Retrograde regulation of GABA transmission by the tonic release of oxytocin and endocannabinoids governs postsynaptic firing

Abstract

The probability of neurotransmitter release at the nerve terminal is an important determinant of synaptic efficacy. At some central synapses, the postsynaptic, or target, neuron determines neurotransmitter release probability (P(r)) at the presynaptic terminal. The mechanisms responsible for this target-cell dependent control of P(r) have not been elucidated. Using whole-cell patch-clamp recordings from magnocellular neurosecretory cells in the paraventricular and supraoptic nuclei of the hypothalamus, we demonstrate that inhibitory, GABA synapses specifically onto oxytocin (OT)-producing neurosecretory cells exhibit a low P(r) that is relatively uniform at multiple synapses onto the same cell. This low P(r) results from a two-step process that requires the tonic release of OT from the postsynaptic cell. The ambient extracellular levels of neuropeptide are sufficient to activate postsynaptic OT receptors and trigger the Ca2+-dependent production of endocannabinoids, which act in a retrograde manner at presynaptic cannabinoid CB1 receptors to decrease GABA release. The functional consequence of this tonic inhibition of GABA release is that all inhibitory inputs facilitate uniformly when activated at high rates of activity. This causes inhibition in the postsynaptic cell that is sufficiently powerful to disrupt firing. Blockade of CB1 receptors increases P(r) at these synapses, resulting in a rapid depression of IPSCs at high rates of activity, thereby eliminating the ability of afferent inputs to inhibit postsynaptic firing. By playing a deterministic role in GABA release at the afferent nerve terminal, the postsynaptic OT neuron effectively filters synaptic signals and thereby modulates its own activity patterns.

Figure 1.

P_r at GABA synapses is target-cell dependent. A, MNCs were filled with biocytin and identified with post hoc immunocytochemical processing. Top depicts a biocytin-filled cell that was positive for OT neurophysin (arrow) but negative for VP. Bottom depicts a biocytin-filled cell that was negative for OT but positive for VP (arrow). B, Examples of synaptic currents obtained from immuno-identified OT and VP cells in both SON and PVN in response to paired stimulation. OT cells in both nuclei exhibited PPF (PPR, 1.42 ± 0.09; n = 22), whereas VP cells exhibited PPD (PPR, 0.64 ± 0.04; n = 16). Data are summarized in bar graph. C, Two independent afferent pathways were stimulated while a recording was obtained from a single MNC. Independence was tested by first stimulating one pathway (S1; a) and then confirming that previous stimulation of the second pathway (S2) had no effect on the amplitude of the evoked response from S1 (b, c). The S1 pathway exhibited PPF (d). D, Stimulation of S2 resulted in a PPR that was not different from the PPR obtained from S1 (p > 0.05; n = 16 cells), as expected if P_r is governed by the postsynaptic cell. E, There was a strong correlation between PPR in S1 and PPR in S2.

Figure 2.

Control of P_r by retrograde transmitters. A, Exogenous application of the OT receptor antagonist MC (10 μm) increases amplitude of IPSCs in OT cells (1.33 ± 0.17 as a fraction of control; n = 12; p < 0.01) but not VP cells (1.11 ± 0.33 as a fraction of control; n = 7; p > 0.05). Representative traces are left, and summary of effects of MC on evoked responses are right. B, MC decreases PPR in OT cells (control PPR, 1.47 ± 0.19; MC PPR, 0.90 ± 0.08; n = 12; p < 0.01) but has no effect on PPR in VP cells (control PPR, 0.64 ± 0.04; MC PPR, 0.74 ± 0.17; n = 7; p > 0.05). Representative scaled traces are left, and summary of effects of MC on PPR are right. C, Exogenous application of the CB₁ receptor antagonist AM-251 (5 μm) increases the amplitude of IPSCs in OT cells (1.30 ± 0.13 as a fraction of control; n = 14; p < 0.01) but not VP cells (0.90 ± 0.13 as a fraction of control; n = 7; p > 0.05). Representative traces are left, and summary of effects of AM-251 on evoked responses are right. D, AM-251 decreases PPR in OT cells (control PPR, 1.87 ± 0.19; AM-251 PPR, 0.96 ± 0.08; n = 14; p < 0.01) but has no effect on VP cells (control PPR, 0.57 ± 0.12; AM-251 PPR, 0.57 ± 0.1; n = 7; p > 0.05). Representative scaled traces are left, and summary of effects of AM-251 on PPR are right. E, IPSC amplitude was decreased by application of either OT (1 μm; 0.68 ± 0.14% as a fraction of control; n = 9; p < 0.01) or the CB₁R agonist WIN (1 μm; 0.54 ± 0.09 as a fraction of control; n = 6; p < 0.01). Representative traces are left, and summary of effects of OT or WIN on evoked responses are right. F, PPR was also increased in the presence of either OT (control PPR, 1.37 ± 0.17; OT PPR, 2.19 ± 0.26; n = 9; p < 0.01) or WIN (control PPR, 1.39 ± 0.12; WIN PPR, 2.05 ± 0.29; n = 6; p < 0.01). **p < 0.01; ^#p < 0.01.

