Genetic identification of AChE as a positive modulator of addiction to the psychostimulant D ...

13 Pages·2006·0.54 MB·English

Genetic Identification of AChE as a Positive
Modulator of Addiction to the Psychostimulant
D-Amphetamine in Zebrafish
JovicaNinkovic,1AnjaFolchert,1YuriV.Makhankov,2StephanC.F.Neuhauss,2
IngeSillaber,3UweStraehle,4LaureBally-Cuif1*
1 Zebrafish Neurogenetics Junior Research Group, Institute of Virology, Technical University-Munich,
Trogerstrasse 4b, D-81675, Munich, Germany and GSF-National Research Center for Environment
and Health, Department Zebrafish Neurogenetics, Institute of Developmental Genetics, Ingolstaedter
Landstrasse1,D-85764Neuherberg,Germany
2BrainResearchInstitute,UniversityofZurichandETHZurich,Winterthurerstrasse190,
CH-8057Zurich,Switzerland
3MPIforPsychiatry,Kraepelinstrasse2-8,D-80804Muenchen,Germany
4 University of Heidelberg and Institute of Toxicology and Genetics, Forschungszentrum Karlsruhe,
Postfach3640,D-76021Karlsruhe,Germany
Received28July2005;accepted12October2005
ABSTRACT: Addiction is a complex maladaptive adult zebrafish show robust CPP induced by the psy-
behaviorinvolvingalterationsinseveralneurotransmit- chostimulant D-amphetamine. We further show that
ternetworks.Inmammals,psychostimulantstriggerele- this behavior is dramatically reduced upon genetic
vated extracellular levels of dopamine, which can be impairment of acetylcholinesterase (AChE) function in
modulated by central cholinergic transmission. Which ache/þ mutants, without involvement of concomitant
elementsofthecholinergicsystemmightbetargetedfor defectsinexploratoryactivity,learning,andvisualper-
drug addiction therapies remains unknown. The formance.Ourobservationsdemonstratethatthecholi-
rewarding properties of drugs of abuse are central for nergicsystemmodulatesdrug-inducedrewardinzebra-
the development of addictive behavior and are most fish,andidentifygeneticallyAChEasapromisingtarget
commonlymeasured bymeans oftheconditionedplace forsystemictherapiesagainstaddictiontopsychostimu-
preference (CPP) paradigm. We demonstrate here that lants.Moregenerally,theyvalidatethezebrafishmodel
to study the effect of developmental mutations on the
molecular neurobiology of addiction in vertebrates.
ThisarticleincludesSupplementaryMaterialavailableviathe
'2006WileyPeriodicals,Inc.JNeurobiol66:463–475,2006
Internet at http://www.interscience.wiley.com/jpages/0022-3034/
Keywords: zebrafish; reward; acetylcholinesterase;
suppmat.
Correspondenceto:L.Bally-Cuif([email protected]). amphetamine;conditionedplacepreference;behavior
Contractgrantsponsor:EU6thFramework;contractgrantnum-
ber:LSHC-CT-2003-503466(L.B.C.,U.S.,S.C.F.N.).
Contractgrantsponsor:ETHCommission;contractgrantnum- INTRODUCTION
ber:0-20941-01(Y.V.M.).
Contractgrantsponsor:SchweizerNationalfonds;contractgrant
number:PP00A-68868. Addiction, the uncontrollable compulsion to take
'2006WileyPeriodicals,Inc. drugsinspiteoftheirnegativeeffectonnormalbrain
Published online 6 February 2006 in Wiley InterScience (www.
function, is a widespread and costly brain disorder in
interscience.wiley.com).
DOI10.1002/neu.20231 modern societies. Yet, its neural mechanisms remain
463
464 Ninkovic etal.
incompletely understood, and pharmacological treat- Here we address the relevance and generality of
ments,ifavailable,areinmostcasesineffective. AChE inhibition on drug-induced behavior. The ze-
Themesocorticolimbicdopamine(DA)system,orig- brafish, often proposed as an alternative to mamma-
inating in the midbrain ventral tegmental area (VTA) lian models, is increasingly used as a new genetic
and innervating the nucleus accumbens (NAc) in the systeminsocialbehaviorstudies,becauseofitsame-
ventral striatum, was directly implicated in the re- nability to large-scale genetic screens (Gerlai et al.,
wardingeffectsofmostdrugsofabuseinhigherver- 2000;DarlandandDowling,2001;Guo,2004;Lock-
tebrates(EverittandWolf,2002;Laaksoetal.,2002; wood et al., 2004). We demonstrate that adult zebra-
White, 2002; Wise, 2002). Indeed, psychostimulants fish show a clear CPP induced by a widely used psy-
increase extracellular levels of DA in the NAc, for chostimulant, D-amphetamine, and that this effect is
example,byblockingtheDAtransporter(Amaraand significantly attenuated upon genetic impairment of
Sonders, 1998; Jones et al., 1998; Gainetdinov et al., AChE activity in achesb55/þ mutants (Behra et al.,
2002;Saaletal.,2003),therebyaffectingthehomeo- 2002). Our results provide the first genetic demon-
stasisofthebrainrewardsystem.However,addiction strationthatAChEisapromisingtargetfortherapeu-
is an intricately complex dysfunction of the reward ticapproachestoaddiction,andvalidatethezebrafish
pathwayandlikelyinvolvesaplethoraofneurotrans- to study the consequences of developmental muta-
mitter responses, themselves dependent on parame- tions and the neuronal pathways underlying this con-
ters that can vary upon drug administration, such as dition in vertebrates. The experimental paradigms
progressive neuroadaptative mechanisms. To date, developed here are the first sufficiently robust to use
additional monoamine systems (norepinephrine, sero- the excellent genetic model zebrafish to conclusively
tonin)andmodulatorypathways[acetylcholine(ACh), screenformutationsaffectingreward.
