doi:10.1152/jn.00180.2013 111:1746-1758, 2014. First published 5 February 2014;J NeurophysiolChen, Hui-hui Jiang, Ze Mi, Bo Hu, Jun Zhang and Ying XiongZhi-ru Zhu, Fenglian Xu, Wei-gang Ji, Shuan-cheng Ren, Fang Chen, Peng-zhiactivation-induced plasticity in the rat auditory cortexSynaptic mechanisms underlying thalamic

You might find this additional info useful...

57 articles, 23 of which can be accessed free at:This article cites /content/111/9/1746.full.html#ref-list-1

including high resolution figures, can be found at:Updated information and services /content/111/9/1746.full.html

can be found at:Journal of Neurophysiologyabout Additional material and information http://www.the-aps.org/publications/jn

This information is current as of May 14, 2015. 

American Physiological Society. ISSN: 0022-3077, ESSN: 1522-1598. Visit our website at http://www.the-aps.org/.(monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2014 by the

publishes original articles on the function of the nervous system. It is published 12 times a yearJournal of Neurophysiology

on May 14, 2015

Dow

nloaded from on M

ay 14, 2015D

ownloaded from

Synaptic mechanisms underlying thalamic activation-induced plasticity in therat auditory cortex

Zhi-ru Zhu,1 Fenglian Xu,2 Wei-gang Ji,3 Shuan-cheng Ren,1 Fang Chen,1 Peng-zhi Chen,1

Hui-hui Jiang,1 Ze Mi,1 Bo Hu,1 Jun Zhang,1 and Ying Xiong1

1Department of Physiology, Third Military Medical University, Chongqing, People’s Republic of China; 2Department of CellBiology and Anatomy, the Hotchkiss Brain Institute, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada;3Department of Chemistry, Third Military Medical University, Chongqing, People’s Republic of China

Submitted 11 March 2013; accepted in final form 31 January 2014

Zhu ZR, Xu F, Ji WG, Ren SC, Chen F, Chen PZ, Jiang HH,Mi Z, Hu B, Zhang J, Xiong Y. Synaptic mechanisms underlyingthalamic activation-induced plasticity in the rat auditory cortex. JNeurophysiol 111: 1746–1758, 2014. First published February 5,2014; doi:10.1152/jn.00180.2013.—Electrical stimulation of ventraldivision of medial geniculate body (MGBv) neurons evokes a shift ofthe frequency-tuning curves of auditory cortical (AC) neurons towardthe best frequency (BF) of the stimulated MGBv neurons (frequency-specific plasticity). The shift of BF is induced by inhibition ofresponses at the BF of the recorded AC neuron, with coincidentfacilitation of responses at the BF of the stimulated MGBv neuron.However, the synaptic mechanisms are not yet understood. Wehypothesize that activation of thalamocortical synaptic transmissionand receptor function may contribute to MGBv stimulation-inducedfrequency-specific auditory plasticity and the shift of BF. To test thishypothesis, we measured changes in the excitatory postsynaptic cur-rents in pyramidal neurons of layer III/IV in the auditory cortexfollowing high-frequency stimulation (HFS) of the MGBv, usingwhole cell recordings in an auditory thalamocortical slice. Our datashowed that in response to the HFS of the MGBv the excitatorypostsynaptic currents of AC neurons showed long-term bidirectionalsynaptic plasticity and long-term potentiation and depression. Phar-macological studies indicated that the long-term synaptic plasticitywas induced through the activation of different sets of N-methyl-D-aspartate-type glutamatergic receptors, �-aminobutyric acid-type re-ceptors, and type 5 metabotropic glutamate receptors. Our data furtherdemonstrated that blocking of different receptors with specific antag-onists significantly inhibited MGBv stimulation-induced long-termplasticity as well as the shift of BF. These data indicate that thesereceptors have an important role in mediating frequency-specificauditory cortical plasticity.

rat auditory cortex; thalamocortical system; plasticity

THE MAMMALIAN SENSORY CORTEX has a considerable capacity toadapt to changing environments (Beaulieu and Cynader 1990;Coq and Xerri 1998; Weinberger 2007). The auditory cortical(AC) area devoted to sound frequency is not immutable andcan be changed by behavioral training, environmental changes,and sensory deprivation. The capacity of the auditory cortex tochange its response properties is known as “AC cortical plas-ticity” (Norena et al. 2006; Weinberger et al. 1990). Owlmonkeys that had been trained over several weeks to discrim-inate sounds significantly expanded the cortical area tuned tothe frequencies of the training tones (Recanzone et al. 1993). Inaddition to experience-induced neural plasticity, changes in the

neuronal receptive fields of the auditory cortex can be inducedby electrical stimulation of the ventral division of the medialgeniculate body (MGBv) (Jafari et al. 2007; Ma and Suga2009). The neuronal receptive fields in the auditory cortexchange reliably toward the receptive fields of the electricallystimulated thalamic neurons, indicating the involvement ofthalamocortical (TC) frequency-specific plasticity. Such recep-tive field changes can last for weeks.

Cortical receptive fields are formed by organized arrays ofTC afferents that project onto neurons in the thalamorecipientlayer III/IV of the auditory cortex as well as intracorticalconnections. Sound information is transmitted from the thala-mus to the auditory cortex via TC afferents (Feldman andBrecht 2005; Liu et al. 2007). Studies have revealed that TCprojections link those MGBv and AC neurons that have asimilar characteristic frequency, which suggests that TC inputdirectly mediates cortical response to characteristic frequency(Miller et al. 2001). It is thus likely that TC projections providean anatomic basis for the plasticity of cortical neurons. Inaddition, several studies have shown that synaptic plasticitydrives cortical map reorganization (Allen et al. 2003; Buono-mano and Merzenich 1998; Feldman 2009). Long-term poten-tiation (LTP) and long-term depression (LTD), two fundamen-tal forms of synaptic plasticity, can be induced in TC synapses(Feldman 2009; MacDonald et al. 2006; Malenka and Bear2004). The importance of TC LTP and LTD in experience-dependent cortical plasticity and cortical map reorganizationhas been proven (Gagolewicz and Dringenberg 2011; Harding-ham et al. 2003; Liu et al. 2013). However, different stimula-tion protocols are usually employed to induce LTP and LTD(Habib and Dringenberg 2010). We recently investigated thepossibility of coexistence of LTP and LTD in AC neurons uponstimulation of the MGBv, using the same high-frequencystimulation (HFS) protocol described in the present study (Zhuet al. 2013). Our choice of HFS parameters is based onpreviously established work (Cai et al. 2013; Duffy andNguyen 2003). Specifically, we have demonstrated that HFS ofthe MGBv produces both LTP and LTD as well as the shift ofbest frequency (BF) in AC neurons. The evoked potentiationand depression depended on the frequency of the electricalstimulation of the MGBv correlated with the BF of MGBv andAC neurons, respectively. However, the underlying synapticmechanisms are still not known.

To this purpose, we made whole cell patch-clamp recordingsof the synaptic plasticity in AC pyramidal neurons of layerIII/IV after HFS of MGBv neurons in a TC slice preparation.

Address for reprint requests and other correspondence: Y. Xiong, Dept. ofPhysiology, Third Military Medical Univ., Chongqing 400038, People’s Re-public of China (e-mail: xiongying2001n@163.com).

J Neurophysiol 111: 1746–1758, 2014.First published February 5, 2014; doi:10.1152/jn.00180.2013.

1746 0022-3077/14 Copyright © 2014 the American Physiological Society www.jn.org

on May 14, 2015

Dow

nloaded from

Our results indicate that HFS of thalamic neurons inducesdifferent forms of long-term synaptic plasticity (rapid and slowLTP/LTD), which involve activation of different sets of glutama-tergic, �-aminobutyric acid(GABA)ergic, or endocannabinoid(eCB) receptors. Inhibition of these receptors prevented theinduction of TC plasticity as well as the shift of BF in ACneurons. implying the essential role of receptor activity inmediating TC frequency-specific plasticity in the rat audi-tory cortex.

