Center for Neural Science · New York University

DIFFERENT INDUCTION PROTOCOLS RECRUIT NMDA AND L-TYPE CALCIUM CHANNEL-DEPENDENT LTP IN THE LATERAL AMYGDALA (LA): CORRELATION WITH FEAR MEMORY.

E.P. Bauer,* G.E. Schafe and J.E. LeDoux. Center for Neural Science, NYU, NY, NY, 10003.

LTP at sensory input synapses to the LA is a candidate mechanism for memory storage during fear learning. NMDA receptors may be required for LTP at cortical synapses to LA (Huang and Kandel, 1998), and are necessary for acquisition of fear conditioning (Gewirtz and Davis, 1997). This latter finding is confounded, however, by their contribution to baseline synaptic transmission (Weisskopf and LeDoux, 1999) and effects on fear expression (Maren et al., 1996). We have recently shown that calcium influx through L-type voltage-gated calcium channels (VGCCs) is necessary for LTP at thalamic input synapses to LA in vitro (Weisskopf et al., 1999). In the present study, we induced LTP in vitro by pairing trains of presynaptic stimulation with depolarizations of the postsynaptic cell to open dendritic VGCCs (30Hz, 10 stimuli, paired with 1nA, 5 ms depolarizations, given 15 times) (Markram et al., 1997). The LTP induced with this protocol was blocked by bath application of the VGCC-blocker Verapamil (50mM), but was insensitive to D-APV (50mM). Verapamil had no effect on baseline synaptic transmission. We next induced LTP using a 30Hz tetanus (100 stimuli, given twice). Here, LTP was blocked by D-APV, but was unaffected by Verapamil. Since LTP induction via co-occurrence of activity in the pre- and post-synaptic cell is believed to be more consistent with the mechanisms of plasticity underlying natural learning, we next examined the role of VGCCs in the acquisition of fear conditioning. Rats were given bilateral infusions of Verapamil (4 mg/side) or vehicle 30 min prior to Pavlovian fear conditioning. Consistent with the LTP findings, results showed that fear memory was impaired 24 hr later. Further, infusions of Verapamil prior to fear expression had little effect. Collectively, findings suggest that while NMDA receptors may be involved in both plasticity and routine transmission, VGCCs play a selective role in synaptic plasticity and memory in LA. Supported by an NSF Graduate Fellowship and NIMH grants: 46516, 38774, 00956.

H.T. Blair* and J.E. LeDoux. Center for Neural Science, New York University, New York, NY 10003.

Convergent auditory and nociceptive inputs to LA may drive Hebbian synaptic plasticity that underlies memory storage during auditory fear conditioning (LeDoux 1995, Ann Rev Psych 46:209-35). Supporting this theory, it has been shown that LA neurons respond to both auditory and nociceptive stimuli in anesthetized rats (Muramoto et al. 1993, Neurosci 52:621-36; Romanski et al. 1993, Beh Neurosci 107:444-50), and that fear conditioning to a tone CS potentiates auditory evoked responses in LA (Quirk et al. 1995, Neuron 15:1029-39; Rogan et al. 1997, Nature 390:604-7). Here, for the first time, we recorded responses of LA neurons to auditory and nociceptive stimuli during auditory fear conditioning in freely moving rats. Auditory fear conditioning was conducted in a chamber where hungry rats searched for small food pellets dropped randomly by an overhead dispenser. The CS was a train of 20 white noise pips (80 dB), each 250 ms in duration, presented at a rate of 1 Hz (total CS duration=19.25 s). The US, which began 300 ms after the last CS pip, was a train of 8 shock pulses (2.5 mA), each 2 ms in duration, delivered to the eyelid at a rate of 5 Hz (total US duration=1.4 s). Training consisted of 6 habituation (CS alone) trials, 6 sensitization (unpaired CS/US) trials, 16 acquisition (paired CS-US) trials, and 16 extinction (CS alone) trials. Freezing behavior was measured to verify that conditioning was successful in all rats. Neurons (n=25) were recorded from LA and the amygdalostriatal transition area in 5 rats; 5 cells responded to the CS alone, 3 responded to the US alone, 7 responded to both CS and US, and 10 had no clear firing correlate. Of the 12 CS responsive cells, 3 increased their CS response during acquisition, and all 3 were shock responsive; 2 cells decreased their CS response during acquisition, and neither was shock responsive. These preliminary findings are consistent with the theory that convergent auditory and nociceptive inputs may support Hebbian learning in LA. Supported by MH12341 to H.T.B., and MH38774, MH46516, & MH00956 to J.E.L.