GABA, glutamate] have been proposed to contribute
to the manifestation of the addictive state, by sharing
molecular components with the DA system, by influ- METHODS
encingtheactivityofVTAorNAcneurons(Gasparet
al.,2003; Auclair etal.,2004;Kelley, 2004),orbyas AnimalCareandMaintenance
yetunresolvedmechanisms.
Recentemphasiswasplacedonthecholinergicsys- Adultzebrafishwerekeptinthefishfacility,aspreviously
tem.Inseveralinstances,blockingmuscariniccholiner- described(Kimmeletal.,1995).Atleast2daysbeforeeach
assay,thefishweremovedtoanisolatedroomundermain-
gic transmission in rodents increases response to psy-
tenanceconditionsandfeedingscheduleidenticaltothefish
chostimulants in the self-administration or conditioned
facility (14/10 h day/light cycle, two feedings per day at
place preference(CPP) tests (Gerberet al., 2001; Ichi-
8:00amand2:00pm—exceptinthecaseoflearningtests,
kawa et al., 2002), paradigms classically applied to see below—water temperature 288C). Thus, we kept envi-
evaluateaddictivebehavior.TheNAcisdenselyinner-
ronmentalvarianceataminimumforallbehavioralassays.
vated by cholinergic interneurons; theactivity of these TheABstrainwasbredinourfishfacilityformorethan20
neurons is decreased by DA release (Alcantara et al., generations. achesb55/þ (Behra et al., 2002), initially on
2003)andtheirablationincreasesthesensitivitytopsy- ABO background, were crossed for at least three genera-
chostimulants and the reinforcing effects of cocaine tions to our AB strain before all experiments. They were
(Hikida et al.,2001). Further, other components of the identified in random brother-sister crossing, giving rise to
25%achesb55/achesb55immobileembryos,asdescribedpre-
cholinergic system, like M5 muscarinic receptors on
viously (Behra et al., 2002). All experiments reported in
VTA DA neurons or nicotinic ACh receptors on DA
this article were done on 3- to 6-month-old females (how-
terminals in the striatum, are implicated in the regula-
ever, we observed no difference in the behavior of males
tion of addictive behavior (see Tzschentke, 1998).
comparedtofemales;notshown).Carewastakentoalways
While most studies agree that cholinergic activity can
testthesamefishatthesametimeofthedayoverthesuc-
modulate DA transmission,and therebythepropensity cessivedaysofeachexperiment.
for addiction, a reliable pharmacological target of this
systemfordrugaddictiontherapyhastobedetermined.
Using pharmacological inhibitors, Hikida et al. (2003) PlacePreference(PP)Determination
suggested that inhibiting acetylcholinesterase (AChE)
Thetestingapparatuswasa3L,rectangulartankcontaining
activity, which terminates ACh action at the synapse,
2Lofwaterandplacedinanisolatedcabinetwithtopillu-
can block cocaine- and morphine-induced CPP in the
mination.Thewaterlevelwaskeptto10cmfromthetank
mouse.ThusAChEinhibitorsmightbepromisingther- bottomtominimizestress.Distinctvisualcuesdividedthe
apeutic agents, although their specificity and range of experimental tank into two halves: a dark half colored in
efficiencyremaintobedetermined. brown and a light half colored in white with two frighten-
JournalofNeurobiology.DOI10.1002/neu
AChEandAmphetamine-Induced RewardinZebrafish 465
ing,blackcirclesplacedatthebottomofthetank.Afteran 1(B) and 2, and see Fig. 3(B,C)]. The significance of all
initial introduction in the testing apparatus, each fish was comparisons was established using independent samples
separately accommodated to the new environment for 2 Student’s t test and all graphics were generated using Ori-
entire days (days 1 and 2) and was afterwards recorded in gin v.7 (OriginLab). All manipulations were performed
one 15 min trial on day 3 using Noldus Ethovision v.2.3 according to authorization N8AkZ 209.1/211-2531-20/02
system(NoldusInformationTechnology,Netherlands).The from the Government of Upper Bavaria and the German
preferred compartment was defined as the compartment in InstituteforDrugControl(Bundesopiumstelle).
which afish spentmost timeonday 3, and thebasal level
ofPP(PPi),servingasabasisforallfurthercomparisons,
LearningandMemoryTests
wascalculatedasthepercentageoftimethatthefishspent
in the preferred compartment during the 15 min recording
Learning and memory tests were conducted in a T-maze,
onday3.PPilevelsbetweenachesb55/þandtheirwild-type
followingaproceduremodifiedfromDarlandandDowling
siblings were measured in the set-up described above and
(2001). The experimental tank was composed of one long
compared using independent samples Student’s t test; all
(46cm)andtwoequallyshort(30cm)arms.Allarmswere
graphics were generated using Origin v.7 (OriginLab,
6 cm wide and 15 cm deep with a water level of at least
Northampton,USA).
