Induction mechanisms for L-LTP at thalamic input
synapses to the lateral amygdala: requirement of
Ok kyung Lee, Chang-Joong Lee1 and Sukwoo ChoiCA
1Department of Neuroscience, Ewha Institute of Neuroscience (EIN), School of Medicine, Ewha Womans University, Jongno-Gu, Jongno-6-Ga,
70, Ewha Dong-Dae-Mun Hospital, Seoul 110 -783; 2Department of Biology, Inha University, Inchon 402-751, South Korea
Received 4 February 2002; accepted 4 February 2002
L-LTP (late-phase long-term potentiation) at thalamo-amygdala sy-
saturating concentrations before and during repeated tetanus. By
napses is thought to be critical for auditory fear conditioning, but it
contrast, the mGluR1antagonist CPCCOEt (80 mM) failed to show
has not been clear what kinds of surface receptors and channels
any e¡ects on L-LTP induction. Neither D-AP5 nor MPEP produced
are involved in the induction phase of the L-LTP. Here we report
any signi¢cant e¡ects on potentiated synaptic responses when
that the NMDA receptor antagonist D-AP5 (50 mM), the L-type
applied after L-LTP had been established. Thus, our data suggest
calcium channel antagonist nifedipine (30 mM) and the metabotro-
that NMDA receptors, L-type calcium channels and mGluR5 are
pic glutamate receptor 5 antagonist MPEP (10 mM) prevented
involved in L-LTP induction in the thalamo-amygdala pathway.
L-LTP induction when each antagonist was separately applied at
c 2002 Lippincott Williams & Wilkins.
Key words: Amygdala; Brain slices; Fear conditioning; L-LTP; L-type calcium channel; NMDA receptor; Rat
been explored up until recently. A study of E-LTP (early
LTP is believed to be critical for learning and memory, and
phase LTP) induced by pairing pre- and postsynaptic
tremendous efforts have been made to find a link between
activity reveals that the LTP induction is dependent on L-
LTP and learning [1–3]. One of the best examples for the link
type voltage-gated calcium channels, but not on NMDA
between LTP and learning in the mammalian brain may be
receptors . In addition, an enduring form of LTP (L-LTP)
cued conditioning, a form of fear conditioning that requires
induced by multiple trains of high-frequency stimulation at
the lateral amygdala. This form of fear conditioning is
thalamic input synapses to the lateral amygdala has been
produced by the pairing of a neutral tone as a conditioned
shown to be dependent on protein synthesis, and is
stimulus (CS) with a shock as an unconditioned stimulus
mediated by protein kinase A and mitogen-activated
(US). These two stimuli converge onto the lateral amygdala,
protein kinase (see Fig. 7 in ).
and the coincidental presentation of the CS and US is
Since L-LTP has an enduring phase, it may be more
thought to induce fear conditioning by potentiating the
relevant to study L-LTP as a cellular substrate for condi-
synaptic strength of the CS pathway by a long-term
tioned fear memory. Especially, receptors and channels
potentiation-like mechanism [4,5]. The CS alone then could
involved in induction mechanisms for L-LTP at thalamic
produce a sufficient excitation of the lateral amygdala to
input synapses to the lateral amygdala has not been clearly
elicit conditioned fear. The CS comes into the lateral nucleus
defined yet. Therefore, we have examined a possible role of
of the amygdala via two routes: directly from the medial
NMDA receptors, L-type calcium channels and group I
geniculate nucleus and indirectly from the auditory cortex
mGluRs in the induction of L-LTP at thalamic input
[6,7]. Although the synapses of both of these projections
undergo long-term potentiation, the in vivo and in vitrostudies linking amygdala LTP to fear learning have involvedthe thalamic pathway to the lateral amygdala [4,5]. How-
ever, studies examining amygdala LTP using in vitro
Brain slices were prepared using techniques described
preparations have focused mainly on the cortical inputs to
previously [15,16]. Sprague–Dawley rat (3–5 weeks old)
were decapitated. The isolated whole brains were placed in
Mechanisms for LTP (early phase or late phase LTP) in the
an ice-cold (0–41C) modified artificial cerebrospinal fluid
thalamic input synapses to the lateral amygdala have not
(aCSF) solution. The composition of modified aCSF was as
follows (in mM): 175 sucrose, 20 NaCl, 3.5 KCl, 1.25
Consistent with their results, the field potential in our
NaH2PO4, 26 NaHCO3, 1.3 MgCl2, 11 D-(þ)-glucose. Coronal