N.S. Burghardt*, C. R. Farb, G.M. Sullivan and J.E. LeDoux. Center for Neural Science, NYU, NY, NY, 10003.

Serotonin (5-HT) has been shown to modulate the processing of sensory input to the LA (Stutzmann et al., 1998; Rainnie, 1999). To gain a better understanding of how 5-HT mediates its effects, we used antisera directed against the 5-HT1A, 2A, 2C, 3 and 7 receptor subtypes and evaluated their regional, cellular and subcellular localizations within the LA. Each antiserum produced unique distributions of perikaryal and neuropilar staining. Using light microscopy, 5-HT1A-immunoreactivity (-ir) was seen in the parikarya and processes of many cells throughout the LA. The morphological features of these cells are characteristic of excitatory cells. 5-HT2A-ir appeared as diffuse neuropil staining in the dorsal and lateral regions of the LA. Only a few lightly stained cell bodies with morphological characteristics of interneurons were observed. 5-HT2C-ir was seen throughout the LA as a dense plexus of labeled processes, with no observable cell body staining. 5-HT3-ir prominently stained many axons and punctate processes. Some cell bodies with morphological features of inhibitory interneurons were also labeled. This finding is consistent with our dual label experiments, which indicate that the 5-HT3 receptor is expressed in GABAergic cells in LA. The 5-HT7 immunolabel was present in many cell bodies and processes, with the most robust staining in the dorsal LA. More detailed analyses using electron microscopy revealed that the immunoreaction product of all receptor subtypes was seen in somata, dendrites and dendritic spines receiving asymmetric (excitatory) contacts, and glia. Often, the dendritic immunolabel was also seen away from the synaptic site. In addition to these sites, 5-HT3 receptor antiserum frequently labeled presynaptic axon terminals. The presence of 5-HT receptor immunolabel in a heterogeneous population of cells, on dendrites receiving asymmetric synapses, on glia, and on presynaptic axon terminals, suggests serotonergic modulation of neurotransmission in the LA is complex. Supported by NIMH grants: 58911 and 00956.

L.R.Johnson *, J. E. LeDoux. CNS, New York University, New York, NY. 10003.

Auditory fear conditioning is dependent on auditory signaling from the medial geniculate (MGm) and the auditory cortex (TE3) to principal neurons of the lateral amygdala (LA). Local circuit GABAergic interneurons are known to inhibit LA principal neurons via fast and slow IPSP's. Stimulation of MGm and TE3 produces excitatory post-synaptic potentials in both LA principal and interneurons, followed by inhibitory post-synaptic potentials. Manipulations of D1 receptors in the lateral and basal amygdala modulate the retrieval of learned association between an auditory CS and foot shock. Here we examined the effects of D1 agonists on GABAergic IPSP's evoked by stimulation of MGm and TE3 afferents in vitro. Whole cell patch recordings were made from principal neurons of the LA, at room temperature, in coronal brain slices using standard methods. Stimulating electrodes were placed on the fiber tracts medial to the LA and at the external capsule/layer VI border dorsal to the LA to activate (0.1-0.2mA) MGm and TE3 afferents respectively. Neurons were held at -55.0 mV by positive current injection to measure the amplitude of the fast IPSP. Changes in input resistance and membrane potential were measured in the absence of current injection. Stimulation of MGm or TE3 afferents produced EPSP's in the majority of principal neurons and in some an EPSP/IPSP sequence. Stimulation of MGm afferents produced IPSP's with amplitudes of -2.30 0.53 mV and stimulation of TE3 afferents produced IPSP's with amplitudes of -1.98 1.26 mV. Bath application of 20mM SKF38393 increased IPSP amplitudes to -5.94 1.62 mV (MGm, n=3) and -5.46 0.31 mV (TE3, n=3). Maximal effect occurred <10mins. A small increase in resting membrane potential and decrease in input resistance were observed. These data suggest that DA modulates both the auditory thalamic and auditory cortical inputs to the LA fear conditioning circuit via local GABAergic circuits. Supported by NIMH Grants 00956, 46516, and 58911.