10 cm. One of the two short arms opened into a deep
(20cm)squaretank(23(cid:1)23cm)withblackwallsandarti-
ficial grass to offer a favorable habitat for the zebrafish.
D-Amphetamine-InducedCPP Mostofthefishtestedspentthemajorityoftheirtimeinthis
compartmentoncetheyfoundit.Twodaysbeforetesting,a
AfterPPdetermination(day3),eachfishwasweighed(typ-
groupof10fishwasrestrictedwiththetransparentsliderto
ically, weights varied between 0.5 and 3 g per fish) then
thelongarmoftheT-mazefor1hperdaytoaccommodate
intraperitoneally injected with D-amphetamine (40 (cid:1)g of
tothetestingenvironment.Afteraccommodation,eachfish
D-amphetamine;Sigma-Aldrich A0922, Germany) and
was placed alone at the beginning of the long arm and the
methylene-blueasatracer(3(cid:1)gpergramoffishin110mM
time needed to find the deep compartment was recorded.
NaCl, if not otherwise indicated) and immediately con-
Afterreachingthedeepcompartment,thefishgottheirdaily
finedtothenonpreferred compartmentfor45min(day4).
feeding.Eachfishwastestedonceaday.Weconsideredthat
Restrictiontothiscompartmentwasachievedusingatrans-
a fish had learned the task when its time to find the target
parent slider such that visual contact with the preferred
compartment varied less than 10% upon consecutive trials.
compartment was possible, to minimize the difference
The learning ability of fish with different genotypes was
between the conditioning and the measuring tank. The
comparedusingindependentsamplesttestandallgraphics
experimental tank and conditions were otherwise identical
weregeneratedusingOriginv.7(OriginLab).
to the ones used for PP determination, and each fish was
tested alone. After 45 min, the fish was removed from the
experimentaltankandkeptina1.5Ltankonacolor-neu- BrainD-AmphetamineLevels
tralbackground.Onday5,thefishwasinjectedintraperito-
neally with a saline solution (3 (cid:1)g of methylene-blue per Thebrainswereremoved,putonice,weighed,andhomo-
gram of fish in 110 mM NaCl), then restricted for 45 min genized in 100 (cid:1)L of plasma-spiegel solution (0.12%
into the preferred compartment. Between each injection H PO , pH 3.5 adjusted with 6 M NaOH). Amphetamine
3 4
session,theexperimentaltankwascleanedwith70%etha- extractionanddeterminationofamphetamineconcentration
nolandrinsedwithfishfacilitywater.TheD-amphetamine werecarriedoutaccordingtoanHPLCprotocoldeveloped
treatmentwasrepeatedondays6and8andthesalinetreat- byLabmed(Dortmund,Germany).
ment on day 7. PP was then measured again on day 9.
Repeatingthesalinetreatmentonday9,followedbymeas-
AChEandAChAssays
urement of PP on day 10, lead to identical results (not
shown). Unless otherwise specified, conditioning was esti-
The brains were removed, put on ice, and weighed. AChE
matedasthechangeinPPbeforeandaftertreatment,rela-
activity of brain extracts was determined according to Ell-
tivetothePPbeforetreatment,asfollows:%ofchange¼
manetal.(1961)andmeasuredastheamountofAChthat
100(cid:1)(Ppi(cid:2)PPf)/PPi,where%ofchangeistherelative
isbrokendownbyAChEpergramofbraintissuepertime
changeinPP,PPiisthepercentageoftimespentintheini-
unit. ACh concentration in brain extracts was measured
tially preferred compartment (measured at day 3, see
with the Amplex Red Acetylcholine/Acetylcholinesterase
above), and PPf the percentage of time spent in this same
AssayKit(MolecularProbes,Eugene,OR)accordingtothe
compartmentaftertreatment(measuredatday9or10).For
manufacturer’sinstructions.
