NOREPINEPHRINE-INDUCED POTENTIATION IS NOT BLOCKED Br THE LOCAL DIFFUSIONOF THE NMDA CHANNEL BLOCKER, KETAMINE,
AT CONCEHTl<A.TIONSTHAT PREVENT TETANICALLr-INDUCED LONG-TERM POTENTIATION IN THE DENTATEGYRUS
IlIYIYll
BY
o LYNN MARIE FRIZZELL
A thesissubmitted to the School of Graduate Studiesin partial fulfilment of the
requirementsfor the degree of MasterofScience
Departmentof Psycholoqy MemorialUn iv ersityof Newfoundland
June 1994
St.John's Newfoundland
. +.
NaliOnalUbrary 0'"""",,~~"""'"
~~5UM
Kt A1;N4
~ubl~~naliooale Direcliondes acquisitionset des$8rvices bib/iog~s
""M_
~f(lrWaITIlEAU11IOR HASGRANrnDAN IRREVOCABLE NON·EXCLUSIVE LICENCE ALLOWING TIlE NATIONAL LmRARYOFCANADATO REPRODUCE, LOAN,DiSTRIBUTEOR SELLCOPIESOFIIISIHER TIlESISBY ANYMEANS ANDINANYFORMOR FORMAT, MAKING
nus
TIlESIS AVAILABLETOINTERESTED PERSONS.TIlEAU11IOR RETAINS OWNERSlDP OFTIlECOPYRIGHTIN IIISIHER TIlESIS.NEITIIERTIlE TIlESISNOR SUBSTANTIALEXTRACTSFROMIT MAYBE PRINTED OROTIlERWlSE REPRODUCEDwrrsour IIISIHER PERMISSION.
L'AUTEURA ACCORDE UNEUCENCE IRREVOCABLE ETNONEXCWSIVU PERMETTANTALABmLlOTIlEQUE NATIONALE DU CANADADE REPRODUIP.E,PRETER, DlSTRmUER OUVENDREDES COPIESDE SA TIlESE DE QUELQUEMANIEREET SOUSQUELQUEFORMEQUECESOIT POURMETTREDESEXEMPLAIRESDE CEITE TIlESEALADISPOSITIONDES PERSONNEINTERESSEES.
L'AUTEUR CONSERVE LAPROPRIETE DUDROITD'AUTEUR QUIPROTEGE SATIlESE.NILATIlESENIDES EXTRAITS SUBSTANTJELS DECELLE·
CINEDOIVENT ETREIMPRIMES OU AUTREMENT REPRODUITSSANSSON AtJrORlSATION.
ISBN 0-612-018 54- 7
Canada
Anenduring increase 1n the size of the perforant path- evoked dentate qyl'US response follows brief tetanic stimulation. Thb Long-Term Potentiation (LTP) is considered a cellular substrateof memory. A similar phenomenon, Norepinephrine -InducedPotentiation CNEP), results from exogenous application of HE or from endogenous HE release following Locus Coeruleus(Le) stimulation. It has been postulated that NE promotes LTP induction in a manner consistent witha role in arousal and attention. It has bee ndocumentedboth.in Y..1..tnl:andl.n Y.iY2that LTP induction requires the NMDA receptorand that NEP induction requires activation of the e-edrenecatc receptor. It.fact , i t hasbeen shown in the slice that LTP and NEP dependon bothrQceptors. This suggestsa codependence or syn ergy between NEP and LTP in the slice. However,the samehas not been demonstratedin~. In tAct, LTPisnot eliminated by HE depletioninY.1Ys;l, thus the 8-receptoractivationis probably not required. It was the aim of the present study to determinethe role of the NMDA receptor in NEP.in~.
REP was elicited by glutamatestimulation of the Lc, Two recording pipettes were placedin the dentate gyrus. One pipette contained saline and the other, loc'atedless than lmm away, contained the NMDA channel blocker, ketamine.
Diffusion of ketalllinesignificantly attenuated LTP as compared to that at the saline site. However, NEP was
11
either unaffected or enhanced&II compared to that at the saHne dte. KEPin~does not depend on the HHDA channel. 8-receptor activation, and the cAHP-mediatel1 biochemical cascade triggered, appears to be necessary and sufficienttorthe induction of NEP. It NE promotes attention lind memory through the 8-receptor in the dentate gyrus, it does 80 independently of the NMDA receptor.
iii
Acknowledqelllcnt.
Thank-you to .y aupervisor, Dr. carolyn Harley for thh opportunity, for all that you have tauqht lIle and for your encouraqelllent and enthusiaslIl. Thank-youto .y coauaittee lIlellbera, and. teachers,Dr. Robert Adamecand Dr.John zvene, Thank-you to all the staff and ff.llowatudents(Itour bUSy laboratory,pastand present, for ta kingtime tohe l p . Thank-you to lilymot he r , Judy, my husbandBob Antle and all my family andfriends for your infiniteenc ou r age me nt and support. Last but not least, thank-youto NSERe and The HUH Alumni Associationfor financial support received.
iv
ABSTRACT AClCNOWLEDGEMEN'l'S LISTorTABLES LIST OF FIGURES INTRODUCTION KETHODS RESULTS DISCUSSION REFERENCES
CONTENTS
page ii tv vi vii
.7 6.
76
, ..
LIST OF TABLES
Table 1. The effect of increasing concentrations of ketallline on the magnitude of LTP of the popspike as compared to a saline control: A pilot of 6 animals.
Table 2. The effect of ketamineon basel!r,e evoked pot,entialparameters and magnitude of potentiation ascompared to sallne ccnecct,
v1
LIST OF FIGURES
Figure1.Transverse view of the hippocampus. Shading represents the<::e l 1 layer8 and includes illustrations of the pyramidal cells of Ammon's Horn and the granule cells of the dentate gyrus.
Figure2.Asagittal view of the rat brain illustratingthe location of the Locusccerureus recording/ejecting pipette and of the two dentate gyrus pipettes. The similarity in the two dentate gyrusevoked potentials is also evident.
FIqure3.The magnitude of LTP evident on the pEPSP (A) and popspike (B), at the ketamine site, is signifIcantly attenuated compared to that at the saline site.
Figure 4.,The r eisno between-Elitedifference In REP of the pEPSP (A) but REP of the popspike (8) is significantly enhanced by ketamine.
Figure 5. The 10 availablesites from the 11 animals are illustrated. All recording/ejecting siteswere within 400PD!of the Locus cceeureus ,
vii
He bb (19 49 ) pr oposedthat an increasein theeffici en cy atsynaptictrans .bsionwould cons t i t ut e lea rningat the cel lUla r lev el. Kore than 20 year safter Hebb first pr e s e nted his th e ory, Blis9and Lomo, inthe ana e s thetize d rabbit andBliss and Gard ne r-M e dwin,in tha awake rabbit, in a pai r of 1973 stUdies, de mo nstr ate daphenomenon wh i c hqav B He bblsthe o ry 8u~. ta nce (Gus tafasonand Wiqstr6111, 1988) . These st udiesshowed thata br i e f hiqh fre que nc ytrai nof st bm l a tion, delivered to the in put pat hway, produced a sign ifica nt lindenduring increas e in the effic i e nc y of a group ofsynapses inthe den t ate gyrusof the hippocam pus• ....tte r the te t anus , the qr a nu l e cellsof the dentateshowe d gr e ater sensit i vi ty tosubseque nt low freque ncy(s Ingl e pul s e) stimula t i o nofth at input pathway . The effect could be Bustained fo r lIIa nyho ur s. Bli s s andLomo arbitra ril y defined 30 lIinute. as sufficicmt todifferentiatethis phenome no nfrOD augmen ta tion, facilitationand post- te tanic potentiat i on. Bliss and Lomo (1973) , calledthis chang ein cellulllr beha v iour long-lastingpot-Inti. tion, nowIlor e commonl yknown as Long -T e m Po t. e n t iation, (LTP).
