Pain-related increase of excitatory transmission and decrease of inhibitory transmission in the central nucleus of the amygdala are mediated by mGluR1.

Abstract

Neuroplasticity in the central nucleus of the amygdala (CeA), particularly its latero-capsular division (CeLC), is an important contributor to the emotional-affective aspects of pain. Previous studies showed synaptic plasticity of excitatory transmission to the CeLC in different pain models, but pain-related changes of inhibitory transmission remain to be determined. The CeLC receives convergent excitatory inputs from the parabrachial nucleus in the brainstem and from the basolateral amygdala (BLA). In addition, feedforward inhibition of CeA neurons is driven by glutamatergic projections from the BLA area to a cluster of GABAergic neurons in the intercalated cell masses (ITC). Using patch-clamp in rat brain slices we measured monosynaptic excitatory postsynaptic currents (EPSCs) and polysynaptic inhibitory currents (IPSCs) that were evoked by electrical stimulation in the BLA. In brain slices from arthritic rats, input-output functions of excitatory synaptic transmission were enhanced whereas inhibitory synaptic transmission was decreased compared to control slices from normal untreated rats. A non-NMDA receptor antagonist (NBQX) blocked the EPSCs and reduced the IPSCs, suggesting that non-NMDA receptors mediate excitatory transmission and also contribute to glutamate-driven feed-forward inhibition of CeLC neurons. IPSCs were blocked by a GABAA receptor antagonist (bicuculline). Bicuculline increased EPSCs under normal conditions but not in slices from arthritic rats, which indicates a loss of GABAergic control of excitatory transmission. A metabotropic glutamate receptor subtype 1 (mGluR1) antagonist (LY367385) reversed both the increase of excitatory transmission and the decrease of inhibitory transmission in the arthritis pain model but had no effect on basal synaptic transmission in control slices from normal rats. The inhibitory effect of LY367385 on excitatory transmission was blocked by bicuculline suggesting the involvement of a GABAergic mechanism. An mGluR5 antagonist (MTEP) inhibited both excitatory and inhibitory transmission in slices from normal and from arthritic rats. The analysis of spontaneous and miniature EPSCs and IPSCs showed that mGluR1 acted presynaptically whereas mGluR5 had postsynaptic effects. In conclusion, mGluR1 rather than mGluR5 can account for the pain-related changes of excitatory and inhibitory synaptic transmission in the CeLC through a mechanism that involves inhibition of inhibitory transmission (disinhibition).

