5-HT3a are required in LTD and AMPA receptor internalization 2014

5-HT 3A RECEPTORS ARE REQUIRED IN LONG-TERM DEPRESSION AND AMPA RECEPTOR INTERNALIZATIONY.YU,a D.-Q.CAO,a H.-Y.XU,a M.SUN,d Z.-L.HUANG,b W.-H.YUNG,c N.LU a *AND Y.HUANG a *aDepartment of Physiology and Pathophysiology,Shanghai Medical College,Fudan University,Shanghai,China b Department of Pharmacology,Shanghai Medical College,Fudan University,Shanghai,Chinac School of Biomedical Sciences,Faculty of Medicine,The Chinese University of Hong Kong,Shatin,Hong Kong,Chinad R&D China,GlaxoSmithKline,Pudong,Shanghai 201203,ChinaAbstract—5-Hydroxytrytamine (serotonin)type 3A receptors (5-HT 3A Rs),as the only ligand-gated ion channels in the serotonin receptor family,are known to regulate neuronal excitation and release of GABA in hippocampal interneu-rons.However,their physiological role in glutamatergic syn-aptic plasticity remains unclear.Here,we show that deletion of the 5-HT 3A R gene in transgenic mice abolished N-methyl-D -aspartate (NMDA)receptor (NMDAR)-dependent long-term depression (LTD)induced by low-frequency stimulation (LFS)in hippocampal CA1synapses in slices,whereas the metabotropic glutamate receptor (mGluR)-dependent LTD did not change in the 5-HT 3A R knockout mice.In addition,5-HT 3A Rs disruption inhibited a -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs)internaliza-tion,without altering basal surface levels of AMPARs.However,the deletion of 5-HT 3A Rs did not lead to loss of synapses and structural alteration of dendritic spines.Fur-thermore,the concentrations of 5-HT and 5-hydroxyindole-acetic acid (5-HIAA)in the hippocampus were not affected by the deletion of 5-HT 3A Rs.These observations revealed an important role of 5-HT 3A Rs in NMDAR-dependent long-term depression,which is critical for learning behaviors.Ó2014IBRO.Published by Elsevier Ltd.All rights reserved.Key words:5-HT 3A receptor,long-term depression,AMPA receptor internalization,dendritic spines.INTRODUCTIONSynaptic plasticity is believed to be a crucial cellular substrate for the information storage underlying learning and memory.As a long-lasting form of activity-dependent synaptic plasticity in the hippocampus,long-term depression (LTD)is proposed to decrease synaptic efficacy in the non-potentiated synapses that do not participate in a memory trace,thus enhances signal-to-noise ratio in memory storage,facilitating the creation of spatial memory and is important for the clearing of old memory traces (Nicholls et al.,2008;Malleret et al.,2010).a -Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs),together with N-methyl-D -aspartate (NMDA)and kainate receptors,mediate the majority of fast excitatory synaptic transmission in the brain and are involved in the expression of synaptic plasticity (Lin et al.,2000;Kessels and Malinow,2009).The induction of LTD requires membrane potential depolarization,NMDA recep-tor (NMDAR)activation and trafficking of AMPARs.AMPARs are composed of GluR1-4subunits (Wenthold et al.,1996).GluR2is constitutively recycled in and out of synapses,which helps in maintaining a stable level of AMPARs (Luthi et al.,1999;Lee et al.,2002;Mameli et al.,2007).Although a decrease in presynaptic neuro-transmitter release may also play a role in LTD,it is thought that LTD results mainly from a decrease in postsynaptic receptor density.There are numerous studies (Sheng and Hyoung,2003;Davidson et al.,2009)indicating that a common postsynaptic mechanism of synaptic plasticity involves a net change in the number of AMPARs.NMDAR-dependent LTD is primarily mediated by the endocytosis of GluR2-containing AMPARs on the postsyn-aptic membrane (Beattie et al.,2000;Collingridge et al.,2004;Shepherd and Huganir,2007).Serotonin 5-hydroxytrytamine type 3A receptors (5-HT 3A Rs)(Engel et al.,2013),the only ligand-gated cation channels in the serotonin (5-HT)receptor family,are