While LTP can beprod uc e d elsewhe r e in thebrai n , nowhere is th is form of syna pt i c pl asticity4E1ea sU y, reliably and dra ma tic.:'!II ll y produc e d as in thehippocampus
(CollingridlJ e and Bliss , 1987). In fact, LTP canbe produc e d in any ofthe thre e synaptic areasof the hi ppocampa l Itt r isynapticci r c u it-. Althoughthe hippocampus
(He) ha. beentraditionallyreferredtoaa thetr b y naptic circui tand bee n thoug h t of asa.implerelBYst ation, it i.
nov Jtnovn tha t the HCh .uc h1Il0re coa pl ex. In the yor ds ot Swans on et d. (1987 ), "whilethe hippocampal fona t i on cont a in s the si mp lest cortica l fie lds•••it receives , processes andtransllit s the most complexarra y ot intonationotany corti cal region". Thus,while theHeis st ruc t ur a llysImple andthebul k othippoc ampalthroughput foll oW8 the tris ynllopti cloop, directcor tica l input islIIad e to all three areas at theHC and a greatdealof intrinsic recipr ocalinn Qrv a tionoccurs among the s eregions,both within the loc al cIrcui t andalong thelongitudinal axis (Amaral and witte r, 1989, Yeckel andBerge r, 1990 ). Thus, conve rg i ng vith in the Heisa compl exarr ayat info rma t i on, includingsensoryin to rma tion frolll pri maryand as s oc i ation cort ices asveIlu modulato ryinpu tsfrom subcortica l sites such .s thecholinergicbas a l for e b r ai n , theserotone rgic raph enUClei, th e dopaminergic ve nt r alteqmentalareaand the norad rene rg i c locuscoerul e us. Rather tha n acting as a sllDple::8 1a y station, the He functionsinthe synthes is , mod u lationand elabo r a tionotinc oming into rmation, before
"relaying" it back to thecortex . Thus , while the cortex is the putativesite for permanentmemory st ora qe , the He is thought integral in forming the memor i e s (Morris, 1990) .
Given the proposed integrative function of th9 HC and the properties of LTP, it became a cellular model of mG.mory (Morris, 1990). Initially, LTP attracted favour by virtue of the tact thatitIS properties fitted the predictions of Hebb, that it was discoveredin the HC, a structure then known to be vital to new learninq (Milner, 1972) and b..cause the time course of LTP was such that it was "potentially useful for memory storage" (Bliss and Lomo, 1973). The HC has since proven critical forso me kinds of learning and memory, 'ds pe cia lly spatial learning and memory. The principle cells of the He fire in response to an animal's position relative to a constellation of external cues, (O'Keefe and Nadel, 1978), and the integrity of these cells is necessary torlearnin~and memory at spatial maze tasks (McNaughton et al., 1989). Furthermore, preventing the induction of LTP prevents spatial learning (Morris et al., 1986). Such findings have made LTP the most widely accepted and exploited model of som1! kinds of learning and memory.
Animportant ramification of the use of LTP as a cellular model of memory is that it has led researchers to uncover- other putative modulators of hippocampal plasticity.
The tocus of the present study is the modulOlotor norepinephrine, (NE). Bringing HE to the forefront of research into hippocampal plasticity was Neuman and Harley's 1983 discovery of the phenomenon of norepinephrine-induced
lonq-Iastinq potentiation (NE-LLP) inthe anaesthetized rat.
As in hiqh-frequency induced L'I'P (HFT-LTP),a long lasting cha nge in the size or dentate granule cell responses was induced, butnot by HFI' electrical stimulation.
Iontophoreticapplication of HE to the granule cell body layer produced a potentiation that, like HFTeLTP,outlasted the induction st i mulation and was often qreater than 30 minutes in duration.
As with LTP, norepinephrine-inducedpotentiation(HEP) is now recognized as an important player in hippocampal plasticity. The ev i de nc e , to be reviewed,shows that HE is releaseddurlt:gar ous e d andattentive behavioural states.
Theef f e c t of HE on sensoryevoked unit and population responses is described as enhancement or facilitation which functionally meansthat the responsesnow have more physiologJCal impa ct. Thus, theory holds that the role of NE is to prime or enable the Heto perform itIS integrating and consolidatingfunctions on thp. arousing,attention- gettinq input. The processesotNEP and LTP are thoughtto undl9rl1eth~sefacilitatinq and memory-makingfunctionH.
Due to a number of discrepanciesbetween the properties of LTP and NEP, it isnotre t certain whether HE promotes other forms of plasticity, such as LTP, via the i.Ti' mechanism,or i tit promotes LTPby some other mechanismor finally, if NEP is a mechanistically distinct form of potentiation.
The 111m of the present research was to investigatethe degree to whichthe mechanisms of NEP and LTP convergein Y.1Y2. The impetus for this invr.stlgation was the tinding that, in theinYi.tDlpreparation,NEP and LTP depend on the same mechemism (Burgard et al., 1989,stanton et al., 1989). To date thishas riot been demonstratedinYm. The purpose of the present experiment ....as to determ.ine , inY..1..Y..g, whether REP depends on the LTP induction mechanismor is capableot inducing a potentiation independently ofHFT-L'l'P.
The remainderof the introductio nprovides an overv iew of hippocampal anatomy and physiology,with particular attention to thedentate gyrus. The anatomy and physiology of the norepinephrine-containingnucleus locus coeruleus (LC) isreviewed with emphasis on suggested LC function;
namely. a general role in mediatingcortical arousal. Emphasis is placed on whatisknown of NE innervation of the dentate gyrusand on the effects of HE on dentate gyrus physiology. The effect of NE on unitresponses, popul."Uon responsesand on the L'l'P of populationresponseswill be ex;;mined. What is understoodof NEP and LTP will be det."iled as necessary . sInc e eninY.1.Y.gana l ys i s ot the degreeot convergence ot me chani s mbet....een NSP "'nd LTP is the aim of the presentst udy , comparisons between LTP and NEP will conclude the introduction and set the stage for the present research. As the present W'ork is an extensionotin
xllDlfindings, differences betweeninn.t.xgandin rlY2 preparations are highlighted.th r ough out.
The ten hippocampal tormation includes four simple cortical regions, the entorhinalcortex (Re),the subicular comp l e x , (Wh i ch includes the pre- and parasubiculUlll and subiculus proper),th€<dentate gyrus and Almcon's Horn, (Amaral and Witter,1989) . The term hippocampus refers only to the lattertwo simplecor ti c e s . Thehippocampal formation, in it's three dimensional extent, formsaC&hape originating justcaudalot the septumand terminatinginthe temporal lobe just caudal and medial to the amygdala. ThUs the long axisofthe hippocampalformation is descriptively called the septotemporal axis. (Amaral and Witter, 1989, sveeeenet.at. , 1987).
Investigators of the hippocampal fanation often conceptualize it in the transverseplane, perpendicUlar to the septotemporal axis, asa lamella 01." slice. Figure1 depicts a traversesection of the HC and illastratesthe three synaptic regions and theirmairi connections. Viewed transversely, the HClooks like two interlocking C shapes, the larger (more lateral) being the pyramidal cellsof Ammon'sHorn, includingareas CAl and CAJ and the smaller (more medial) being the granule cells of the dentate gyrus.
Cells of the EC funnel cortical inputinto the HC via their axona,knowncollectively as the perforant path (PP),which
synapse onto the dendrItes of the granule cells of the dentate gyrus. The granule cells, 1n turn, project their mossy fibres to synapse on the pyra'lddal cells of area CA3.
Finally, the Schaffer collateralsot the CA] pyramidal cells make synaptic contact with the pyramidal cells of CAL Traditionally, it was thought that all hippocampal throughput follo....ed this trisynaptic route and did so within II thin transverse section, and this belief led to the development of theinili.r..2preparation. While not disputingthe usefulness of the slice preparation to certaIn inquiry, nor the usefulness of the term "trisynaptic circuit" to describe the main intrahippocampal connections, Amaral and Witter have taken exception to this classical view of the HC. They have shown that Intrahlppocampal projections, with the exception of the dentate mossy fibres, do not project primarily within a lamella, but in fact, innervate targets along a large portion of the septoti!mporal axis. Consideration of this pointiscrucial where comparisons are made between tbe findings yieldedinYi.t..t:Q versusinilig, since the slieG preparationisdevoid of those projections. For the present work, thla consideration is central.
The first synapse of the classical hipPoclIImpal triad, the perforant path(PP)-dentate gyrus (OG) synapse, where LTPand NE-LLP were first discovered is the locus for the
' 1 ' ;~ram~~al
coli':,
Ammon's Hom-s' "
./ '\ '
./ ~~"
".. : Baskelce,(
....-p., . - . \,,,-:..:.:
L ; : , " . . . . ~~
_.,-,.~~,. -\ ,
G.ronUI.cell' • . . ··' ~3
~~
DentateGyrusFi gu r e 1; Transverse view of the hippocampus. Shading represents the cell layers and includes illustrations of the pyramidalcells of Ammon' 8 Horn and the granule cella of the dentate gyrus.
present study. The perforant path (PP) fibres arise primarily froID. layer two cells of theEC, the source of cortical input to theHC. The ECisdivided into hteral (LEC) andmed ial (KEC) portions, retlecting the preference shown ror one part or tbe otber by some corticalinputs.
The ECreceivesprimary sensoryand multilIlodal associatlonal Informationfrom olfactory bulb and areas tbat receive a direct projection from olfactory bulb, and from temporal, prefrontal, ci ngulat e and insul arregions. si nc e tbe pre subl.culr1lllterminatesexclusivelyin theHEC, inputs that travel vi a the presubiculum , suc h as that from visual co r t e x areas17 and ~8balso terminate exclusivelyin tbeHEC.
conversely, the bulk of olfactory input has been sholtln to terminatepreferentiallyand directly In theLEC (Swanson at.al.,1987).