Full text

Page 1

RESEARCHOpen Access
Pain-related increase of excitatory transmission
and decrease of inhibitory transmission in the
central nucleus of the amygdala are mediated by
mGluR1
Wenjie Ren, Volker Neugebauer*
Abstract
Neuroplasticity in the central nucleus of the amygdala (CeA), particularly its latero-capsular division (CeLC), is an
important contributor to the emotional-affective aspects of pain. Previous studies showed synaptic plasticity of
excitatory transmission to the CeLC in different pain models, but pain-related changes of inhibitory transmission
remain to be determined. The CeLC receives convergent excitatory inputs from the parabrachial nucleus in the
brainstem and from the basolateral amygdala (BLA). In addition, feedforward inhibition of CeA neurons is driven by
glutamatergic projections from the BLA area to a cluster of GABAergic neurons in the intercalated cell masses (ITC).
Using patch-clamp in rat brain slices we measured monosynaptic excitatory postsynaptic currents (EPSCs) and poly-
synaptic inhibitory currents (IPSCs) that were evoked by electrical stimulation in the BLA. In brain slices from
arthritic rats, input-output functions of excitatory synaptic transmission were enhanced whereas inhibitory synaptic
transmission was decreased compared to control slices from normal untreated rats. A non-NMDA receptor antago-
nist (NBQX) blocked the EPSCs and reduced the IPSCs, suggesting that non-NMDA receptors mediate excitatory
transmission and also contribute to glutamate-driven feed-forward inhibition of CeLC neurons. IPSCs were blocked
by a GABAA receptor antagonist (bicuculline). Bicuculline increased EPSCs under normal conditions but not in slices
from arthritic rats, which indicates a loss of GABAergic control of excitatory transmission. A metabotropic glutamate
receptor subtype 1 (mGluR1) antagonist (LY367385) reversed both the increase of excitatory transmission and the
decrease of inhibitory transmission in the arthritis pain model but had no effect on basal synaptic transmission in
control slices from normal rats. The inhibitory effect of LY367385 on excitatory transmission was blocked by bicu-
culline suggesting the involvement of a GABAergic mechanism. An mGluR5 antagonist (MTEP) inhibited both exci-
tatory and inhibitory transmission in slices from normal and from arthritic rats. The analysis of spontaneous and
miniature EPSCs and IPSCs showed that mGluR1 acted presynaptically whereas mGluR5 had postsynaptic effects. In
conclusion, mGluR1 rather than mGluR5 can account for the pain-related changes of excitatory and inhibitory
synaptic transmission in the CeLC through a mechanism that involves inhibition of inhibitory transmission
(disinhibition).
Background
Pain has a strong emotional component and is signifi-
cantly associated with anxiety and depression. The
amygdala plays a key role in emotional learning and
memory as well as in affective disorders [1-4] and is
also important for the emotional-affective dimension of
pain and pain modulation [5-8]. Pharmacologic inhibi-
tion of amygdala hyperactivity has been shown to
decrease nocifensive and affective responses in animal
pain models [5,8-13]. Conversely, pharmacologic activa-
tion can produce pain behavior even in the absence of
tissue injury [14-17].
The amygdala consists of several anatomically and
functionally distinct nuclei [2,18]. The laterocapsular
division of the central nucleus (CeLC) has been termed
* Correspondence: voneugeb@utmb.edu
Department of Neuroscience & Cell Biology, The University of Texas Medical
Branch, Galveston, Texas 77555-1069, USA
Ren and Neugebauer Molecular Pain 2010, 6:93
http://www.molecularpain.com/content/6/1/93
MOLECULAR PAIN
© 2010 Ren and Neugebauer; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.