spe-cifically expressed in cholecystokinin (CCK)interneurons (Ferezou et al.,2002).These receptors are permeable to Na +,Ca 2+and K +.Activation of 5-HT 3A Rs located on somatodendritic regions affects neuronal excitation,and those expressed on axon terminals modulate neurotrans-mitter release.In addition,CCK interneurons,which have 5-HT 3A Rs expression,have inhibitory synaptic projection to pyramidal cells (Freund and Katona,2007).Alteration of activity and GABAergic transmission of interneurons may affect synaptic plasticity and cognition (Gong et al.,2009;Han et al.,2012).Converging evidence shows that/10.1016/j.neuroscience.2014.07.0700306-4522/Ó2014IBRO.Published by Elsevier Ltd.All rights reserved.*Corresponding authors.Address:Department of Physiology and Pathophysiology,Shanghai Medical College,Fudan University,130Dong’an Road,Shanghai,China.Fax:+86-21-54237098.E-mail addresses:luning7@ (N.Lu),yinghuang@ (Y.Huang).Abbreviations:AMPARs,a -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors;CCK,cholecystokinin;DHPG,3,5-dihydroxyphenylglycine;fEPSP,field excitatory postsynaptic potential;GABA,gamma-aminobutyric acid;HPLC,high-pressure liquid chromatography;5-HT 3A Rs,serotonin 5-hydroxytrytamine type 3A receptors;5-HIAA,5-hydroxyindoleacetic acid;5-HT,5-hydroxytrytamine;KO,knock-out;LTD,long-term depression;LFS,low-frequency stimulation;MnR,median raphe nucleus;mGluR,metabotropic glutamate receptor;NMDA,N-methyl-D -aspartate;NMDAR,NMDA receptor;RH,raphe-hippocampal.Neuroscience 278(2014)105–1121055-HT3Rs modulate neuronal activity,LTP,and memoryprocesses.For example,systemic injection of the selec-tive5-HT3receptor antagonist ondansetron had beenreported to facilitate the induction of LTP,increase the fre-quency of theta electroencephalogram rhythm,andenhance the retention of memory in hippocampus-depen-dent tasks.In addition,blockade of5-HT3receptorsin vivo results in a reduction of thefiring activity of hippo-campal interneurons and an increase of thefiring of hip-pocampal pyramidal neurons(Maeda et al.,1994;Fontana et al.,1995;Staubli and Xu,1995;Reznic andStaubli,1997;Bhatnagar et al.,2004;Naghdi andHarooni,2005).However,the physiological role of5-HT3A Rs in LTD still remains unclear,and whether5-HT3A Rs are involved in the trafficking pattern of theAMPARs internalization is unknown.In this study,using5-HT3A R gene knock-out(KO)mice or specific5-HT3A R inhibitor,we examinedpharmacological and genetic perturbations of5-HT3A Ractivity on LTD in CA1region in the hippocampus.Wefound that NMDAR-dependent LTD was impaired in5-HT3A Rs KO hippocampal slices,while5-HT3A Rs arenot necessary for metabotropic glutamate receptor(mGluR)-dependent LTD.We also examined AMPARinternalization and found that5-HT3A Rs were essentialfor this process.Besides,5-HT3A Rs disruption does notimpair structural plasticity in hippocampal CA1pyramidal neurons.Thus,5-HT3A Rs may serve animportant function in LTD and AMPAR internalization.EXPERIMENTAL PROCEDURES Animals5-HT3A R KO mouse strain was used in this study.Allanimal experiments were conducted with protocolsapproved by the Animal Care and Use Committee of theFudan University.Only male mice were used and allanimals were on C57BL/6background. Electrophysiological recordingHippocampal slices were prepared from16to20-day-old5-HT3A R KO and wild-type mice(Bortolotto et al.,2011).In brief,after decapitation under isoflurane anesthesia,the brain was rapidly removed and placed in ice-cold artifi-cial cerebrospinalfluid(ACSF,saturated with95%O2/5%CO2)which was composed of(in mM)NaCl(119),NaHCO3(26.2),NaH2PO4(1),KCl(2.5),CaCl2(2.5),MgSO4(1.3)and d-glucose(11).Hippocampal slices(400l m)were prepared in oxygenated ACSF with a vibroslice,and main-tained in ACSF at30°C for at least1h before use.Theslices were submerged in the recording chamber,and per-fused with ACSF at a rate of2–4ml/min at32°C.All electrophysiological