MATERIALS AND METHODS

Thalamocortical Slice Preparation and Maintenance

Sprague-Dawley rats (P21–28) were used in this study. Animalswere obtained from the Laboratory Animal Center at the ThirdMilitary Medical University in China, and all protocols and proce-dures were approved by the University Animal Care and Use Com-mittee. Acute primary TC slices containing the auditory cortex andportions of the MGBv in the thalamus were prepared as described byCruikshank et al. (2002). After halothane anesthesia, each animal wasdecapitated and the brain was removed quickly. The brain wassubsequently submerged in cold artificial cerebrospinal fluid (ACSF)containing (in mM) 125.0 NaCl, 2.5 KCl, 25.0 NaHCO3, 1.25KH2PO4, 1.2 MgSO4, 2.0 CaCl2, and 10.0 dextrose, bubbled with95% O2-5% CO2, with a pH of 7.4. The brain was blocked, and anoscillating tissue slicer (VT1000, Leica, Wetzlar, Germany) was usedto cut 500-�m-thick sections. Sections were obtained from the lefthemisphere by using a slicing angle of 15°. Slices were initiallyincubated for a minimum of 90 min at room temperature (22–24°C) inACSF and were then transferred to and submerged in a recordingchamber, where they were perfused continuously with oxygenatedACSF at room temperature.

Electrophysiological Recording and Stimulation

Whole cell patch-clamp recordings were obtained from cell bodiesof layer III/IV thalamorecipient neurons in the rat auditory cortex.Pyramidal neurons are the principal thalamorecipient neurons in thethalamorecipient layers in the auditory cortex of many mammalianspecies (Richardson et al. 2009; Smith and Populin 2001). The cellbodies of these neurons were found with an upright microscopeequipped with Leica differential interference contrast optics, a �40water immersion objective, and an infrared video imaging camera.Data acquisition was conducted with Axopatch 200B (Axon Instru-ments, Union City, CA) or EPC10 amplifiers (HEKA Elektronik,Lambrecht/Pfalz, Germany). The signal was stored for off-line anal-ysis with Clampex 9.0 software (Molecular Devices) or Pulse/Pulse fitv.8.74 (HEKA Elektronik) and IGOR Pro v.4.03 (WaveMetrics).Pipettes (4–8 M�) for whole cell recording were pulled on a hori-zontal micropipette puller (P-97, Sutter Instrument) from filamentedcapillary glass and were filled with a pipette solution containing (inmM) 125 CsMeSO3, 2 CsCl, 10 HEPES, 0.1 EGTA, 4 MgATP, 0.3NaGTP, and 10 phosphocreatine, with 5 mM QX-314 (pH 7.4,adjusted with CsOH, 290–295 mosM). Membrane voltages werecorrected for an estimated liquid junction potential of 10 mV. Cellsrequiring firing property characterization were recorded with a pipettesolution containing (in mM) 130 potassium gluconate, 5 KCl, 2MgCl2, 4 MgATP, 0.3 GTP, 10 phosphocreatine, and 10 HEPES (pH7.3 with KOH). Neurons were given at least 5 min after gigaohm sealformation and patch rupture to stabilize before data collection. Seriesresistance was compensated by 80% and was continually monitoredthroughout the experiment. Cells were discarded if the series resis-tance changed by �15%. We monitored the peak amplitude of a briefhyperpolarizing test pulse (10 ms, �5 mV) given after thalamicstimulation to ensure consistent series resistance during the entire

experiment, and only recordings that remained stable over the periodof data collection were used.

Auditory pyramidal neurons were held at �70 mV to recordexcitatory postsynaptic currents (EPSCs). Inhibitory inputs wereblocked with the GABAA-type receptor blocker bicuculline in someof the experiments. A Grass S88 stimulator (Astro-Medical, WestWarwick, RI) and a constant-current stimulus isolation unit (GrassInstruments) were used to evoke EPSCs by stimulating the MGBv. Abipolar stimulating electrode was used to deliver monophasic stimuluspulses (0.2 ms, at 0.05 Hz), and the stimulation intensity was adjustedto evoke EPSCs that were 30% of maximal evoked amplitudes. Incontrol experiments, “baseline” EPSCs were recorded for 10 min atthe test stimulation intensity. Long-term plasticity was induced byswitching the amplifier to current-clamp mode and delivering four 1-s100-Hz trains at 20-s intertrain intervals. The criteria that we used toobtain the percentage of neurons for different types of plasticity are asfollows: 1) If the increase in EPSC amplitude to its maximum valueis shorter than 5 min and then stabilizes or continues to increase 35min after HFS, it is considered a long-lasting LTP. 2) If the increasein EPSC amplitude to its maximum value is longer than 5 min andthen starts to return toward baseline within 35 min after HFS, it isconsidered a short-lasting LTP. 3) Similarly, we consider LTD thatinduces a quick decline to minimum EPSC amplitude within 5 minand then quick return toward the baseline to be a fast LTD. 4) Weconsider LTD that induces a gradual decrease in EPSC amplitudewithin 30 min, which then stabilizes, to be a slow LTD. When furtherpharmacological experiments were needed, neurons were alwaysallowed to recover to 95% of their baseline level before exposure topharmacological reagents. It usually took �10 min to see a 95%recovery while cells were still in good condition.

GABAergic miniature inhibitory postsynaptic currents (mIPSCs)were recorded in the presence of ionotropic glutamate receptor block-ers, including CNQX (20 �M) and APV (50 �M) in the superfusingACSF. One micromolar tetrodotoxin (TTX) applied to the bathsolution completely blocked voltage-dependent Na� channels within5 min. Neurons were voltage-clamped at �10 mV for mIPSC record-ing.

Recording of Receptive Field in Auditory Cortex

The protocols for acoustic stimulus and electrical stimulation in theMGBv and auditory cortex have been provided in previous publica-tions (Zhu et al. 2013). Briefly, the acoustic stimuli were a series ofwhite noise or pure tone bursts at 5-ms rise/decay time for 60-msduration. A multibarrel glass electrode with a carbon fiber in thecentral barrel was advanced to the right auditory cortex. Four sur-rounding barrels were connected to the Neuro Phore System (6400Advanced, Dagan) for microiontophoretic injection of different recep-tor antagonists. Signals from the electrode were fed to a RA16PAmultichannel preamplifier (Tucker-Davis Technologies) and filteredwith a band pass of 0.3–10 kHz. Single units were isolated andanalyzed with BrainWare software (Tucker-Davis Technologies). Theexcitatory responses of AC neurons to acoustic stimuli were firstmeasured with the computer-controlled frequency/amplitude scan.The receptive fields (areas in which acoustic stimulation leads to aresponse in an AC or MGBv neuron), frequency-threshold curves, andBF were subsequently derived from the collected data. HFS (four 1-s100-Hz trains at 20-s intertrain intervals) was delivered to the tungstenelectrode placed in the MGBv. Changes in AC responses after elec-trical stimulation of the MGBv in the presence or absence of differentreceptor antagonists were recorded.

Anatomic Localization and Histology

Tract tracing with DiI. The fluorescent lipophilic tracer DiI wasused to examine fiber pathways in recorded slices. The slices werefixed by immersion in 4% paraformaldehyde for 3 days. DiI was

1747SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

delivered into the MGBv under a dissecting microscope with a brokenmicropipette. The slices were then placed in small plastic cups andcovered in warm liquid agarose (3%) that was made with PBScontaining 0.1% sodium azide. After the agarose solidified, the cupswere filled to the top with PBS, sealed, and placed in the dark at 37°Cfor 1 mo to allow sufficient DiI diffusion to occur. The slices weresubsequently sectioned on a vibratome at 50 �m. After three rinseswith PBS, the slices were mounted on gelatin-coated slides andcoverslipped.

Pyramidal cell identification. Biocytin hydrochloride was dis-solved in the pipette solution in a number of recordings after thecompletion of data collection (final concentration: 0.5%) to identifyauditory pyramidal cortical neurons. Slices with biocytin-filled cellswere first transferred to 4% paraformaldehyde for 48 h and were thencryoprotected. The sections were incubated in avidin-biotin-HRPcomplex (ABC kit, Vector Labs) at 37°C for 3 h. They were thenrinsed in a phosphate buffer and incubated with nickel/cobalt-inten-sified diaminobenzidine (DAB). The sections were then mounted andcoverslipped. Neurons were visualized with a fluorescence micro-scope (Olympus, Tokyo, Japan).

Drugs

All reagents were obtained from Sigma-Aldrich, with the exceptionof QX-314 (Tocris, Ellisville, MO). Bicuculline and CNQX weredissolved in dimethyl sulfoxide. Other drugs were dissolved in ACSF.All drugs were prepared as concentrated stock solutions and wereadded immediately either to the recording ACSF or to the internalsolution at working concentrations (1:1,000 dilution) and then frozenat �20°C until use.

