem 1999272149ndash155

24 Parihar AS Groebe DR Scott VE Feng J Zhang XF War-rior U Gopalakrishnan M Shieh C-C Functional analysisof large conductance Ca2-activated K channels ion fluxstudies by atomic absorption spectrometry Assay Drug DevTechnol 20031647ndash654

25 Scott CW Wilkins DE Trivedi S Crankshaw DJ Amedium-throughput functional assay of KCNQ2 potassiumchannels using rubidium efflux and atomic absorption spec-trometry Anal Biochem 200315319251ndash257

Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552

ondary levels Traditional methods developed for HTSof ion channels such as binding ion flux and fluores-cent probes measure ion channel activity indirectlyPatch-clamp electrophysiology is regarded as the ldquogoldstandardrdquo for measuring ion channel activity and phar-macology Patch-clamp allows for the direct real-timemeasurement of ion channel activity but in its traditionalformat is low-throughput and requires a high degree ofoperator skill Hence drug screening assays for ion chan-nels in comparison to those for enzyme and receptor tar-gets have compromised data quality for throughput Re-cently a number of new screening technologies have beendeveloped and improved for ion channel assays and arepoised to change this13ndash15 A summary of the currentlyavailable screening technologies is listed in Table 3

Demands on Ion Channel Screening Assays

The current demands for ion channel screening in thedrug discovery process can be grouped into three mainareas primary screening assays secondary screening as-says and ion channel safety assessment

Primary screening assays (HTS)

HTS has seen a tremendous advance during the last10 years and remains the first critical step in the dis-covery of lead chemical structures for novel drug tar-gets To increase the probability of success for findingnew leads from HTS many companies have investedheavily in expanding both the diversity and quality oftheir compound libraries For most mid- and large-sizedcompanies the library collection has grown to 400000ndash1000000 (or more) compounds The standard para-digms used to screen these libraries have evolved to au-tomated 384-well or higher-density single-compoundtest formats Minimal throughput of 30000 (ideally100000) compoundsday has become the requisite An important consideration in screening compound li-braries of this size is the materials used in assay devel-opment (including dead volume during robotic screen-ing positive and negative controls etc) which cantypically account for 30ndash60 of the total reagent andconsumable requirements for completion of a single tar-get screen

Zheng et al544

TABLE 1 ION CONCENTRATIONS INSIDE

AND OUTSIDE OF A CELL

Concentration

Ion Extracellular Intracellular

Na 145 mM 12 mMK 45 mM 140 mMCa2 18 mM 01ndash02 M (cytosol resting cell)

100 M (cytosol stimulated cell)Mg2 15 mM 08 mMCl 116 mM 4 mMpH 74 71

TABLE 2 CLASSIFICATION OF ION CHANNELS

Channel type Activator Ion permeability TM domains a

Ligand-gated ion channelsIP3R IP3 Ca2 6-TMCNG cAMP Na K Ca2 6-TMnAChR ACh nicotine Na K Ca2 4-TM5-HT3 5-HT Na K Ca2 4-TMGABAAC GABA Cl 4-TMGlycine GABA Cl 4-TMNMDA Glutamate NMDA Na K Ca2 3-TMAMPA Glutamate AMPA Na K Ca2 3-TMKainate Glutamate Na K Ca2 3-TMP2X P2Z ATP Na 2-TM

Voltage-gated ion channelsK channels membrane potential K 6-TMNa channels membrane potential Na 24-TMCa2 channels membrane potential Ca2 24-TMCl channels membrane potential Cl 12-TM

TM transmembrane IP3 inositol trisphophate IP3R IP3 receptor CNG cyclic nucle-otide-gated cAMP cyclic adenosine monophosphate ACh acetylcholine nAChR nico-tinic ACh receptor 5-HT 5-hydroxytryptamine (serotonin) GABA -aminobutyric acidNMDA N-methyl-D-aspartate AMPA -amino-3-hydroxy-5-methylisoxazole-4-propi-onic acid

aThe pore-forming subunit of these channels

5315_16_p543-552 11504 1235 PM Page 544


TA

BL

E3

SC

RE

EN

ING

TE

CH

NO

LO

GIE

SF

OR

ION

CH

AN

NE

LT

AR

GE

TS

Ass

ay t

ype

Thr

ough

put

For

mat

Cos

t pe

r w

ella

Com

men

ts

Mem

bran

e bi

ndin

gM

ediu

mh

igh

96-w

ellb

Med

ian-

high

Lim

ited

by

liga

nd a

vail

abil

ity

and

stru

ctur

e N

ot f

unct

iona

l38

4-w

ell

(SP

A)

Med

ian

Ele

ctro

phys

iolo

gyV

olta

ge-c

lam

ppa

tch-

clam

pV

ery

low

Sin

gle

cell

bC

lass

ic a

nd g

old

stan

dard

IonW

orks

HT

Low

-med

ium

384-

wel

lbH

igh

Low

sea

l re

sist

ance

(G

)