recordings were performedwith an Axopatch-200B amplifier(Molecular Devices)atthe sampling rate of10kHz andfiltered at5kHz.Datawere acquired and subsequently analyzed using thesignal4.08software.To induce LTD in the CA1regionin the hippocampus,a bipolar stimulating electrode waspositioned in the CA1stratum radiatum where terminalsof the Schaffer collaterals(SC)are found.Electrical stimulation consisting of low-intensity,square-wavepulses(0.1ms)was delivered every30s in recordings.The induction protocol of NMDAR-dependent LTDconsisted of900pulses at a1-Hz stimulation(LFS)(Jiao and Li,2011)at the same intensity as the control.mGluR-dependent LTD was induced with the applicationof3,5-dihydroxyphenylglycine(DHPG,a group I-selectivemGluR agonist)(Tokay et al.,2009)to perfusion solutionfor10min.Baseline responses were recorded for15minprior to delivering either low-frequency stimulation(LFS)(1Hz,15min)or drug(DHPG100l M,10min)perfusion. Responses were subsequently recorded for an additional60min to monitor long-term changes in synaptic transmis-sion.The synaptic strength was determined by the slopefrom10%to90%of the rising phase of thefield excitatorypostsynaptic potential(fEPSP).The magnitude of LTDwas quantified as the normalized average slope of thefEPSP taken from the last15min of recording.Hippocampal neuron culturesHippocampal neurons were obtained from neonatal WT and KO mice at1–2days of age using a modified published protocol(Davidson et al.,2009).Freshly dis-sected hippocampal tissues were cut into fragments and digested with0.125%trypsin for20min at37°C,then ter-minated by DMEM containing10%bovine serum.Disso-ciated neurons were seeded at a density of4Â105cells/ ml on poly-l-lysine(0.1mg/ml)-coated dishes in neuroba-sal medium containing2%B27supplement.The cells were incubated in a humidified incubator(5%CO2and 95%O2)at37°C.Medium was changed every3days by replacement of half the medium.All experiments were performed using8-to10-day-old cultures.GluR2internalization assayThe internalization assay was performed as described (Lee et al.,2002;Li et al.,2004;Jiao and Li,2011).Briefly, hippocampal neurons at8-to10-day-old cultures were incubated with an monoclonal antibody against N-termi-nus of GluR2for10min at37°C,then stimulated with NMDA(50l M)for10min.Neurons were thenfixed for 5min at room temperature in4%paraformaldehyde/4% sucrose,and surface-remaining antibody-labeled GluR2 was saturated by incubation with Cy3secondary antibody for1h at room temperature.Neurons were then perme-abilized for1min with100%methanol atÀ20°C,and internalized GluR2was stained with Alexa-488secondary antibody.Fluorescence images were acquired on a Zeiss inverted LSM510confocal microscope with40Âobjec-tive.Images were quantified by Image-Pro Plus6.0soft-ware.The normalized internalized index=integrated fluorescence intensity of internalized GluR2/integrated fluorescence intensity of internalized GluR2plus surface GluR2,while the integratedfluorescence intensity equals averagefluorescence intensityÂarea.Quantification was performed in a genotype-and treatment-blind manner.106Y.Yu et al./Neuroscience278(2014)105–112Dendritic spine density and spine morphology Dendritic spine density and morphology(Li et al.,2004;Jiet al.,2005,2010;van der Velden et al.,2012)in CA1neurons of6–8week-old mice were examined byquantifying spines in neurons labeled withfluorescentdye1-1-dioctadecyl-3,3,30,30-tetramethylindocarbocya-nine perchlorate(DiI,Invitrogen,Carlsbad,CA,USA).Coronal brain slices(300l m)were cut on a vibratome in ACSF,same as which for electrophysiological experi-ments.Slices were then immediatelyfixed in4%parafor-maldehyde/4%sucrose solution for one night at4°C andwashed thoroughly with PBS three times.Neurons werelabeled by touching the tip of the glass micropipettefilledwith DiI with the membrane of neuron for8min.Sliceswere then rinsed with PBS