percentagesoutsidetherange30–70%,thecalculationof%
of change was followed by arcsine transformation (as rec-
ommended in Hogg, 1995) [Figs. 1(C) and 2(C)]. In all Electroretinography(ERG)
experiments, a comparison with scoring the percentage of
change in absolute values (PPi (cid:2) PPf) was performed, ERGs were performed on adult zebrafish as previously
leading to identical conclusions [see Supplemental Figs. described (Makhankov et al., 2004). Briefly, the recording
JournalofNeurobiology.DOI10.1002/neu
466 Ninkovic etal.
electrode was positioned on the surface of the cornea. A light. Before exposure to light, they were adapted in com-
chlorodizedsilverwirewasfixedontheoppositenostrilas pletedarknessforatleast15min.A1slightstimuluswas
areferenceelectrode.Toensureoxygenationoftheanimal, chosen with an interstimulus interval of 10 s. Illumination
MESAB medium (Sigma–Aldrich, Buchs, Switzerland) wasincreasedin1.0logunitstepsovertherangefrom(cid:2)5
wasflushedbygravityforcesoverthegillsbyaplastictube logunit(0.5cd/m2)to(cid:2)1logunit(5000cd/m2).Unattenu-
inserted into the mouth. The fish were dark-adapted for at atedlightintensitywasmeasuredatthediffusersurfaceof
least 30 min before mounting the electrode under dim red thelightguideoverthefishheadbyalightmeter(Tekronix
Figure 1
JournalofNeurobiology.DOI10.1002/neu
AChEandAmphetamine-Induced RewardinZebrafish 467
J17; Texas Instruments, USA). ERG responses were aver- the rewarding effect of amphetamine with an initially
agedthreetoseventimesdependingonthesignal-to-noise nonpreferredenvironment,recognizablebyvisualcues.
ratio. Statistical analysis was performed using SPSS v.11 To establish the ideal visual parameters, we built on
(SPSSInc.,Chicago,IL).
previous reports demonstrating PP of adult zebrafish
foronesideofatwo-colortank(Serraetal.,1999;Ger-
laietal.,2000;DarlandandDowling,2001).
RESULTS
Reliability and robustness of a CPP test means it
must meet a number of requirements. First, before
Amphetamine-CPPinZebrafish
beingconfrontedwiththedrug,theanimalhastohabit-
Amphetamine can trigger a conditioned change in PP uatetothetestconditionsuntilitsPPtoonesideofthe
in higher vertebrates, but has only been tested in tele- tank becomes stable. This level (PP after habituation,
osts (goldfish) in an associative paradigm (Lett and orPPi)istheonlyreliablebasetomeasurechangesin
Grant,1989).Thuswefirstinvestedinsetting-upareli- PP following administration of the drug. Second, a
able amphetamine-conditioned CPP test in zebrafish. two-colorcompartmenthastobedevelopedwherePPi
Our test was based on remembering the association of isnotexcessivesoastopermitreversaluponadminis-
Figure 1 Establishment of reliable test conditions to measure D-amphetamine-induced reward in
zebrafishadults.(A)Placepreferencemeasurementset-up(leftpanel,viewedfromtop)andrepresen-
tativevideo-recordedroutefollowedbyawild-typefishinthisset-upatday3(rightpanel,example
with 58% of the timespent in the brown compartment). (B) Habituation curve. Percentage of time
spentinthebrownversusthewhitecompartment[see(A)]in15minmeasurementsover5consecu-
tivedays (one trialperday).Eachbarrepresentstheaverage offive fish,withstandard errors indi-
cated.Thevariationsinthepreferenceoffishforthebrownsidebetweenconsecutivedayswerecom-
paredusingindependentsamplesStudent’sttest.Notethatthetimespentinthepreferredcompart-
mentsignificantlydecreasesuntilday3,afterwhichtimeitremainsstableatavalueof55–70%(until
atleastday10,notshown).Thisinitialdecreasemightreflectincreasedexploratoryactivityofthefish
asitgetsusedtothetesttank.Thepercentageoftimespentinthepreferredcompartmentatday3(PP
after habituation, later referred to as PPi) is taken as a basis in all subsequent measurements. (C)
Change in place preference induced by the intraperitoneal administration of 40 (cid:1)g/g of D-amphet-
amine (left) paired with the nonpreferred compartment, compared to the administration of saline
(right)inidenticalconditions.The‘‘changeinplacepreference’’(yaxis)ismeasuredastherelative
differenceintimespentinthepreferred(nonamphetamine-paired)compartmentbefore(PPi)andafter
(PPf) drug exposure (in percentage of PPi). This change was compared between n amphetamine-
treatedandcontrolfishusingindependentsamplesStudent’sttestfollowedbyarcsinetransformation
(Methods),standarderrorsareindicated[seeSupplementalFig.1(A),middleandbottompanels,fora
survey of the response of individual fish]. Note that amphetamine-injected fish significantly revert
theirplacepreferencetochoosetheamphetamine-pairedcompartmentafterdrugexposure,whilefish
injectedwithsalinedonotreverttheirpreference.Identicalconclusionsarereachedwhenscoringthe
absolutedifferencebetweenPPiandPPf[seeSupplementalFig.1(B)].(D)Changeinplaceprefer-
ence(yaxis,left,redcurve)andmortality(percentageoffishdyingduringtheprocedure,yaxis,right,
dottedblackcurve)asafunctionofthedoseofamphetamineinjected(xaxis,salineinjectionsindi-
catedasthezerodose).Eachpointisbasedonthetestofatleast15fish,withstandarderrorsindi-
cated;statisticalsignificancebetweentheresponsesatdifferentdoseswascalculatedusingindepend-
entsamplesStudent’sttest(seevaluesonthegraphandtablebelow).Ourexperimentalconditions
producea dose/change responsecurve following a Gaussian distribution similarto that observedin
mammals (Tzschentke,1998), where doses above a certainthreshold (here50 (cid:1)g/g) induce a toxic
response.(E)Changeinplacepreference[yaxis,left,redcurve,sameasin(D)]andbrainamphet-
amine levels (y axis, right, dotted black curve) as a function of the dose of amphetamine injected
(x axis). Each measurement of brainamphetamine levels is averagedfrom threefish, with standard
errorsindicated;statisticalsignificancebetweenthelevelsofamphetamineinthebrainatconsecutive
dosesinjectedwascalculatedusingindependentsamplesStudent’sttest(seevaluesinblackonthe
graph). Note that the level of amphetamine received in the brain is a linear function of the dose
injectedwithintherangeofvaluesusedinourexperiments.Notealsothatitkeepsincreasinglinearly
abovethethresholddoseinducingmaximumCPP,confirmingthatthedecreasedCPPabovethisdose
isduetoatoxiceffectofamphetamineratherthanitseliminationfromthezebrafishorganism.