R. Lamprecht and J.E. LeDoux. Center for Neural Science, New York University, New York, NY 10003.

We are studying molecular events in amygdala, and interconnected brain regions, underlying long-term fear-conditioning memory. In the present study we have monitored the alteration in phosphorylation of proteins extracted from amygdala and auditory thalamus following different stages of fear conditioning. Adult rats were habituated to the conditioning chamber. In the next day the animals were divided into four groups: 1) Control group: exposed to conditioning chamber; 2) CS group: exposed to tone (5kHz 75dB for 30 sec); 3) US group: received mild footshock (1.5 mA for 1 sec); 4) Fear conditioned group: 30 sec tone (5kHz 75 dB for 30 sec) terminated with a footshock (1.5mA for 1 sec). The brains were removed at various time points following the behavioral manipulations, frozen rapidly, and the amygdala and the thalamus (which contains the auditory thalamus) were dissected. The brain tissue was homogenized and subjected to SDS-PAGE followed by Western blots with anti-phosphoserine, anti-phosphothreonine or anti-phosphotyrosine antibodies. Tyrosine phosphorylation of a ~60kDa protein in the thalamus was significantly increased 30 minutes following fear conditioning training when compared to the control, CS and US groups (p<0.0003; p<0.008; p<0.005 respectively; n=8 for control group, n=8 for the CS group, n=9 for the US group and n=9 for fear conditioning group). Tyrosine phosphorylation of the 60 kDa protein was also significantly increased 60 minutes (n=6 p<0.05), but not 10 minutes, 6 hrs and 24 hours, following fear conditioning when compared to the control group. Tyrosine phosphorylation of the 60 kDa in the CS and US groups was not significantly increased when compared to the control group. Fear conditioning experience can thus increase tyrosine phosphorylation of a 60 kDa protein in the auditory thalamus.

M.A.P. Moita, H.T. Blair, G.M. Sullivan*, and J.E. LeDoux. Center for Neural Science, New York University, New York, NY 10003.

During fear conditioning, the hippocampus may construct a representation of the environment where training takes place (Kim & Fanselow 1992, Science 256: 675-677; Phillips & LeDoux 1992, Behav Neurosci 106: 274-285). The hippocampus is also involved in trace fear conditioning to an auditory CS (McEchron et al. 1998, Hippocampus 8:638-46). Here we studied hippocampal involvement in fear learning by recording hippocampal neurons during auditory fear conditioning. Hungry rats foraged randomly for small food pellets in two familiar chambers: box A, a neutral chamber where no stimuli were presented, and box B, where auditory fear conditioning occurred using the same CS and US as in the accompanying abstract by Blair & LeDoux. Conditioning in box B occurred in three parts: habituation (HAB = 9 trials, CS alone), acquisition (ACQ = 16 CS-US pairings), and extinction (EXT = 16 trials, CS alone). Baseline activity in box B was recorded before/after each conditioning session, and in box A before HAB, after ACQ, and after EXT. In 5 rats we recorded 51 hippocampal neurons, of which 11 were theta cells and 40 were complex spike (CSP) cells. 38/40 CSP cells had place fields in box B; of these, 18 were stable throughout the experiment, and 15 "remapped" their fields during conditioning, despite remaining stable in box A. This suggests that conditioning may induce some hippocampal cells to re-represent the training environment. 27/51 cells responded to the CS (19/38 CSP cells; 8/11 theta cells), of which 10 also responded to the US (3/19 CSP cells; 7/8 theta cells). Importantly, 26/27 cells that responded to the CS were not CS responsive during HAB, but only during ACQ/EXT. This suggests that hippocampal cells acquire conditioned responses to auditory CSs, in agreement with previous reports (Berger & Thompson 1982, Behav Brain Res 4:63-76). These preliminary results suggest that fear conditioning induces changes in spatial and non-spatial firing properties of hippocampal neurons, which may be related to the role of the hippocampus in some fear learning paradigms. Supported by PRAXIS XX1 (Portugal) to MAPM, MH12341 to HTB, and MH38774, MH46516, & MH00956 to JEL.