and mountedfocal microscope(Zeiss LSM510,was used to acquire a stack ofing=0.5l m)of the apical dendrites fromneurons using a63Âoil objective at a1.0-tion.Regions with dense DiI staining inneurons could not be distinguished wereDendritic spine morphology wasImageJ software.Two or threeimaged and at least three cells perdendrites were secondary branchesdendritic tree.The dendritic protractionsrecognizable stalk and with a size largerwhich was traced and measured withsoftware,were counted as spines.defined as the number of spines per50was calculated as spine numbermeasured length(in l m)andQuantification of dendritic spine lengthperformed as per the followingwas measured from the dendritic shaftspine tip.The width of a spine wasof the spine head perpendicular to theMeasurement of5-hydroxytrytamine5-hydroxyindoleacetic acid(5-HIAA)Tissue concentrations of5-HT andperformed using a sensitivechromatography(HPLC)system(Kim ettissues of hippocampus were isolatedold WT and KO mice and transferred toacid according to the tissue weight(10sues were homogenized and15,000Âg for10min at4°C.Aliquot(10natant was injected to the HPLCC18,150mmÂ4.6mm,5l m;Agilent,phase consisted of10%methanol andsolution,as follows(in mM):75NaH20.025EDTA,100l l/L TEA,at aflowand at pH value of3.5.The level ofwas detected using an electrochemicalvoltage setting at0.7V and theHPLC system was connected to aall compounds by comparing the areawith the area of reference standards withsoftware(Chromatography Station for Statistical analysisData are expressed as mean±SEM.The statistical software Sigmastat(SPSS)and Origin software were used for data parisons were performed using unpaired student’s t-test between two groups or one-way ANOVA among three orfive groups as appropriate.Probability values of<0.05were considered statistically significant.RESULTSNMDAR-dependent LTD is impaired in5-HT3A Rs knockout hippocampal slices5-HT3A Rs disruption impairs LFS-induced LTD in the campus.(A)Field EPSPs were recorded in the CA1region.LFS at1Hz)was delivered to induce LTD.The fEPSPs slope normalized to the baseline value before LTD induction.(n=8 for WT mice,n=7for ondansetron(25nM)-treated WTInsert in A:fEPSPs from pre-LFS recordings(1,gray) post-LFS recordings obtained at the point of60min(2,Summary of the data showed the amplitude of LTD observedKO and antagonist-treated slices.Data are represented±SEM;⁄p<0.05,⁄⁄p<0.01.Y.Yu et al./Neuroscience278(2014)105–112107however,LFS failed to induce LTDbaseline,n=8slices from eightcompared with WT,Fig.1A and B).In5-HT3A Rs antagonist,ondansetron(25LFS-induced LTD to88.6±5.0%(n=7mice,p<0.05for WT+ondansetronthe CA1region in hippocampal slicesmice,suggesting that5-HT3A Rs areNMDAR-dependent LTD.5-HT3A Rs are not necessary forLTDTo determine whether5-HT3A Rs are alsoforms of LTD,we measured(mGluR-LTD),which was induced bymGluR agonist DHPG100l M for10min)in the CA1region ofhippocampi.DHPG,via activation ofable to induce a form of LTD that isdistinct from NMDAR-dependent LTD.LTDs were similar between WT and78.3±5.0%of baseline,n=10slicesKO:83.1±4.1%of baseline,n=8slicesp>0.05).These results suggest that5-HT3A Rs specifically impairs(see Fig.2).5-HT3A Rs are essential for AMPARThe above results indicate that5-HT3A RsNMDAR-dependent LTD.Givenendocytosis is one of critical post-synapticin this form of LTD(Lin et al.,2000;Malinow,2009),we next examinedare required in AMPAR endocytosisfeeding assay to analyze the endocytosis of AMPAR sub-unit GluR2.Cultured hippocampal neurons(10daysin vitro)prepared from WT and KO mice were stimulatedwith NMDA(50l M for5min),a method to induce‘‘chem-ical LTD’’that shares the molecular mechanism with elec-trically induced LTD(Beattie et al.,2000;Jiao and Li2011).As shown in Fig.3A,B,NMDA-induced GluR2internalization was significantly reduced in the hippocam-pal neurons from5-HT3A Rs KO