This medial/lateral informational divergenceis also reflectedtopographically in the pattern of termination in the dentate qyrus. The PP fibres oriqinating trom tbe NEC synapse in the middll!! one third of the granule cell dendritic molecular layer while thosearI singin the LEe innervatethe outer one third. Furthermore, a functional dIfference is also ob servable , LTPcan be produced in the dentato gyrus in response to HFT stimUlation of eitheL·the medial or the lateral PP (Bliss and Lynch, 1988) but, NEP is produced only in the medial PP. Cells respondingto the
lateral PP exhibit HE-induced lonq-lastinq depressions (Dahl and Sarvey, 1989). This isjust one point at "'hich the phenomena of LTP and HEP diverge.
The EC a180 serves as the ultimate output target of the HC. output from the LEe reaches the entire extrahippoclIlllpal cortical mantlfJ including the olfactory, temporal aUditory and somatosensory cortices. Innervation of these areas from the MEC is present but sparse. Both parts of the entire EC
"inne rva t e motor and medial prefrontal cortex. There are also, ncvever , direct projections to cortical regions from subicular complex and Ammon'sHorn and projections from AmmonIs Horn that bypass thesub i c u l a r complexto terminate in the EC (swanson ee , a1., 1987). Thus the exceptions to the trisynaptic circuit theory and the lateral/medial topographical divergence discussed in relation to hippocampal input, also apply to hippocampal output.
The s1D:plest of the cortical areas of the h.ippocampal formation, the dentate gyrus, contains three main layers;
the molecular (dendrItic) layer, the granule cell layer and the polymorph layer or hUus and two main populations of cells; the principle output cells, called the granule cells, and the local circuit interneurons. It is the bodies of the granule cells that define the C shape of the dentate. As mentioned, the granule cells receive input from the perforant path in a topographical manner where LPP and HPP
10
terIlllnate in the outer and middle thirds or the molecular l<!lyer, respectively. Found in the inner thirdot the molecular layer are the terminationsot hllar Ineerneurene , granule cell recurrent collaterall!l and c01lllllissural tlbres from contralateral interneuronE!. The collective output of the granule cells,the 1l10SSy fibres. make en passage exci tatory synaptic contact with hilar interneurone berore terIllinatlng In the CAl area.
The interneurons of the dentateg'Yrus are scattered throughoutthe three layers but do concentrate in the hilus.
The four main interneuron types are the GABAergic basket cells, the glutamatergic mossy cell s, the somatol:>tatin i1lllllunoreac:tive(-IR) aapiny cell" which are also 90%
GABAerglc and the calretinin-IR spiny cells, for which the neurotransmitter remains unidentified.
The basketcellli6.r8found in the molecular layer but are more dense in the hilus and especially the subqranular zone of the granule cell layer (Ribak and Seress, 1983). Basket cells receive excitatory'PP input directlyin the molecular layer and respond to this input at a lower threshold than do the granule cells (Scharfman,1991). Thus PP excitation can inhibit gran\llie cells through a feedtorward circuit beforedlrEtctlyexciting them. BaSkdt cellsalso receive recurrent excitatory input from the granulecell mossy fibre collaterals thus placing the
11
granulece l t- und er feedb a c kinhi b i to ry<.:ont r ol as veIl (Sch ar ban, 199 1).
The 1I0esycellsof thehilus are qluhlDatergicand.
there f oreexcitatory(Sc harf man, 1991). These interneu!'~"!J rece ivedire ctPPinpu t at1Il01ecula r la yerdendrites and also re s pond to this inputat alo wer thr esholdthan dothe granu l e cella (Sc h arf man, 1991). Si nc e theseinterneurons ar e excitatoryandsendaxons backintothegra nUlecell dend ri t iclayl!r , feedfo n.'a r dexci ta tionissugge st e d. The ev i d e nc e sugge sts, howe v er, tha t the moss y cell s probably act throughothe r interneurons sincethe i r activity is associatedwithqranul e cell inhibition(Misgeldet aI., 1992 ). Moss y ce ll s also receive mos sy fibr e excitatory input and therefore IDa ymediatefeedback effect s.
TheeOlllato s t a tin- I R aspi nycells , located in the hilus , have de ndri t es in the JIIOlecul ar lay e r and receivedirect pp exc i t a t o ry inpu t aswellasmos9Y fibre input (Leranth , 1990). Giventhat 90' are GABAerq i c , thesece lls pr o babl y inhi b it the gr anule cells throughfe ed forvardand feedback mechanisllIs.
Thecalretinin-I R sp iny cellsdonot proje ctdendrites Into the lIolecular layer and thereforeprobablydo not receivedire ctPP input. These cellsdo re ceive excitatory lIlossy fibreactivationand termInateon the granulecell dendrites andthere foreare pla cedtomed iat e feedback
12
effects on l y. No tran.a itter has be e nidenti!1ed ,and it has yet to be shownwhet herthisci r cu i t b inh i b i t ory or excitatory (Miett ine n , 1992) .
I tis noteworthythat wh i le basketcell inne rvationb concentrated nearthepare ntcell, innerva tionof granule cellsat di sta nc e salong these pt ot empo r el axlealso exieta.
Mossycells make nocont actneartheparent cell and in tact mak e contact onl yat distanc es sev eral mil limetresfr omthe pare nt ce l l (Amaral and Witter , 1989 ). Theseareexa mp l es of theelementa of den tate circ uit rythat are missing from th e.inill1".Qprepa ra t ion and which de mandconsiderationin comparisons madebetwe e nfindi ngsyi elded from.in:d.txsI:
vers usin~pr ep ara t i ons.
In BUJII" built Int o the denta t e local circui t ry is the
aeeneto requlate gran u le cell exc i tab i litythroug htonic inh i bit i o n (through sp ontaneous activation) andfeedforward an d feedbackinhibitio n and excit ation. Influe nCRII on these local circuitneur ons have been prop os ed ae themechan i •• by which someputativeneurol\Od.ulators l114ymodifyhi ppoc a mp al cellu lar exc itabilityandrespons ivity (Custafssonlind Wiqstrolll. , 1988). In tac t ,it hasalreadybeen de monstrated thatdisinhibitIon can produce apote ntiationin the dentate gyrusat rat, inthe slice, which is long-lasting (Burgard and Sarvey, 1991).
13
Unlike the cognitive cortical input, which funnels into the He by way of the EC, brainstem and forebrain inputs, such as the noradrenerqic locus coeruleus (LC), terminate directly and dittuseiythroughout the HC. It was due to this diffuse patternot innervation that NE was first consIdered. a modulator. purportedly involved in some generalized function.
The main source of NE in the CNS is the nucleus locus coeruleus. Forming a distinct clusterof neurons in the dorsal ponti...e brain stem at the ventrolateral edge of the fourthve nt r i c l e , the LC also contains peptides, such as neuratensin and vasopressin and neurotransmitter-related enzymes, such as dopamine-8:-hydroxylase and tyrosine hydroxylase (SuUn and Jacobowitz, 1991). The LCisin a fibre rich area where axons of neiqhbourinq modUlatory neuron clusters course nearby (Foote et 0111., 1991). This fact was an important consideration in selecting the methodology of the present stUdy.
The LCisthought to receive itls strongest affarent innervation trom two other bralnstem nuclei, the nucleus paraqigantaceUularis, (PGl) and the nucleus preposltus hypoqloss! (PrH) (Aston-Jones et al., 1991a). The PGi provides a powerful, most probably qlutamaterqic, input to the LC and itisthe OOi that activates the LC In response to some kinds of sensory stimUlation, such as tailor
14
footpinch (Aston-JoneB et a1., 1991111). The PGi isalso active in cardiovascular and respiratory functions and receives input from areae involvedin autonomic action.
which prompted the proposal that the LC may be the cognitive cOlllpl€llllent tosympa the t i c functioning (Aston-Joneset a1., 1991b). In contrast to the OOi, the PrH provides GABAergic inhibitoryinput totc (Aston-Jones €It1111., 1991b). In cat and primate this nucleus has been implicatedin thecontrol of eye movements, a role that ha s obvIous importa n ce to an attentiona1 system, (Aston-Jones €It a1., 1991a, 1991b) . The diffuse output fromI.ereaches sensoryprocessingareasof the cortex, tectum, cerebellum and the dorsal horn of the spInal cord. There is,however, little LC innervation of th e brainstem. Noncoerulean nuclei provideHE to brainstem and motor nuclei (Grzannaand Fritschy, 1991).
Within the dentate gyrus, HE innervation is densestin the hilus where HE terminals may form axoscmaticsynapses (swanson et a1., 1981) which suggests termination on interneurons. By teninating on interneurona, HE would be positioned to achieve modulatory effects through influences on the local feedback and feedforward circuitry.