experimental condition had a constant and short latency of
slices (400 mm) containing the amygdala were cut using a
about 5 ms, followed high frequency (50 Hz) stimulation
vibratome (Campden, UK), and were incubated in aCSF
reliably and without failure, and it could be blocked by
continuously bubbled at room temperature with 95% O2/
kynurenic acid (5 mM), a non-selective glutamate receptor
5% CO2 for 3 h before recordings. Just before transferring
antagonist (data not shown, see also ). These findings
the slice to the recording chamber, the cortex overlying the
suggest that the field potential measured in the present
amygdala was cut away with a scalpel so that, in the
study reflects glutamatergic, monosynaptic responses at
presence of picrotoxin, cortical epileptic burst discharges
thalamic input synapses to the lateral amygdala. As shown
in the previous studies, we also included picrotoxin in our
The recording chamber was continuously superfused
recording solution to block feedforward GABAergic inputs
with aCSF (30–321C) at a flow rate of 1-2 ml/min. The aCSF
to principal neurons in the lateral amygdala .
contained (in mM): 120 NaCl, 3.5 KCl, 1.25 NaH2PO4, 26
L-LTP at thalamic input synapses to the lateral amygdala
NaHCO3, 1.3 MgCl2, 2 CaCl2, 11 D-(þ)-glucose. Picrotoxin
has been shown to be induced by 3 trains of tetanus
(10 mM) was included in all experiments to minimize fast
(100 Hz, 1 s duration) . Therefore, we have examined the
GABAergic transmission . The slices were incubated in
effect of 3–5 trains of tetanic stimulation delivered at
the recording chamber 30 min before the start of record-
variable intervals to obtain maximal L-LTP. Five trains at
1 min intervals were found to be most effective for L-LTP
To record field potentials at thalamic input synapses to
induction. One successful example showed L-LTP lasting up
the lateral amygdala, we placed a bipolar stimulating
electrode in the thalamic afferent fibers innervating the
In order to determine whether L-LTP induction at
lateral amygdala, which is located in the ventral part of the
thalamic input synapses to the lateral amygdala depends
striatum, just above the central nucleus of the amygdala, just
on NMDA receptors, L-type voltage-gated calcium channels
medial to the lateral amygdala (see Fig. 7 in ). A trunk of
or group I mGluRs, we applied antagonists for each
the thalamic afferent fibers appeared to be well isolated
candidate molecule before and during L-LTP induction.
from other structures and it could be easily visualized under
Each antagonist was applied for a total of 20 min, 15 min
our microscope. A stimulating electrode was located
prior to repeated tetanus, and an additional 5 min during
specifically on the trunk to elicit the field potential. The
repeated tetanus. To enable a more reliable comparison, we
recording electrode (41.0 MO) was filled with 0.9% NaCl
obtained a pair of recordings for control and antagonist-
and placed in the dorsal subregion of the lateral amygdala.
treated slices from the same animal. We first examined the
Synaptic responses were elicited at 0.017 Hz. L-LTP was
effect of the NMDA receptor antagonist D-AP5 (50 mM) on
induced by five trains of tetanic stimulation (100 Hz, 1 s at
induction of L-LTP in field potential recording experiments.
1 min intervals) with the same intensity and pulse duration
D-AP5 prevented a potentiation of field potentials by five
as the test stimuli. For the baseline field potential recording
trains of tetanus. L-LTP could be induced in paired control
50% of the maximum amplitude was used. The range of
slices (164 7 5.5% of control in the amplitude of field
stimulus intensity and duration for each pulse is 0.1–0.3 mA
potential at 2.5–3 h post-tetanus, n ¼ 6), but not in the
presence of D-AP5 (Fig. 1a; 50 mM, 95 7 3.3% of control in
Extracellular field potentials were amplified using a DP-
the amplitude of field potentials 2.5–3 h post-tetanus,
301 amplifier (Warner Instrument Co., CT) and the output
p o 0.0001, paired t-test, n ¼ 6). Exposure to D-AP5 com-
was digitized with a DIGIDATA 1322A interface (Axon
instruments Inc., Foster City, CA). The digitized signals
In order to determine whether L-type voltage-gated
were stored and analyzed with a PC computer using
calcium channels are involved in the induction of L-LTP,
pClamp 8 (Axon Instruments Inc., Foster City, CA).