JournalofNeurobiology.DOI10.1002/neu
Figure 2 Loweredamphetamine-inducedCPPinachesb55/þheterozygotes.Inallcases,fishwithdif-
ferentgenotypeswerecomparedusingindependentsamplesStudent’sttest,andstandarderrorsareindi-
cated.(A,B)AlteredbrainlevelsofAChEactivity(A)andACh(B)inachesb55/þheterozygotes(left)
comparedtotheirwild-typesiblings(þ/þ,middle)orABcontrols(right).Eachvalueisanaverageofn
fish.BrainlevelsofAChEarereducedbynearly50%(A)andbrainlevelsofAChareincreasedby40%
(B)inachesb55/þheterozygotescomparedtotheirwild-typesiblings,whichdonotdifferfromABcon-
trols.(C)Changeinplacepreferenceinducedby40(cid:1)g/gamphetamineinachesb55/þ heterozygotes
(red,left),theirwild-typesiblings(þ/þ,red,middle),andABcontrols(red,right),alsocomparedtothe
effectofsalineinjectionsinachesb55/þ(black,right)andAB(black,left).Thechangeinplacepreference
(yaxis)ismeasuredandstatisticallyevaluatedasinFigure1(C),andeachvalueisanaverageofnfish.
Amphetamine-inducedchangeinplacepreferenceissignificantlydecreased(1.6times)inachesb55/þ
heterozygotescomparedtotheirwild-typesiblings,whichdonotdifferfromABcontrols,andsaline
injectionshavenoeffect.Identicalconclusionswerereachedwhenscoringthechangeinplaceprefer-
enceinabsolutevalues(SupplementalFig.2).(D)Compareddose-responsecurveinachesb55/þhetero-
zygotes(dottedline)andABcontrols(redline).ThepercentageofchangeinPP(yaxis)isrepresented
asafunctionofthedoseofamphetamineinjected(xaxis).Eachpointistheaverageofatleast15fish.
Statisticalsignificancebetweentheresponsesofthetwogenotypesatagivendosewascalculatedusing
independent samples Student’s t test (standard errors indicated, see significance values on the table
belowthegraph).achesb55/þheterozygotesrespondedtoamphetaminefollowingaGaussiancurveof
similaramplitudetothatofwild-typefishbutshiftedtowardsloweryvalues(centerpositionforwild-
typefish:65.3;forachesb55/þfish:62.3).Thus,thedoseof40(cid:1)g/gamphetamineselectedasoptimalfor
wild-typefishisalsooptimalinachesb55/þheterozygotes,whichrespondedwithlowerCPPthanwild-
typetothecompleterangeofamphetaminedosestested.
AChEandAmphetamine-Induced RewardinZebrafish 469
tration of an optimal dose of drug. In higher verte- wild-type fish significantly reverted their PP when
brates, conditions where PPi to one side approaches administered with 40 (cid:1)g/g amphetamine (n ¼ 22)
55–65%workbest(Tzschentke,1998).Third,thedose [Fig. 1(C)], while 97% of control fish (i.e., injected
of drug has to be determined as the best compromise only with a saline solution) failed to do so [n ¼ 21;
betweenahighresponseandanacceptablerateofani- p<0.001;Fig.1(C);seealsoSupplementalFig.1].
mal survival, as determined on a dose/response-sur-
vivalcurve.Inaddition,tomakesurethatthedrugacts
via its impairment of brain function, one has to ascer-
ReducedAmphetamine-InducedCPP
tainthatthelevelofdrugreceivedinthebrainisinpro-
inZebrafishAChEMutants
portiontothedoseadministered.