K Nader, GE Schafe* & JE LeDoux, Center for Neural Science, NYU, NY, NY, 10003.

"New" memories are initially labile and sensitive to disruption before being consolidated into stable long-term memories. A wealth of data indicates that this consolidation process involves the synthesis of new proteins in neurons. The lateral and basal nuclei of the amygdala (LBA) are believed to be a site of memory storage in fear learning. Intra-LBA infusion of the protein synthesis blocker anisomycin (ANISO) shortly after training prevents consolidation of this memory. We retested the idea that when consolidated memories are reactivated, they return to a labile state. Rats were trained with 1 CS-US pairing. The following day, 1 CS presentation was given followed immediately by intra-LBA infusions of ANISO. The next day, ANISO infused rats showed greatly impaired freezing compared to controls. This result was obtained regardless of whether reactivation occurred 1 or 14 days after conditioning. The behavioral deficit was not due to ANISO-induced amygdala damage or late waves of protein synthesis because if memory reactivation was omitted prior to ANISO infusions, then freezing was normal the next day. Consistent with a time limited role for protein synthesis production in consolidation, delay of the infusion until 6 hours after memory reactivation produced no amnesia. Lastly, animals demonstrated intact freezing 4 hr but impaired freezing 24 hr after reactivation and ANISO. This pattern of results is similar to the effects of protein synthesis on consolidation: intact short-term memory and impaired long-term memory. These data demonstrate that consolidated fear memories, when reactivated, return to a labile state that requires de novo protein synthesis to be reconsolidated. These findings are not predicted by traditional theories of memory consolidation. Supported by Grant MH00956, MH46516, MH38774. We'd like to acknowledge the WM Keck Foundation.

S.M. Rodrigues*; J.E. LeDoux*; J.H. Morrison** . *Center for Neural Science, New York University, New York, NY, USA . **Neurobiology of Aging Laboratories, Mount Sinai School of Medicine, New York, NY, USA

Emotionally relevant sensory information about the external world reaches the amygdala primarily be way of its lateral nucleus (LA). More specifically, anatomical and physiological evidence suggests that distinct divisions of the LA play separate functional roles in processing the information involved in fear conditioning to auditory stimuli. Previous studies indicate that NMDA receptors are key for synaptic transmission in the LA and possibly in fear conditioning. Moreover, the overexpression of the NR2B subunit of NMDA receptors in transgenic animals has been shown to elicit stronger fear responses as compared to those of wild-type animals. The present study utilized immunohistochemical methods to analyze the distribution of the NR2B subunit of the NMDA receptor in the LA of the rat amygdala. The resultant staining revealed a higher density of the NR2B subunit in the LA's lateral portion in comparison to that found in the medial portion. In addition, the dorsolateral division (Ld1) displayed a higher abundance of the NR2B subunit in relation to the levels detected in both the ventrolateral and medial divisions of the LA. Furthermore, co-localization analysis of the NR1 and NR2B NMDA receptor subunits by immunofluorescence revealed that virtually all functional NMDA receptors in Ld1 express the NR2B subunit. The NR1-NR2B complex contributes to prolonged excitatory postsynaptic potentials and thus increases the time window for detecting synaptic coincidence. Ld1 receives the earliest information about auditory stimuli and it is also the site of auditory and somatosensory convergence. These findings, which display a prominent expression of the NR2B subunit in the Ld1 when contrasted to other divisions of the LA, help explain how this unit is primed to play a role in the transmission of sensory information during the process of emotional learning. Supported by: MH00956, MH46516 and MH58911

