Synaptic correlates of fear extinction in the amygdala - CBMEG

Synaptic correlates of fear extinction in the amygdala - CBMEG

a r t i c l e sSynaptic correlates of fear extinction in the amygdalaTaiju Amano 1,2 , Cagri T Unal 1,2 & Denis Paré 1© 2010 Nature America, Inc. All rights reserved.Anxiety disorders such as post-traumatic stress are characterized by an impaired ability to learn that cues previously associatedwith danger no longer represent a threat. However, the mechanisms underlying fear extinction remain unclear. We found that fearextinction in rats was associated with increased levels of synaptic inhibition in fear output neurons of the central amygdala (CEA).This increased inhibition resulted from a potentiation of fear input synapses to GABAergic intercalated amygdala neurons thatproject to the CEA. Enhancement of inputs to intercalated cells required prefrontal activity during extinction training and involvedan increased transmitter release probability coupled to an altered expression profile of ionotropic glutamate receptors. Overall,our results suggest that intercalated cells constitute a promising target for pharmacological treatment of anxiety disorders.It is commonly believed that understanding the mechanisms underlyingfear extinction will ultimately lead to improvements in thetreatment of anxiety disorders 1,2 . Consistent with this, the approachused by clinicians to treat anxiety disorders is similar to that usedto extinguish conditioned fear responses in the laboratory. In bothcases, the subject is repeatedly presented with the feared object (orconditioned stimulus, CSt) in the absence of adverse consequences(or unconditioned stimulus, USt), leading to fear extinction.The main amygdala output for fear responses is the central medialnucleus (CEm). Indeed, amygdala projections to the periaqueductalgray, controlling behavioral freezing 3 , originate from CEm 4 . On theinput side, the lateral amygdala is the main target of thalamic andcortical structures conveying CSt information to the amygdala 5,6 .Although the lateral nucleus is a critical site of plasticity for conditionedfear 7,8 , it does not project to CEm 9–11 . However, it can influenceCEm indirectly via the basolateral nucleus (BLA) 9–11 . Consistent withthis, post-training BLA lesions abolish conditioned fear responses 12 .However, many neurons in the basolateral complex continue to firestrongly to the CSt after extinction training 13–15 , highlighting thecentral paradox of extinction. How does extinction training block fearexpression despite the persistence of CSt-evoked responses in BLA?BLA can influence CEm via direct glutamatergic projections 9–11,16and through indirect di-synaptic routes involving the glutamatergicexcitation of GABAergic intercalated (ITC) 16 or central lateral (CEl)neurons 9–11 that project to CEm 10,11,16,17 . Thus, an increased recruitmentof ITC or CEl neurons by BLA inputs about the CSt mightaccount for the reduction of fear expression after extinction, despitethe persistence of CSt-evoked responses in BLA. Moreover, giventhat ITC cells receive a strong excitatory projection from the infralimbiccortex 18 , a prefrontal region required for the consolidationof extinction 2 , the increased recruitment of ITC cells by BLA inputsmight depend on infralimbic activity. Thus, we sought to test whetherextinction training alters the responsiveness of ITC and CEA neuronsto BLA inputs and to assess whether such changes are dependent oninfralimbic activity. We found that extinction was associated with aninfralimbic-dependent potentiation of BLA inputs to ITC cells thatled to an increased inhibition of fear-output CEm neurons.RESULTSTo test whether extinction depends on increased levels of feedforwardinhibition in CEm, we first compared the responses of CEm neuronsto BLA inputs (Fig. 1a) in coronal slices of the amygdala obtainedfrom rats that were previously subjected to just fear conditioning(n = 16) versus rats that were fear conditioned and trained on extinctionthe next day (n = 14; Fig. 