Because previous studiesdid not agree onthe pref- In mammals, lowered brain ACh signaling has been
erence of zebrafish adults for brightness or darkness associated with an increased propensity to get ad-
(Serra et al., 1999; Gerlai et al., 2000; Darland and dicted to psychostimulants (Gerber et al., 2001; Ichi-
Dowling,2001),wefirsthadtofindappropriatecondi- kawa et al., 2002). To determine whether the choli-
tionsyieldingreliableresults.Wethusdesignedanum- nergic system was involved in a similar regulatory
ber of two-color 3 L tanks and tested more than 20 pathway in zebrafish, we measured the sensitivity of
wild-type adults in each. We selected a light brown zebrafish with genetically impaired ACh metabolism
versus white contrast [Fig. 1(A)], where a test on 50 towards the rewarding effects of amphetamine. Ze-
wild-type adults demonstrated that more than 95% of brafish achesb55 mutants harbor a point mutation in
these fish spent between 55 and 70% of their time on the AChE-encoding gene, resulting in the production
thebrownsidefollowing2daysofhabituation(ondays of a nonfunctional AChE enzyme (Behra et al.,
1and2;notshown).Wefurtherobservedthattheper- 2002). Because AChE is the only ACh-degrading
centage of time spent in this preferred compartment enzyme in zebrafish, achesb55/achesb55 homozygous
remainedstableafterday3[Fig.1(B)].Wethusmeas- embryos are completely deficient in ACh hydrolysis
ured PP at day 3 as PPi and started our conditioning and die of progressive paralysis at early larval stages
procedureatday4.Insuchconditions,morethan90% (Behra et al., 2002). achesb55/þ heterozygotes, how-
oftheanimalstestedsignificantlyrevertedtheirPPfol- ever, reach adulthood without obvious locomotor or
lowing three intraperitoneal administrations of 10 (cid:1)g any morphological defects (not shown; n > 50). We
amphetamine per gram of body weight (on days 3, 5, found that AChE activity was decreased by nearly
and 7), each time paired with 45 min in the nonpre- 50% in the brains of achesb55/þ heterozygotes com-
ferred compartment and separated by daily injections pared to their wild-type siblings or our AB control
of saline (on days 4 and 6) paired with the preferred strain [Fig. 2(A)] (n ¼ 9; p < 0.05), resulting in a
compartment [Supplemental Fig. 1(A), top panel]. 1.4-fold increase in their level of brain ACh [Fig.
Control fish injected every day with a saline solution 2(B)](n¼9;p<0.02).Immunohistochemicalanaly-
alternativelypairedwiththewhiteandbrowncompart- sesfailedtorevealdefectsinthelocationandorgani-
mentsfailedtoreverttheirPP[SupplementalFig.1(A), zation of dopaminergic-, serotonergic-, and cholines-
bottompanel]. terase-positive brain neuronal clusters in achesb55/þ
Next,wedeterminedtheoptimaldoseofdrugper- heterozygotes, suggesting that, in spite of their per-
mitting robust CPP at acceptable toxicity. In a dose- turbed levels of AChE and ACh, these fish do not
response study, we observed that doses of amphet- suffer from grossly abnormal neuroanatomy (not
amine below 5 (cid:1)g/g did not lead to a reproducible shown). These results suggest that achesb55/þ mu-
change in PP, while doses above 50 (cid:1)g/g progres- tants are a valuable genetic model to test the effects
sivelyreducedthischangeduetoimmobilizationand ofincreasedbrainlevelsofACh.
stress of the animal and/or death [Fig. 1(D), each To measure the impact of lowered AChE activity
point >15 fish]. Similar observations have been on drug-induced effects, we assessed the behavior of
reported in mammals (Tzschentke, 1998). Thus, we achesb55/þ adults upon administration of 40 (cid:1)g/g
selected a dose of 40 (cid:1)g/g, which triggers robust amphetamine in the conditions described above. We
change in PP but low mortality [Fig. 1(D)]. We fur- found that achesb55/þ heterozygotesexhibit asignifi-
ther verified that at this value, the dose of amphet- cantly lowered change in PP compared to their wild-
amine received in the brain, measured by gas-HPLC type siblings: while the average change in siblings is
on brain extracts within 10 min after injection, was a 24%, it is 14% in achesb55/þ, representing a 46%
linear function of the dose injected [Fig. 1(E), each reduction in the response [Fig. 2(C), red crossed bar]
point ¼ 3 fish]. In these conditions and using the (n ¼ 49; p < 0.05; see also Supplemental Fig. 2).
experimental paradigm described above, 95% of Again, siblings did not differ from controls of the
JournalofNeurobiology.DOI10.1002/neu
470 Ninkovic etal.
Figure 3
JournalofNeurobiology.DOI10.1002/neu
AChEandAmphetamine-Induced RewardinZebrafish 471
wild-typeABstrain[Fig.2(C),redstippledbars],and of PP (a high PPi may interfere with conditioning);
achesb55/þ fish injected with saline, like AB fish, (ii) the capacity of the animal tested to learn and
failed to modify their PP [Fig. 2(C), black bars]. A remember which compartment was associated with
dose-response curve confirmed that the dose of thedrug;and(iii)thecapacityoftheanimaltestedto
amphetamine administered (40 (cid:1)g/g) was also trig- appreciate the visual cues identifying the drug-paired
geringthemostrobustrewardingeffectstolerablefor compartment. To rule out an involvement of these
a reasonable survival rate in achesb55/þ fish [Fig. factors on the lowered CPP response of achesb55/þ
2(D); each point > 15 fish; note that achesb55/þ fish heterozygotes, we assessed their basal levels of PP,
respondwithasignificantlylowerCPPthantheirsib- memory,andvisualperformance.