Norepinephrine tet"lllinals have also been reported in close apposJ.tion to dendrites in the molecularlayer (Swansonet al., 1981) ",hereHE can mediatp. direct effects on granulo ce l l s.
15
eur iou.ty, tbeadrenerq icre ceptordistri bution pa t te rn , in ten_of dens ity, bears little res e mbl a nc e to the pattern of HE innervation {c rutcbe r and Dav ia , 1980, Booze et al., 19U}. Whilea-r ecep t ore .ore closely corr e l a t e withHE ta r-inaIsites, in the Heacre of theHE ef t ect. studiedtodat e seemto bemed i a t ed by B-re cept ors (seqal et aI. , 1991). The a-rec eptor, UkeHE, ismost abund a nt In the hilue andlDolecu l er layer. The S··rec e pt ors, J,oweve r , arenot Illost abunda nt inthehilus, but are found in themolecu l arlayer. B-rec optornumbers arerep ort ed ly equ a l in thedentate and Alblllon1sHorn, however theaI- rece p tor s aretwi ceas numerous indenta teasin CAl (Crutcherand Dav ia ,1980). It is note 'Worthy tha t the rC-NE ay s te1ll is underautorequla tory con t r ol via th e actionof presynapticGa-receptore, 'Whichexiston HEte r1llin ala.
Antag oni s motthe se receptors increases HE release'Wh ile activa tion decreas esNEre l ease.
since the 8-receptor me diates most otthe HE effects atudiedto date , it J.iJapo r ta nt toconsider hOW'the 8- receptoroperatell. The 8-receptorisnot associat ed. 'With an ion channel but with a GTP-bindinq (G-) pr otein'Which activates an intr ac ellular se cond me s s enger sys tem. When HE attaches to the transmitterrecognit ionsiteon the &- receptor, the G-protein isactivated, in turn activati ng adenylate cyclase . Adenylate cycla s econvertsATP to cAMP.
16
ItiscAHP that accumulates intracellularly and acts as the second messenger that achieves HE effects. Second messongera in general, including cAMP in HE effect., can initiate phosphorylation of membrane-bound receptor proteins and ion channel proteins (Sarvey et a1., 1989,Abrams and Kandel, 1988). By altering membrane associated receptors and channels, second messengerssuch as cAMP producechanges in cellular excitability and respClnaivity that underlie the actions of neuromodulators such asNE.
When is NE released? LC neuronslireactive during arousal, as defined behaviourally or by the presence of the
",aking EEG pattern. LC neurons are spontaneously active, (Williams et 41., 1991) and this basal firingisaltered by input, for example frOID the PGi and PrH. Increases (excitation) and decreases (inhibition) in LC activity correlate with behavioural state and environmental events
(Aston-Jones at aI., 1991a). Thus, activityof LC cells varies overthe sleep-waking cycle, showing a continuum of increasing excitability corresponding to a continuum of increasing behavioural alertness or attentiveness (Aston- Jonea et aI., 1991b). The LC cells are active during wakinq, quiet during slow wave sleep and silent during paradoxical sleep. The continuum of LC activity is even seen within the ",aking state, increasing by degree with increasing attention to the environment. During the waking
17
stat., automatic behaviours, suc h as grooming, are associated with a low level of activity in ther.c. A stimulus that interrupts automatic behaviour and induces an orienting,or attending, response is associated with the greatest ar:tivation of the rc. The kinds of stimuli that best activllte theLe are unexpected or are novel or are meaningful, in that thl)y hold some environmental significance. (e.g. a stimulus that predicts reward), and
"typically cause both a sympathetic and behavioural orienting response- (Aston-Joneset aI., 1991a). Thus. the greatest release of HE denotes the presence, in the environment, of important, novel or unexpected etimuli. HE release indexesbr~aviouralstate.
Studies correlating LC firing and behavioural state have rElcently gained experimental support. Administration of the LC-activatinq cholinergic agonist, bethanechol, to rats under surgical levels of halothane anaestf'..sfe produced activation of a majority of LC neurons and the appearance of 8 waking pattern in the hippocampal and frontal cortex EEG record. When bethanechol failed to induce LC activationno waking EEG appeared. It was concluded that LC activation was crucial in mediatingcortical activation,however it was also speCUlated that I.e-HE is sufficient but may not be necessary for such activation (Foote et al., 1991).
18
The correlational and experimental link between I.e tiring and behavioural state supports the early contention, based on little more than 'diffuse anatomy', thatthe I.e-HE system 111 involvedin the global function of mediating attention or arousal. It followsthat the LC-HE system should also influe ncetarget functionsthat depend on arousal and attention to the environment. particubrlythose functions thatdepend on attention tonov e l or me aningful enviroJUllentalevents. The impo r t a nc e of suc h a fnnctionto the HC, a target structure involved in learning and me mory , isobvious. The tc re sponds intens elytone wand meaningful stimUli, Which may becauseof their meaningand novelty , be worthyof remembrance. Perhapsthe I.e-HE systom functions to prepare the He for thisimportant input, pos siblyprbling or even mediating the mechanismsthat underlie learning and memory. The evidence to date suggests induction of NEP may be the llIeans by which the I.e-HE system promotes learning and memory in the HC.
Intheir 1991 review, Sara and segal presentevidence supporting a role for the I.e-NE attentional enhancementin enhanced hippocampal functio n. They proposed that the learning situation is optimal for LC activationbecause, in a learningsituation, environmental stimuli are novel and the significance (or meanIngfulness) of stimuli are changinq. As stated, I.e neuronsar e most active When
19
stimuli are newor meaningful or when a shift in attention isrequired. Accordingly, it has been sholffl in one experi.ent thst in awake, chronic LC-recorded rats, in learning' situations where a reward was first attached to a low tone and then reverseet and associated with a high tone, the greatest LC activation was correlated with the shifts in CS-reward pairing. That is, LC activation was seen at the point where the low tone and reward were first paired and Again when the reward was first unexpectedly withheld. The LC was activated again, immediately and vigorously, when the r9versal occurred and the hJgh tone was now paired with reward. This reversal task illustrates the elements of novelty and shifts in attention that activate the rc. Also, this experimer.tillustrateshow the LC responds to the meaningfulnessata stimulus. The significance of the stimUlUS, and thus the propensity for th.e LC to fire, increased at the points where the relationship between the stimulus andfoodrewaret shifted, where something could be learned. A number of other investigations (reviewed in Sara anet segal, 1991) have shown that the LC activation that occurs In the learning situation is also important for the learning. Volriou8 pharmacological manipulations demonstrate that increasinq tc tiring facilitates learning in tasks that require shifts in attention, a response to novelty or successive discrimination reversals. These effects all
20
appear to de pend on thepos t syna p t ic 8-adre nerqic receptor.
In fac t, the fac U i t a t i nqeffectis. bicted by A-agonist.
and b preven tedbya- b lockade . Ins ofar
a.
the HC b invo l ve din thesekinds of learni nqtask. ,thisevidflnce sugge sts tha t theLC-NE systell is iIIlportant in promoting le a r ningand.emoryintheHC. Stud yof thecellul ar effec t s of HE, towhich thedi s c u ssionwill nowturn,al so suppo rts thispro posal.In aseri es of ear ly expe ri ments in whichthe effe ctot app li ed or synapt i cally releas ed HEonsing l e CAl cell ti ri ng waseval uated HE behaved IJke an inhi bitory neurotra n s mitter, (Sega l andBloom, 1974 " , 19 74b, 19 76).
Howeve r , ot he r re s earch has shown thatNEexc i t e s the inhi b itoryinterneu r on a (Rose and Panq, 1989 )the reby pre s u.ably increasi ng theinhIbitorytone of the dentate . Theseef fe cts can.not;be easily re c onc ile dwi t h the putative role of HEin promoti ngarousal, attention and learn i ng . CUr rent controve rsysurroundi ng thediffere n tiati onbet ween the phys iol ogical_iqnatu resof the vari ous interneurons questionsinterneuron excit a t i on findi ng_ butalsoprec ludes an accuratere portotHEeffects on interneurons at this til:le, furt!ler researchis needed . Hore recent
investigations have als';re ve aledthat the effectsotHEon granulecella are morecomp l e x than simpleinhibition. It has been showntha t in granulecells , the effect of HE
21
dependeupo n the receptorac tiva t e d . At lov level., NE excite. qranulecell. via the &-receptor, a result.t-icked by 8-aljlonbt.whU.at hiqher levd., NE, at the (II-receptor mediates inhibition, a. do Ql-aqoniat. (Rose and Panq, 1989, Lacaille and SchvartzJcroln, 1988). Itisprobablethat, at the level.otME administeredin Dost experiments, the (11- receptorinh i bit o ry effect mas ks the lov dose 8-lllediated exc i t ato ry efCect. He nceth e early conclusionthat HE acts
8San inhibitory neuro t ra nsm i t te r. It ha sbeensU9qested thouqh , that the a-mediated exc i t a t i o n , at least in CAl pyramida l cells, isprobably the predomInanteffect se en vithendoge nous release(Muelleret 81., 1982). At theve ry least it canbeconcludedthat the neteffectot HE on epceeenecaecellularactivity in thedentate may depend upo n the balance of(1- and 8-receptoractivation(Stanton et al. , 1989)•
The net ettecteeNE on stimulus-evokedneuronal activityalso dependson the balanceof(1- and&-r e c e ptor activationand hasbeencha r ac t e r ize d as an increase in siqn41tonoise ratio (Foote at al., 1991,Mc Co rmic k et al., 1991, Har l ey, 1987). This relative increaseappears to occur a. 4resultof little orno suppre ssionofthe response signal4qalnst more suppres sionotthe sp o ntaneous baCkgroundactivi t y or asa resultof enhancementot the signalag ai ns t le e s or no enhancementof the background.
