Presynaptic GABA(B) receptors regulate experience-dependent development of inhibitory short-term plasticity

Abstract

Short-term changes in synaptic gain support information processing throughout the CNS, yet we know little about the developmental regulation of such plasticity. Here we report that auditory experience is necessary for the normal maturation of synaptic inhibitory short-term plasticity (iSTP) in the auditory cortex, and that presynaptic GABA(B) receptors regulate this development. Moderate or severe hearing loss was induced in gerbils, and iSTP was characterized by measuring inhibitory synaptic current amplitudes in response to repetitive stimuli. We reveal a profound developmental shift of iSTP from depressing to facilitating after the onset of hearing. Even moderate hearing loss prevented this shift. This iSTP change was mediated by a specific class of inhibitory interneurons, the low-threshold spiking cells. Further, using paired recordings, we reveal that presynaptic GABA(B) receptors at interneuron-pyramidal connections regulate iSTP in an experience-dependent manner. This novel synaptic mechanism may support the emergence of mature temporal processing in the auditory cortex.

Figure 1.

iSTP matures after hearing onset. A, PPR shifts away from depression. Left, Representative IPSCs from a pre-hearing neuron (P10), a post-hearing neuron (P17), and a late post-hearing neuron (P27). IPSCs were evoked in cortical L2/3 pyramidal cells by extracellular stimuli applied to L4 (ISI = 80 ms). Right, Average PPR (means ± SEM) of IPSCs at ISIs ranging from 40 to 3000 ms for neurons recorded from pre-hearing (P8–11), early post-hearing (P17–22), and late post-hearing (P25–30) animals. Pre-hearing neurons displayed depression at ISIs up to 2000 ms while post-hearing neurons did not. Black asterisk indicates p < 0.05 for pre-hearing versus post-hearing; gray asterisk indicates p < 0.05 for pre-hearing versus late post-hearing (number of neurons sampled at ISIs of 40–120, 200–3000: pre-hearing, 14–18, 7–8; early post-hearing, 23–26, 7–14; late post-hearing, 12–15, 8–9). B, iSTP maturation is revealed by stimulus trains. Left, Representative IPSCs from pre-hearing (P9), early post-hearing (P17), and late post-hearing (P27) neurons during a train of 10 stimuli. Right, Average ISPC₁₀/ISPC₁ values (means ± SEM) at ISIs from 40 to 300 ms are shown for neurons recorded from pre-hearing, early post-hearing, and late post-hearing animals. IPSCs from pre-hearing neurons are robustly depressed by the end of the 10 pulse train; post-hearing and late post-hearing neurons show significantly less depression. Black asterisk indicates p < 0.05 for pre-hearing versus post-hearing; gray asterisk indicates p < 0.05 for pre-hearing versus late post-hearing (number of neurons sampled: pre-hearing, 6–12; early post-hearing, 8–13; late post-hearing, 13).

Figure 2.

Maturation of iSTP depends on auditory experience. A, Hearing loss leads to a decrease in PPR. Left, Representative IPSCs from a control post-hearing (P19), a CHL (P21), and a SNHL neuron (P19). IPSCs were evoked by paired extracellular stimuli applied to L4 and recorded in cortical L2/3 pyramidal cells (ISI = 80 ms). Right, Average PPR (means ± SEM) at ISIs ranging from 40 to 3000 ms for post-hearing (P17–22), CHL (P17–22), and SNHL (P17–22) neurons. SNHL and CHL neurons displayed depression at a range of ISIs. Orange asterisk indicates p < 0.05 for post-hearing versus CHL; red asterisk indicates p < 0.05 for post-hearing versus SNHL (number of neurons sampled at ISIs of 40–120, 200–3000: post-hearing, 23–26, 7–14; CHL, 8–12, 6–9; SNHL, 27–42, 12–19). B, Disrupted iSTP is revealed by stimulus trains. Left, Representative IPSCs from post-hearing (P17), CHL (P20), and SNHL (P18) neurons during a train of 10 stimuli. Right, Average ISPC₁₀/ISPC₁ (means ± SEM) at ISIs from 40 to 300 ms for neurons recorded from post-hearing, CHL, and SNHL neurons. IPSCs from SNHL and CHL neurons were depressed by the end of the 10 pulse train; post-hearing animals displayed significantly less depression. Orange asterisk indicates p < 0.05 for post-hearing versus CHL; red asterisk indicates p < 0.05 for post-hearing versus SNHL (number of neurons: post-hearing, 8–13; CHL, 6–11; SNHL, 8–12).