lingsoverarangeofdosesreachingatleast60(cid:1)g/g]. We used the PP test to measure the PPi of
We conclude that the cholinergic system strongly achesb55/þheterozygotes,asdescribedabove:after2
modulatesaddictivebehaviorinzebrafish. days of habituation, the percentage of time spent in
the preferred compartment was recorded. Although,
like wild-type fish, most achesb55/þ heterozygotes
still preferred the brown side (not shown), we found
GeneticImpairmentofAChEFunction
thattheirpreferenceforthiscompartmentwassignifi-
IsNotAssociatedwithanAbnormally
cantly decreased compared to that of their wild-type
HighInitialPP,withLoweredMemory,
siblingsorABcontrols[Fig.3(A);n¼54;p<0.05].
orwithVisionDefects
Thus,achesb55/þfishdisplay arelative greater initial
Three parameters unrelated to reward per se might preferencefortheconditioningcompartment,making
artificially bias the CPP response: (i) the initial level it unlikely that their reduced change in PP following
Figure 3 LowerPPi,increased learning potential, andunaltered visionin achesb55/þ adults.In
allcases,fishwithdifferentgenotypeswerecomparedusingindependentsamplesStudent’sttest,
and standard errors are indicated. (A) Preference of achesb55/þ heterozygotes (left) for one com-
partmentofthetesttank,comparedtothatoftheirwild-typesiblings(middle)orABcontrols.The
percentageoftimespentinthepreferredversusnonpreferredcompartmentofthetankafter2days
ofhabituationismeasured(PPi).Barsrepresentanaverageresultfornfish.NotethatPPiissignif-
icantly smaller (p < 0.05) in achesb55/þ fish. (B,C) Correlation between the percentage of time
spent in the initially preferred compartment before conditioning (PPi, x axis) and the absolute
changeinplacepreferenceafterconditioning(PPf(cid:2)PPi,yaxis)inwild-type(B)andachesb55/þ
(C)fish.Eachdotrepresentsanindividualfish.Alinearregressionanalysisshowsthatthecorrela-
tionfactorbetweentheabsolutechangeinPPandPPiisbelow0.2forbothgenotypes(r¼0.17,n
¼31,p<0.3forABandr¼0.19,n¼33,p<0.3forachesb55/þ),indicatingthatthesevaluesare
not linked.(D,E) Learning capacity of achesb55/þ heterozygotes compared to their wild-type sib-
lings(þ/þ)andABcontrols(AB).(D)Timeneededtoreachadeepchambertarget(yaxis)asa
functionofthetrialnumber(onetrialperfishperday).Eachpointistheaverageofatleast30fish,
with standard error indicated. Both achesb55/þ and þ/þ fish decreased their time upon trial, thus
progressively learning to find the target (see Supplementary Movies online). Note also that they
started from comparably high values (>100 s) and reached a plateau at comparably low values
(20s)afterenoughtrials.Thusourresultsarenotbiasedbyadifferentialspeedofswimming.How-
ever,achesb55/þheterozygotesreachedthisplateauaftertrial3(arrow)whileþ/þsiblingsneeded
six trials (arrowhead). (E) Compared number of trials required to learn the position of the target
chamber(yaxis)inthedifferentgenotypes;eachbaristheaverageofnfish(seebars).Weconsid-
eredthatonefishhadlearnedwhenitstimetofindthetargetvariedlessthan10%uponconsecutive
trials; the first trial of this plateau was then considered as ‘‘learning trial’’. Note that achesb55/þ
heterozygotes(left)learnedinsignificantlylesstrialsthantheirwild-typesiblings(middle)orcon-
trolfish(right)(p<0.05).(F,G)Functionalityoftheretinalnetworkofachesb55/þfishcompared
totheirwild-typesiblings(þ/þ),measuredbyelectroretinography.Typically,theevokedresponse
consisted of an initial negative deflection (a-wave) followed by a large, positive component (b-
wave)afterapplyingalightstimulus(Dowling,1987).(F)Amplitudeoftheelectricresponsefora
stimulationof5000cd/m2(top)and0.5cd/m2(bottom)inonefishofeachgenotype(color-coded).
Note that the curves are similar in both genotypes. A range of stimulation values was tested
betweentheseextremesandshowedasimilarresponseinachesb55/þandþ/þfish.(G)Compared
amplitudeoftheb-waveresponseinachesb55/þfishcomparedtotheirwild-typesiblings(þ/þ).