mice(WT:128.5±6.2%,n=38;KO:86.9±5.4%,n=25,p<0.001).Similarly,GluR2internalization was also inhibited by treatment withondansetron(50nM)in neurons from WT mice(WT+ondansetron:88.9±6.1%,n=39,p<0.001versus WT in NMDA).However,the basal internalizationof GluR2,which was measured in the absence of NMDAtreatment,was not altered in KO neurons.Thus5-HT3A Rsare required for NMDA-induced AMPAR internalization.5-HT3A Rs deletion does not affect spine density and morphology in hippocampal CA1pyramidal neurons Having established the role of5-HT3A Rs in AMPAR internalization during‘‘chemical LTD’’,we then examined whether the absence of5-HT3A Rs alters the spine density and spine growth.The density of spines in the secondary dendrites of pyramidal neurons from KO mice averaged36.0±2.0per50l m,as compared with 34.7±2.2per50l m in WT neurons(p>0.05, Fig.4A,B).To determine whether spine growth is different in hippocampal neurons from WT and KO mice, we examined the length of spine and the width of spine head.There was no difference in the length of spine between neurons in slices from WT(0.96±0.03l m) and KO(1.00±0.05l m)mice(Fig.4C).Besides, average spine head width did not differ between WT (0.48±0.03l m)and KO(0.49±0.02l m)mice (Fig.4D)either.These results suggest that dendritic spine density and spine growth are not sensitive to 5-HT3A Rs disruption in CA1pyramidal neurons.The concentration of5-HT and5-HIAA does not change in5-HT3A Rs knockout HippocampusIt has been previously reported(Zhong et al.,2008)that serotonin facilitates the induction of LTD of glutamatergic synaptic transmission in pyramidal neurons.Besides, (Janusonis et al.,2006)and Shifrovich found that the level of5-HT was lower in the brain in5-HT1A receptor-KO mice at3days after birth.To determine whether the difference of LTD and AMPAR internalization could possibly be caused 5-HT3A Rs deletion does not change mGluRs-dependenthippocampus.(A)Field EPSPs were recorded inThe fEPSPs slope was normalized to the baselineLTD induction.mGluRs agonist DHPG(100l M,10the bath solution to induce mGluRs-dependent LTDn=10for WT mice).Insert in A:fEPSPs from pre-DHPG-recordings(1,gray)and from post-DHPG-treated recordingsat the point of60min(2,black).(B)Summary ofthe amplitude of LTD observed in KO and WT mice. represented as mean±SEM.108Y.by abnormal release of5-HT in5-HT3A Rs KO hippocam-pus,we measured the concentrations of5-HT and its metabolite5-HIAA in the hippocampus.Fig.5shows the levels of5-HT and5-HIAA in the hippocampus of WT and KO pared with WT mice(5-HT:26.5±2.1pg/ mg,n=10;5-HIAA:277.0±29.4pg/mg,n=10),the levels of5-HT and5-HIAA in adult hippocampus did not differ in KO mice(5-HT:23.2±3.0pg/mg,n=10,p>0.05; 5-HIAA:260.2±26.7pg/mg,n=10,p>0.05).DISCUSSIONAs the unique ligand-gated ion channels in the serotonin receptor family,5-HT3A Rs are known to regulateFig.3.Deletion or inhibition of5-HT3A Rs impairs NMDA-inducedAMPAR internalization in hippocampal neurons.(A)Cultured hippo-campal neurons from WT or5-HT3A Rs KO mice were treated with50l M NMDA,50l M NMDA+50nM ondansetron or untreated.(B) Bar graph showing the internalization index of GluR2antibodystaining detected in untreated(n=25for WT,n=43for KO),NMDA-treated(n=38for WT,n=25for KO)and NMDA/ondan-setron co-treated group(n=39for WT),Internalization index=inte-gratedfluorescence intensity of internalized GluR2/(integrated fluorescence intensity of internalized GluR2+surface GluR2),nor-malized to levels detected in untreated WT neurons.Data are represented as mean±SEM;⁄⁄⁄p<0.001,##p<0.01,NS:no significance.Fig.4.5-HT3A Rs deletion does not change dendritic spine growth of mature neurons.