Statistical Analyses

Data are presented as means � SE. Statistical analysis was per-formed with a one-way analysis of variance (ANOVA) test with SPSS18.0 software and Student’s t-test. Significance was set at the level ofP 0.05.

RESULTS

HFS of MGBV Induced Bidirectional Plasticity in RatAuditory Cortex

To determine the changes in synaptic plasticity in responseto thalamic stimulation, we made whole cell patch-clamprecordings of EPSCs from thalamorecipient layer III/IV pyra-midal neurons in a slice that contained auditory TC pathways.We first performed tract-tracing studies using the lipophilic dyeDiI to ensure that the TC pathway in the slice was complete.Figure 1A shows the results for a slice in which DiI was placedin the MGBv. The dye followed a pathway from the MGBvtoward projections along the medial edge of the lateral genic-ulate, after which it curved posterior-laterally around the hip-pocampus and moved toward the auditory cortex. On reachingthe auditory cortical border, the fibers ascended to layersIII–IV, where they formed a dense terminal plexus (Fig. 1B).We identified pyramidal AC neurons by their pyramidal-shaped cell bodies and long apical dendrites extending towardthe pial surface (Fig. 1C), as described in previous reports(Rose and Metherate 2005). A subset of these cells wereinjected with biocytin for further morphological identification.All labeled cells had the morphological features of pyramidalneurons: a pyramidal soma and a prominent apical dendrite(Fig. 1D). Pyramidal cells are clearly identified by their relativebroad spike width, slow spike afterhyperpolarization (AHP),low steady-state firing rate, and spike adaptation (Fig. 1, E and

F, Table 1). After characterization of the intrinsic properties ofthe neurons, EPSCs were evoked by thalamic stimulation.Neurons were examined for each current level in steps of 50�A. Figure 1G shows examples that were classified as pre-sumed EPSCs of pyramidal cells. TC stimulation can reliablyevoke EPSCs in layer III/IV pyramidal neurons (Fig. 1, G andH, Table 2).

After we identified pyramidal neurons, we conducted exper-iments to characterize the long-term synaptic plasticity of TCsynapses that connect with regular spiking neurons in theauditory cortex. HFS (4 trains of 100 Hz for 1 s with an intervalof 20 s) was applied to the layer when the amplitude of theEPSC was stable for �10 min. The amplitude of an individualEPSC was normalized to the averaged amplitude of the EPSCsthat had occurred during the 10-min baseline recording. Ourdata showed that HFS of the MGBv reliably induced five typesof long-term plasticity at 218 of 249 (87.6%) of the thalamicinputs. Specifically, 21 TC inputs (8.4%) showed long-lastingLTP, 101 inputs (40.6%) showed short-lasting LTP, 24 inputs(9.6%) showed rapid LTD, 72 inputs (28.9%) showed slowLTD, and 31 inputs (12.5%) showed no change after stimula-tion (Fig. 2). Figure 2A shows the average time course of theEPSC amplitude changes induced by 21 long-lasting potenti-ation inputs. HFS initially elicited significant increase in theamplitude of EPSCs, which then plateaued. The normalizedmean maximum EPSC amplitude increased to 210.4 � 19.8%of the baseline (n 21). Representative traces in Fig. 2A,bottom, are averaged EPSCs from five consecutive responsestaken before, immediately after, and 30 min after LTP induc-tion. Figure 2B shows the time course of EPSC amplitudechanges averaged from 101 short-lasting potentiation inputs.Potentiation of the EPSC amplitudes increased during the first10 min after HFS and then returned gradually to the baseline.The mean maximum EPSC amplitude increased to 156.2 �24.5% of the baseline (n 101). Representative traces in Fig.2B, top, are averaged EPSCs from 5 consecutive responsestaken before, immediately after, 10 min after, and 30 min afterLTP induction. Figure 2C shows the time course of EPSCamplitude changes averaged from 24 rapid depression inputs.HFS initially induced a significant decrease in EPSC amplitudeand then stabilized. The mean minimum amplitude of the post-HFS EPSCs decreased to 69.1 � 6.6% of the baseline (n 24).Figure 2D shows the time course of EPSC amplitude changesaveraged from 72 slow depression inputs. EPSC amplitudesgradually decreased during the first 40 min after HFS and thenstabilized. The mean minimum amplitude of the post-HFSEPSCs decreased to 65.2 � 7.2% of the baseline (n 72). Theremaining 31 of the 249 thalamic inputs (12.4%) did notchange after HFS. The EPSC amplitudes for the inputs that didnot show plasticity were stable throughout the testing period,and the mean post-HFS response amplitude was 99.0 � 2.5%of the control (n 31; Fig. 2E).

Pharmacological Characteristics of Synaptic PlasticityElicited by HFS of MGBv

We characterized these different forms of synaptic plasticityin the presence of different pharmacological agents. We firstexamined the involvement of N-methyl-D-aspartate (NMDA)receptors in thalamic HFS-induced long-lasting LTP. We re-corded HFS of MGBv-induced LTP prior to (as a control) and

1748 SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

after application of D-APV (50 �M), a selective antagonist ofthe NMDA receptor. The mean maximum EPSC amplitudeafter HFS was 128.7 � 13.6% of the baseline, and the meanEPSC amplitude 30 min after HFS was 101.7 � 7.23% ofbaseline (n 7; Fig. 3A). The long-lasting LTP was signifi-cantly suppressed in the presence of APV (P 0.05, n 7;Fig. 3B). We next examined whether the short-lasting LTPinduced by thalamic stimulation relies on the activation ofGABAA receptors, because activated AC neurons and stimu-lated MGBv neurons have been shown to drive IPSCs in ACneurons by activating interneurons (Xu et al. 2010). TheGABAA antagonist bicuculline (10 �M) was applied in ACSF.Interestingly, bicuculline enhanced the magnitude of the short-lasting LTP that was obtained and turned short-lasting LTPinto long-lasting LTP. For slices treated with bicuculline, thenormalized mean maximum EPSC amplitude was 284.9 �22.1% of the baseline. The mean EPSC amplitude 30 min afterthe stimulation was 191.0 � 27.6% of the baseline (n 7; Fig.

3C). The magnitude of the LTP obtained in the presence ofbicuculline was significantly larger than that of the control(P 0.05, n 7 for bicuculline, n 20 for control; Fig. 3D).Similarly, bath application of D-APV (50 �M) also preventedthe induction of short-lasting LTP (P 0.05, n 6 for APV,n 20 for control; Fig. 3D). These results suggested that theinduction of short-lasting LTP depends on the activation ofboth NMDA receptors and GABAA receptors. We then exam-ined whether the activation of NMDA receptors is also neces-sary for the induction of rapid LTD. Our results showed thatAPV did not block induction of rapid LTD. The mean ampli-tude of EPSCs 30 min after HFS was 86.3 � 8.7% of thebaseline with APV and 82.7 � 6.7% of the baseline withoutAPV (n 5; Fig. 4A). Previous studies have shown that someforms of the long-term plasticity of auditory TC synapses aremetabotropic glutamate receptor dependent. In particular, type5 metabotropic glutamate receptors (mGluR5) play an impor-tant role in modulating AC plasticity (Blundon et al. 2011;

Fig. 1. Intrinsic properties of pyramidal cells in layers III/IVof the auditory cortex. A: anterograde DiI tracing along thethalamocortical (TC) pathway in the slice. The main route ofdye movement was anterior-lateral, initially along the edgeadjacent to the hippocampus and then across the striatum tothe temporal cortex. MGBv, ventral division of the medialgeniculate; LG, lateral geniculate; Hip, hippocampus; AC,auditory cortex. B: tracings of labeled axons. C: infrared(IR)-differential interference contrast (DIC) image of pyra-midal neurons from layers III/IV. Red arrow points to cellbody, green arrow shows apical dendrite, and yellow arrowindicates position of recording electrode. D: biocytin labelingrevealed pyramidal cells locating in layers III/IV of theauditory cortex, with a long apical dendrite extending tolayer I. E: the voltage response was induced by hyperpolar-ization and depolarization in the same cell shown in C.Current used: �0.15, 0, and �0.15 nA. Resting membranepotential �71 mV. F: average firing rate vs. current curvesfor the pyramidal cell population (n 16). G: example ofexcitatory postsynaptic currents (EPSCs) evoked in pyrami-dal cell pairs following stimulation in the MGBv. The stim-ulation current varied from 0 to 400 �A in steps of 50 �A.H: average EPSC amplitudes for different stimulus intensi-ties (n 22).