JournalofNeurobiology.DOI10.1002/neu
472 Ninkovic etal.
amphetamine administration was due to their differ- such as anxiety, addiction, and social interactions
enceinPPicomparedtowild-typefish.Insupportof (Engeszeretal.,2004).Todate,twostudies,focusing
this interpretation, we further verified that, for both respectively on ethanol and cocaine, used zebrafish
achesb55/þ and wild-type fish, the absolute change in adults to approach the neurogenetics of drug addic-
PP following amphetamine administration was unre- tion (Gerlai et al., 2000; Darland and Dowling,
latedtotheinitialPPilevel[Fig.3(B,C)]:thecorrela- 2001). Importantly, in contrast to these early works,
tion factor between the absolute change in PP and we failed to reproduce efficient drug administration
PPi in a linear regression analysis was below 0.2 for by dissolving drugs in water (Darland and Dowling,
bothgenotypes.Theseresultsfurthersupportourcon- 2001):insuchcases,inourhands,theamountofdrug
clusion that the lowered change in PP of achesb55/þ reaching the brain was undetectable (as measured by
fish upon amphetamine injection was not related to HPLC, see Methods) (not shown). Rather, we suc-
theirbasalpreferencelevels. cessfully used intraperitoneal injections. We also
Wenextassessedthelearningcapacitiesofachesb55/þ foundthatahabituationphaseofatleast2daystothe
heterozygotes.WebuiltasimpleT-mazeassaywhere test conditions was necessary to obtain reliable
thefishhadtoreachadeepchamber,afterwhichper- assessmentofPPandCPP.Finally,ourmethodelimi-
formancethey received afoodtreat (seeMethodsfor nates scoring behaviors unrelated to the reward path-
details).Onetrialwasperformedperfishperday.We way per se but rather reflecting changes in the stress
observed that the time taken to find the chamber status of the tested individual, or impaired drug
decreased upon trials in both achesb55/þ heterozy- uptake or transport to the brain. It was crucial, for
gotesandwild-typesiblings,showingthatbothpopu- instance,toverifytheamountofdrugreceivedbythe
lations were capable of learning [Fig. 3(D), and Sup- brainforeachsubjectshowingatypicalbehavior,and
plementalMovie1on-line].Weconsideredthelearn- to verify the normality of other parameters involved
ingphasetobeoverwhenthetimeneededtofindthe such as memory and vision. In a CPP test using the
compartment changed less than 10% upon consecu- so-called ‘‘biased’’ place conditioning procedure
tive trials. We observed that achesb55/þ fish needed (i.e., with a preexisting bias for one area of the test
less trials to reach this point than their wild-type sib- arena, as we used here), anxiety might also affect
lingsorABfish[Fig.3(D),arrowandarrowhead,and placereversal.Hereweruleouttheinfluenceofsuch
Fig.3(E);n ¼20;p<0.005],whilethe timeneeded a phenomenon by demonstrating that the PPi of
to reach the chamber in the initial trial and after achesb55/þ mutants is not greater than that of their
learning was comparable in all cases [Fig. 3(D), see siblings [Fig. 3(A)], which should permit condition-
time for trial 3 in achesb55/þ, for trial 6 inwild-type, ing. Finally, because achesb55/þ and wild-type fish
and see Supplemental Movies 2 and 3 online]. The differintheirinitialplacepreferencelevel(PPi),ver-
location of the chamber was memorized for over 2 ifying that the PPi value does not influence CPP is
weeks after a pause in testing in both genotypes (not anotherimportantcontrol[Fig.3(B,C)].Together,the
shown).Weconcludethat,inthistest,achesb55/þfish reliabilityofourset-upwasinstrumental indetecting
learn faster than controls. Thus, their reduced CPP is psychostimulant-induceddifferencesintheCPPpara-
unlikely to result from an inability to learn and digm between different genetic backgrounds such as
remember. achesb55/þheterozygotesandwild-type.Thissuccess
Finally, we assessed the visual performance of demonstrates that zebrafish adults can be used to
achesb55/þ fish by ERG. The ERG b-wave was used screen for the effect of developmental mutations on
toevaluatephysiologicalfunctioninvivo.Atwo-way adult reward-related behavior and therefore to iden-
repeated measure ANOVA indicated that the b-wave tify dominant modulators of behavior related to
amplitude of achesb55/þ heterozygotes was not sig- addiction,aswehereidentifyAChE.
nificantlydifferentfromtheirwild-typesiblings[Fig. Another major result of our study is the implica-
3(F,G);two-wayANOVA,n¼3each;F(df¼2)<1, tion of the cholinergic system in the modulation of
p < 0.7]. These ERG tests together with the T-maze the rewarding properties of amphetamine, of PP, and
assaysuggestthatvisiondefectsareunlikelytoaccount ofcognitivecapacitiesinzebrafish. Becausethispar-
forthereducedCPPinachesb55/þ. allels the situation in mammals, our results provide
thefirstvalidationofthezebrafishmodelforstudying
the neurotransmitter and molecular pathways that
DISCUSSION underlie the process of addiction in vertebrates. This
result is of high significance given the demonstrated
Thezebrafishbehavioral repertoireis complexand is amenabilityofthezebrafishsystemtogeneticscreens
increasingly studied for the modeling of parameters andthemolecularmappingofmutations,atalevelto
JournalofNeurobiology.DOI10.1002/neu