1b). This data was compared with thatobtained in naive rats (n = 11) and rats presented with the CSt and UStin an unpaired fashion (n = 10; Fig. 1b). We examined the behaviorof these rats (Fig. 1c) and then analyzed the effects of the trainingprocedures on the responsiveness of CEm neurons to BLA inputsin vitro (Fig. 1d,e). The individuals carrying out the in vitro experimentsand scoring the rats’ behavior were blind to group identity.Analysis of percent time spent freezing (Fig. 1c) confirmed thatrats that were only subjected to fear conditioning (Fig. 1c) versusfear conditioning and extinction (Fig. 1c) exhibited nearly identicallevels of conditioned freezing by the end of the fear-conditioningsession (day 2). Although rats from the unpaired groups (Fig. 1c) didnot receive paired CSt-USt presentations on day 2, they did expresssignificant freezing levels (paired t test, habituation versus last CSt,P = 0.002), which presumably represents contextual freezing. On day 3in a different context, only rats from the extinction group (Fig. 1c)received CSt presentations. We measured freezing in the unpaired orfear-conditioned rats during corresponding 30-s periods and foundsignificantly higher freezing levels in the extinction group than inthe other groups (ANOVA, F 2,59= 37.83, P = 0.0001; Bonferronicorrectedpost hoc t tests, P ≤ 0.0001).We anesthetized the rats 24 h after exposure to the extinction trainingcontext and prepared coronal sections from their amygdala. Weobtained patch recordings from samples of 10–16 CEm cells per groupand compared their responsiveness to electrical stimuli deliveredat a standard position in BLA (Fig. 1a). We carried out these tests1 Center for Molecular and Behavioral Neuroscience, Rutgers, The State University of New Jersey, Newark, New Jersey, USA. 2 These authors contributed equally to thiswork. Correspondence should be addressed to D.P. ( 3 December 2009; accepted 11 January 2010; published online 7 March 2010; doi:10.1038/nn.2499nature NEUROSCIENCE VOLUME 13 | NUMBER 4 | APRIL 2010 489

a r t i c l e s© 2010 Nature America, Inc. All rights reserved.CEm. This potentiation depended on an increased transmitter releaseprobability and an altered expression profile or phosphorylation levelof ionotropic glutamate receptors at BLA synapses onto ITC cells.Finally, the extinction-related enhancement in the efficacy of BLAinputs to ITC cells was dependent on activity in the infralimbic cortexduring extinction training.Reduced fear depends on increased CEm inhibitionDespite years of investigations, there is still uncertainty regarding thenature of CEA control over conditioned fear. CEm output neuronsare thought to be GABAergic, raising the question of whether conditionedfear responses are generated by an increase or a decrease inthe CSt-evoked responses of CEm neurons. One study in rabbits 21reported that fear conditioning reduced the CSt responsiveness ofCEA neurons with physiologically identified projections to the brainstem.In contrast, two other studies in rats (C.E. Chang, J.D. Berke &S. Maren, Soc. Neurosci. Abstr. 478.14, 2008) and mice 22 reported theopposite. However, the latter conclusion is supported by the resultsof stimulation, lesion and inactivation studies in which proceduresthat increased or decreased CEA activity were generally found to causeaugmented or reduced fear expression, respectively 3 .Our results provide additional support for this notion. Indeed, weobserved that treatments such as extinction and conditioned inhibition,which cause a reduction in fear responsiveness, were associatedwith persistently increased levels of BLA-evoked inhibition in CEmneurons (Supplementary Fig. 8). However, different populations ofGABAergic neurons were responsible for this increased inhibition inconditioned inhibition versus extinction. In conditioned inhibition(Supplementary Fig. 8), the BLA responsiveness of CEl, but not ITC,neurons was increased relative to that seen in fear-conditioned andnaive animals, whereas in extinction (Supplementary Fig. 8), BLAstimuli