we examined the effect of the L-type voltage-gated calcium
Drugs used were D-AP5, nifedipine, picrotoxin and
channel antagonist nifedipine (30 mM) on the induction of L-
kynurenic acid from Sigma-Aldrich (St. Louis, MO). MPEP
LTP. Five trains of tetanus resulted in the induction of L-LTP
(2-methyl-6-(phenylethyl)-pyridine) and CPCCOEt (7-(hy-
in paired control slices (169 7 5.8% of control in field
droxyimino)cyclopropa[b]chromen-1a-carboxylate ethyl es-
potential amplitude 2.5–3 h post-tetanus). In contrast, pre-
ter) were from Tocric Cookson (Ballwin, MO). Drugs were
treatment of the L-type voltage-gated calcium channel
made up in stock solutions and diluted more than 1000
antagonist nifedipine partially prevented the induction of
times into aCSF. Picrotoxin, CPCCOEt and nifedipine were
L-LTP (Fig. 1b; 122 7 4.6% of control in field potential
amplitude at 2.5–3 h post-tetanus, p o 0.05 paired t-test,n ¼ 6). The potentiation in the presence of nifedipine wasmaintained for 3 h (p o 0.05, paired t-test), suggesting
that NMDA receptors alone can support tetanus-induced
L-LTP has been defined as a form of LTP that has an
L-LTP. Neither nifedipine (30 mM) nor D-AP5 (50 mM) had
enduring phase (4 3 h). In order to achieve a stable
any effects on baseline synaptic responses in this pathway
recording over 3 h we chose to measure the field potential
evoked when the thalamic fibers onto the lateral amygdala
An antagonist for NMDA receptors has been shown to
were stimulated (see Materials and Methods). In the
be effective in reducing conditioned fear when applied
previous study , the field potential at thalamic input
after conditioning . This result raises the possibility
synapses to the lateral amygdala has been characterized,
and both the E-LTP and L-LTP were studied at this synapse.
mediated by NMDA receptors. Therefore, we examined
INDUCTION MECHANISMS FOR L-LTP AT THALAMIC INPUT SYNAPSES
Involvement of NMDA receptors and L-type voltage-gated calcium channels in the L-LTP induction. (a) L-LTP at thalamic input synapses onto the
lateral amygdala was completely blocked by the NMDA receptor antagonist D-AP5 (50 mM, n ¼ 6; closed circles). (b) L-LTP at thalamic input synapsesonto the lateral amygdala was partially inhibited by the L-type voltage-gated calcium channel inhibitor nifedipine (30 mM, n ¼ 6; closed circles). Please notethat L-LTP was maintained even in the presence of nifedipine. (c) The potentiated synaptic responses during L-LTP were not altered by D-AP5 (50 mM,n ¼ 6).The averaged data traces taken before (left) and 3 h after (right) tetanus were shown at the top of the ¢gure. Calibration ¼ 3 ms, 0.2 mV.
the effect of an antagonist for NMDA receptors on the
significant effects on baseline synaptic responses in this
maintenance phase of L-LTP. We applied D-AP5 (50 mM)
to the slices when L-LTP had been stably established (2–2.5 h post-tetanus). However, we failed to observe anysignificant effects of D-AP5 on the potentiated synapticresponses during L-LTP (Fig. 1c; 104.6 7 3.4% of control
in the amplitude of field potentials, p 4 0.9, paired t-test,
In the present study, we have found that L-LTP induction at
n ¼ 4), suggesting that most of the potentiated synaptic
thalamic input synapses onto the lateral amygdala is
responses are mediated by non-NMDA receptors, most
dependent on activation of NMDA receptors, L-type
voltage-gated calcium channels and mGluR5. mGluR1 does
Since D-AP5 (50 mM) completely blocked L-LTP induction,
not appear to be involved in the induction of L-LTP.
activation of L-type calcium channels alone during repeated
Compared to the previous studies of LTP in this pathway
tetanus does not appear to support tetanus-induced L-LTP.