"
The net eftectot thb complex,.odulatory actioni . to incr€la8€1 th€l phy .iolO<jicalstrength of theresponse (Foote
€It al.,1991). Footeet aI., (1975 ) werethef'iratto re portthis lIodul a to ry effect of HE. These re searchers foun d that respons •.l!Iot squirrel.onkey auditory co rte x neuronsto species-specificcalls wereinhibited by NE but less80 than spontan eousbac kgroundactivity, yieldinqan inc r e ase in signalto no iseratio . Asimilar effecthas bee nobserved in ne uronsof ratsomatos en sorycortex which res p on d toel ec t r i c a l stimulationof thohindlimb with in i t i a l activatio nthen an extended pau se . Bethanechol activation of the Le, whenpairedwithhi ndlimbstimUlation, resultedin asupprossi on of thespo nt a ne ousactivityand , to a less erdegree, of the activationandthe pause. The net effe ctvasanincr e a s e inthe dur ationand overall lIag n i t ude oftheactivation aga insta supp r essed backgrOUnd (Foo t e et al., 1991). It has alllo beenob served tha t HE enhancestheinhibitory and excitatoryresponsesof cerebellarcells elicited by GADAand glutamate, respectively. This a-cAMPmediatedeffectveIlill u s t r a t e s the modulatory natureof NE as di stincttromstraightforward inhibitionor excitation(Woodward €It aI., 1991). The impact of HE lIIodulation has also be en illustratedby oriontationdetectorsotthe vis ual cortex. Atter HE, orientationdete ctorsbec ome ec re ac t i v a t ed by optimal
23
stimuli, those closest to proper orientation. and less activated by stimuli that are lesa than optimal (KlIsllmatsu and Heqqelund, 1982). This narrowing of theresponse reflectathe strengthening of the focus on that stimUlUs, an effect useful to an attentionalsystem. The evidence shows that the consequence of HE releasein concert with sensory activation is to ensure the meaningful ornove l stimulus- induced sensorysi gnal more physiologicalimpact. These findingssupport the contentionthat HE functions in drawing thean i mal ' s attentionto meaningfulor novel stimUli.
Anatomicallyand physiologically, in its correlation with behaviour andin its effects on responses in target cells, the LC-HE systemdoessee m to Gubs ervee generalized arousing, activating function (Foote et al.,1991). What followsisanana l ysi s of the phy siologicaleffects of NE release on tho responseof a particular population of cells, the granule cells of the dentate gyrus. I tis well documentedthat HE, like high frequencyelectrical stimulation, produces a potentiationof dentate granule cell POPulation responses. The discussion will now turn to the physiology of the dentate gyrus evoked population response, the potentiationof this response, inthe form of LTP and HEP, and how potentiationis measured and what it reflects.
When the medial pertorant path is stimulated by single pulses the granule cells of the dentaterespondwith an
24
evoked population potential. This potential encompasses a number or aeasurable components which are used in the assessment or synaptic transmission and cellular excitability and of LTP or NEP. The three aain component.
are the population excitatory post synaptic potential (pEPSP), the population spike (popspike) and the popspike onset latency (latency). The pEPSP "is a measureof the currentwhich flows into the cells (thesi nk ) in the synaptic reqion of themo l e cu la r layerandout of the cells {the source] In the cell bodylayer" (Blis sand Gardner- Med",in, 1973.p. 373). Thusduringinitial synaptic activation a positive going wave is rec ordedin thecel l body layer. During cell firing. the sink and source reverse and so the recorded wave reverses. Thus, when recordedin thecell layer, the popspike appearsasa ne g a ti ve deflection superimposed on the pEPSP. The popspike amplitude retlec~Bthenumbe r and synchronyof cellstiring and therefore r.eflects the strength of synaptictransmission and the excitability of the popUlationof cells sampled.
The latency reflects the rapidity with which the cells are brought to threshold and thus ca n also reflect synaptic strength and the excitability or the receivingpopUlation (Blis8and Lema, 1973).
When synaptic efficiency is increased, such as might be expected when the input is new or meaning-fuland requires
25
remembrance, it i. predicted that incom1ng currentbecomes larger and 011 a result it is also expected that more cells tire and perhaps do 80 acre quickly and in greater untecn, ThUs, increases in synaptic etticacy and cellular excitability are retlected in increases in pEPSP and popspike amplitude and decreases in the latency. These are exactly the kindsat chang-es seen in LTP and NEP. Following- HFT or LC stimulation the slope of the pEPSP and/orthe amplitude of the popspike increa s e dramatically, and in LTP the latencydecrea s es. The popspikeincreaseis often greater than the change that can be accounted for by the pEPSP increase and in fact it can increasewithout any comparable change in the pEPSP. The fact that the pEPSP and popspike are dltferentiable and can bepotentia-ced independently b.plicates independent mechanisms and suggests that perhaps both pre- and post-synaptic mechanisms are involved in LTP and NEP (Bliss and Loma, 1973). Though simplified, what is currently known of the mechanisms of LTP and KEP will now be presented in brief.
Subsequent to the discovery of LTP. researcherssought a mechanis_ of LTP induction that acted essentially as a detector af coincidental pre- and postsynaptic acti.vation in accordance with this "es s entiallyHebbian" induction rule: irA El'ynapse will sustain an increment of potentiation if, and onlyif. it 18 active at a time whenthe regionotdendrite
2.
on whIchith locatedi • •trongly depolarized" (BU •• and Lynch, 1988, p.12). I th now known that the key to thb teature, the detector, is the N-methyl -D-aspartate (NMDA) receptor. The NKDA receptor is a unique memberatthe excf tatory amino acid receptor group since I t requires the satisfaction of two conditions before the channel opens.
First,';ll u t a ma t e must bind the neurotransmitter rec09'nition site and, unlike non-NMDA excitatory amino acid receptor subtypes, the post-synaptic membrane, on which the receptor is located, must be depolarized. Dp.polarizationrelieves a Mg2+block that exists in the channel at resting membrane potential. Hence, the NMOA receptor requires coincidental pre- and postsynaptic activation betore the channel can open. Due to the sustained activation achieved at the synapse in response to the LTP-inducing 8FT stimulation.
this coincidence can occur. Furthermore, because of this unique property, channels open only where this coincidence occurs and thus LTP is selective to the activated synapses.
Allot-ober important featureatthe NMDA receptor is that in addition to sodium Dnd potassium, it permits significant ca 2+entry. Research",rs have sho'om that the NMDA-mediated accumulation of intracellular free Ca2+J.Bnecessary for LTP induction and that in factisene "critical triqqer t'or LTP", (Nicoll et a1., 1988).
27
When t.TP 18 triggered the pEPSP and/or popspike can increase and the latency can eeeeeeee , However, not all three para.etera vill, necessarUy, exhibit potentiation in any on. expert••nt (811s. and LoIDO, 1973). Furthenore, it ecre than one at these indicators i_ present, thema g n i t Ud e of changes in each are not necessarily comparable..
According to 8US8 and Lomo (1973), in their pioneering study in the rabbit,the latencydecrease was the most reliable indicator of LTP. It isthe currentopinIonat many, however, that the pEPSP Inceeaseas a rstiection of synaptic events, is the truest indIcator of LTP (Gutaffson and Wigetrom, 1988). With the HFT stimulation parameters in current use, chanqes usually occur in all three parameters.
Theae properties of LTP h..Jld true for theinill.x:Q(sarvey . t al.,1989) and J.nUm (ErrInqton et a!., 1987) rat dentate. I tiaalao characteristicin~that the potentiation peaka and aSYJlptotes within the first 30 Ilinutes and invariably, t...'louqh over hours, days and even weeka, decays.Thepattsrn &00 intensity Clf tetanic still.ulation used in an experiment, will, to a great degree, detenine the Ilagnitude of the potentiation.
various phllt'1llllcological methods have shown unequivocally that blockade of the MHDA complex, eitherby receptor antagonists such as APV or CPP, or by channel blockers such as MIt-80l, pcp and Jl:etalline, prevents
2.
tetanically-induced LTP in dentate and CAl, bothin ill1:2 and in~, (Collinqridqe and Bli.s, 1987). rntereatin9ly, Harris and. Cohan (1986) have reported that APV doesnot prevent the inductionof idossyfibre-LTP in area CA3, a NHDA receptor poor reqion.