Figure 3.

Experience-dependent reduction of presynaptic GABA_B receptors. Left, Effects of the GABA_B receptor antagonist, SCH-50911 (10 μm), on IPSCs from representative pre-hearing (P9), post-hearing (P18), and SNHL (P18) neurons. PPR of IPSCs is increased by SCH-50911 in both the pre-hearing and SNHL neurons, but not in the post-hearing neuron. Gray traces were obtained after application of SCH-50911. Right, Bar graph showing the average effect of SCH-50911 on neurons recorded from pre-hearing, post-hearing, and SNHL neurons. The magnitude of the drug effect is represented as ΔPPR (PPR_SCH50911 − PPR_PRE-DRUG) (means ± SEM). This effect was greater in pre-hearing and SNHL animals compared with post-hearing at ISIs of 80, 120, and 500 ms. Blue asterisk indicates p < 0.05 for pre-hearing versus post-hearing; red asterisk indicates p < 0.05 for SNHL versus post-hearing (number of neurons: pre-hearing, 5–6; post-hearing, 6–9; SNHL, 5–12).

Figure 4.

Experience-dependent development of IPSCs evoked by FS and LTS interneurons. A, IPSCs at FS- and LTS-pyramidal connections from pre-hearing, post-hearing, and SNHL animals. IPSCs were evoked by FS or LTS interneuron spikes and recorded in pyramidal cells (V_HOLD = −60 mV). Representative spiking responses of FS and LTS interneurons to current injection (1500 ms) are shown below. FS interneurons displayed characteristic tonic spiking and LTS interneurons showed characteristic spike adaptation in response to suprathreshold current injection. Amplitude of current injection is indicated above traces. B, Table of synaptic properties of FS- and LTS-pyramidal connections from pre-hearing (P8–12), post-hearing (P17–22), and SNHL animals (P17–22) (means ± SEM). FS-evoked IPSC amplitudes showed a robust increase in amplitude and decrease in amplitude variance during development, but this was prevented by SNHL. In addition, SNHL FS-evoked IPSCs showed immature kinetics compared with post-hearing controls, including a significantly smaller rising slope, a longer rise time and latency to peak, and larger variance of the latency to peak. Conversely, at LTS-pyramidal connections, IPSC amplitudes significantly decreased during development, but not in SNHL animals. *p < 0.05, **p < 0.01. IPSC amp, IPSC amplitude; Amp CV, amplitude coefficient of variation; Rise time, IPSC 20–80% rising time; Rise slope, IPSC 20–80% rising slope; Lat to peak, latency from spike to IPSC peak; Peak CV, latency to peak coefficient of variation. The numbers of connections sampled (n) are indicated in the table.

Figure 5.

Experience-dependent development of iSTP occurs at LTS-pyramidal connections. A, IPSCs recorded at FS- and LTS-pyramidal connections evoked by two spikes in pre-hearing, post-hearing, and SNHL neurons (ISI = 80 ms). B, Average PPR (means ± SEM) of IPSCs evoked by FS or LTS interneurons from pre-hearing (P8–12), post-hearing (P17–22), and SNHL (P17–22) animals at a range of ISIs. At FS-pyramidal connections, pre-hearing, post-hearing and SNHL neurons showed similar paired-pulse depression across all ISIs (number of connections sampled: pre-hearing, 12; control, 16–22; SNHL, 12–14). At LTS-pyramidal connections, pre-hearing and SNHL neurons showed a significant reduction in PPR compared with post-hearing neurons at ISIs from 40 to 500 ms (number of connections sampled at ISIs of 40–120, 200–3000: pre-hearing, 12–13, 12; post-hearing, 17–21, 12–14; SNHL, 13–14, 12–13). Blue asterisk indicates p < 0.02 for pre-hearing versus post-hearing; red asterisk indicates p < 0.01 for SNHL versus post-hearing.