1749SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

Kudoh et al. 2002). We therefore examined the involvement ofmGluR5 in rapid LTD and found that bath application of theselective mGluR5 antagonist MPEP (10 �M) significantlyinhibited rapid LTD. The mean amplitude of EPSCs 30 minafter HFS was 92.5 � 8.7% of the baseline with MPEP and66.0 � 9.9% of the baseline without MPEP (P 0.05, n 5;Fig. 4B).

We explored the role of eCB type I receptors (eCB1Rs) inrapid LTD by examining the effects of the eCB1R antagonistAM251 because previous studies have indicated that eCBs, theendogenous ligands for cannabinoid receptors, play a role inmGluR5-dependent LTD (Freund et al. 2003). We found thatAM251 at 10 �M also blocked rapid LTD. The mean ampli-tude of EPSCs 30 min after HFS was 93.3 � 8.7% of thebaseline with AM251 and 79.5 � 11.5% of the baselinewithout AM251 (P 0.05, n 4; Fig. 4C). The histogram ofminimum amplitudes of EPSCs in Fig. 4D indicates that rapidLTD may be mGluR5 (P 0.05, n 5) and eCB1R (P 0.05, n 4) dependent. Finally, we examined the possibleroles of glutamatergic and GABAergic receptors in slow LTD.The mean minimum amplitude of EPSCs 30 min after HFS was78.4 � 7.1% of the baseline with APV and 56.6 � 7.4% of thebaseline without APV (n 7; Fig. 5A). We found that the slowLTD was suppressed significantly (P 0.05, n 7; Fig. 5D)in the presence of 50 �M APV. However, the selective mGluRantagonist MPEP (10 �M) had no significant effect on slowLTD (P 0.72, n 6; Fig. 5D). The mean minimumamplitude of the EPSCs 30 min after HFS was 59.7 � 8.3% ofbaseline with MPEP and 58.5.5 � 7.1% of the baseline withoutMPEP (n 6; Fig. 5B). Finally, we investigated the effectsof bicuculline, an antagonist of GABAA receptors, on slowLTD. Our results show that the slow LTD was almostcompletely suppressed by 5 �M bicuculline (n 6; Fig.5C), which suggests the involvement of GABAA receptorsin slow LTD.

These experiments as a whole indicate that long-lastingLTP involves activation of NMDA receptors and short-lasting LTP involves activation of both NMDA and GABAA

receptors. Rapid LTD involves activation of both mGluR5and eCB1Rs, and slow LTD involves activation of bothNMDA and GABAA receptors. A summary showing thefour types of plasticity and their pharmacological profiles isprovided in Table 3.

Inhibition of Glutamate and GABAA Receptors PreventsMGBv Activation-Induced BF Shift in AC Neurons

We next asked whether these forms of TC synapticplasticity and their responsible receptors could contribute toMGBV stimulation-induced shift of BF in the auditorycortex, as demonstrated by our recent in vivo studies usinganesthetized rats (Zhu et al. 2013). To answer this question,we examined the changes in the frequency-tuning curve ofAC neurons before and after blockade of glutamate andGABAA receptors (see MATERIALS AND METHODS). In Fig. 6A,a cortical neuron was tuned to 23.0 kHz. When the thalamicneuron was electrically stimulated, the frequency-tuningcurve of the cortical neuron sharply tuned to 21.0 kHz. The shiftedBF returned to the control BF �90 min after the stimulation.Application of APV to AC neurons before the stimulation reducedthe auditory responses and abolished the BF shift after HFS ofthe MGBv (Fig. 6B). Specifically, in seven cortical neuronsexamined, the BF shift in APV-treated neurons was 0.43 �0.29 kHz, which was significantly smaller (60% less) than thatwithout APV (P 0.05, n 7; Fig. 6C). In Fig. 6D, a corticalneuron was tuned to 26.0 kHz. When the thalamic neuron waselectrically stimulated, the frequency-tuning curve of the neu-ron tuned to 29.0 kHz. Application of bicuculline and MPEP toAC neurons before the stimulation broadened the frequency-tuning curve and increased the auditory responses. Moreover,

Table 1. Summary of type-dependent neuronal intrinsic properties

LTPlong LTPshort LTDr LTDs NC

n 7 20 8 10 10Rin, M�* 244.10 � 7.66 262.06 � 8.02 188.83 � 6.83 201.06 � 7.72 292.65 � 8.26Vthres, mV* �37.81 � 0.76 �44.63 � 0.49 �37.78 � 0.61 �27.53 � 0.52 �35.42 � 0.69RMP, mV �58.23 � 0.65 �63.53 � 0.72 �67.09 � 0.79 �63.68 � 0.61 �58.85 � 0.68AP amp, mV 108.40 � 0.53 107.58 � 0.46 117.06 � 0.61 112.86 � 0.69 118.21 � 0.73AP width, ms* 1.61 � 0.02 1.78 � 0.02 1.44 � 0.01 2.69 � 0.03 1.50 � 0.02Adaptation ratio* 0.41 � 0.04 0.29 � 0.08 0.35 � 0.06 0.42 � 0.04 0.49 � 0.09Max rate, Hz* 109.29 � 1.53 78.74 � 1.51 109.28 � 2.68 125.78 � 1.34 87.7 � 1.32

Values are means � SE for n cells sampled. LTPlong, long-lasting long-term potentiation; LTPshort, short-lasting LTP; LTDr, rapid long-term depression; LTDs,slow LTD; NC, no change; Rin, input resistance; Vthres, voltage threshold to spike; RMP, resting membrane potential; AP amp, action potential amplitude; APwidth, action potential half-width; adaptation ratio, average of the last 2 interspike intervals (ISIs) divided by the average of the first 2 ISIs; Max rate, maximumspike rate evoked by �300-pA current injection. *P 0.05 (1-way ANOVA).

Table 2. Summary of type-dependent synaptic properties

LTPlong LTPshort LTDr LTDs NC

n 7 20 8 10 10Average onset latency, ms* 3.96 � 0.31 7.01 � 0.43 7.86 � 0.52 6.81 � 0.40 7.63 � 0.55Onset latency variability jitter* 0.38 � 0.12 0.55 � 0.11 0.51 � 0.13 0.73 � 0.16 0.63 � 0.12Amplitude, pA* �47.36 � 0.55 �78.07 � 1.08 �62.06 � 0.79 �105.23 � 0.81 �21.73 � 0.48Rise time, 20–80%, ms* 2.61 � 0.11 5.06 � 0.21 2.97 � 0.61 3.03 � 0.16 4.08 � 0.66Width, ms* 38.12 � 2.40 41.58 � 3.31 38.33 � 2.71 32.88 � 2.63 42.92 � 5.60

Values are means � SE for n cells sampled. *P 0.05 (1-way ANOVA).

1750 SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

application of bicuculline and MPEP abolished the BF shiftafter the onset of MGBv activation (Fig. 6E). In six corticalneurons studied, the BF shift was significantly smaller thanthat without bicuculline and MPEP (P 0.05, n 6; Fig.6F). These experiments indicate that application of APV,MPEP, and bicuculline to AC neurons can inhibit the shift incortical BF induced by HFS of the MGBv. These datafurther support essential roles for these receptors and TCsynaptic plasticity in frequency-specific plasticity of the ratauditory cortex.