elicited stronger responses in ITC, but not CEl neurons. Thecell type–specific alterations in BLA responsiveness, seen as a functionof group identity, suggest that changes in neuronal excitability at thestimulation site were not responsible for our results. Consistent withthis, ex vivo studies that examined the impact of prior training on fearconditioning alone, fear conditioning and extinction, or unpairedpresentations of the CSt and USt found no substantial training-relatedchange in the input resistance of BLA neurons (see ref. 23).Mechanisms of increased CEm inhibition after extinctionThe finding that BLA stimuli elicited more inhibition in CEm neuronsfrom fear-extinguished rats than from naive or fear-conditioned ratsis consistent with earlier results indicating that extinction depends, atleast in part, on the strengthening of an inhibitory process 24 . Severalfactors suggest that ITC neurons are important contributors to thisincreased inhibition. First, ITC cell masses contain one dominantcell type that uses GABA as a transmitter and these cells project toCEm 10,25,26 . Second, it was previously shown that the inhibition elicitedby BLA stimuli in CEm neurons is blocked by prior pressureapplication of nonNMDA glutamate receptor antagonists in ITC cellclusters 16 . Finally, ITC lesions 27 or pharmacological inhibition of BLAinputs to ITC cells 28 interfere with fear extinction.In principle, several pre- and postsynaptic mechanisms could leadto an enhanced inhibition of CEm neurons by BLA inputs. In additionto the increased efficacy of BLA synapses onto ITC cells, therecould be a facilitation of GABA release by ITC cells themselves, anincreased expression or altered phosphorylation state of GABA Areceptor subunits in CEm neurons, and/or a change in intracellularchloride homeostasis in CEm cells. Arguing against this last possibility,however, the reversal potential of IPSPs elicited in CEm neuronsby pressure application of a GABA Aagonist did not differ betweenfear-extinguished and fear-conditioned rats. However, there is evidenceof postsynaptic contributions to the enhanced inhibition ofCEm neurons in extinction. Indeed, it was reported that extinctiontraining causes an increase in the expression of alpha 2 GABA Areceptorsubunits in CEA 29 . Thus, these considerations suggest that extinctionprobably engages a variety of control mechanisms to regulatefear expression.How could extinction facilitate the recruitment of ITC cells byCSt-related BLA inputs? Extinction has been shown to depend onNMDA-dependent synaptic plasticity in the amygdala 30–32 . Moreover,stimulation of the infralimbic cortex, which sends a robust glutamatergicprojection to ITC cells 18 , accelerates extinction 33 and inhibitsCEm neurons 34 . These observations, coupled with the fact that BLAinputs to ITC cells undergo NMDA-dependent long-term potentiationwhen paired with sufficient depolarization 35 , suggest that convergenceof BLA and infralimbic inputs to ITC cells during extinctiontraining leads to the NMDA-dependent potentiation of BLA inputsto ITC cells. As a result, the GABAergic output of ITC cells onto CEmneurons would be increased, leading to a reduction of conditionedfear. Given recent data indicating that infralimbic neurons exhibitincreased bursting and CSt-evoked responses following extinctiontraining 33,36 , plasticity in ITC cells might be further facilitated afterextinction training, during a consolidation phase.ConclusionsOverall, our results suggest that fear extinction depends, at least inpart, on an increased inhibition of fear-output CEm neurons. Thisincreased inhibition is caused by an enhanced recruitment of ITC cellsby BLA inputs. Moreover, these changes require infralimbic activityduring extinction training, suggesting that the infralimbic cortexdrives extinction-related plasticity in the amygdala. Because someanxiety disorders are associated with a fear-extinction deficit 20 andhypoactivity of the infralimbic cortex 37,38 , our results suggest thatpharmacological manipulations that enhance the