Page 2

the “nociceptive amygdala” because it receives nocicep-
tive-specific information from the spinal cord and brain-
stem (external parabrachial area, PB) and the vast
majority of CeLC neurons respond exclusively or prefer-
entially to noxious stimuli [5,8,19]. Synaptic plasticity of
PB inputs to the CeLC has been shown in models of
arthritic pain [20-23], visceral pain [24] and chronic
neuropathic pain [25] and is associated with pain-related
central sensitization of CeLC neurons [21,26-31]. Highly
processed multimodal, including nociceptive, informa-
tion reaches the CeLC from thalamus and cortex
through the lateral-basolateral (LA-BLA) network [5,8].
The LA-BLA circuitry is critical for the emotional eva-
luation of sensory stimuli and for acquisition and conso-
lidation of aversive associations [2,3,32,33]. Our previous
studies showed pain-related synaptic plasticity of excita-
tory transmission at the LA-BLA and BLA-CeLC
synapses [10,20,23]. The BLA can influence CeA pro-
cesses via direct glutamatergic projections and through
indirect disynaptic routes involving GABAergic neurons
in the intercalated cell masses (ITC) that project to the
CeA [2,32,34]. Activation of inhibitory ITC neurons and
subsequent inhibition of CeA neurons has been sug-
gested to play an important role in fear extinction
[2,35]. However, the role of synaptic inhibition of CeLC
neurons in pain-related plasticity remains to be deter-
mined and was addressed in this study.
Another focus of this study was on the involvement of
group I metabotropic glutamate receptors (mGluRs) in
synaptic inhibition of CeLC neurons, because our pre-
vious studies showed that these receptors are important
modulators of excitatory synaptic transmission in the
CeLC [20]. Group I mGluRs comprise mGluR1 and
mGluR5 subtypes and are involved in neuroplasticity
associated with normal brain functions as well as in
neurological and psychiatric disorders [36-39] and in
pain mechanisms [40-42]. Group I mGluRs play a criti-
cal role in pain-related central sensitization of amygdala
neurons [20,27] and in amygdala-mediated pain beha-
viors [9,15,43]. Using patch-clamp recordings in brain
slices from arthritic rats (kaolin-carrageenan model) and
from controls, we measured and compared pain-related
changes in inhibitory and excitatory transmission from
the BLA to the CeLC and the contribution of group I
mGluRs to these changes.
Results
Pain-related increase of excitatory transmission and
decrease of glutamate-driven inhibitory transmission in
CeLC neurons
Excitatory and inhibitory postsynaptic currents (EPSCs
and IPSCs, respectively) were evoked in CeLC neurons
by electrical stimulation in the BLA (Figure 1A and 1B).
Monosynaptic EPSCs recorded in voltage-clamp at
A
Glutamatergic
GABAergic
B
Recording
ITC
BLA
Stimulation
DEC
NBQX
CeLC1 mm
Bicuculline
FFHHGG
40
60
EPSCs
IPSCs
of reponses
Number o
1
failure rate
IPSC f
**
55
10
**
ncy (ms)
Laten
510
0
20
Latency (ms)
0
3rd2nd
HFS
**
0
IPSCEPSC
Figure 1 Monosynaptic excitatory and polysynaptic inhibitory
synaptic transmission in CeLC neurons. (A) Coronal brain slices
containing the right amygdala were obtained from normal rats and
arthritic rats 4-6 h after injections of kaolin and carrageenan (K/C)
into the left knee joint. Magnified area shows position of the patch-
clamp electrode ("Recording”) in the CeLC; stimulation electrode in
the BLA to activate direct glutamatergic projections and indirect
disynaptic connections that involve GABAergic neurons in the
intercalated cell masses (ITC). Diagrams are from [66]. (B) Biphasic
synaptic responses were evoked at different holding potentials (-70,
-30, and 0 mV). (C) Individual traces (average of 8-10) of synaptic
responses evoked at -70 mV (downward deflections, EPSCs;
inhibited by NBQX, 10 μM) and at 0 mV (upward deflections, IPSCs;
blocked by bicuculline, 10 μM). Scale bars, 50 pA, 30 ms. (D)
Monosynaptic EPSCs, but not polysynaptic IPSCs, follow high-
frequency stimulation (20 Hz; 6 individual traces each). Scale bars, 50
pA, 30 ms. (E) Individual EPSCs and IPSCs evoked with twice-
threshold stimulation (30 sweeps each). Latencies of IPSCs were
longer and more variable. Calibration: 50 pA, 3 ms. (F) Distribution
of EPSC and IPSC latencies measured from stimulus artifact to onset
of synaptic current in one neuron (n = 100 events). (G) Bar
histograms show average latencies (means ± SE) of EPSCs and IPSCs
in 10 neurons. ** P < 0.01, paired t-test. (H) Bar histograms show
the number of neurons that did not respond to the second or third
high-frequency stimulus (HFS, 20 Hz). The failure rate is normalized
for each neuron and averaged across the sample of neurons (n =
15; 0, no failure; 1, no IPSC). ** P < 0.01, compared to 1ststimulus
(no failure), Dunnett’s multiple comparison test.
Ren and Neugebauer Molecular Pain 2010, 6:93
http://www.molecularpain.com/content/6/1/93
Page 2 of 14