Thalamic Activation May Induce Inhibitory Potentiation inAuditory Cortex by Acting on Presynaptic Sites

Our study underscores the important roles of not only thewell-known excitatory glutamate but also the inhibitoryGABA in thalamic activation-mediated frequency-specificcortical plasticity. These data and other evidence accumu-lated in recent years strongly support the hypothesis thatinhibitory plasticity participates in most AC computations(Collingridge et al. 2004; Xu et al. 2010). Despite theadvances in our understanding of glutamate transmission andreceptor functions in mediating long-term cortical plasticity,neither the role nor the precise synaptic (pre- compared withpostsynaptic) mechanism of GABA-mediated plasticity is wellknown. We therefore next investigated whether presynaptictransmitter release and/or postsynaptic receptor responsivenesswas enhanced at GABAergic synapses of AC neurons afterthalamic HFS. mIPSCs in pyramidal cortical neurons were

therefore recorded in the presence of TTX. A stable baseline ofmIPSCs was first recorded for 5 min as a control, followed bya period of TTX washout for 15 min. HFS was then applied tothe MGBv and TTX was subsequently washed in (Fig. 7A). Torecord mIPSCs, pyramidal cells were voltage-clamped at �10mV and ionotropic glutamate receptor-induced spontaneousEPSCs were blocked with CNQX (10 �M) and APV (50 �M)in ACSF. Under these experimental conditions, mIPSCs wererecorded as outward currents (Fig. 7B). These events wereblocked by 10 �M bicuculline (data not shown), which indi-cates that they are mediated by the release of GABA. Thefrequency of mIPSCs 25–30 min after HFS was significantlyincreased in thalamorecipient pyramidal neurons of the audi-tory cortex (from 40.33 � 4.49 to 89.65 � 6.53 events in 1min, n 6, P 0.05; Fig. 7, B–D). The amplitude of themIPSCs was not altered after conditioning HFS (baseline85.38 � 7.08 pA compared with 90.16 � 8.12 pA at 25–30min after HFS, P � 0.05; Fig. 7, B–D). Neither the frequencynor the amplitude of mIPSCs was affected in three neurons thatdid not receive thalamic input from the MGBv (data notshown). A change in the frequency of spontaneous postsynap-tic current indicates a presynaptic mechanism, whereas achange in spontaneous postsynaptic current amplitude indi-cates a change in postsynaptic responsiveness (Van der Kloot1991). These data suggest that the MGBv activation-mediatedfacilitation of the GABAergic synaptic transmission may occurat the presynaptic site.

Fig. 2. Long-term changes of synaptic transmission induced byhigh-frequency stimulation (HFS) of the MGBv: time course ofEPSC amplitude changes with TC inputs, which representslong-lasting long-term potentiation (LTP) (A, n 21), short-lasting LTP (B, n 101), rapid long-term depression (LTD)(C, n 24), slow LTD (D, n 72), or no change (E, n 31)after HFS of the MGBv. The amplitude of individual EPSCswas normalized to the averaged amplitude of EPSCs during the10-min baseline recordings that had been made immediatelybefore HFS of the MGBv. A: representative traces in inset areaveraged EPSCs from 5 consecutive responses taken before(1), immediately after (2), and 30 min after (3) HFS of theMGBv. B: representative traces in inset are averaged EPSCstaken before (1), immediately after (2), and 6 min (3) and 30min (4) after HFS of the MGBv.

1751SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

DISCUSSION

The frequency-specific plasticity of AC neurons inducedby MGBv activation has been studied in mice and bats, butthe underlying cellular mechanisms are not well understood.Our present study of MGBv activation-mediated long-termplasticity in the rat auditory cortex demonstrates that differ-

ent forms of long-term plasticity of TC inputs depend on theactivation of different sets of ionotropic and metabotropicglutamate receptors, GABAA receptors, and eCB receptorsand that MGBv activation induces an enhancement ofGABAergic synaptic transmission that involves presynapticmechanisms.

Fig. 3. Effects of glutamate and GABAergicreceptor antagonists on LTP induced by HFSof the MGBv. A: the ability of HFS of theMGBv to induce long-lasting LTP was blockedby bath application of the NMDA receptorantagonist D-APV (50 �M; n 7). B: sum-mary of data obtained immediately after and30 min after the long-lasting LTP inductiondepicted in A. *P 0.05. C: blockage ofGABAA receptors with 10 �M bicuculline(BIC) changed short-lasting LTP into long-lasting LTP. The ability of HFS to induceshort-lasting LTP was blocked by bath appli-cation of the NMDA receptor antagonist D-APV (n 20 for control, n 7 for BIC, n 6 for APV). D: summary of data obtained at 30min after the short-lasting LTP induction de-picted in C (right) and at the moment when theEPSC was at its maximum (left). *P 0.05.

Fig. 4. Rapid LTD in TC synapses. A: timecourses of rapid LTD recorded in normalmedium (control) or in the presence of 50 �MD-APV. APV (50 �M) did not affect rapidLTD, which suggests that there is no involve-ment of NMDA receptors in rapid LTD (n 5). Vertical bars indicate SEs. B: MPEP (10�M), an antagonist of type 5 metabotropicglutamate receptors (mGluR5), prevented theexpression of rapid LTD, which suggests theinvolvement of an mGluR5 receptor (n 5).C: AM251 (10 �M), an antagonist of endo-cannabinoid (eCB) receptors, prevented theexpression of rapid LTD, which suggests theinvolvement of an eCB1 receptor in rapidLTD (n 4). D: normalized minimum am-plitudes of the EPSCs after HFS of the MGBv

in the absence (control) or presence of 50�MAPV, 10 �M MPEP, or 10 �M AM251.*P 0.05.

1752 SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

Long-Lasting LTP May Represent Monosynaptic PotentiationAfter Thalamic HFS

In the present study, NMDA receptor-dependent rapid andlong-lasting LTP was elicited in rat AC neurons after HFS ofthe MGBv (Fig. 3A). Similarly, NMDA receptor-dependentLTP has been found in the mouse auditory cortex (Liu et al.2013). Because thalamorecipient neurons in the cortex areinterconnected, stimulation of TC afferents may polysynapti-cally activate intracortical feedforward excitatory synapses(Blundon et al. 2011; Speechley et al. 2007). It is unclearwhich specific synaptic pathway showed rapid LTP after HFSof the MGBv. The short onset latency response and low jitter ofonset latency are two principal criteria to identify monosynap-tic TC EPSC (Cruikshank et al. 2002; Rose and Metherate2005). Our study demonstrates that thalamic stimulation-evoked EPSCs had a short onset latency of 3.96 � 0.31 ms andlittle onset latency jitter (0.38 � 0.12 ms) in the inputs thatexpressed rapid and long-lasting LTP. These EPSCs exhibitedlittle change in latency with increased stimulus intensity andfixed latencies with high-frequency repetitive stimulation,which suggests activation of monosynaptic TC connections

(Table 2). We thus assume that the rapid and long-lasting LTPin the present study results from monosynaptic TC inputs thatare based on an NMDA receptor-dependent mechanism.

Short-Lasting LTP May Represent Polysynaptic PotentiationAfter Thalamic HFS

In the present study, the EPSCs evoked in the inputs thatshowed short-lasting LTP had long latencies and high variabil-ity jitter (Table 2), suggesting the activation of polysynapticconnections (Berry and Pentreath 1976; Blundon et al. 2011;Cruikshank et al. 2002; Rose and Metherate 2005; Speechley etal. 2007). In addition, the sensitivity of short-lasting LTP toantagonists of both NMDA and GABAA receptors indicatedthat pyramidal neurons in the auditory cortex receive bothglutamatergic inputs and GABAergic inputs from interneurons(Fig. 3C). Previous studies have reported that the presence ofGABAergic interneurons may contribute to the induction ofshort-lasting LTP (Chapman et al. 1998; Grover and Yan1999). Our data show that the blockage of GABAA receptorsfacilitates the induction of NMDA receptor-dependent short-lasting LTP. We therefore speculate that synapses that ex-

Table 3. Summary of four types of plasticity and their pharmacological profiles

LTPlong LTPshort LTDr LTDs

n 7 7 5 7Amplitude* 2.43 � 0.32 1.77 � 0.30 0.44 � 0.08 0.63 � 0.12Slope* 4.29 � 0.30 0.35 � 0.32 1.81 � 0.1 0.01 � 0Recovery time, min 22.12 � 3.08 25.03 � 4.79Pharmacological profile APV� APV� APV� MPEP�

BIC� MPEP� APV�AM251� BIC�

Parameter values are means � SE for n cells sampled. Amplitude, amplitude of LTP (LTD); slope, slope of LTP (LTD) as measured at half-amplitude of LTP(LTD); recovery time, time required for excitatory postsynaptic current (EPSC) to recover from LTD/LTP to 85% of its amplitude; �, had an effect on LTP(LTD); �, had no effect on LTP (LTD); BIC, bicuculline. *P 0.05 (1-way ANOVA).

Fig. 5. Slow LTD in TC synapses. A: D-APV(50 �M) abolished slow LTD, which suggeststhe involvement of NMDA receptors (n 7).B: time courses of slow LTD recorded innormal medium (control) or in the presence of10 �M MPEP. MPEP (10 �M) did not affectslow LTD, which suggests the absence of in-volvement of mGluR5 receptors in slow LTD(n 6). C: time courses of slow LTD recordedin normal medium (control) or in the presenceof 10 �M BIC. Slow LTD was almost com-pletely suppressed (n 6) when 10 �M BICwas applied. D: maximal changes in the am-plitudes of EPSC after HFS of the MGBv in theabsence (control) or presence of 10 �MMPEP, 50 �M APV, or 10 �M BIC. *P 0.05.