excitability of ITCcells by exploiting their unusual profile of receptor expression 39,40could prove useful for treating anxiety disorders.MethodsMethods and any associated references are available in the online versionof the paper at Supplementary information is available on the Nature Neuroscience website.AcknowledgmentsThis work was supported by National Institute of Mental Health grant RO1MH083710 to D.P.AUTHOR CONTRIBUTIONST.A. and C.T.U. performed all of the electrophysiological experiments and mostof the analyses on ITC and CEA cells, respectively. T.A. performed the behavioraltraining and C.T.U. scored the behavior. D.P. designed the experiments, wrote thepaper and contributed to data analysis.COMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.Published online at and permissions information is available online at Myers, K.M. & Davis, M. Mechanisms of fear extinction. Mol. Psychiatry 12,120–150 (2007).2. Quirk, G.J. & Mueller, D. Neural mechanisms of extinction learning and retrieval.Neuropsychopharmacology 33, 56–72 (2008).nature NEUROSCIENCE VOLUME 13 | NUMBER 4 | APRIL 2010 493

a r t i c l e s© 2010 Nature America, Inc. All rights reserved.3. Davis, M. The role of the amygdala in conditioned and unconditioned fear andanxiety. in The Amygdala: A Functional Analysis (ed. Aggleton, J.P.) 213–287(Oxford University Press, Oxford, 2000).4. Hopkins, D.A. & Holstege, G. Amygdaloid projections to the mesencephalon, ponsand medulla oblongata in the cat. Exp. Brain Res. 32, 529–547 (1978).5. LeDoux, J.E., Cicchetti, P., Xagoraris, A. & Romanski, L.M. The lateral amygdaloidnucleus: sensory interface of the amygdala in fear conditioning. J. Neurosci. 10,1062–1069 (1990).6. Romanski, L.M. & LeDoux, J.E. Information cascade from primary auditory cortexto the amygdala: cortex in the rat. Cereb. Cortex 3, 515–532 (1993).7. Amorapanth, P., LeDoux, J.E. & Nader, K. Different lateral amygdala outputs mediatereactions and actions elicited by a fear-arousing stimulus. Nat. Neurosci. 3, 74–79(2000).8. Maren, S., Yap, S.A. & Goosens, K.A. The amygdala is essential for the developmentof neuronal plasticity in the medial geniculate nucleus during auditory fearconditioning in rats. J. Neurosci. 21, RC135 (2001).9. Krettek, J.E. & Price, J.L. A description of the amygdaloid complex in the rat andcat with observations on intra-amygdaloid axonal connections. J. Comp. Neurol.178, 255–280 (1978).10. Paré, D., Quirk, G.J. & LeDoux, J.E. New vistas on amygdala networks in conditionedfear. J. Neurophysiol. 92, 1–9 (2004).11. Pitkänen, A., Savander, V. & LeDoux, J.E. Organization of intra-amygdaloid circuitriesin the rat: an emerging framework for understanding functions of the amygdala.Trends Neurosci. 20, 517–523 (1997).12. Anglada-Figueroa, D. & Quirk, G.J. Lesions of the basal amygdala block expressionof conditioned fear but not extinction. J. Neurosci. 25, 9680–9685 (2005).13. Herry, C. et al. Switching on and off fear by distinct neuronal circuits. Nature 454,600–606 (2008).14. Hobin, J.A., Goosens, K.A. & Maren, S. Context-dependent neuronal activity in thelateral amygdala represents fear memories after extinction. J. Neurosci. 23,8410–8416 (2003).15. Repa, J.C. et al. Two different lateral amygdala cell populations contribute to theinitiation and storage of memory. Nat. Neurosci. 4, 724–731 (2001).16. Royer, S., Martina, M. & Paré, D. An inhibitory interface gates impulse trafficbetween the input and output stations of the amygdala. J. Neurosci. 19,10575–10583 (1999).17. Petrovich, G.D. & Swanson, L.W. Projections from the lateral part of the centralamygdalar nucleus to the postulated fear conditioning circuit. Brain Res. 763,247–254 (1997).18. McDonald, A.J., Mascagni, F. & Guo, L. Projections of the medial and lateralprefrontal cortices to the amygdala: A Phaseolus vulgaris leucoagglutinin study inthe rat. Neuroscience 71, 55–75 (1996).19. Rescorla, R.A. Conditioned inhibition of fear resulting from negative CS-UScontingencies. J. Comp. Physiol. Psychol. 