(A)Examples showing dendritic spines of hippo-campal pyramidal neurons in slices from WT and KO mice.Right: spine length and spine head width.Scale bars=10l m(low magnification,left),2l m(high magnification,middle)and1l m(highmagnification,right).Quantification of(B)spine densities(number of spines per50l m of dendrite length),(C)spine length(l m)and(D) spine head width(l m).neuronal excitation and release of GABA in CCK interneurons.It has been suggested previously that 5-HT3Rs are involved in memory and long-term potentiation of synaptic plasticity in the hippocampus (Maeda et al.,1994;Passani et al.,1994;Staubli and Xu,1995;Alkadhi et al.,1996;Reznic and Staubli, 1997).In the present study,we found that in the CA1 region of the mouse hippocampus,deletion of5-HT3A Rs impaired NMDAR-dependent LTD but not mGluR-depen-dent LTD.In addition,GluR2internalization was reduced in cultured hippocampal neurons from5-HT3A Rs knockout mice.These effects could be mimicked by5-HT3A Rs antagonist ondansetron.However,5-HT3A Rs disruption did not change the spine density or spine morphogenesis in hippocampal CA1pyramidal neurons.Moreover,in transgenic mice,the loss of5-HT3A Rs did not change the concentrations of5-HT and its metabolite5-HIAA in the hippocampus.Thus,5-HT3A Rs may serve an impor-tant role in NMDAR-dependent LTD and AMPAR internalization.The serotonergic raphe-hippocampal(RH)pathway that originates within the midbrain median raphe nucleus (MnR)is a key modulator for hippocampal mnemonic function(Hensler,2006).Electrical stimulation of the RH pathway produces a rapid and robust modulation of hippocampal activity.Anatomical evidence(Freund et al., 1990)shows that MnR projections form classical synapses onto GABAergic interneurons in the hippocampus,poten-tially providing a substrate for a fast neuromodulation of hippocampal functions.In addition,fast transmissions of EPSPs on these interneurons are mediated by synapses formed by RH afferents.5-HT3Rs regulate AMPAR/ NMDAR-mediated EPSPs,whichfinally modulate pyrami-dal cell membrane potential as well as its activity and could influence synaptic plasticity and the mnemonic functions (Varga et al.,2009).Consistently,we show here that 5-HT3A Rs play an essential role in NMDAR-dependent LTD in the hippocampus.As one of LTD mechanisms is through regulation of AMPAR-mediated fast synaptic transmission(Man et al.,2000),5-HT3A Rs,together with AMPA/NMDA receptors,may influence the induction of NMDAR-dependent LTD and therefore the process of memory formation that requires this form of synaptic plas-ticity.More specifically,given that LTD in the hippocampus may allow acquisition of new information by suppressing interference and involving in the process of spatial mem-ory,these results implicate an important role of5-HT3A Rs in memory.5-HT3A Rs are specifically expressed in CCK interneurons.The loss of5-HT3A Rs mainly affected firings of these interneurons and the GABAergic inhibitory synaptic inputs to pyramidal neurons,thus disrupted LTD in pyramidal cells.As GABA A receptor-mediated hyperpolarization is thought to maintain the voltage-dependent Mg2+block of NMDAR channels, GABAergic interneurons affect synaptic plasticity,such as LTP and LTD,depending on activation of NMDA receptors(Collingridge et al.,1988).Thus it is not surpris-ing that5-HT3A Rs knockout impaired LFS-induced NMDA-dependent LTD.In addition,it has been reported that mGluR-dependent LTD at Schaffer collateral-CA1 synapses is unaffected by GABA A R-mediated synaptic transmission(Rohde et al.,2009).This is a possible rea-son for the specific change in NMDAR-dependent LTD but not mGluR-induced LTD.Chemical LTD,which is evoked by NMDA,shares the same molecular mechanisms with that of electrically induced LTD(Beattie et al.,2000;Jiao and Li,2011).It has been suggested previously that serotonin facilitates LTD in prefrontal cortex via enhancement of AMPAR internalization(Zhong et al.,2008).In this study,we found that5-HT3A Rs disruption reduced NMDA-induced inter-nalization of GluR2.Pharmacological treatment with 5-HT3antagonist ondansetron reduced the internalization of AMPARs responding to NMDA.These results are