1753SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

hibit short-lasting LTP receive either more synaptic inhibi-tion or less excitatory drive and therefore have a highthreshold for producing LTP. Interestingly, our experimentsshow that short-lasting LTP depends on both NMDA andGABAA receptors. However, the induction of long-lasting

and short-lasting LTP in the hippocampus and visual cortexalso depends on different NMDA receptor subtypes (Castro-Alamancos and Connors 1996; Volianskis et al. 2013). Inthe future, it will be interesting to investigate whetherdifferent subtypes of NMDA receptors are present that

Fig. 6. Blockade of glutamate and GABAergic receptors inhibited best frequency (BF) shifts of cortical neurons evoked by HFS of the MGBv. A: HFS of theMGBv evoked a BF shift [electrical stimulation (ES)] of an auditory cortical (AC) neuron. B: 0.05 �l of 10 mM APV (pH 7.0; dissolved in 0.9% saline) appliedto auditory cortex reduced the responses and inhibited BF shifts of the cortical neuron. C: changes in the BF shift after HFS of the MGBv in the absence (control)or presence of APV. *P 0.05, n 7. D: HFS of the MGBv evoked a BF shift (ES) of an AC neuron. E: 0.05 �l of 1 �M BIC and 0.5 �M MPEP (pH 7.0;dissolved in 0.9% saline) applied to auditory cortex increased the responses and inhibited BF shifts of the cortical neuron. F: changes in BF shift after HFS ofthe MGBv in the absence (control) or presence of BIC and MPEP. *P 0.05, n 6.

Fig. 7. HFS of the MGBv increased the frequency but not the amplitude of miniature inhibitory postsynaptic current (mIPSC) in AC neurons. A: experimentaldesign for the recording of mIPSCs. B: representative example of mIPSCs recorded from an AC neuron that received synaptic input from the MGBv underbaseline conditions and 25–30 min after HFS. Calibration: 200 ms, 50 pA. C: cumulative probability curve of interevent intervals (left) for mIPSCs from theneuron shown in A was shifted 25–30 min after HFS compared with baseline, which indicates enhanced transmitter release. In contrast, no change in thecumulative probability curve of amplitudes (right) of mIPSCs was observed. A total of 70 events for baseline and 200 events that occurred at 25–30 min afterHFS were plotted. D: summary of the change in frequency and amplitude of mIPSCs from 10 thalamorecipient AC neurons. *P 0.05; n.s., not significant(P � 0.05).

1754 SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

might contribute to the induction of short-lasting LTP in ourmodel.

Rapid LTD May Represent eCBR-DependentmGluR-Dependent LTD

In the present study we found that the rapid LTD follow-ing HFS of the MGBv was dependent on the activation ofmGluR5 but not of ionotropic NMDA receptors (Fig. 4).Several previous studies have also shown the importance ofmGluR-dependent LTD in the auditory cortex, visual cortex,striatum, cerebellum, and hippocampus (Blundon et al.2011; Calabresi et al. 1992; Kudoh et al. 2002; Luscher andHuber 2010). Studies have further demonstrated thatmGluR-initiated LTD involves the activity of a retrogrademessenger, the eCB, that is synthesized and released from apostsynaptic cell by the activation of mGluRs (Gerdemanand Lovinger 2003). Upon release, eCBs act presynapticallyto regulate the probability of glutamate release and to alterrelease mechanisms after mGluR stimulation (Grueter et al.2007). On the basis of these and our present findings, wespeculate that HFS of the MGBv led to the sequentialactivation of mGluR5 in pyramidal neurons and the activa-tion triggered the release of neuromodulators such as eCBs(Freund et al. 2003; Huang et al. 2008). The eCBs in turnactivated receptors in those distant stimulated or nonstimu-lated synapses, and the synaptic transmission in those syn-apses was inhibited, leading to rapid LTD (Serrano et al.2006).

Slow LTD May Represent Thalamocortical PolysynapticDepression After Thalamic HFS

APV, but not MPEP, effectively blocked the induction ofslow LTD in the present study (Fig. 5, A and B), whichindicates that NMDA receptors play an important role in theinduction of slow LTD. Interestingly, our data further demon-strate that slow LTD was also suppressed significantly bybicuculline, which indicates the involvement of GABAA re-ceptors (Fig. 5C). Similar involvement of GABA receptoractivity has been found in the mouse auditory cortex (Liu et al.2013). Previous reports have shown that the inhibition levelmediated by GABA receptors determines the direction ofsynaptic plasticity during HFS-induced LTD in the neocortexby modulating postsynaptic depolarization (Artola et al. 1990).Low levels of postsynaptic depolarization lead to LTD,whereas stronger depolarization leads to LTP (Artola et al.1990; Ngezahayo et al. 2000). Our overall conclusion is thatsynapses in pyramidal neurons are activated when HFS isapplied to the MGBv. However, inhibitory interneurons werealso driven by axon collaterals of MGBv neurons or by acti-vated AC neurons (Fig. 7). HFS may provide an optimaldepolarization level for LTD induction by concurrently acti-vating GABAergic and glutamatergic transmission. It is there-fore not surprising that HFS of the MGBv can induce a slowtype of LTD in AC neurons.

A Possible Synaptic Mechanism Underlying MGBvStimulation-Induced Auditory Plasticity

Serial section reconstruction of single MGBv axon in theauditory cortex revealed that a single MGBv axon may projectonto many AC neurons, and a single neuron in the auditory

cortex may receive inputs from a large number of thalamicneurons (Cetas et al. 1999; de Venecia and McMullen 1994;Hashikawa et al. 1995). Paired recordings of neurons in theMGBv and the auditory cortex reveal that TC projections linkMGBv and AC neurons with similar BFs (Miller et al. 2001).There are also abundant cortico-cortical connections within theauditory cortex (McMullen and de Venecia 1993). In thepresent study, cells that expressed short-lasting LTP are clearlydifferentiated from other cells by their low thresholds, adapta-tion ratio, and maximum firing rates, as well as their large andbroad EPSCs. Conversely, cells that expressed slow LTD arecharacterized by their high thresholds, broad spikes, as well aslarge and narrow EPSCs (Tables 1 and 2). These differencesbetween the plasticity classes in both intrinsic and synapticproperties suggest that they may represent a separate class ofneurons or receive a separate class of synaptic inputs. On thebasis of previous findings that AC receptive field plasticity ishighly frequency specific (Weinberger 2007; Zhu et al. 2013)together with the present data showing the importance of inhib-itory and excitatory synaptic transmission in AC frequency-spe-cific plasticity (Fig. 6), we developed a possible TC synapticmechanism model for frequency-specific AC plasticity that isinduced by focal thalamic stimulation.

Specifically, in Fig. 8A, AC neuron A2 receives excitatory TCinputs from thalamic neuron M2, which has the same BF as A2,and A2 also receives collateral projection inputs from such otherthalamic neurons as M1 and M3, which have BFs that differ fromthat of A2. In addition to the TC inputs, pyramidal neurons in theauditory cortex also receive excitatory inputs from other glutama-tergic pyramidal neurons, such as A1 and A3, and inhibitory inputsfrom GABAergic interneurons, such as I1 and I2.

If A2 has the same BF as the stimulated M2, the TC synapsebetween M2 and A2 will show NMDA receptor-dependent LTPafter activation of M2 by focal HFS (Fig. 8A, Fig. 3A). Enhance-ments of the excitatory postsynaptic potential trigger an increaseof A2’s response to tone. If, on the other hand, the BF of therecorded AC neuron is different from the BF of the stimulatedMGBv neuron, the underlying synaptic mechanism is more com-plex. We stimulated M1 in the MGBv and recorded the activity ofneuron A2 in the auditory cortex (Fig. 8A). After HFS, synapsesof collateral projections from M1 to A2 showed NMDA-depen-dent LTP. The facilitated cortical neuron A1 may also enhance theresponses of A2 through intracortical excitatory projections. Thesetwo mechanisms cause the recorded AC neuron to show facilita-tion at or near the BF of the stimulated thalamic neuron. Wehypothesize that there are two possible reasons for inhibition at ornear the BF of the recorded AC neuron. First, stimulated MGBvneurons and facilitated AC neurons can decrease the responses ofAC neurons through inhibitory interneurons. As shown in Fig. 8A,the inhibitory interneuron I1 was also driven by axon collateralsfrom M1 or A1 when focal HFS was applied to the thalamicneuron M1. The action of interneuron I1 produced membranehyperpolarization of A2. Synapses from M2 to A2 may be acti-vated by the limited spread of electrical stimulation. HFS mayprovide an optimal depolarization level for LTD induction byconcurrently activating GABAergic and glutamatergic transmis-sion. A second possibility is that heterosynaptic suppression isinvolved in the inhibition. As shown in Fig. 8C, HFS leads toglutamate release, which triggers NMDA-dependent LTP in the

1755SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

activated synapse. Glutamate triggers a simultaneous heterosyn-aptic depression (Fig. 8C).