67, 504–509 (1969).20. Milad, M.R. et al. Presence and acquired origin of reduced recall for fear extinctionin PTSD: results of a twin study. J. Psychiatr. Res. 42, 515–520 (2008).21. Pascoe, J.P. & Kapp, B.S. Electrophysiological characteristics of amygdaloid centralnucleus neurons during Pavlovian fear conditioning in the rabbit. Behav. Brain Res.16, 117–133 (1985).22. Ciocchi, S., Herry, C., Muller, C. & Luthi, A. Fear conditioning– and extinction-inducedneuronal plasticity in the central amygdala. FENS Abstr. 4:057.10 (2008).23. Lin, H.C., Mao, S.C. & Gean, P.W. Block of gamma-aminobutyric acid-A receptorinsertion in the amygdala impairs extinction of conditioned fear. Biol. Psychiatry66, 665–673 (2009).24. Harris, J.A. & Westbrook, R.F. Evidence that GABA transmission mediates contextspecificextinction of learned fear. Psychopharmacology (Berl.) 140, 105–115(1998).25. Paré, D. & Smith, Y. Distribution of GABA immunoreactivity in the amygdaloidcomplex of the cat. Neuroscience 57, 1061–1076 (1993).26. Paré, D. & Smith, Y. The intercalated cell masses project to the central and medialnuclei of the amygdala in cats. Neuroscience 57, 1077–1090 (1993).27. Likhtik, E., Popa, D., Apergis-Schoute, J., Fidacaro, G.A. & Pare, D. Amygdalaintercalated neurons are required for expression of fear extinction. Nature 454,642–645 (2008).28. Jüngling, K. et al. Neuropeptide S–mediated control of fear expression andextinction: role of intercalated GABAergic neurons in the amygdala. Neuron 59,298–310 (2008).29. Heldt, S.A. & Ressler, K.J. Training-induced changes in the expression of GABAAassociatedgenes in the amygdala after the acquisition and extinction of Pavlovianfear. Eur. J. Neurosci. 26, 3631–3644 (2007).30. Falls, W.A., Miserendino, M.J.D. & Davis, M. Extinction of fear-potentiated startle:blockade by infusion of an NMDA antagonist into the amygdala. J. Neurosci. 12,854–863 (1992).31. Lee, H. & Kim, J.J. Amygdalar NMDA receptors are critical for new fear learningin previously fear-conditioned rats. J. Neurosci. 18, 8444–8454 (1998).32. Walker, D.L., Ressler, K.J., Lu, K.T. & Davis, M. Facilitation of conditioned fearextinction by systemic administration or intra-amygdala infusions of d-cycloserine asassessed with fear-potentiated startle in rats. J. Neurosci. 22, 2343–2351 (2002).33. Milad, M.R. & Quirk, G.J. Neurons in medial prefrontal cortex signal memory forfear extinction. Nature 420, 70–74 (2002).34. Quirk, G.J., Likhtik, E., Pelletier, J.G. & Pare, D. Stimulation of medial prefrontalcortex decreases the responsiveness of central amygdala output neurons. J. Neurosci.23, 8800–8807 (2003).35. Royer, S. & Pare, D. Bidirectional synaptic plasticity in intercalated amygdalaneurons and the extinction of conditioned fear responses. Neuroscience 115,455–462 (2002).36. Burgos-Robles, A., Vidal-Gonzalez, I., Santini, E. & Quirk, G.J. Consolidation of fearextinction requires NMDA receptor-dependent bursting in the ventromedial prefrontalcortex. Neuron 53, 871–880 (2007).37. Bremner, J.D., Elzinga, B., Schmahl, C. & Vermetten, E. Structural and functionalplasticity of the human brain in posttraumatic stress disorder. Prog. Brain Res.167, 171–186 (2008).38. Shin, L.M., Rauch, S.L. & Pitman, R.K. Amygdala, medial prefrontal cortex, andhippocampal function in PTSD. Ann. NY Acad. Sci. 1071, 67–79 (2006).39. Herkenham, M. & Pert, C.B. Light microscopic localization of brain opiate receptors:a general autoradiographic method which preserves tissue quality. J. Neurosci. 2,1129–1149 (1982).40. Jacobsen, K.X., Hoistad, M., Staines, W.A. & Fuxe, K. The distribution of dopamineD1 receptor and mu-opioid receptor 1 receptor immunoreactivities in the amygdalaand interstitial nucleus of the posterior limb of the anterior commissure: relationshipsto tyrosine hydroxylase and opioid peptide terminal systems. Neuroscience 141,2007–2018 (2006).494 VOLUME 13 | NUMBER 4 | APRIL 2010 nature NEUROSCIENCE

