Effects of stress on AMPA receptor distribution and function in the basolateral amygdala

Abstract

Stress is a growing public health concern and can lead to significant disabilities. The neural response to stressors is thought to be dependent on the extended amygdala. The basolateral amygdala (BLA) is responsible for associations of sensory stimuli with emotional valence and is thought to be involved in stress-induced responses. Previous behavioral and electrophysiological experiments demonstrate that, in response to stress, changes occur in glutamatergic neurotransmission within the BLA and, in particular in transmission at AMPA receptors. Given the established role of AMPA receptors in memory and synaptic plasticity, we tested the hypothesis that stress produces alterations in the distribution of these receptors in a way that might account for stress-induced alterations in amygdala circuitry function. We examined the subcellular localization of GluR1 subunits of the AMPA receptor and the electrophysiological characteristics of BLA principal neurons in an animal model of unpredictable stress. Compared to controls, we demonstrated an increase in the ratio of labeled spines to labeled dendritic shafts in the BLA of rats 6 and 14 days post-stress, but not 1 day post-stress. Furthermore, the frequency of mini-EPSCs was increased in stressed animals without a change in general membrane properties, mini-EPSC amplitude, or in paired pulse modulation of glutamate release. Taken together, these data suggest that the shift of GluR1-containing AMPA receptors from dendritic stores into spines may be in part responsible for the persistent behavioral alterations observed following severe stressors.

Figure 1

Electron micrograph of GluR1 immunoreactivity in the BLA from a control rat (A) and a 6-day post stress rat (B). Arrowheads indicate GluR1-immunoreactive spines; arrows indicate GluR1-immunoreactive dendritic shafts; and open arrowheads indicate unlabeled spines. Note that there are non-labeled spines and dendritic shafts intermingled with the labeled ones. Scale bar in (B) is 500nm.

Figure 2

Bar chart of the distribution of GluR1 immunoreactivity in different types of neuronal elements, pooled from control subjects. GluR1 is primarily located in dendritic shafts and spines with very little observed in axonal elements or glia. Bars indicate mean with standard errors (n=16).

Figure 3

GluR1 immunoreactive spine to dendritic shaft ratio is increased following exposure to stress. The proportion of GluR1 labeled spines to GluR1 labeled dendritic shafts was calculated in control subjects and in 3 groups of subjects exposed to 4 daily 30 minute footshock sessions (1-, 6-, and 14-day post shock groups). There was a significant effect of time post stress on the spine to dendrite ratio (F(3,30)=15.582, p<0.0001). The labeled spine to dendrite ratio is increased in the 6- (p<0.0001) and 14-day (p=0.0055) post stress groups as compared to the control group, but not in the 1-day (p=0.81) post stress group. Asterisks indicate a significance level of p<0.05.

Figure 4

Stress increases the frequency of mEPSCs. Spontaneous mEPSCs of BLA principal neurons were recorded with intracellular application of Picrotoxin (100 μM), bath application of CGP36742 (2 μM) and TTX (1 mM). (A) and (B) Representative traces showing mEPSCs of BLA principal neurons from Control and Stress animals respectively. (C) and (D) Group data indicate unpredictable shock stress increased the frequency but not the amplitude of mEPSCs. (Control: n=14, Stress: n=17; * p < 0.05). (E) and (F) Cumulative data plots showing that unpredictable shock stress decreased the inter-event interval but not the amplitude of mEPSCs.

Figure 5

(A) Raw data show the effects of unpredictable shock stress on the PPR of eEPSCs. eEPSC amplitude was measured using paired stimuli pulses separated by 50 ms, in slices from control and stress animals respectively. (B) A plot chart showing the group data for the effects of unpredictable shock stress on PPR using paired stimuli pulses separated by 20, 50, 100 and 200 ms. Note: The PPR was not changed (n=14, P=0.79) in slices from stress animals (triangles) compared to control animals (squares). Error bars indicate S.E.M.