con-sistent with our electrophysiological data and support the notion that5-HT3A Rs modulate LTD by its effect on AMPAR internalization.There is evidence that most excitatory synaptic transmission occurs at dendritic spines,whereas spine number and morphology are readouts of the functional state of glutamatergic synapses(Toni et al.,1999;Meng et al.,2002;D’Agostino et al.,2013).While a previous study(van der Velden et al.,2012)reported changes in apical dendritic complexity of cortical layer2/3pyramidal neurons in mice lacking5-HT3A Rs,in the present study, we found that5-HT3A R deletion did not influence spine density in apical densities and morphogenesis of CA1 pyramidal neurons.As5-HT3A receptors are specifically expressed in CCK interneurons but not pyramidal neurons,one possible explanation for this phenomenon5-HT3A Rs disruption does not affect the concentration5-HIAA in the hippocampus.The concentration of(A)5-HT5-HIAA in the hippocampus of KO(n=10)and WTData are represented as mean±SEM.278(2014)105–112is that the loss of5-HT3A receptors may affect only inter-neurons but not pyramidal cells.Although a decrease of presynaptic neurotransmitter release may also play a role in LTD(Yang and Calakos, 2013),it is thought that LTD mainly results from a decrease of postsynaptic receptor density.NMDAR-dependent LTD is induced by low-frequency stimulation, which may cause the slight depolarization of the postsyn-aptic cells.Due to the depolarization,activation of volt-age-dependent calcium channels leads to calcium influx and the calcium-dependent protein phosphatase cas-cade.The rise in calcium and calcineurin activity in post-synaptic cells triggers the AMPA endocytosis and NMDAR-dependent LTD(Beattie et al.,2000).The inhib-itory GABAergic system can regulate NMDAR activation and plasticity of the postsynaptic neurons(Wagner and Alger,1995).Given our unpublished data showing that, despite membrane depolarization,5-HT3A R activation enhanced spike-frequency accommodation and sup-pressedfirings of5-HT3A R interneurons via calcium-dependent potassium channels,we propose that the loss of5-HT3A Rs expressed on CCK-positive interneurons may result in diminished spike-frequency accommodation and increase offiring frequency in5-HT3A R interneurons. As a consequence,the inhibitory inputs from5-HT3A R cells to pyramidal neurons would be enhanced,thus disrupting the membrane depolarization in pyramidal neurons.This will lead to failure of calcium influx and Ca2+-dependent phosphatase pathway,which is neces-sary for AMPAR internalization and LTD.In conclusion,this study provides direct evidence of a critical role of5-HT3A Rs in hippocampal LTD and AMPAR trafficking.As LTD is known to be a cellular mechanism for learning and memory,elucidating the effect of 5-HT3A Rs on behavior is essential for further study in understanding of the role of these receptors in cognitive function and memory.Acknowledgment—This work was supported by National Natural Science Foundation of China(31070931)and(81170237).REFERENCESAlkadhi KA,Salgado-Commissariat D,Hogan YH,Akpaudo SB (1996)Induction and maintenance of ganglionic long-term potentiation require activation of5-hydroxytryptamine(5-HT3) receptors.J Physiol496(Pt2):479–489.Beattie EC,Carroll RC,Yu X,Morishita W,Yasuda H,von Zastrow M, Malenka RC(2000)Regulation of AMPA receptor endocytosis bya signaling mechanism shared with LTD.Nat Neurosci3:1291–1300.Bhatnagar S,Nowak N,Babich L,Bok L(2004)Deletion of the5-HT3 receptor differentially affects behavior of males and females in the Porsolt forced swim and defensive withdrawal tests.Behav Brain Res153:527–535.Bortolotto ZA,Amici M,Anderson WW,Isaac JT,Collingridge GL (2011)Synaptic plasticity in the hippocampal slice preparation.Curr Protoc Neurosci:6–13.Chapter6.Collingridge GL,Herron CE,Lester RA(1988)Synaptic activation of N-methyl-D-aspartate receptors in the Schaffer