© 2010 Nature America, Inc. All rights reserved.ONLINE METHODSAll procedures were approved by the Institutional Animal Care and Use Committeeof Rutgers State University, in compliance with the Guide for the Care and Use ofLaboratory Animals (US Department of Health and Human Services).Behavior. We used Coulbourn conditioning chambers (25 × 29 × 28 cm, with aluminumand Plexiglas walls). Their appearance was altered by introducing varioussensory clues. For context A, we used the chamber described above. For contextB, we introduced a black plexiglass floor washed with peppermint soap into thechamber. The conditioning chambers were placed in sound-attenuating boxes witha ventilation fan and a single house light. Male Sprague-Dawley rats (4–6 weeksold) were randomly assigned to one of five groups. The rats of the naive groupwere left in their home cage until the electrophysiological experiment. On day 1,rats belonging to the other four groups underwent habituation to the trainingchamber (context A) for 20 min. For the fear-conditioned group, rats were presentedwith four tone CSts (4 kHz, 80 dB, 30 s) on day 2, each terminating witha foot shock (USt, 0.5 mA, 1 s). The inter-trial intervals were pseudo-randomlydrawn from intervals ranging between 80–180 s (in 10-s increments). On day 3, therats were placed in context B, but were not presented with the CSt or USt. The ratsof the extinction group were treated similar to the fear-conditioned rats, with theexception that they received 20 CSt presentations in context B on day 3. For theunpaired group, the rats were treated similar to the fear-conditioned rats, withthe exception that the CSt and USt presentations on day 2 were unpaired. For theunpaired + CSt group, the rats were treated similar to the unpaired rats, with theexception that they received 20 unpaired CSts in context B on day 3.In a separate experiment, two other rat groups were implanted with infusioncannulas just above the infralimbic cortex under isoflurane anesthesia andin sterile conditions. After recovery, they underwent the same protocol as theextinction rats, with the exception that they received infralimbic infusions ofvehicle (0.2 µl per hemisphere) or muscimol (0.02 µg per 0.2 µl) 10 min beforeextinction training.The behavior of the rats was recorded with a video camera mounted on topof the conditioning chambers. The conditioned response that we monitoredwas behavioral freezing, quantified off-line by an observer that was blind tothe rats’ condition. In all groups, the electrophysiological experiments wereconducted on day 4.Electrophysiology. The animals were deeply anesthetized with ketamine, xylazineand pentobarbital (80, 12 and 60 mg per kg of body weight, intraperitoneal,respectively). The brain was extracted and cut in 400-µm-thick slices in ice-coldoxygenated aCSF with a vibrating microtome. The aCSF contained 126 mMNaCl, 2.5 mM KCl, 1.25 mM NaH 2 PO 4 , 1 mM MgCl 2 , 2 mM CaCl 2 , 26 mMNaHCO 3 and 10 mM glucose, pH 7.3, 300 mOsm. Prior to recordings, sliceswere kept in an oxygenated chamber for at least 1 h at 24 °C, and then transferredone at a time to a recording chamber perfused with oxygenated aCSF at a rate of7 ml min −1 . The temperature of the chamber was gradually increased to 32 °Cbefore the recordings began. Whole-cell patch recordings were performed undervisual control with pipettes (6–10 MΩ) pulled from borosilicate glass capillariesand filled with a solution containing 130 mM potassium gluconate, 10 mMN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 10 mM KCl, 2 mM MgCl 2 ,2 mM ATP-Mg, 0.2 mM GTP-tris(hydroxy-methyl)aminomethane (pH 7.2,280 mOsm) and 0.2% neurobiotin (wt/vol) for post hoc morphologicalidentification of recorded cells and to ascertain whether the recordings wereobtained from the intended amygdala nucleus. To minimize inter-animalvariability in the position of stimulation and recording sites, we took the followingprecautions. First, a strict coronal slicing angle was consistently used. Second, allof the experiments were conducted using slices at a particular anteroposteriorlevel at which the intra-amygdaloid segment of the stria terminalis clearly delineatesCEm. Only two 400-µm slices per hemisphere