Figure 6.

Altered iSTP at LTS-pyramidal connections revealed by stimulus trains. A, IPSCs recorded at FS- and LTS-pyramidal connections evoked by trains of 10 spikes in pre-hearing, post-hearing, and SNHL neurons (ISI = 80 ms). B, Average IPSC amplitude (means ± SEM) normalized to the first IPSC (ISPC_N/ISPC₁) evoked by each of 10 stimuli at FS- and LTS-pyramidal connections (ISI = 40, 80, and 120 ms as indicated). At FS-pyramidal connections from pre-hearing (P8–12), post-hearing (P17–22), and SNHL (P17–22) animals, IPSCs became depressed during the 10-spike train (number of connections sampled: pre-hearing, 12; post-hearing, 16–20; SNHL, 12–13). At LTS-pyramidal connections, IPSCs from post-hearing animals were generally not depressed during the 10-spike train (number of connections sampled, 9–12). In contrast, IPSCs from pre-hearing and SNHL animals showed robust depression, corresponding to a significant decrease in the ISPC_N/ISPC₁ ratio (number of connections sampled: pre-hearing, 12–13; SNHL, 10–12). Blue asterisk indicates p < 0.05 for pre-hearing versus post-hearing; red asterisk indicates p < 0.05 for SNHL versus post-hearing.

Figure 7.

Experience-dependent development of GABA_B receptor function occurs at LTS-pyramidal connections. A, Effect of the GABA_B receptor antagonist, SCH-50911 (10 μm), on IPSCs at FS- and LTS-pyramidal connections from pre-hearing, post-hearing, and SNHL neurons (ISI = 80 ms). PPR of IPSCs was not affected by SCH-50911 at FS-pyramidal connections. PPR of IPSCs was increased by SCH-50911 at LTS-pyramidal connections from pre-hearing and SNHL animals. This effect was not seen at LTS-pyramidal connections from post-hearing animals. Gray traces were obtained after the application of SCH-50911. B, The average magnitude (means ± SEM) of the drug effect, ΔPPR (PPR_SCH50911 − PPR_PRE-DRUG) on IPSCs from pre-hearing (P8–12), post-hearing (P17–22), and SNHL animals (P17–22; ISI = 80 ms). ΔPPR was significantly greater in pre-hearing and SNHL animals compared with post-hearing animals at LTS-pyramidal connections (number of connections sampled: pre-hearing, 9; post-hearing, 7; SNHL, 7; *p < 0.05), but not FS-pyramidal connections (number of connections sampled: pre-hearing, 8; post-hearing, 12; SNHL, 10).

Figure 8.

Auditory experience regulates GABA_B receptor-mediated inhibition of calcium currents. Effects of the GABA_B receptor agonist, baclofen (100 μm; green trace), on calcium currents evoked by voltage steps (200 ms) in LTS cells. Left, Baclofen produced no effect on the peak calcium current in LTS cells from post-hearing animals. Right, Baclofen decreased the amplitude of the peak calcium current in LTS cells from SNHL animals. Insets show spiking responses of LTS interneurons from post-hearing and SNHL neurons to current injection (1500 ms) obtained immediately after rupture. Numbers indicate the average magnitude of baclofen-induced suppression of calcium currents (means ± SEM). The percentage decrease in the peak calcium currents by baclofen was significantly greater in SNHL LTS cells compared with post-hearing LTS cells (number of LTS cells sampled: post-hearing, 12, SNHL, 6; p = 0.02).