A single fear-inducing stimulus induces a transcription-dependent switch in synaptic AMPAR phenotype

Abstract

Changes in emotional state are known to alter neuronal excitability and can modify learning and memory formation. Such experience-dependent neuronal plasticity can be long-lasting and is thought to involve the regulation of gene transcription. We found that a single fear-inducing stimulus increased GluR2 (also known as Gria2) mRNA abundance and promoted synaptic incorporation of GluR2-containing AMPA receptors (AMPARs) in mouse cerebellar stellate cells. The switch in synaptic AMPAR phenotype was mediated by noradrenaline and action potential prolongation. The subsequent rise in intracellular Ca(2+) and activation of Ca(2+)-sensitive ERK/MAPK signaling triggered new GluR2 gene transcription and a switch in the synaptic AMPAR phenotype from GluR2-lacking, Ca(2+)-permeable receptors to GluR2-containing, Ca(2+)-impermeable receptors on the order of hours. The change in glutamate receptor phenotype altered synaptic efficacy in cerebellar stellate cells. Thus, a single fear-inducing stimulus can induce a long-term change in synaptic receptor phenotype and may alter the activity of an inhibitory neural network.

Figure 1

An olfactory stimulus altered synaptic AMPA receptor subtype and expression of GluR2 mRNA in stellate cells. A. A natural olfactory stimulus, fox urine caused fear (measured as a freezing response). % freezing was calculated during the 3 min control and 5 min fox urine exposure period (n = 5; *, P < 0.01). B. Inhibition of EPSCs at −60 mV by IEM-1460 (100 µM), a selective Ca²⁺-permeable AMPAR blocker (n = 3; P < 0.05). C. Cumulative distribution of decay time constant of EPSC at −60 mV of individual synaptic events from 5 cells under each condition (Kolmogorov-Smirnov test, P < 0.0001, control vs. 3 hr). D. Synaptic currents and I–V relationship when spermine was included in the pipette in stellate cells. Control cells displayed an I–V relationship with pronounced inward rectification, suggesting the presence of GluR2 lacking receptors. Mice were exposed to fox urine, a fear-inducing olfactory stimulus, for 5 min. Cerebellar slices were prepared 15 min after olfactory stimulus, synaptic currents were recorded within 2 hour after fox urine exposure and exhibited an inwardly rectifying I–V relation (n = 4). Three hours following fox urine exposure the synaptic current in stellate cells showed a near linear I–V relationship (n = 5, P < 0.01), indicating that it was mediated mainly by GluR2-containing AMPARs. When slices were prepared 15 hours after fox urine exposure the synaptic current still showed a near linear I–V relationship (n = 4). E. Rectification index (*, P < 0.01; **, P < 0.005). F. The GluR1, 2 and 3 mRNA level in individual stellate cells was determined using real time single cell RT-PCR. GFP positive neurons (~8 µm in diameter) were isolated from the cerebellar cortex of GAD-65 GFP mice (control, 23 cells from 4 animals; 3 hr following fox urine exposure, 18 cells; 15 hr following fox urine exposure, 15 cells; *, P < 0.05). Scale bars, 200 µM (top) and 20 µM (bottom). Error bars show ± s.e.m.

Figure 2

β-adrenergic receptors mediated the olfactory stimulus-induced change in synaptic AMPA receptor subtype. Mice were injected with propranolol or saline (as control) 15–30 min prior to the fox urine exposure. A. A natural olfactory stimulus, fox urine caused fear (measured as a freezing response). % freezing was calculated during the 3 min control and 5 min fox urine exposure period (n = 5; *, P < 0.01). B. Slices were prepared 15 hours after fox urine exposure. Synaptic currents and I–V relationship of EPSCs in stellate cells from the mice pre-injected with saline (n = 6) and propranolol (n = 6). C. Cumulative distribution of EPSC amplitude at +40 mV and decay time constant of EPSC at −60 mV of individual synaptic events from 6 cells under each condition (Kolmogorov-Smirnov test, P < 0.0001, 15 h vs. propranolol). D. Rectification index (***, P < 0.001). Error bars show ± s.e.m.

Figure 3

Noradrenaline induced a change in synaptic AMPA receptor phenotype. A. Average sEPSCs displayed an inwardly rectifying I–V relationship in control, and became more linear following noradrenaline treatment (control, n = 4; noradrenaline treatment, n = 8). Cerebellar slices were incubated with kynurenic acid (1 mM) and picrotoxin (100 µM) in the absence (control) or presence of noradrenaline (10 µM, 3 h). Following each treatment noradrenaline and kynurenic acid were washed out prior to recordings of sEPSCs. B. The decay time of sEPSCs at −60 mV increased following noradrenaline treatment (Kolmogorov-Smirnov test, P < 0.0001). Cumulative distribution of decay time constant of EPSC at −60 mV of individual synaptic events from 4 control cells and 8 noradrenaline treated cells. C. Noradrenaline also induced a change in the I–V relationship of evoked EPSCs at the parallel fibre to stellate cell synapse (control, n = 4; noradrenaline treatment, n = 5). D. Summary of rectification index of EPSCs. Cerebellar slices were incubated with 10 µM noradrenaline for 3 h, 0.5 h (+2.5 h in picrotoxin and kynurenic acid control, n = 5; picrotoxin and kynurenic acid control, n = 4). E. CNQX (10 µM) and cyclothiazide (100 µM) evoked inward currents of comparable amplitude in control (n = 4) and noradrenaline treated cells (n = 5; two way ANOVA test, P = 0.39). (*, P < 0.05; **, P < 0.005). F. Noradrenaline (10 µM) increased the frequency (left panel) and duration (right panel) of spontaneous action potentials in stellate cells at 36°C (frequency, n = 5, P < 0.05; duration, n = 5, P < 0.005). Error bars show ± s.e.m.