NMDA blockadestudieshave also revealedthat the processes underlying inductionand mainte nanceare differentiable. OnceLTP hasbeen inducedthe involvement of the NHDA receptor ce a s e s. If the NMDA receptoris blocked after inductionLTP expressionis not reversed, The processesturned on byNMDA-mediatedca2+influx ar e the sameprcceeeeeresponsible for theexpressionand maintenance of LTP (Collingridqe and Bliss, 19B7, Muller et al., 1988). Itisbeyond the scope of this stUdy to evaluate the many putative mechanisms of LTP maintenance, exceptto say that various pre- ar.dpostsynapticmechanisms have been implicated and that an as yot unidentitied retrograde messenger may coordinate theSQ trans synaptic mechanisms.
Norepinephrine induces an LTP-like potentiation of the dentate pEPSP slopeand/o r popspikeamp l i tUde,that, unUke IJrP, requires no concurrent HFTst imu l a t i o n. Early on i t was shown that electrical st i mu l a t i o n of LC could induce a potentiation of the popspike amplitude recordedin the somal layer of the dentate qyrus in the anae"thetlzed rat (Assaf
29
et aL, 1979, Harleyet al., 1982). Althoughthe Ha r leyet al. (1982) reportindicated. that the po t entiatIon couldbe lOIl9-la.tinq ,withoutp~operpharJIll col09Ical inve stigation the respo n.ibil1ty to r the ettect.co u ldnot be unequivocallyas s i g ne dtoNE. Intactit haa sinc e be e n shownthat elec trica l I.estimulation can prod uc ea pot e nt iatio nditte re nt trom NEP , proba b l ybecaus eot ext r ac oe ru learstimulation (Har leyet al. , 1989, but see Was hb u rn and Hois es , 1989 ) . Ne uman andHar ley's 1983 study was the tirst toillustratethat NE could lnduce a potent iat i o not theevo ked popUl ation pote nt ia l that mimicked LTP, lr~tha t it couldbe lon g-la sting , (>:'t 'l .inute" )• In th eden t ate qyrusotth e anaesthetiz e d rat , HE, lo ntophore sedtor1-8 minu tli! s, produceda potentiatio n ot thepop_pikaamplitude recordedat 41of 54sit e s in the cellbodylayer. At 16at 41si tes the po te ntiation was lo ngo- l a s ti ng. Thtaphe nome nonwa s te rmed NE-LLP, to differentiate it tro. 8FT-inducedLTP. NE-LLP develo ped gradually vi tha delayat 30seconds to 8 mi nutes . Like LTP, HE-LLP peaked and thenplateaued (t o .meanof130- 140' atcontrol),vithin the first30 minutesand inv a ria bl y showeddecay(Neumanand Harley, 1983)•
.Based on earlie r ,preliminary work (Harleyand NSUlllan,
1980), NeWllan and Harley (19B3) tentatively sugge s t e d that the B-receptor, and not the a-receptor, likely med i a t e d NEP,
30
LacaiUeandHarley (1915) wereable to continthb speculationand extend previous finding.by delllonstrating NEP in the dentate of thsinill1:gd iee. A 10 .inute bath application of the threshold level of NE(10~")on the aUce produc ed potentiation of the popspikeallpl1tude(131\ of control) in74\ atthe sUces teilted. an increasecompa r abl e tothat found withion topho r e d llin~ (130-140 \. NeWlan and Harley, 1983). In24 \ ofslice s potentiation ....as 10ng- lastir:t; (>30 minutes) as comparedto 40\inrl2..g (NeuMan and Harley, 198 3 ). In 53\ of theslic e s, a pEPSP increasewas rClported (117\ of control)and in74 ' at slices a latency decrease was seen (94' of control), effect. not reported.1.n llY2 (NeWllanand Harley, 1983). In2/38 and3/38 slices, the pEPSPand latency effects ....ere long-lasting. The pEPSP , latericy and popspikecha~eswereall .i.icked by a-agonist.
and blockedby 8-antagonists,but not by Q-agonistsand antagonists. Intact theincr e ase s weremore reliably produced bythe a-agonistIsoproterenolthan by HE. These Investiqatlonsshow that NE, In the de ntateqyrus. can Inducea potentiation ofthe popspikeelone or tog e t he r with thepEPSP which canbe long-lasting , likR LTP. Lacaille and HarleyalsoInd i c a t e d that both the short- and 10ng-last1:lg HEP appear todll!pend on the 8- a drener g i c receptur, dnceno potentiation, short- orlong-lasti ng, was Induced in the presence of 6-receptor bl oc kers.
31
stanton andSa rve y (1985b), using thein:Y.J.n:Q preparation, as didLacdlleandHa rley, we r e able to producea mare reliableMEP. Using30minutesupertus i o not
50 ~"ME,popspike potentiation (16' \ of contro l )was induced in all 8 slI c e s tested. Moreove r, inall 8ca s e s, HEP was lo nq - l a s t i nq, compare d with just24'at calles withlOpH, 10 minute exposure (taca1l1eand Harley, 19 8 5). The magn i tude of potentiationdurinq drug exposurewa s alao larqerwi t h 50pK, 30 minute (165\ ofco nt r o l) thanwith a 10#H, 10 minute superfusion(1 30-140 t ) . Thus the like l ihood of inducinq a long-lastingpopspike pote ntiation increaseswith prolonqedHEl"xpos ure to a higher conce n t ra t I o n of HEand may dependonthe initia l mag ni tudeof thepotentiati o n. While Stantonandsarve yma keno ment i o n of the pEPSP or li!ten cy in this stud y, in a later exper i ment i t was re port ed that 30.inut e 50~HNE perfusio nprod uc edlong-lasti nq NEP of the popspikeand the dendriti callyrecordedpEPSP (stanton and Sarve y, 1987 ) .
Return ing to theinY.b!.gpreparat i on,Ha r l ey and Milwa y (1986) establishedanother me thod forind ucing long - lasti ng dentateMEP. Usingpre ssure-ejecti on of 100-150nL of0.5H I-glutamateint o the rc, in anaesthetizedrats, an attempt was made to determinewhether lo ng-las t i n g NEPcouldresult fr omendogenously-released, physiological levels of NE.
This lIlethodof LC stimulation....as chosenoverelectrical
'2
since glutamate se19ctively stimulates cell bodieo and spares tlbres of passage (GOodchild et aI., 1982). The results showed that glutamate, within 300,1111 of the Le, produced an activation of the LC which invariably resulted in potentiation or the popspike. Enhancement1141'of control) occurred with a mean delay or1.1minutes and lasted 8.4 minutes on average. In 37t of the trials, the potentiatiol1 was lonqwlasting (>20 minutes). In 13/20 anbals the pEPSP increased (127\ control) in conjunction with the popspike, at a mean delay of 0.8 minutes and lasting an average of 3.2 minutes. In the majority of anbals, no change in popspike onset latency was t"eported.
The percent enhancement, the proportion of effects maintained for the long term and the inconstancyof PEPSP effects, are comparable with results obtained from experiments utilizing local NE methods. Also consistent with earlier reports was the finding that the LC-induced enhancement of the popspike depl!nded on the B-receptor.
Glutamate ejection in the LC at 20, 40 and 60 lIinutes after 30 mg/kg Lp. propranolol injections, either produced greatly attenuated popspike potentiation or none at all.
Interestingly, pEPSP changes ....ere not blocked by propranolol but were shortenecl from 4.2 to 2 minutesii\duration. since the propranolol injection alone lengthened the popspike onset latency. it ....as possible that NEP attenuatiO:1 or
33
blockade va. secondary to a peripheral effect. A8a.result LC-NEP va. tested against intradentata8-receptor blocker•.
Harley and Evans (U88) utilized pressure ejection of 8-blockers fro. pipettes located withinthe dentate, to test the role of 8-receptors inLC-induced NEP. Prior to tholol or propranolol,popepike enhancement (115-200'of control) was evidentin 14/14experimentsat a mean delay of 1.5 minutes. In 10/14 experiment s a concurre ntpotentiation of the pEPSPslope (105-11 5' control )wa s reported . In half of thocas es , NEP was long-lasting. Intradentateeje ctionof suffi cienttimol olor propranolol preventedI.e-induced effects, bothsho rt- and long-lasting, on pEPSP and popsplke. In agreement withthe prior LC stimulation study, (Harley and Hilway, 1986), but in contradiction to the slice fi nd i ngs (Lacaille and Harley, 1985)I.no effect on popsp:i.ke onset latency was reported. Thelack of 1ate.ncy effects also contrasts \I!thLTP, in whichla t e ncy decreasesoccur virtually invariably. The magnitudes of pot entiation effects and latenciesto their occurrences and proportions of long-term effects are,however, comparab l e toprevious reports, .bot hin~andin~.
