D2-like dopamine receptor-mediated modulation of activity-dependent plasticity at GABAergic synapses in the subthalamic nucleus

Abstract

Reciprocally connected glutamatergic subthalamic nucleus (STN) and GABAergic external globus pallidus (GP) neurons normally exhibit weakly correlated, irregular activity but following the depletion of dopamine in Parkinson's disease they express more highly correlated, rhythmic bursting activity. Patch clamp recording was used to test the hypothesis that dopaminergic modulation reduces the capability of GABAergic inputs to pattern 'pathological' activity in STN neurons. Electrically evoked GABA(A) receptor-mediated IPSCs exhibited activity-dependent plasticity in STN neurons, i.e. IPSCs evoked at frequencies between 1 and 50 Hz exhibited depression that increased with the frequency of activity. Dopamine, the D(2)-like dopamine receptor agonist quinpirole and external media containing a low [Ca(2+)] reduced both the magnitude of IPSCs evoked at 1-50 Hz and synaptic depression at 10-50 Hz. Dopamine/quinpirole also reduced the frequency but not the amplitude of miniature IPSCs recorded in the presence of tetrodotoxin. D(1)-like and D(4) agonists were ineffective and D(2/3) but not D4 receptor antagonists reversed the effects of dopamine or quinpirole. Together these data suggest that presynaptic D(2/3) dopamine receptors modulate the short-term dynamics of GABAergic transmission in the STN by lowering the initial probability of transmitter release. Simulated GABA(A) receptor-mediated synaptic conductances representative of control or modulated transmission were then generated in STN neurons using the dynamic clamp technique. Dopamine-modulated transmission was less effective at resetting autonomous activity or generating rebound burst firing than control transmission. The data therefore support the conclusion that dopamine acting at presynaptic D(2)-like receptors reduces the propensity for GABAergic transmission to generate correlated, bursting activity in STN neurons.

Figure 2. Reduction in extracellular [Ca²⁺] leads to a reduction in synaptic depression

A, IPSCs evoked at 50 Hz in a representative neuron under control conditions (Aa; extracellular [Ca²⁺]= 1.6 mm), in the presence of reduced extracellular [Ca²⁺] (Ab; extracellular [Ca²⁺]= 0.6 mm) and upon return to control conditions (Ac). B, population data arising from 6 neurons. Plots of IPSC amplitude against IPSC number. Ba, the amplitudes of total (phasic + tonic) IPSCs were reduced throughout the period of evoked transmission by lowering extracellular [Ca²⁺]. Bb and c, total (b) and phasic (c) IPSC amplitude expressed as a percentage of the amplitude of the first evoked IPSC. Ba–c, in the presence of reduced extracellular [Ca²⁺] synaptic depression was reduced and facilitation was observed. *P < 0.05.

Figure 1. GABA_A receptor-mediated synaptic transmission exhibits activity-dependent depression

A–C, IPSCs evoked by stimulation at a frequency of 1 Hz, 10 Hz and 50 Hz for a duration of 20 s, 5 s and 2 s, respectively, in a representative neuron. D, total IPSC amplitude expressed as a percentage of the amplitude of the first IPSC plotted against time for each frequency of stimulation (1 Hz: □ 10 Hz: ○ 50 Hz: ▵; n = 26, 38 and 40 cells for stimulation at 1, 10 and 50 Hz, respectively). E, mean amplitude of the ‘tonic’ current measured immediately prior to each stimulus artifact plotted against time for 10 and 50 Hz stimulation. A sizeable current was detected in 14 of 26 cells at 10 Hz (○) and in all neurons at 50 Hz (▵). F, phasic IPSC amplitude (total IPSC amplitude – tonic IPSC amplitude) expressed as a percentage of the amplitude of the first IPSC plotted against time for each frequency. Grey lines depict fits to the depression of phasic IPSCs at each frequency.

Figure 3. Dopamine reduces activity-dependent depression

A and B, IPSCs evoked at 1 (panels a), 10 Hz (panels b), and 50 Hz (panels c) in a representative neuron under control conditions (A) and in the presence of dopamine (B; 10 μm). Dopamine reduced the amplitude of evoked IPSCs at each frequency tested but only reduced synaptic depression at 10 and 50 Hz. Ca–c, total IPSC amplitude (pA) plotted against IPSC number for the sample population. Ca, the amplitude of IPSCs evoked at 1 Hz was consistently reduced by dopamine throughout the sequence of stimulation (n = 7). Cb and c, at 10 (b) and 50 (c) Hz, the amplitude of evoked IPSCs was significantly reduced in the early but not the later phase of the stimulation sequence (10 Hz: n = 8; 50 Hz: n = 10). Da–c, total IPSC amplitude expressed as a percentage of the first evoked IPSC, plotted against IPSC number for the sample population (a, 1 Hz, n = 7; b, 10 Hz, n = 8; c, 50 Hz, n = 10). Dopamine significantly reduced synaptic depression at 10 and 50 Hz and unmasked synaptic facilitation at 50 Hz. *P > 0.05.