Every neuron in the neocortex has thousands of synapses.Activation of only a few hundred of these synapses canevoke cell firing (Crochet et al. 2005). As shown in Fig. 8B,neuron A2 in layers III/IV of the auditory cortex hasthousands of synapses. A tone at a frequency within thereceptive field of A2 might activate only a portion of thesesynapses. Another tone at a different frequency can activeanother group of these synapses. These synapses may showdifferent forms of long-term plasticity after focal thalamicHFS. Changes in postsynaptic potential that respond to acertain frequency should be the summation of the LTP andLTD that are mediated by that set of activated synapsesactivated by the tone at this frequency. The summativechange in postsynaptic potential in response to a certainfrequency may therefore be LTP, LTD, or no change (Fig.8B, top).

However, one would need to be aware that the excitability ofstimulated neurons can additionally be modified by conditionalstimulation. Bidirectional changes in the intrinsic excitabilityof CA1 hippocampal pyramidal neurons are often observedafter induction of synaptic LTP or LTD (Li et al. 2004). It ispossible that the intrinsic excitability of the stimulated MGBvneurons is modified by HFS, which in turn affects the auditorysynaptic plasticity. Future studies will be required to investi-gate these possibilities.

In addition, previous studies have revealed neuronal plasticityin the auditory cortex in both awake and behaving animals ofdifferent species (Banerjee and Liu 2013; David et al. 2012; Guoet al. 2013; Jaramillo and Zador 2011). A very recent study alsoconfirmed that in mice under anesthesia the LTP and LTD

induced by thalamic stimulation can contribute to the remodelingof receptive fields in the auditory cortex (Liu et al. 2013). Ourresults further demonstrated that the plasticity evoked by repeti-tive thalamic stimulation could be generated through a fundamen-tal TC circuit alone without inputs from other sources. Otherfactors like attention or cross-modality effect in awake animalsmay modulate the characteristics of neuronal plasticity generatedby this TC pathway through different mechanisms.

ACKNOWLEDGMENTS

We thank Dr. Yutian Wang at the University of British Columbia (Van-couver, BC, Canada) for his valuable suggestions concerning the writing ofthis manuscript. We are also grateful to Dr. Enquan Gao at WashingtonUniversity School of Medicine (St. Louis, MO), Tara Janes at the Universityof Calgary, and Yi Zhou at Third Military Medical University for carefulreading and editing of the manuscript.

GRANTS

This work was supported by grants from the National Natural Foundationof China (31070970, 31100759, 31271177) and the Scientific Foundation ofChongqing (CSTC2011BB5039) to Z.-R. Zhu and Y. Xiong. We also ac-knowledge the support by a grant from the Natural Sciences and EngineeringResearch Council of Canada (NSERC) to F. Xu.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: Z.z. and X.Y. conception and design of research; Z.z.,J.W., R.s., C.f., and J.h. performed experiments; Z.z., X.F., and H.b. analyzeddata; Z.z. and C.f. prepared figures; Z.z. and X.F. drafted manuscript; Z.z.,X.F., C.p., M.Z., and Z.J. edited and revised manuscript; Z.z., X.F., and Z.J.

Fig. 8. Hypothetical model for illustrating the role of the TC system in frequency-specific plasticity of AC neurons. A: AC neuron A2 receives TC inputs fromM2, which has the same BF as A2. A2 also receives TC inputs from MGBv neurons that have BFs differing from those of A2, such as M1 and M3. In additionto the TC inputs, A2 receives excitatory intracortical inputs from other AC neurons, such as consecutive A1 and A3. In addition, A2 also receives inputs fromsuch GABAergic inhibitory interneurons as I1 and I2. Spectral information therefore converges on A2 by the means of a combination of TC and intracorticalpathways. B: amplification of neuron A2 (dashed box in A) to show the distribution of its synapses. A2 receives impulses from thousands of synapses, and a givenfrequency activates only a portion of these synapses. Activation of this portion of synapses causes neuron A2 to fire. Another frequency in the receptive fieldcan activate a different portion of synapses to A2. Different forms of changes in synaptic transmission in A2 can be induced by HFS of the MGBv. Changes inpostsynaptic potential to a certain stimulus frequency should be a summation of LTP and LTD that are mediated by that set of synapses activated by the frequency.C: simplified schematic diagram for long-lasting changes of synaptic transmission. Left: a condition where one of the synapses is activated by HFS. HFS leadsto the release of presynaptic vesicles containing glutamate (Glu) that activate NMDA receptors, which results in a potentiation of this activated synapse.Postsynaptic depolarization that is induced by HFS is small in some cases because of the inhibition exerted by interneurons. Right: Glu simultaneously triggersheterosynaptic depression. This begins with the activation of an mGlu5 receptor, which then triggers a greater release of a neuromodulator (indicated by greencircles). This modulator can diffuse and activate its receptors in distal nonactivated synapses, which depresses the target synapse.

1756 SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

approved final version of manuscript; X.F., J.W., C.p., H.b., and X.Y. inter-preted results of experiments.

REFERENCES

Allen CB, Celikel T, Feldman DE. Long-term depression induced by sensorydeprivation during cortical map plasticity in vivo. Nat Neurosci 6: 291–299,2003.

Artola A, Brocher S, Singer W. Different voltage-dependent thresholds forinducing long-term depression and long-term potentiation in slices of ratvisual cortex. Nature 347: 69–72, 1990.

Banerjee SB, Liu RC. Storing maternal memories: hypothesizing an interac-tion of experience and estrogen on sensory cortical plasticity to learn infantcues. Front Neuroendocrinol 34: 300–314, 2013.

Beaulieu C, Cynader M. Effect of the richness of the environment on neuronsin cat visual cortex. II. Spatial and temporal frequency characteristics. BrainRes Dev Brain Res 53: 82–88, 1990.

Berry MS, Pentreath VW. Criteria for distinguishing between monosynapticand polysynaptic transmission. Brain Res 105: 1–20, 1976.

Blundon JA, Bayazitov IT, Zakharenko SS. Presynaptic gating of postsyn-aptically expressed plasticity at mature thalamocortical synapses. J Neurosci31: 16012–16025, 2011.

Buonomano DV, Merzenich MM. Cortical plasticity: from synapses to maps.Annu Rev Neurosci 21: 149–186, 1998.

Cai X, Kallarackal AJ, Kvarta MD, Goluskin S, Gaylor K, Bailey AM, LeeHK, Huganir RL, Thompson SM. Local potentiation of excitatory syn-apses by serotonin and its alteration in rodent models of depression. NatNeurosci 16: 464–472, 2013.

Calabresi P, Maj R, Pisani A, Mercuri NB, Bernardi G. Long-term synapticdepression in the striatum: physiological and pharmacological characteriza-tion. J Neurosci 12: 4224–4233, 1992.

Castro-Alamancos MA, Connors BW. Short-term synaptic enhancement andlong-term potentiation in neocortex. Proc Natl Acad Sci USA 93: 1335–1339, 1996.

Cetas JS, de Venecia RK, McMullen NT. Thalamocortical afferents ofLorente de No: medial geniculate axons that project to primary auditorycortex have collateral branches to layer I. Brain Res 830: 203–208, 1999.

Chapman CA, Perez Y, Lacaille JC. Effects of GABAA inhibition on theexpression of long-term potentiation in CA1 pyramidal cells are dependenton tetanization parameters. Hippocampus 8: 289–298, 1998.

Collingridge GL, Isaac JT, Wang YT. Receptor trafficking and synapticplasticity. Nat Rev Neurosci 5: 952–962, 2004.

Coq JO, Xerri C. Environmental enrichment alters organizational features ofthe forepaw representation in the primary somatosensory cortex of adult rats.Exp Brain Res 121: 191–204, 1998.

Crochet S, Chauvette S, Boucetta S, Timofeev I. Modulation of synaptictransmission in neocortex by network activities. Eur J Neurosci 21: 1030–1044, 2005.