[13,14], our findings reveal unique characteristics of L-LTP;
This would be either because calcium influx through L-type
(1) tetanus-induced L-LTP at thalamo-amygdala synapses
calcium channels during tetanus is too weak to induce L-
depends upon activation of NMDA receptors, (2) L-type
LTP or because calcium influx especially through NMDA
calcium channel-dependent LTP has an enduring phase
receptors during tetanus is required for the L-LTP induction.
(4 3 h). (3) L-LTP induction requires mGluR5 activation.
In order to determine whether enhanced activity of L-type
LTP at thalamic input synapses to the lateral amygdala
calcium channels helps to achieve L-LTP upon blockade of
has been proposed as a cellular substrate for conditioned
NMDA receptors, we examined L-LTP induction in the
fear [4,5]. One approach to test the hypothesis would be to
presence of BAY K 8644 (1 mM), a potentiator for L-type
compare pharmacological and physiological characteristics
calcium channels, as well as D-AP5 (50 mM), BAY K 8644 did
of LTP and conditioned fear. NMDA receptors, L-type
not have significant effects on baseline synaptic responses
voltage-gated calcium channels and mGluR5, which have
(Fig. 2a; n ¼ 3), whereas L-LTP could be induced. In the
been shown to be involved in the induction phase of fear
presence of 1 mM BAY K 8644 and 50 mM D-AP5 (Fig. 2b;
conditioning [18–22], appear to mediate L-LTP induction.
155.0 7 4.7% of control in the amplitude of field potentials at
Thus, L-LTP and conditioned fear share some of induction
2.5–3 h post-tetanus, p o 0.05, paired t-test, n ¼ 4). The
mechanisms with each other, supporting the proposal that
magnitude of the L-LTP with BAY K 8644 and D-AP5 was
L-LTP is a cellular substrate for conditioned fear.
similar to that in paired control slices (p 4 0.3, paired t-test
Perhaps the most critical finding in the present study is
at 2.5–3 h post-tetanus, n ¼ 4; L-LTP in paired con-
that mGluR5, but not mGluR1, is involved in the induction
trol ¼ 171.0 7 6.0% of control in the amplitude of field
of L-LTP. In the previous study , MPEP showed a
potential at 2.5–3 h post-tetanus, p o 0.01, paired t-test,
selective effect on mGluR5, but not on other glutamate
n ¼ 4). One special feature of L-LTP induced in the presence
receptors including NMDA receptors at the concentration
of D-AP5 and BAY K 8644 was that the potentiation after
used herein. No effects of the mGluR1 antagonist CPCCOEt
repeated tetanus developed slowly over B1 h, implying that
on L-LTP induction further suggest that MPEP selectively
the early component of L-LTP depends upon activation of
antagonized mGluR5, but not mGluR1, in our experiments.
NMDA receptors. Thus, our data suggest that L-LTP can be
It is worthwhile to note that we observed the blocking effect
achieved by enhanced calcium influx through L-type
of 80 mM CPCCOEt on the induction of striatal LTD ,
calcium channels without intervention of NMDA receptor
suggesting that 80 mM CPCCOEt is sufficient to block
mGluR1 at least in case of striatal slices. Activation of
Next we examined group I mGluRs (mGluR 1 and 5) on
mGluR5 during L-LTP induction can stimulate PI hydrolysis
L-LTP induction. We first tested the mGluR5 antagonist
that leads to activation of PKC and increases in intracellular
MPEP (10 mM) on induction of L-LTP. L-LTP could be
calcium levels, which could contribute to the induction of
induced in paired control slices (185 7 6.8% of control in the
amplitude of field potential at 2.5–3 h post-tetanus, p o 0.05,
Another interesting finding in this study is that L-LTP
n ¼ 5), but not in the presence of MPEP (Fig. 3a; 10 mM,
induction at thalamic input synapses to the lateral amygdala
113 7 5.4% of control in the amplitude of field potentials at
depends upon both the NMDA receptors and L-type
2.5–3 h post-tetanus, p 4 0.3, paired t-test, n ¼ 5). Similar to
voltage-gated calcium channels. The involvement of L-type
the experiment using D-AP5, exposure to MPEP completely
voltage-gated calcium channels in L-LTP induction would
prevented L-LTP induction. We next examined the effect of
be expected because of its involvement in the pairing-
the mGluR1 antagonist CPCCOEt on L-LTP induction.
induced E-LTP induction shown in the previous study .