collateral–commissural pathway of rat hippocampus.J Physiol399:283–300. Collingridge GL,Isaac JT,Wang YT(2004)Receptor trafficking and synaptic plasticity.Nat Rev Neurosci5:952–962.D’Agostino G,Kim JD,Liu ZW,Jeong JK,Suyama S,Calignano A, Gao XB,Schwartz M,Diano S(2013)Prolyl endopeptidase-deficient mice have reduced synaptic spine density in the CA1 region of the hippocampus,impaired LTP,and spatial learning and memory.Cereb Cortex23:2007–2014.Davidson HT,Xiao J,Dai R,Bergson C(2009)Calcyon is necessary for activity-dependent AMPA receptor internalization and LTD in CA1neurons of hippocampus.Eur J Neurosci29:42–54.Engel M,Smidt MP,van Hooft JA(2013)The serotonin5-HT3 receptor:a novel neurodevelopmental target.Front Cell Neurosci 7:76.Ferezou I,Cauli B,Hill EL,Rossier J,Hamel E,Lambolez B(2002) 5-HT3receptors mediate serotonergic fast synaptic excitation of neocortical vasoactive intestinal peptide/cholecystokinin interneurons.J Neurosci22:7389–7397.Fontana DJ,Daniels SE,Henderson C,Eglen RM,Wong EH(1995) Ondansetron improves cognitive performance in the Morris water maze spatial navigation task.Psychopharmacology(Berl) 120:409–417.Freund TF,Gulyas AI,Acsady L,Gorcs T,Toth K(1990) Serotonergic control of the hippocampus via local inhibitory interneurons.Proc Natl Acad Sci USA87:8501–8505.Freund TF,Katona I(2007)Perisomatic inhibition.Neuron56:33–42. Gong N,Li Y,Cai GQ,Niu RF,Fang Q,Wu K,Chen Z,Lin LN,Xu L, Fei J,Xu TL(2009)GABA transporter-1activity modulates hippocampal theta oscillation and theta burst stimulation-induced long-term potentiation.J Neurosci29:15836–15845.Han J,Kesner P,Metna-Laurent M,Duan T,Xu L,Georges F, Koehl M,Abrous DN,Mendizabal-Zubiaga J,Grandes P,Liu Q, Bai G,Wang W,Xiong L,Ren W,Marsicano G,Zhang X (2012)Acute cannabinoids impair working memory through astroglial CB1receptor modulation of hippocampal LTD.Cell 148:1039–1050.Hensler JG(2006)Serotonergic modulation of the limbic system.Neurosci Biobehav Rev30:203–214.Janusonis S,Anderson GM,Shifrovich I,Rakic P(2006)Ontogeny of brain and blood serotonin levels in5-HT receptor knockout mice: potential relevance to the neurobiology of autism.J Neurochem 99:1019–1031.Ji Y,Lu Y,Yang F,Shen W,Tang TT,Feng L,Duan S,Lu B(2010) Acute and gradual increases in BDNF concentration elicit distinct signaling and functions in neurons.Nat Neurosci13:302–309.Ji Y,Pang PT,Feng L,Lu B(2005)Cyclic AMP controls BDNF-induced TrkB phosphorylation and dendritic spine formation in mature hippocampal neurons.Nat Neurosci8:164–172.Jiao S,Li Z(2011)Nonapoptotic function of BAD and BAX in long-term depression of synaptic transmission.Neuron70:758–772. Kessels HW,Malinow R(2009)Synaptic AMPA receptor plasticity and behavior.Neuron61:340–350.Kim HS,Jang HJ,Cho KH,Hahn SJ,Kim MJ,Yoon SH,Jo YH,Kim MS,Rhie DJ(2006)Serotonin inhibits the induction of NMDA receptor-dependent long-term potentiation in the rat primary visual cortex.Brain Res1103:49–55.Lee SH,Liu L,Wang YT,Sheng M(2002)Clathrin adaptor AP2and NSF interact with overlapping sites of GluR2and play distinct roles in AMPA receptor trafficking and hippocampal LTD.Neuron 36:661–674.Li Z,Okamoto K,Hayashi Y,Sheng M(2004)The importance of dendritic mitochondria in the morphogenesis and plasticity of spines and synapses.Cell119:873–887.Lin JW,Ju W,Foster K,Lee SH,Ahmadian G,Wyszynski M,Wang YT,Sheng M(2000)Distinct molecular mechanisms and divergent endocytotic pathways of AMPA receptor internalization.Nat Neurosci3:1282–1290.Luthi A,Chittajallu R,Duprat F,Palmer MJ,Benke TA,Kidd FL, Henley JM,Isaac JT,Collingridge GL(1999)Hippocampal LTD expression involves a pool of AMPARs regulated by the NSF–GluR2interaction.Neuron24:389–399.Maeda T,Kaneko S,Satoh M(1994)Inhibitory influence via5-HT3 receptors on the induction of LTP in mossyfiber-CA3system of guinea-pig hippocampal slices.Neurosci Res18:277–282.Y.Yu et al./Neuroscience278(2014)105–112111。