Harley and Evans (1988) also reportedthat, as with the NMDA receptor in LTP, blockade of the 8-receptor, after the induction of KEP, did'not;attenuate or reverse HE:"'· Thus, like the NMDA receptor in LTP, the 8-receptor is essential
34
to the induction of HEP butisnot needed for the maintenance or expression of it. Thi. 1. consi.tent with the.1D~experillents where it haa been repeatedly shown tha t HEP outlasts NE washout. NEP, once induced by&- receptor activation, persists in the absence of further activation.
Just as ongoing 8-receptor activation i. not needed for NEP expression and maintenance, II recentexperill'~ntshows that lasting I.e cellular activation is also not necessary to sustainNEP, for the short or long term (Harley and Sara, 1992). Since glutamate-pre ssureejection proved reliable in selective Le activation and a-dependent NEP induction.
Harley and Sara investigated the effectot glutamate ejection on LC unit activity. Cellular activity was recorded from the same location am glutamate ejection, within the Le. For LC activaHon, pressure-ejection of SO- 100nL of O.SHglutamatewas used In some experiments While in the remainder, 100nL 1l1crosyringe injections were Ilade through implanted cannulae. Dentate evoked responseswere also recorded from the anaesthetized rats. Glutamate ejection into the rc produced an immediate and dramatic increase inI.ecellular activity. lIarley and Sar.a reported thatthe effect was llIost easily observed as an audible 'rush' on the audio-llIonitor. Thls rush of unit activity was very brier, lasting no lonqer than 250-400 .,8. The burst
35
wall followed. by an equally abrupt and dramaticai l en ce, a pause in activitythat lasted on average 4.6lIins before returning to baseline. with all 34 glutamate ejections, popspikepotentiation was observed at a llIean latency of 34 seC8 post-injection. The maximalpotentiationseen was 158\
otcontrol. The meanduration of potentiation was 4.4mins, although long-lastingHEP was also seen. HE-UP occurred more often\iith cannulae than with pipette-ejected glutamate . Potentiationof the pEPSP was evident in22/ 34 experime ntsata late ncy of 15 sees. Popspikeonsetlatency was typically unaffected,although an increasewas reported in 5/34ca s e s. With theexce p t io n of latencyincreases, these results are consistentwith previousrepo rts .
Although the duration of the pOflt-glutamatepause in LC activity (4.6JIlins) and the eeen duraUon of popspike potentiation (4.4mina) were quite comparable, there was in tact no relationship between them in individual animals thus.
refut i ng any suggestion that LC inactivationinduces NEP.
Furthermore, injections of the Q2-agonist, clonidine, which dlences the LC tor3- 5 minutes,produced no NEP (Harley and Sara, 1992).
Harley and Sara conclUdedthat REP induction depends upon brief, intense activationof icneurons and that a critical numberat LC neurons eceebe recruited. This id ea was supported by the observation that when NEP was induced,
36
there was always an audible rush ot LC units upon delivery of gluta1llate and vhen there vas no rush, there vas no NEP.
A glutamate..induced activation without concomitant NEP occurred vhen .maller amountsotglutamate vere ejected.
Larger voluml"l8productld larger unit responses, more reliable NEP and lIlore trequent NE-LLP. This valli revealed when ejectionotsmall qlntamate volumes induced no NEP in a given location, but an increase in vetuse resulted in NEP in that same location. Also, thecannula method, which resultedin larger localvolumes than pre ssure-ejection, produced larger and longer-lasting NEP. Based on these observations i t vas suggested that the likelihoodot inducing larqer And therefore 10nger"laBting eftects deperrde on the numberatLC units recruited. This is reminiscent of earlierin timfindings showing that 10 Ilinutesotl0l-lJlI NE supertusion produced HE-LLP24\ot the th\e (Lacaille and Harley, 1985) but 30 minute 50l-lm NE exposure produced a larger KEP that vas long"lasting100\ot the time (Stanton and Sarvey, 1985b). Since longer lengthsotexposure to higher concentrationsotNEin Y.1..t.I.2(Stanton and Sarvey, 1985b, LacAllle and Harley, 1985), and IS larger numberot activated LC unitsin~, all precHct larger and longer- lasting NEP,and since NEP is B-dependent, it may be concluded that the lIlaqnitude and duration ot NEP is, ultimately, determlnecl by the deqree, duration or numberot
37
&-recepton activated. Thie h a180 .upported by the subtractive ettect of propranolol. In a dose-dependent fashion, propranolol,by decreaeinq the nu.ber of participating a-receptors, decrea••• the .agnitude and duration of HlP. Largo enough doses block NEP altoqether.
The evidence seells to sho" that a-receptor activation is responsible for the inductiononly of KEP, bei tshort- or long-lasting. Given that the duration of the glutamate- inducedtccellular burst persisted just 250- 400 ms, "hile popsp1)ce enhancoment was 4.4 minutes in duration, ongoingLC activity, just Ukeongoing a-rec eptoractivity, is obviously not re sponsiblefor NEPexp ress i onand .aintenance. The duration of the effect seellls to dependon the strenqth or duration of induction .s supported by the finding. that an increase in HEvol ume or in intensityof LC activation results in larqer and longer NEP.
InSUlU&ary, the resultsof HEPinvestigations sU9q e s t that activation of the Le, such as that provided by novel 01' noxious stbuU,produces a brief and Ineenee LCce llu l a r response wbicb translates into a eaeefve but brief influx of HE into the dentate. Given sufficientLC activation, 8- adrenergic receptorsindUCe potentiation in dentate cells.
The deqree of LC sctivation determines the maqnitude and duration of NEP. It may also be that the numbers of a- receptors activated and the volume ofcAMPaccumulational s o
38
intluencea the likel ihoodof inducinqlarger and longer effects. continued NE release and a-receptoractivationar_
not needed to sustain KEP. Perhaps a-receptoractivation tr i gge r s KEP lIlaintenance but a-receptor involvementceases after induction, as does theNMDA receptor in LTP. It is not clear, just aa in LTP, what the mechanismsare that maintain HEP. It is known only that a-activated CAMP accumulationmay inducephosphorylationof proteina. The importance of proteinsy nt h es is to HEPisunderli ned by the findinqtha t inhibitorswhichpreventproteinsynthesis ef ec preventthe maintenanceof the latephas eof KEP (Stanton and Sarvey 1985b).
If, in the dentategyrus, LTP is the cellularsubstrl.lte of memoryand the role of HE is to promote attention and memory, by prblinq or enabling the induction of a plasticity mechanism,then it might be expectedthat HE \lould promote LTP. In fact, stanton and sarvey and their colleagues have compiled convincingevidence frolll a series of 10Y.1..tx.2 experiment:sthat NEPvi a the a-receptorand LTP via the NKOA receptor may actuallyact synerqisticallyto promote dentate plasticity .
The first step taken In determiningthe role of HE In LTP was a study of the effectof HE depletion on the integrity of LTP. Stanton and seevey found that in slices prepared trom HE-depleted animaln the frequoncy and
J'
amplitude or popspike LTP was greatly decreased (19858, 1987) A8 w•• pEPSP LTP lIleasl'l1red in the dendrites (1987).
They.concluded that without HE, 8FT-induced LTP was .virtually eliminated· (stanton and Sarvey, 1985a). Taking the experiment one step further, Stanton and Sarvey (1985a) examined the role of the 8-receptor in the apparenttlE lIlodulation of LTP. By blocking the noradrenergic 8- receptor, the same receptor that mediatesHEP induction, the frequency and the magnitude of LTP induction was attenuated in • manner similar to HE depletion. since HE activation of the 8-receptor, through adenyl ate cyclase, promotes the conversion of ATP to cAMP, a three fold increase in dentate cAMP content following bath application of NE waM expected (stanton and Sarvey, 1985c). However, in this uaee study it was found that LTP-inducing HFT stimulation also produced a comparable 2.5told increase in cAMP levels. Moreover, in slices prepared from HE~·depletedanimals it was found that
~ddin9HE to the,bath still produced cAMP inoreases, showin9 that the mechanisms underlying the increase were intact, however HFT-induced increases were abolished aU998stlnq an LTP dependency on NE-8 activation. Forthermore, in HE depleted slices, in which LTP could not be induced, the additionot torsJcolin, a direct adenylate cyclase activator, restoredHFT-in~uc8dLTP (stanton and saney, 1985a).