Although applied at a saturating concentration (80 mM),
However, it is surprising to see the effect of an antagonist for
CPCCOEt failed to block L-LTP induction (Fig. 3b; L-LTP
NMDA receptors on the L-LTP induction since it did not
in control ¼ 166 7 5.0%, n ¼ 5; L-LTP in CPCCOEt ¼
show any effects on the induction of E-LTP induced by
173 7 7.4%, n ¼ 5). We also examined the effect of MPEP
pairing . One possibility is that the repeated tetanic
on L-LTP maintenance. We applied MPEP (10 mM) to the
stimulation used herein produces a more localized depolar-
slices when L-LTP had been stably established (2.5–3 h post-
ization around the synapse, which would be sufficient for
tetanus). MPEP failed to show any significant effects on the
maximal activation of NMDA receptors. However, such a
potentiated synaptic responses during L-LTP (Fig. 3c;
local depolarization would allow at most a partial activation
p 4 0.2, paired t-test, n ¼ 3), supporting the suggestion that
of extrasynaptic L-type voltage-gated calcium channels.
mGluR5 is involved in the induction phase of L-LTP.
Thus, one can expect that NMDA receptors play a more
Neither MPEP (10 mM) nor CPCCOEt (80 mM) had any
important role in initiating calcium entry during the tetanus
INDUCTION MECHANISMS FOR L-LTP AT THALAMIC INPUT SYNAPSES
Restoration of L-LTP by exposure to BAY K 8644 in the presence of D-AP5. (a) BAY K 8644 (1 mM) alone did not alter baseline synaptic transmis-
sion (n ¼ 3). (b) L-LTP was induced by repeated tetanus before and during exposure to 1 mM BAY K 8644 and 50 mM D-AP5 (n ¼ 4; closed circles). Pleasenote that the potentiation after repeated tetanus develops slowly. The averaged data traces taken before (left) and 3 h after (right) tetanus are shown atthe top of the ¢gure. Calibration ¼ 4 ms, 0.2 mV.
than L-type voltage-gated calcium channels do. By contrast,
Restoration of L-LTP induction by BAY K 8644 upon
it is possible that the massive postsynaptic depolarization
blockade of NMDA receptors supports the idea that L-type
used for the pairing-induced E-LTP in the previous study is
calcium channels play a major role in the induction of
so effective at raising calcium levels via L-type calcium
NMDA receptor-independent LTP . Furthermore, our
channels that it obviates a need for calcium entry through
data clearly indicate that L-type calcium channel-dependent
LTP has an enduring phase in this pathway. Although an
Involvement of mGluR5, but not mGluR1, in L-LTP induction. (a) L-LTP at thalamic input synapses onto the lateral amygdala was completely
inhibited by the mGluR5 inhibitor MPEP (10 mM, n ¼ 5; open circles). (b) The mGluR1 antagonist CPCCOEt (80 mM, n ¼ 5; open circles) failed to block L-LTP induction at thalamic input synapses onto the lateral amygdala. (c) The potentiated synaptic responses during L-LTP were not altered by MPEP (10 mM,n ¼ 3).The averaged data traces taken before (left) and 3 h after (right) tetanus are shown at the top of the ¢gure. Calibration ¼ 5 ms, 0.2 mV.
INDUCTION MECHANISMS FOR L-LTP AT THALAMIC INPUT SYNAPSES
exogenous compound, BAY K 8644 was used in the present
of amygdala L-LTP. We conclude that activation of these
experiment, L-type calcium channel activity could be
receptors and channels is necessary for the induction of
enhanced by a variety of endogenous signal molecules such
L-LTP at thalamic input synapses to the lateral amygdala.
as G-proteins and protein kinases, so that, at some instances,L-type calcium channels might contribute to L-LTP induc-tion more than NMDA receptors.