Page 3

-70 mV were mediated by non-NMDA receptors
because they were completely blocked by NBQX (10
μM, Figure 1C). IPSCs recorded at 0 mV holding poten-
tial were blocked by a GABAA receptor antagonist
(bicuculline, 10 μM, Figure 1C). EPSCs, but not IPSCs,
followed high-frequency (20 Hz) synaptic stimulation
reliably (Figure 1D). EPSCs had a fixed latency whereas
the latencies of IPSCs showed larger variability (Figure
1E and 1F). Average latency of EPSCs (from stimulus
artifact to onset of synaptic current) was significantly
shorter than that of IPSCs (see individual example in
Figure 1F; data are summarized in Figure 1G, n = 10
neurons; P < 0.01, paired t-test). Failure to follow high-
frequency synaptic stimulation was significant in the
sample of neurons (n = 15; P < 0.01, Dunnett’s multiple
comparison test, Figure 1H). The data show that CeLC
neurons receive monosynaptic excitatory and polysynap-
tic inhibitory inputs from the BLA.
To determine pain-related changes of excitatory and
inhibitory synaptic inputs, input-output (I/O) relation-
ships were obtained by measuring peak amplitudes of
EPSCs and IPSCs as a function of afferent fiber stimulus
intensity for each neuron (see examples in Figure 2A
and 2B). CeLC neurons were recorded in slices from
normal untreated rats and in slices from arthritic rats
(4-6 hr after induction of arthritis in one knee joint; see
Methods). Compared with controls (n = 11 neurons), I/
O function of monosynaptic EPSCs at the BLA-CeLC
synapse increased significantly in the arthritis pain
model (n = 12 neurons, F1,231= 30.49, P < 0.0001, main
effect of treatment, two-way ANOVA; Figure 2A),
whereas the I/O function of IPSCs decreased signifi-
cantly (control, n = 11 neurons; arthritis, n = 12 neu-
rons; F1,231 = 22.15, P < 0.0001, main effect of
treatment, two-way ANOVA; Figure 2B). The increase
of excitatory transmission relative to inhibitory trans-
mission in the arthritis pain model is reflected in the
significantly increased EPSC/IPSC ratio (P < 0.001,
unpaired t test; Figure 2C).
Next we tested the hypothesis that polysynaptic inhibi-
tory transmission is glutamate- driven. IPSCs were
inhibited by a non-NMDA receptor antagonist (NBQX,
10 μM; Figure 3) in slices from normal animals (n = 5
neurons, F1,88= 24.11, P < 0.0001, main effect of drug,
two-way ANOVA; Figure 3A) and in slices from
arthritic rats (n = 5 neurons, F1,88= 36.18, P < 0.0001,
main effect of drug, two-way ANOVA; Figure 3B). The
effect of NBQX on I/O functions of inhibitory transmis-
sion was not significantly different in arthritis compared
to normal conditions (P > 0.05, unpaired t-test; Figure
3C). The pharmacological profile (blockade by NBQX)
and synaptic characteristics (longer and more variable
latencies and inability to follow high-frequency stimula-
tion) indicate that IPSCs recorded in CeLC neurons are
A
Normal Arthritis
5050
100
Normal
Arthritis
amplitude (pA)
EPSC a
0.0 0.51.0
0
Stimulus intensity (mA)
B
Normal
Arthritis
NormalArthritis
50
100
IPSC amplitude (pA)
C
0.0 0.5 1.0
0
Stimulus intensity (mA)
1.5
***
0.5
1.0
EPSC / IPSC
0.0
Arthritis
Normal
Figure 2 Increased excitatory and decreased inhibitory
transmission in CeLC neurons in a model of arthritic pain. (A)
Input-output (I/O) functions of monosynaptic EPSCs (recorded at -70
mV) increased significantly (P < 0.0001, two-way ANOVA, see
Results) in slices from arthritis rats (n = 12 neurons) compared with
control slices from normal rats (n = 11 neurons). Traces show EPSCs
evoked with stimulus intensities of 0.3, 0.6 and 0.9 mA in one CeLC
neuron from a normal rat and in another CeLC neuron from an
arthritic rat. Scale bars, 50 pA, 10 ms. (B) I/O function of IPSCs
(recorded at 0 mV) decreased significantly (P < 0.0001, two-way
ANOVA, see Results) in slices from arthritic rats (n = 12 neurons)
compared with slices from normal rats (n = 11 neurons). Individual
traces show IPSCs evoked with stimulation intensities of 0.3, 0.6 and
0.9 mA. Scale bars, 50 pA, 10 ms. (C) The ratio of EPSCs and IPSCs
evoked with a stimulation intensity of 1 mA increased significantly
in slices from arthritic rats (n = 12 neurons) compared to controls
(n = 11 neurons). *** P < 0.001, unpaired t-test. (A-C) Symbols and
error bars represent means ± SE.
Ren and Neugebauer Molecular Pain 2010, 6:93
http://www.molecularpain.com/content/6/1/93
Page 3 of 14