Finally, i t ,,"as also demonstrated that preventinq protein 40
synthesis, the likely consequence of cAMP accumulation, prevents the maintenance of HEP and LTP. By addinq the protein synthes!8 inhibitor, emetine, to t.he bath, the late phases 01:both HEP and LTP were eUminated. While the lIIaintenance IIIflChan181l1": involved in LTP and NEP havenot been identified, i thas been shown that induction of both HEP (Lynch and Bliss, 1986) and LTP (Lynch et aI., 1905) if.
followed by increasedgluta mate release . In sum, these findings ledStan ton and S" rve y toconc lude that. theH£-8 st i mu lated cAMP inc r ea seis a ne c e s s ary st ep in the induction of HFT-LTP in theinY.i..tl:gdentate gyrus. Thus the cooperationof the NMDA- and a-receptor-assoclated mechanisms, when activatedby HE or HFT stimulation, appear to be sufficientand nec essaryror LTP induction10Yitm.
The results of recent investigations lend support. and insight into the means by WhichHE may promote LTP via the NMDA mechanism. working in the slice, investigators beve IIhown that bathapplie d HE (Lacaille and schwartzkroin, 1988 and Stanton et a1., 1989) and a-agonists (Lacailleand Schwartzkroin, 1988) induce granulecell depolarization. It has ej.ec been demonstrated that when HE or isoproterenol wereapplied to the dentate slice at le vels8uffici entto indu c e NEP, a reduction of the after-hyperpolarizationwa.
evident. It was speCUlated that HE-Bactivationreduces a Ca2+-dependent potassiumconductance that underliesth~AMP.
41
The net ettect is a prolonged depolarization (Haas Ilnd Rose, 1987). Through bothC'~these eftects,depolarization and AMP reduction,HE aay enhance the opportunitiestor sathtactionatthe pre-and post-synaptic depolarization coincidence required to triggerLTP.
It has also been demonstrated that HE enhances the influxotCa2+through the NMDA channel (Radhakrlshnan and Albuquerque, 1989, stanton et aL,, 1989) as well as through voltage-gated Ca2+channels (Gray and Johnston,1987, Fisher and Johnston, 1990). Stanton et aL (1989) have shown that HE at concentrations that induced HE-LLP in the slice also induced a facilitation of stimulus-evoked ca 2+influx and potassiulll etnux. These effects were blocked by pretreatment with the NHOA receptor blocker APV. Furthemore, it was directly demonstrated that HE enhanced NMDA-induced changes in extracellular ca 2+and potassium ( r ) concentrations while having no effect on non-NHDA- induced changes. Gray and Johnston (1987) patch clamp- recorded an isolated inward ca2+current in exposed granule CIIU.troll. hippocampal slices. pressure-ejected HE increased the size of the Ca2+inflUX. This effect was mimicked by a a-agonist, a direct adeny1ate cyclase activator and by a cAMP analogue. Altogether these data offer convincingsupport for the notion that NE, via the a- receptor, may enhance the NMDA-mediated cascade of events
42
that induceLTP and therebypromote. LTP . Byenh a nc i ng NMOA-mediatedca2• aCCUlD.ula t i on or byaupp leJIenti nq that accumulation throuqhenh a nceme nt ot voltaqe-qated Ca2•
infl ux , LTPlikel ihood andtlagni t ud e lIay befac ilitatedby NE. Note thatthrouqh actions onvo l tage -gated channel sNE couldpromote LTP withoutNMDA- r e c eptorinteract ion .
It isnotev orthythat just as HE.a y pr omote LTP by enablinq HMOAac t ivationand Ca2+accumula t i on, othe r evidenc eshows th at NMOAre cepto r acti va t ionenha nces HE rel e a s e. Thi s sU9gests thatthe LTPmecha nism ma ypromote REP induction (craiqand White, 1992, Raiter!et aI., 1992, Pittaluqa and Ra ited, 1992aand 1992b ,I'ink and Gothert, 1992) •
The.';'idenca shows thatHE , via the8-r e ce ptor , not onlypro mote s HMDA-mediatedLTP but th a t REPitsel falso depends on theHMO-' re c eptor , at lea st inthesl ice. Just asLTP and taPconverge on the 8-adrenerq icre c eptor, soit has been disc ove red that HEP de pe nds not only on the8- rec eptor, but alsoon the LTPindu ct i on lIe chan i . II, the NMOA receptor(Stanton et al., 1989 , BUr g a r d et aI. , 1989 , Dahl andsa~ey,1990). I twa s demonstr a t ed thatWhile bath ap p lie d HEin du ced reliablepotentiation of the popllpike and pEPSPin the dentateof the slic e , coapp lication of APV produceda dos e-de p endent re duc t i on (1I&M) 01.bloc k ade (10 or 301&M) of REP (Burgardet al., 1989andStant on et a1.,
.3
1989). Stanton et 81. (1989) pointed out that HE produced no potentiation, neither short- nor long-lasting, in the presence of AVY. Dahl and Sarvey(1990) also demonstrated that 1",Misoproterenol-inducedNEPwas completely blockedby HMOAantagonism: in the presenceof 10~KAPV. isopr o t e r e nol induced no short-or long-lastingetfects.
In sum, theinY.1ill results suggest that dentateHEP and LTP convergeonccaa cninduction \IIBc h a nismslthe8- adrenergicandNMOAreceptors. The evidence sh owingNE promotionof LTP andHMOApromotion of NErelease sugg e s t, as previouslyproposed, asyne r g y betweenNEPand LTP. This same thesis, however. has for the mostpart gone without investigation1D~and preliminary evidenceoffers little support.
Kany otthe similarities betweenLTP andNEPthat have been reported froll slice studies; codependence on the 8- receptor, HMOA receptor, cAMP increases, proteinsynt h e s i s and HE, and increasedglutamate release, have not been demonstrated in the whole animal. While HE depletion
"virtually eliminates" LTPinill.t2 (Stanto;'!and Sarvey, 1985a), the same has notbeen demonstratedinY.1YQ. Bliss et a1. (198J) found that although NE depletion decreased the magnitUdeatpEPSP LTP in the anaesthetized rat,popspike LTP \las unaffected. Robinsonand Racine (1985) report ifldthe opposite in awake rats. They found that popspike LTP was
..
reducecl but pEPSP LTP was actually larger in NE-depleted animala. Altogether,the data suggest that HE 18 not critical to the integrity of LTPinrlE ..it 18inY.1..t..r.:g.
The effect of 8-receptorblockade on LTP induction hae not beentestedinili2,but it would be predicted that ifNE is not needed then the HE receptor sho ul d not be necessary . Fi nally , the role of the NMDA receptor in a-mediatedNEP.in rlY2has not been investigated: that wastheaimof the present study.
Sufficient differ en c esexistbetweenthephe nomena of LTP",nd HEP to suggest that the twoma y be me chanistically distlnct. As thepreced ing re v i e w shows, pEPSPand popspike increases and latencydecrea s esare emblematicof HF'l'-LTP, yetin HEP only the popspike increaseoccurs invariably : pEPSP changes are typical but latency changesare rare. The fact that8FTstimulation producesNHDA-dependentLTP in the lateral perforant path while HE inducesIIdepression there, suggests other important differencesbetween HEPand LTP (Dahl and Sarvey,1989).
Other recentin~studies also question the convergence between HEP and LTP. Whileinv e s tiga t i ng the role of muscarinic receptorsin dentate plasticity, Burgard et al., (1993) proposed thatit NEP andLTP convergeon a common mechanism,muscarine should facilitate NEP as it had been shown to do for LTP. This proposalwas not supported.
45
Bath applicationotmuscarine together with a subthreshold concentration (O.l~M)of the a-agonist isoproterenol, did not produce HlP, aa would be expected i t muscarine acted as facilitator. In tact, muscarine completely blocked the induction of HEP by otherwise effective concentrations of Isoproterenol (lpH). Burgard et a1. (1993) concluded that
"there is at least one fundamental difference between induction of NEP and L'I'P." OffQring tentative support to this contention, Dahl and Li (1994) have shown that bath perfusion of a subthreshold dose of isoproterenol follo....ed by a 30 minute wash and second low dose isoproterenol application induces a a-dependent NE-LLP. Unlike otherJ.n
~findings, this NEP was not blocked by NMDA receptor antagonists. This a-mediated HEP did'lo t require the participation of the NMDA receptor. However, in this same stUdy it was found that HEP induced by a single, effective dose of isoproterenol was NMDA receptor dependent. It is not clear what isdiff~rentabout the single and sequenced applications and Which better reproduces the effects of endogenous HE release.
Given the contlictinginY.i.:t.l:Qevidence, the differences in the characteristics of LTP and HEP bothin x.1x.Qandinn.t..rg, and the fact that theJ.nrl.t.r.:2 preparation is missing some of the vital circuitry present in the whole animal (Amaral and Witter, 1989), it is not
46