It might be odd to see a complete block of L-LTP by D-
1. Barnes CA. Neuron 15, 751–754 (1995).
AP5, compared to a partial block of L-LTP by nifedipine. If
2. Stevens CF. Neuron 20, 1–2 (1998).
both the L-type voltage-dependent channels and NMDA
3. Eichenbaum H. Nature 378, 131–132 (1995).
receptors play a role in L-LTP induction by increasing
4. Rogan M, Staubli U and LeDoux J. Nature 390, 604–607 (1997).
intracellular calcium levels, then specific blockade of either
5. McKernan MG and Shinnick-Gallagher P. Nature 390, 607–611 (1997).
molecule would produce a partial inhibition of L-LTP. One
6. LeDoux JE. Annu Rev Psychol 46, 209–235 (1995).
7. Maren S and Fanselow M. Neuron 16, 237–240 (1995).
possibility for the complete block by D-AP5 but only partial
8. Chapman PF and Bellavance LL. Synapse 11, 310–318 (1992).
block by nifedipine is a synergistic rise in calcium involving
9. Gean P-W, Chang F-C, Huang C-C et al. Brain Res Bull 31, 7–11 (1993).
either sources. Thus, NMDA receptors may be able to raise
10. Brambilla R, Gnesutta N, Minichiello L et al. Nature 390, 281–286 (1997).
calcium enough to potentiate some synapses, whereas L-
11. Huang YY and Kandel ER. Neuron 21, 169–178 (1998).
type calcium channels alone may not. However, additional
12. Li H, Weiss SRB, Chuang D-M et al. J Neurosci 18, 1662–1670 (1998).
calcium coming through L-type calcium channels might
13. Weisskopf MG, Bauer EP, LeDoux JE. J Neurosci 19, 10512–10519 (1999).
14. Huang Y, Martin KC and Kandel ER. J Neurosci 20, 6317–6325 (2000).
synergize with that coming through NMDA receptors to
15. Choi S and Lovinger DM. Proc Natl Acad Sci USA 94, 2665–2670 (1997).
give the full amount needed for maximal LTP.
16. Choi S, Klingauf J and Tsien RW. Nature Neurosci 3, 330–336 (2000).
At present, we do not understand a precise role of NMDA
17. Woodson W, Farb CR and Ledoux JE. Synapse 38, 124–137 (2000).
receptor-dependent, L-type calcium channel-dependent and
18. Lee H and Kim JJ. J Neurosci 18, 8444–8454 (1998).
mGluR5-dependent L-LTP in fear memory, and it remains to
19. Maren S, Aharonov G, Stote DL et al. Behav Neurosci 110, 1365–1374 (1996).
20. Gewirtz JC and Davis M. Nature 388, 471–474 (1997).
21. Bauer EP, Schafe JE, LeDoux JE. Soc Neurosci Abstr 26, 1253 (2000).
22. Schulz B, Fendt M, Gasparini F et al. Neuropharmacology 41, 1–7 (2001).
23. Gasparini F, Lingenhohl K, Stoehr N et al. Neuropharmacology 38, 1493–
We have shown that antagonists for NMDA receptors,
24. Sung KW, Choi S and Lovinger DM. J Neurophysiol 86, 2405–2412 (2001).
L-type calcium channels and mGluR5 blocked the induction
25. Pin J-P and Duvoisin R. Neuropharmacology 34, 1–26 (1995).
Acknowledgements: We thank Dr D.M. Lovinger for his comments on this manuscript.This research was supported by grants from
the KMOST grant M1- 0108 - 00 - 0051 under the neurobiology research program to S.W.C.
obesity reviews Obesity Management Recent advances in adaptive thermogenesis: potential implications for the treatment of obesity S. L. J. Wijers, W. H. M. Saris and W. D. van Marken LichtenbeltDepartment of Human Biology, Nutrition andToxicology Research Institute Maastricht,Large inter-individual differences in cold-induced (non-shivering) and diet-induced adaptive thermogenesis exi
Impact assessment of EU Food Information Regulation (1169/2011) on food manufacturers and retailers Focus on: Foods offered for sale by distance communications Introduction GS1 commissioned Mason, Hayes & Curran to provide an in-depth analysis of EU Regulation 1169/2011 on the provision of food information to consumers (FIR). This regulation introduces a new requi