Page 4

polysynaptic, involving a glutamatergic synapse. The
results are consistent with morphological and functional
evidence for glutamatergic projections from the BLA to
GABAergic interneurons in the ITC [2,35] and suggest
that CeLC neurons receive disynaptic feedforward
inhibition.
Pain-related loss of GABAergic inhibition of excitatory
transmission
We sought to determine if feedforward inhibition of
CeLC neurons modulates excitatory synaptic transmis-
sion from the BLA and if that effect changes in the
arthritis pain state. A GABAAreceptor antagonist (bicu-
culline, 10 μM) significantly increased I/O function of
excitatory transmission at the BLA-CeLC synapse under
normal conditions (n = 5 neurons, F1,88= 8.80, P < 0.01,
main effect of drug, two-way ANOVA; Figure 4A), sug-
gesting GABAergic control of excitatory inputs to the
CeLC. In slices from arthritic rats, however, bicuculline
had no significant effect on EPSCs evoked in CeLC neu-
rons (n = 5 neurons, F1,88= 2.67, P > 0.05, main effect
of drug, two-way ANOVA; Figure 4B). The significantly
decreased facilitatory effect of bicuculline in the arthritis
model (P < 0.05, unpaired t-test; Figure 4C) may suggest
that loss of inhibition contributes at least in part to the
pain-related increase of excitatory transmission (see
Figure 2A).
mGluR1, but not mGluR5, can account for the pain-
related changes of excitatory and inhibitory transmission
in CeLC neurons
Our previous studies showed that mGluR1 and mGluR5
modulate excitatory transmission in the CeLC and upre-
gulation of mGluR1 is associated with pain-related
synaptic plasticity [20]. Here we examined the modula-
tion of inhibitory transmission by mGluR1 and mGluR5.
Confirming the results of our previous study obtained
with a different mGluR1 antagonist (CPCCOEt) [20], a
selective mGluR1 antagonist (LY367385, 10 μM) inhib-
ited excitatory synaptic transmission in CeLC neurons in
slices from arthritic rats (n = 5 neurons, F1,88= 37.10, P <
0.0001, main effect of drug, two-way ANOVA; Figure 5B)
but had no significant effect under normal conditions
(n = 3 neurons, F1,44= 1.03, P > 0.05, main effect of drug,
two-way ANOVA; Figure 5A). The inhibitory effect of
LY367385 in the pain model was significantly different
from that under normal conditions (P < 0.05, unpaired
t-test; Figure 5C). LY367385 increased inhibitory synaptic
transmission in slices from arthritic rats significantly (n =
5 neurons, F1,88= 15.91, P < 0.0001, main effect of drug,
two-way ANOVA; Figure 5E) but had no significant
effect under normal conditions (n = 3 neurons, F1,44=
0.56, P > 0.05, main effect of drug, two-way ANOVA;
Figure 5D). The facilitatory effect of LY367385 in the
A N
A
l
50 50
100
Predrug (normal)
NBQX
amplitude (pA)
IPSC a
NBQX
Normal
0.00.51.0
0
Stimulus intensity (mA)
B A th iti
B Arthritis
5050
100
Predrug (arthritis)
NBQX
amplitude (pA)
IPSC a
NBQX
0.00.5 1.0
0
Stimulus intensity (mA)
CC
0.5
1.0
QX effect on IPSCs
NBQ
malized to predrug)
(norm
0.0
Arthritis
Normal
Figure 3 Inhibitory transmission onto CeLC neurons is driven
by non-NMDA receptors. (A) A non-NMDA receptor antagonist
(NBQX, 10 μM) decreased input-output (I/O) functions of IPSCs
(recorded at 0 mV) in slices from normal rats significantly (n = 5
neurons, P < 0.0001, two-way ANOVA; see Results). Traces show
IPSCs evoked in one CeLC neuron before (Predrug) and during
NBQX. Stimulus intensity, 0.9 mA; scale bars, 50 pA, 10 ms. (B) NBQX
also decreased I/O function of IPSCs (recorded at 0 mV) in slices
from arthritic rats (n = 5 neurons) significantly (n = 5 neurons, P <
0.0001, two-way ANOVA; see Results). Individual traces show IPSCs
evoked in one CeLC neuron before (Predrug) and during NBQX.
Stimulus intensity, 0.9 mA; scale bars, 50 pA, 10 ms. (C) The effect of
NBQX normalized to predrug values (set to 1.0) was not significantly
different between slices from arthritic rats and normal controls (P >
0.05, unpaired t-test). (A-C) Symbols and error bars represent
means ± SE.
Ren and Neugebauer Molecular Pain 2010, 6:93
http://www.molecularpain.com/content/6/1/93
Page 4 of 14