Figure 3.

OT recruits the eCB system to depress GABA release. A, Representative traces show that blockade of CB₁ receptor increases the evoked response (1.30 ± 0.13 of control; n = 4), but there is no additional enhancement during blockade of OTRs (AM-251 plus MC, 1.33 ± 0.08 of control; n = 4; p > 0.05). Middle row shows decrease in IPSC amplitude when OT is applied in the presence of AM-251 (0.59 ± 0.11 as a fraction of control; n = 4; p < 0.01). Bottom row shows effect of exogenous WIN (1 μm) on IPSC in the presence of MC (0.55 ± 0.10 of control; n = 4; p < 0.01). Data are summarized at the right. B, Scaled traces show no effect of MC on PPR when slice is pretreated with AM-251 (AM-251, 1.30 ± 0.13 of control vs AM-251 plus MC, 1.33 ± 0.08 of control; n = 4; p > 0.05; top traces). Similarly, exogenous OT has no effect on PPR in the presence of AM-251 (AM-251 PPR, 0.96 ± 0.08; AM-251 plus OT PPR, 1.12 ± 0.20; n = 4; p > 0.05; middle), but WIN still effectively increased PPR in the presence of MC (MC PPR, 0.90 ± 0.08; MC plus WIN PPR, 1.88 ± 0.33; n = 4; bottom row). The PPR data are summarized at the right. **p < 0.01.

Figure 4.

Increase in postsynaptic Ca²⁺ is necessary for tonic inhibition of GABA release. A, Representative traces showing effects of EGTA alone (10 mm; top), OT (middle), and WIN (bottom) when EGTA was included in the patch pipette. Right summarizes of effects of OT (1.0 ± 0.09 as a fraction of control; n = 6; p > 0.05) and WIN (0.43 ± 0.12 as a fraction of control; n = 5; p < 0.01) on evoked responses when EGTA was included in the patch pipette. B, Representative traces showing (top) effect of EGTA (10 mm) on PPR. EGTA itself decreases PPR (PPR initial 10 min, 1.17 ± 0.05; PPR 20 min, 0.96 ± 0.08; n = 6; p < 0.05). The effects of OT on PPR were blocked when EGTA was included in the patch pipette (control PPR, 0.96 ± 0.08; OT PPR, 0.86 ± 0.07; p > 0.05; middle traces), but WIN was still effective in the presence of EGTA in the postsynaptic cell (WIN, 1.35 ± 0.13; n = 6; p < 0.05 compared with PPR in EGTA; bottom traces). Right summarizes effects of EGTA, EGTA plus OT, and EGTA plus WIN on PPR. **p < 0.01.

Figure 5.

Presynaptic short-term plasticity impacts postsynaptic firing. A, Synapses with PPR >1 exhibit robust facilitation when activated at 20 Hz for 1 s (top trace). After application of AM-251 in this cell, PPD is evident and the synaptic responses exhibit depression during the train. B, The averaged IPSCs from four cells are plotted with control in black circles and AM-251 in white squares. The box depicts the segment of IPSCs used for calculating IPSC_ss shown in the right (control IPSC_ss, 153.7 ± 5.9 pA; AM-251 IPSC_ss, 66.3 ± 3.9 pA; n = 4; p < 0.0001). C, Current-clamp trace from an OT cell shows effects of 20 Hz, 1 s stimulation on firing frequency. Raster plots from five consecutive trials in this cell are shown below. The summary of the effect of 20 Hz stimulation on firing (control spike frequency, 6.34 ± 0.36 Hz; spike frequency during afferent stimulation, 1.47 ± 0.67; p < 0.01; n = 5). D, The effect of 20 Hz stimulation on firing after blockade of CB₁ receptors in the same cell as C is shown. Raster plots below show little effect on firing in individual trials. Immediately below, the summary of the effect of 20 Hz stimulation on firing in the presence of AM-251 (control spike frequency, 6.84 ± 0.19 Hz; spike frequency during afferent stimulation; 4.13 ± 0.13; n = 5; p < 0.001 compared with firing during stimulation in control cells).