Cruikshank SJ, Rose HJ, Metherate R. Auditory thalamocortical synaptictransmission in vitro. J Neurophysiol 87: 361–384, 2002.

David SV, Fritz JB, Shamma SA. Task reward structure shapes rapidreceptive field plasticity in auditory cortex. Proc Natl Acad Sci USA 109:2144–2149, 2012.

de Venecia RK, McMullen NT. Single thalamocortical axons diverge tomultiple patches in neonatal auditory cortex. Brain Res Dev Brain Res 81:135–142, 1994.

Duffy SN, Nguyen PV. Postsynaptic application of a peptide inhibitor ofcAMP-dependent protein kinase blocks expression of long-lasting synapticpotentiation in hippocampal neurons. J Neurosci 23: 1142–1150, 2003.

Feldman DE. Synaptic mechanisms for plasticity in neocortex. Annu RevNeurosci 32: 33–55, 2009.

Feldman DE, Brecht M. Map plasticity in somatosensory cortex. Science 310:810–815, 2005.

Freund TF, Katona I, Piomelli D. Role of endogenous cannabinoids insynaptic signaling. Physiol Rev 83: 1017–1066, 2003.

Gagolewicz PJ, Dringenberg HC. NR2B-subunit dependent facilitation oflong-term potentiation in primary visual cortex following visual discrimi-nation training of adult rats. Eur J Neurosci 34: 1222–1229, 2011.

Gerdeman GL, Lovinger DM. Emerging roles for endocannabinoids inlong-term synaptic plasticity. Br J Pharmacol 140: 781–789, 2003.

Grover LM, Yan C. Blockade of GABAA receptors facilitates induction ofNMDA receptor-independent long-term potentiation. J Neurophysiol 81:2814–2822, 1999.

Grueter BA, McElligott ZA, Winder DG. Group I mGluRs and long-termdepression: potential roles in addiction? Mol Neurobiol 36: 232–244, 2007.

Guo F, Intskirveli I, Blake DT, Metherate R. Tone-detection trainingenhances spectral integration mediated by intracortical pathways in primaryauditory cortex. Neurobiol Learn Mem 101: 75–84, 2013.

Habib D, Dringenberg HC. Surprising similarity between mechanisms mediatinglow (1 Hz)- and high (100 Hz)-induced long-lasting synaptic potentiation inCA1 of the intact hippocampus. Neuroscience 170: 489–496, 2010.

Hardingham N, Glazewski S, Pakhotin P, Mizuno K, Chapman PF, GieseKP, Fox K. Neocortical long-term potentiation and experience-dependentsynaptic plasticity require alpha-calcium/calmodulin-dependent protein ki-nase II autophosphorylation. J Neurosci 23: 4428–4436, 2003.

Hashikawa T, Molinari M, Rausell E, Jones EG. Patchy and laminarterminations of medial geniculate axons in monkey auditory cortex. J CompNeurol 362: 195–208, 1995.

Huang Y, Yasuda H, Sarihi A, Tsumoto T. Roles of endocannabinoids inheterosynaptic long-term depression of excitatory synaptic transmission invisual cortex of young mice. J Neurosci 28: 7074–7083, 2008.

Jafari MR, Zhang Y, Yan J. Multiparametric changes in the receptive fieldof cortical auditory neurons induced by thalamic activation in the mouse.Cereb Cortex 17: 71–80, 2007.

Jaramillo S, Zador AM. The auditory cortex mediates the perceptual effectsof acoustic temporal expectation. Nat Neurosci 14: 246–251, 2011.

Kudoh M, Sakai M, Shibuki K. Differential dependence of LTD on gluta-mate receptors in the auditory cortical synapses of cortical and thalamicinputs. J Neurophysiol 88: 3167–3174, 2002.

Li CY, Lu JT, Wu CP, Duan SM, Poo MM. Bidirectional modification ofpresynaptic neuronal excitability accompanying spike timing-dependentsynaptic plasticity. Neuron 41: 257–268, 2004.

Liu BH, Wu GK, Arbuckle R, Tao HW, Zhang LI. Defining corticalfrequency tuning with recurrent excitatory circuitry. Nat Neurosci 10:1594–1600, 2007.

Liu X, Wang C, Pan C, Yan J. Physiological correspondence dictates corticallong-term potentiation and depression by thalamic induction. Cereb Cortex(September 17, 2013). doi:10.1093/cercor/bht259.

Luscher C, Huber KM. Group 1 mGluR-dependent synaptic long-termdepression: mechanisms and implications for circuitry and disease. Neuron65: 445–459, 2010.

Ma X, Suga N. Specific and nonspecific plasticity of the primary auditory cortexelicited by thalamic auditory neurons. J Neurosci 29: 4888–4896, 2009.

MacDonald JF, Jackson MF, Beazely MA. Hippocampal long-term synapticplasticity and signal amplification of NMDA receptors. Crit Rev Neurobiol18: 71–84, 2006.

Malenka RC, Bear MF. LTP and LTD: an embarrassment of riches. Neuron44: 5–21, 2004.

McMullen NT, de Venecia RK. Thalamocortical patches in auditory neocor-tex. Brain Res 620: 317–322, 1993.

Miller LM, Escabi MA, Schreiner CE. Feature selectivity and interneuronalcooperation in the thalamocortical system. J Neurosci 21: 8136–8144,2001.

Ngezahayo A, Schachner M, Artola A. Synaptic activity modulates theinduction of bidirectional synaptic changes in adult mouse hippocampus. JNeurosci 20: 2451–2458, 2000.

Norena AJ, Gourevitch B, Aizawa N, Eggermont JJ. Spectrally enhancedacoustic environment disrupts frequency representation in cat auditorycortex. Nat Neurosci 9: 932–939, 2006.

Recanzone GH, Schreiner CE, Merzenich MM. Plasticity in the frequencyrepresentation of primary auditory cortex following discrimination trainingin adult owl monkeys. J Neurosci 13: 87–103, 1993.

Richardson RJ, Blundon JA, Bayazitov IT, Zakharenko SS. Connectivitypatterns revealed by mapping of active inputs on dendrites of thalamorecipi-ent neurons in the auditory cortex. J Neurosci 29: 6406–6417, 2009.

Rose HJ, Metherate R. Auditory thalamocortical transmission is reliable andtemporally precise. J Neurophysiol 94: 2019–2030, 2005.

Serrano A, Haddjeri N, Lacaille JC, Robitaille R. GABAergic networkactivation of glial cells underlies hippocampal heterosynaptic depression. JNeurosci 26: 5370–5382, 2006.

Smith PH, Populin LC. Fundamental differences between the thalamocorticalrecipient layers of the cat auditory and visual cortices. J Comp Neurol 436:508–519, 2001.

Speechley WJ, Hogsden JL, Dringenberg HC. Continuous white noiseexposure during and after auditory critical period differentially alters bidi-rectional thalamocortical plasticity in rat auditory cortex in vivo. Eur JNeurosci 26: 2576–2584, 2007.

1757SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from

Van der Kloot W. The regulation of quantal size. Prog Neurobiol 36: 93–130,1991.

Volianskis A, Bannister N, Collett VJ, Irvine MW, Monaghan DT, Fitz-john SM, Jensen MS, Jane DE, Collingridge GL. Different NMDAreceptor subtypes mediate induction of long-term potentiation and twoforms of short-term potentiation at CA1 synapses in rat hippocampus invitro. J Physiol 591: 955–972, 2013.

Weinberger NM. Associative representational plasticity in the auditory cor-tex: a synthesis of two disciplines. Learn Mem 14: 1–16, 2007.

Weinberger NM, Ashe JH, Metherate R, McKenna TM, Diamond DM,Bakin J. Retuning auditory cortex by learning: a preliminary model ofreceptive field plasticity. Concepts Neurosci 1: 91–131, 1990.

Xu H, Kotak VC, Sanes DH. Normal hearing is required for the emergenceof long-lasting inhibitory potentiation in cortex. J Neurosci 30: 331–341,2010.

Zhu ZR, Xu FL, Wu JH, Ren SC, Zhang YH, Hu B, Zhang J, Han L,Xiong Y. Frequency-specific plasticity of the auditory cortex elicited bythalamic stimulation in the rat. Neurosci Lett 555: 30–35, 2013.

1758 SYNAPTIC MECHANISMS OF AUDITORY CORTICAL PLASTICITY

J Neurophysiol • doi:10.1152/jn.00180.2013 • www.jn.org

on May 14, 2015

Dow

nloaded from