Page 5

pain state was significantly different from that under nor-
mal conditions (P < 0.05, unpaired t-test; Figure 5F). The
results show that LY367385 can reverse pain-related
changes of excitatory as well as inhibitory transmission in
the CeLC.
A Normal
80
120
Predrug (normal)
Bicuculline
amplitude (pA)
EPSC a
Bicuculline
0.00.51.0
0
40
Stimulus intensity (mA)
120120
Predrug (arthritis)
BicucullineBicuculline
Bicuculline
B Arthritis
40
80
EPSC amplitude (pA)E
0.00.51.0
0
Stimulus intensity (mA)
2.0 2.0
PSCs
Bicuculine effect on EP
g)
(normalized to predru
C
0 00 0 0.0 0.0
1.0 1.0
*
ArthritisNormal
Figure 4 GABAergic inhibition is lost in a model of arthritic
pain. (A) A GABAAreceptor antagonist (bicuculline, 10 μM)
significantly increased I/O function of EPSCs evoked in CeLC
neurons in slices from normal rats (n = 5 neurons, P < 0.001, two-
way ANOVA; see Results). Individual traces show EPSCs evoked in
one CeLC before (Predrug) and during bicuculline. Stimulus
intensity, 0.9 mA. Scale bars: 50 pA, 10 ms. (B) Bicuculline had no
significant effect on EPSCs recorded in CeLC neurons in slices from
arthritic rats (n = 5 neurons, P > 0.05, two-way ANOVA; see Results).
Traces show EPSCs recorded in one CeLC neuron before (Predrug)
and during bicuculline; they were nearly identical as bicuculline had
no effect. Stimulus intensity, 0.9 mA; scale bars, 50 pA, 10 ms. (C)
Bar histograms show the significantly greater facilitatory effect of
bicuculline on EPSCs (normalized to predrug; set to 1.0) in slices
from normal rats compared to arthritic rats. * P < 0.05, unpaired t-
test. (A-C) Symbols and error bars represent means ± SE.
DA
120
Predrug (normal)
LY 367385
pA)
EPSC amplitude (p
150
Predrug (normal)
LY 367385
A)
IPSC amplitude (pA
NormalNormal
0
40
80
0
50
100
B
120
Predrug (arthritis)
LY367385
A)
EPSC amplitude (pA
0.00.51.0
Stimulus intensity (mA)
120
Predrug (arthritis)
LY367385
A)
IPSC amplitude (pA
0.00.51.0
Stimulus intensity (mA)
Arthritis
E Arthritis
00
40
80
00
40
80
0.00.51.0
Stimulus intensity (mA)
0.00.51.0
Stimulus intensity (mA)
C
F
1.5
PSCs
LY367385 effect on EP
rug)
(normalized to predr
1 51.5
2.0
*
PSCs
LY367385 effect on IP
rug)
(normalized to predr
0 00.0
0.5
1.0
*
0 00.0
0.5
1.0
ArthritisNormalArthritisNormal
Figure 5 Blockade of mGluR1 reverses the pain-related
increase of excitatory transmission and the decrease of
inhibitory transmission in CeLC neurons. (A) A selective mGluR1
antagonist (LY367385, 10 μM) had no significant effect on I/O
functions of EPSCs at the BLA-CeLC synapse in slices from normal
rats (n = 3 neurons, P > 0.05, two-way ANOVA; see Results). (B)
LY367385 inhibited excitatory transmission in slices from arthritic
rats (n = 5 neurons, P < 0.0001, two-way ANOVA; see Results). (C)
The inhibitory effect of LY367385 (normalized to predrug; set to 1.0)
in the arthritis pain model was significantly different from that
under normal conditions. * P < 0.05, unpaired t-test. (D) LY367385
had no significant effect on I/O functions of IPSCs in slices from
normal rats (n = 3 neurons, P > 0.05, two-way ANOVA; see Results).
(E) LY367385 increased inhibitory transmission in slices from arthritic
rats significantly (n = 5 neurons, P < 0.0001, two-way ANOVA; see
Results). (F) The facilitatory effect of LY367385 (normalized to
predrug; set to 1.0) in the arthritis pain model was significantly
different from that under normal conditions. * P < 0.05, unpaired t-
test. (A-F) Symbols and error bars represent means ± SE.
Ren and Neugebauer Molecular Pain 2010, 6:93
http://www.molecularpain.com/content/6/1/93
Page 5 of 14