Pat

chX

pres

s 70

00L

ow16

-wel

lbV

ery

high

G

seal

but

low

thr

ough

put

Ion

flux

ass

ayR

adio

isot

opes

(45

Ca2

22

Na

86

Rb

)L

ow96

-wel

lbM

edia

nL

ow s

igna

l-to

-noi

se r

atio

no

t fo

r H

TS

Rb

flux

ass

ay (

atom

ic a

bsor

banc

e)M

ediu

m96

384

-wel

lbL

owF

or K

ch

anne

lsF

luor

esce

nce

dye

Ca2

dye

Hig

h96

384

-wel

lbL

owndashm

edia

nL

imit

ed f

or C

a2pe

rmea

ble

chan

nels

Mem

bran

e po

tent

ial

Mem

bran

e po

tent

ial

kit

Hig

h96

384

-wel

lbM

edia

nndashhi

ghIC

50sh

ifte

d to

the

rig

ht i

n ce

rtai

n ty

pes

of c

hann

els

FR

ET

-bas

edH

igh

963

84-w

ellb

Low

ndashmed

ian

a Cos

t pe

r w

ell

is c

alcu

late

d on

ly f

or t

he s

peci

al c

onsu

mab

le r

eage

nt (

eg

S

PA

bea

ds

dye

or

spec

ial

plat

ech

ip)

base

d on

the

hig

hest

pla

te d

ensi

ty a

vail

able

b H

eter

ogen

eous

ass

ay (

cell

was

h is

req

uire

d)

5315_16_p543-552 11504 1235 PM Page 545













Zheng et al546

5315_16_p543-552 11504 1235 PM Page 546










5315_16_p543-552 11504 1235 PM Page 547







Zheng et al548

A

B

5315_16_p543-552 11504 1235 PM Page 548





A

B

Ion flux assay





5315_16_p543-552 11504 1235 PM Page 549


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552


TA

BL

E3

SC

RE

EN

ING

TE

CH

NO

LO

GIE

SF

OR

ION

CH

AN

NE

LT

AR

GE

TS

Ass

ay t

ype

Thr

ough

put

For

mat

Cos

t pe

r w

ella

Com

men

ts

Mem

bran

e bi

ndin

gM

ediu

mh

igh

96-w

ellb

Med

ian-

high

Lim

ited

by

liga

nd a

vail

abil

ity

and

stru

ctur

e N

ot f

unct

iona

l38

4-w

ell

(SP

A)

Med

ian

Ele

ctro

phys

iolo

gyV

olta

ge-c

lam

ppa

tch-

clam

pV

ery

low

Sin

gle

cell

bC

lass

ic a

nd g

old

stan

dard

IonW

orks

HT

Low

-med

ium

384-

wel

lbH

igh

Low

sea

l re

sist

ance

(G

)

Pat

chX

pres

s 70

00L

ow16

-wel

lbV

ery

high

G

seal

but

low

thr

ough

put

Ion

flux

ass

ayR

adio

isot

opes

(45

Ca2

22

Na

86

Rb

)L

ow96

-wel

lbM

edia

nL

ow s

igna

l-to

-noi

se r

atio

no

t fo

r H

TS

Rb

flux

ass

ay (

atom

ic a

bsor

banc

e)M

ediu

m96

384

-wel

lbL

owF

or K

ch

anne

lsF

luor

esce

nce

dye

Ca2

dye

Hig

h96

384

-wel

lbL

owndashm

edia

nL

imit

ed f

or C

a2pe

rmea

ble

chan

nels

Mem

bran

e po

tent

ial

Mem

bran

e po

tent

ial

kit

Hig

h96

384

-wel

lbM

edia

nndashhi

ghIC

50sh

ifte

d to

the

rig

ht i

n ce

rtai

n ty

pes

of c

hann

els

FR

ET

-bas

edH

igh

963

84-w

ellb

Low

ndashmed

ian

a Cos

t pe

r w

ell

is c

alcu

late

d on

ly f

or t

he s

peci

al c

onsu

mab

le r

eage

nt (

eg

S

PA

bea

ds

dye

or

spec

ial

plat

ech

ip)

base

d on

the

hig

hest

pla

te d

ensi

ty a

vail

able

b H

eter

ogen

eous

ass

ay (

cell

was

h is

req

uire

d)

5315_16_p543-552 11504 1235 PM Page 545













Zheng et al546

5315_16_p543-552 11504 1235 PM Page 546










5315_16_p543-552 11504 1235 PM Page 547







Zheng et al548

A

B

5315_16_p543-552 11504 1235 PM Page 548





A

B

Ion flux assay





5315_16_p543-552 11504 1235 PM Page 549


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552













Zheng et al546

5315_16_p543-552 11504 1235 PM Page 546










5315_16_p543-552 11504 1235 PM Page 547







Zheng et al548

A

B

5315_16_p543-552 11504 1235 PM Page 548





A

B

Ion flux assay





5315_16_p543-552 11504 1235 PM Page 549


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552










5315_16_p543-552 11504 1235 PM 47







Zheng et al548

A

B

5315_16_p543-552 11504 1235 PM Page 548





A

B

Ion flux assay





5315_16_p543-552 11504 1235 PM Page 549


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552







Zheng et al548

A

B

5315_16_p543-552 11504 1235 PM Page 548





A

B

Ion flux assay





5315_16_p543-552 11504 1235 PM Page 549


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552





A

B

Ion flux assay





5315_16_p543-552 11504 1235 PM Page 549


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552


Electrophysiology





Zheng et al550


A

B

5315_16_p543-552 11504 1236 PM Page 550










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552










Ion channel biology




Conclusions


References









5315_16_p543-552 11504 1236 PM Page 551















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552















West Point PA 19486



















Zheng et al552

5315_16_p543-552 11504 1236 PM Page 552

ion channel drug discovery high throughput assay ... · the radiolabel ligand-binding assay was...

Documents