Figure 4

Increasing the action potential duration in stellate cells induces a change in rectification of the I–V relationship. A. Duration of spontaneous action potentials in cerebellar stellate cells increased during bath application of TEA at room temperature. B. sEPSCs displayed a nearly linear I–V relationship following TEA treatment (control, n = 5; TEA treatment, n = 6). Cerebellar slices were incubated with kynurenic acid (1 mM) and picrotoxin (100 µM) in the absence (control) or presence of TEA (1 mM, 3 h). C. Summary of rectification index of EPSCs. Cerebellar slices were incubated with 100 µM picrotoxin (n = 5; control, n = 4) or with 1 mM TEA for 3 h (at 37°C, n = 3; at room temperature, n = 6). D. The decay time of sEPSCs increased following TEA treatment (Kolmogorov-Smirnov test, P < 0.0001). Error bars show ± s.e.m.

Figure 5

Noradrenaline increases Ca influx during the action potential and Ca entry via L-type Ca channels is required for NA-induced change in AMPAR phenotype. A. Duration of Ca²⁺ currents was enhanced using NA-AP as the voltage command, compared to control (n = 5). Nifedipine (20 µM) blocked most of the Ca²⁺ current using control-AP and NA-AP as the voltage command (n = 5). B. Following 3 hour incubation with noradrenaline and nifedipine sEPSCs displayed an inwardly rectifying I–V relationship (n = 6). C. Summary of rectification index of EPSCs (nifedipine alone, n = 5). (*, P < 0.05; **, P < 0.005). Error bars show ± s.e.m.

Figure 6

Activation of ERK-dependent pathways and gene transcription are required for the noradrenaline and action potential broadening-induced change in sEPSC rectification. A. I–V relationship of sEPSCs remained inwardly rectifying in a MEK1/2 inhibitor, U0126 (2 µM, n = 5) which was included during the noradrenaline treatment. B. I–V relationship was unaltered if U0126 (2 µM, n = 3; 20 µM, n = 5) was present during TEA treatment. C. The presence of actinomycin D (25 µM) during noradrenaline treatment prevented noradrenaline-induced change in sEPSC rectification (n = 5). D. Inclusion of actinomycin D (25 µM) during TEA treatment blocked TEA-induced change in sEPSC rectification (actinomycin D + TEA, n = 5). E. Summary of rectification index (actinomycin D alone, n = 3; U0126 control, n = 4). F. Summary of rectification index (actinomycin D alone, n = 5; DMSO control, n = 4; DMSO + TEA, n = 5; U0126 control, n = 5; PD98059 control, n = 3; PD98059 + TEA, 10 µM, n = 2; 25 µM, n = 5). No difference in sEPSC amplitude was observed between the two different concentrations of each inhibitor and therefore the data was pooled. G. Burst stimulation of parallel fibres induced a change in sEPSC rectification that lasted for 4 hours (before stimulation, RI = 0.32 ± 0.02; 30–60 min after stimulation, RI = 0.62 ± 0.1; 3–4 hours after stimulation, RI = 0.64 ± 0.08; n = 4, P < 0.03). Actinomycin D did not prevent this change (before stimulation, RI = 0.34 ± 0.03; 30–60 min after stimulation, RI = 0.61 ± 0.05; 3–4 h after stimulation; RI= 0.59 ± 0.01; n = 4; P < 0.02). Recordings were made sequentially from a group of stellate cells located in the same region of the molecular layer. (*, P < 0.05; **, P < 0.005). Error bars show ± s.e.m.

Figure 7

Noradrenaline and TEA treatment increased the level of GluR2, but not GluR1 mRNA expression (n = 5). A–H. DIG-labeled RNA antisense probes for GluR1 and GluR2 mRNA (A–C; E–H) or the sense probes (D) were used. A and E. Control. B and C. noradrenaline treatment increased the level of GluR2 mRNA expression and actinomycin D (Act D) blocked the NA-induced increase in GluR2 expression. F. TEA treatment increased the level of GluR2, but not GluR1 mRNA expression. G and H. Actinomycin D (Act D) and U0126 prevented the TEA-induced increase in GluR2 mRNA expression in stellate cells. ML, molecular layer; PL, Purkinje cell layer; GL, granule cell layer; KYNA, kynurenic acid; picrotoxin, picrotoxin. The labeling that has intensity higher than background within an outline of typical stellate cells (~ 8 µm) was selected as regions of interest. Images are typical of n = 5 in each group. I, G and K. The number of labeled stellate cells that express high level of GluR2 mRNA under each condition relative to control. We used the mean of background intensity plus two standard deviations as threshold for positive labeling. (*, P < 0.05, by unpaired Student’s t-test). L, M and N. Cumulative distribution of the labeling intensity of selected regions of interest after background subtraction illustrated the changes in staining intensities of positive stained cells under each treatment condition (the number of ROIs ranged from 104 to 365 under each condition; noradrenaline vs. control or noradrenaline + Act D, TEA vs. control, or TEA + Act D, or TEA + U0126, Kolmogorov-Smirnov test, P < 0.0001). Scale bars, 200 µM (left panels), and 50 µM (right panels).