*G.E. Schafe,**C.M. Atkins*, 2J.D. Sweatt & *J.E. LeDoux. *Center for Neural Science, New York University, NY, NY 10003 and **Division of Neuroscience, Baylor College of Medicine, Houston, TX, 77030.

Recent experiments in our lab have shown that ERK/MAP kinase is transiently activated in the lateral and basal amygdala (LBA) following Pavlovian fear conditioning, a result which is consistent with studies that have shown impairments of fear memory consolidation following either systemic (Atkins et al, 1998) or intracerebroventricular (Schafe et al, 1999) administration of ERK/MAPK inhibitors. To determine whether ERK/MAPK activation in the LBA is obligatory for fear memory consolidation, rats in the present experiments were given localized infusions of U0126, an inhibitor of MEK, an upstream regulator of ERK/MAPK. Rats were infused with different doses of U0126 (0.1 or 1.0 mg/side; 0.5 ml) or 50% DMSO vehicle into LBA 30 min prior to single or multiple-trial auditory Pavlovian fear conditioning. Western blot experiments indicated that administration of U0126 dose-dependently impaired ERK/MAPK activation in the LBA. Accordingly, behavioral experiments showed that U0126 dose-dependently impaired long-term memory (LTM) of fear conditioning assessed 24 hrs after training. However, short-term memory (STM), assessed at 1 hr, was intact. Additional controls showed that the LTM impairment was likely due to impaired memory consolidation processes rather than to sensory or performance deficits. Finally, a time-course analysis of the effects of U0126 revealed that fear memory was intact 1 and 3 hrs following conditioning, but significantly impaired 6 and 24 hrs later. Collectively, results suggest that memory consolidation of Pavlovian fear conditioning is dependent on a MAPK-dependent mechanism in the LBA. Supported by: MH11902, MIMH 57014, MH38774, & MH46516.

A.E. Wilensky*, G.E. Schafe and J.E. LeDoux. Center for Neural Science, New York University, NY, NY, 10003.

Anatomical, behavioral and physiological evidence have all implicated the amygdala in the acquisition of Pavlovian fear conditioning. It is generally accepted that input from the conditioned (CS) and unconditioned (US) stimuli converge in the lateral nucleus (LA) where alterations in synaptic transmission are thought to encode key aspects of the learning. During retrieval, this information is thought to be transmitted to the central nucleus (CE), which controls the expression of conditioned fear responses. Studies of CE function have mostly relied on permanent lesions of the nucleus, which confound effects on learning with effects on expression. We examined the effect of functional inactivation of different nuclei of the amygdala during Pavlovian conditioning. Bilateral injections of muscimol (0.18 nmol/side) or ACSF (0.2 ml) were made into LA or CE before Pavlovian conditioning consisting of two tone-footshock pairings. Findings indicated that infusions aimed at either nucleus impaired acquisition of conditioned fear relative to vehicle controls. All groups showed high levels of postshock freezing after the second tone, suggesting that the differences in fear acquisition did not result from a lack of sensory input from functional inactivation. Examination of focal points of the injections and further behavioral data examining the spread of muscimol into adjacent nuclei will clarify the role of CE vs. LA in the acquisition of Pavlovian conditioning. Supported by NIMH grants: MH46516, MH00956, & MH38774.