met this criterion. Using amicrometric graticule, the BLA-stimulating electrodes were positioned exactly500 µm ventrolateral to the BLA-CEA border, centered on the lateromedial extentof the CEA nucleus. All ITC recordings were obtained from the adjacent BLA-CEA border region. All CEm and CEl recordings were obtained in the ventral400-µm region of these subnuclei. Stimulation of the targeted BLA region elicitedsynaptic responses in all tested ITC and CEA cells. In no instance was there aneed to reposition the stimulating electrodes. BLA stimuli (100 µs, 0.05 Hz)were applied in a range of intensities (0.1–0.5 mA) increasing in steps of 0.1 mA.At each intensity, we obtained independent averages of three subthresholdresponses. For ITC cells, these tests were carried out at a membrane potential of−70 mV, their GABA A reversal potential. The same potential was used for CElneurons. For CEm cells, some of the recordings were performed at a membranepotential of −45 mV to facilitate the measurement of IPSPs. To prevent spikingand contamination of the IPSPs by spike after-hyperpolarizations, we added thelidocaine derivative QX-314 (10 mM) to the intracellular solution in CEm cells.Additional samples of CEm cells were studied at rest or −70 mV. In all cases, measurementsof EPSP slopes were performed in the first 2 ms of the responses.Except for ITC recordings on the nonNMDA to NMDA ratio, all experimentswere performed in control aCSF. In experiments on the nonNMDA to NMDAratio, picrotoxin (100 µM) was added to the aCSF to prevent contamination ofresponses by GABA A IPSCs. Responses were elicited from −80 and + 55 mV involtage-clamp mode. The peak current value at −80 mV was considered to be thenonNMDA component. The current value at 55 mV 300 ms after the stimuluswas considered to be the NMDA component.The reversal potential of GABA A IPSPs was estimated by plotting the amplitudeof the IPSPs evoked by pressure-applied isoguvacine (200 µM in aCSF) as afunction of membrane potential. Linear fits of the data were then performed withthe least-squares method. To study the electroresponsive properties of recordedcells, we applied 500-ms current pulses increasing in 0.02-nA steps. The inputresistance of the cells was estimated in the linear portion of current-voltage plots.The membrane time constant was derived from single exponential fits to voltageresponses in the linear portion of current-voltage relations.Statistical analyses. Statistical analyses consisted of ANOVAs followed byBonferonni-corrected t tests. All values are reported as average ± s.e.m.Histology. At the conclusion of the experiments, slices were placed in a fixative(2% paraformaldehyde and 1% glutaraldehyde (vol/vol) in 0.1 M phosphatebuffer, pH 7.4) overnight. To visualize recorded cells, we washed sections inphosphate-buffered saline (PBS, 0.1 M, pH 7.4), placed them in sodium borohydride(1% in PBS (wt/vol), 20 min) and washed them again repeatedly inPBS. The sections were then incubated for 12 h at 23 °C in 1% BSA (vol/vol),0.3% Triton X-100 (vol/vol), 1% solutions A and B of ABC kit in PBS, washed inPBS, and immersed in a TRIS buffer (0.05M, pH 7.6, 10 min). Neurobiotin wasvisualized by incubation in a TRIS buffer containing 10 mM imidazole, 700 µMDAB and 0.3% H 2 O 2 (vol/vol) for 8–10 min. For the infralimbic experiments,thionin-stained coronal sections at the level of the infusion sites were prepared.We only considered data obtained in rats where the infusion cannulas reachedthe infralimbic cortex. To assess muscimol diffusion, we infused a separate groupof nine deeply anesthetized rats (pentobarbital, 80 mg per kg, intraperitoneal)with fluorescent muscimol using the same volume and muscimol concentrationas in the physiological experiments. We killed the rats by perfusion 10 minlater. These controls revealed that, with the parameters used here, muscimolremained confined to the infralimbic cortex, diffusing ≤1 mm 3 from the centerof the infusion site.doi:10.1038/nn.2499nature NEUROSCIENCE

More magazines by this user
Similar magazines