Figure 4. Dopamine reduces the frequency but not the conductance of mIPSCs in the STN

Aa and b, examples of mIPSCs recorded in voltage clamp at −60 mV under control conditions (a; black trace) and in the presence of 10 μm dopamine (b; grey trace) in a representative neuron. Ba and b, cumulative distributions of intervals between mIPSCs (a) and conductances underlying mIPSCs (b) before (black) and after the perfusion of dopamine (grey) for the neuron illustrated in A. Dopamine caused a significant increase in the intervals between mIPSCs but no alteration in their conductance. Ca and b, composite cumulative distributions of mIPSC intervals (a) and conductances (b) generated from 6 neurons recorded under control conditions (black) and then in the presence of 10 μm dopamine (grey) confirm at the population level the effect of dopamine on miniature GABAergic synaptic transmission. Da and b, line plots illustrating the actions of dopamine on the frequency and conductance of mIPSCs in each neuron. Black horizontal bars represent the mean frequency/conductance. *P < 0.001 (Ba, Ca); *P < 0.05 (Da).

Figure 5. D₂ and/or D₃ dopamine receptors but not D₁-like or D₄ dopamine receptor activation reduces activity-dependent depression

A–C, population plots of total IPSC amplitude (expressed as a percentage of the first IPSC) against IPSC number for 1 (panels a), 10 (panels b) and 50 Hz (panels c) stimulation under control conditions (black) and in the presence of the D₂-like receptor agonist quinpirole (2 μm) (A, red) or the D₄ dopamine receptor agonist PD168077 (1 μm) (B, blue) or the D₁-like dopamine receptor agonist SKF81297 (5 μm) (C, green). Quinpirole significantly reduced activity-dependent depression of IPSCs at 10 (Ab; n = 13) and 50 Hz (Ac; n = 15). Neither PD168077 (Ba–c; n = 6) or SKF81297 (Ca–c; n = 4, 6 and 6 at 1, 10 Hz and 50 Hz, respectively) altered activity-dependent depression. D, summary plots of SSD at 1 (a), 10 (b) and 50 (c) Hz under control conditions (black) and in the presence of quinpirole (red) or PD168077 (blue) or SKF81297 (green). SSD at 10 and 50 Hz was significantly reduced by quinpirole only; *P < 0.05.

Figure 6. The D_2/3 dopamine receptor antagonist sulpiride reverses the action of quinpirole and dopamine on activity-dependent depression

Aa–c, examples of IPSCs evoked at 50 Hz in control conditions (a), in quinpirole (2 μm) (b) and in quinpirole (2 μm) and sulpiride (2 μm; c). B, population plots of IPSC amplitude (expressed as a percentage of the first IPSC) against IPSC number for 3 neurons in the 3 conditions (control: black; quinpirole: red; quinpirole + sulpiride: blue). Ca–d, examples of IPSCs evoked at 50 Hz in control conditions (a), in dopamine (10 μm; b), in dopamine and the D₄ receptor antagonist LY745870 (1 μm; c) and in dopamine, LY745870 and sulpiride (2 μm; d) in a typical neuron. Only sulpiride reversed the effect of dopamine. Da, population plots of total IPSC amplitude (expressed as a percentage of the first IPSC) against IPSC number for the 4 conditions (n = 5; control, black; dopamine, red; dopamine + LY745870, green; dopamine + LY745870 + sulpiride, blue). Db, SSD in the 4 conditions (n = 5). *P < 0.05.

Figure 7. D₂-like but not D₁-like dopamine receptor activation reduces the frequency of mIPSCs

A, examples of mIPSCs recorded in voltage-clamp at −60 mV under control conditions (a; black) and in the presence of 2 μm quinpirole (b; grey) in a representative neuron. B and C, cumulative distributions of intervals between mIPSCs (Ba) and mIPSC conductances for the neuron in A (Bb) and for the sample population (Ca and b; n = 7) under control conditions (black) and in the presence of quinpirole (grey). Da and b, Line plots illustrating the actions of quinpirole on the frequency and conductance of mIPSCs in each neuron. Black horizontal bars represent the mean frequency/conductance. E–H, as for A–D but in the presence of 5 μm SKF81297 (grey). SKF81297 did not modify the frequency (Fa, Ga, Ha) or conductance (Fb, Gb, Hb) of mIPSCs in a representative neuron (E and F) or the sample population (G and H; n = 4). *P < 0.001 (Ba, Ca); *P < 0.05 (Da).

Figure 8. Presynaptic dopaminergic modulation reduces the capability of GABAergic synaptic transmission to reset autonomous activity

A and B, data from a representative neuron that was recorded in the perforated patch configuration. Aa, top, dynamic clamp waveform mimicking GABAergic synaptic transmission at 50 Hz under control conditions. Bottom, resetting of neuronal activity by the control waveform (50 superimposed trials). Ab, raster plot from the neuron in Aa. B, a dynamic clamp waveform mimicking GABAergic synaptic transmission at 50 Hz in 2 μm quinpirole was less effective at resetting autonomous activity (arrangement of panels as for A). C, differences in the latency (Ca), variability (Cb; s.d. of latency) and threshold (Cc; APth) of action potentials following the control and modulated waveform. The variability of the first action potential following the modulated waveform was significantly greater than for the control waveform. *P < 0.05.

Figure 9. Presynaptic dopaminergic modulation reduces the effect of GABAergic inhibition on the threshold of autonomously generated action potentials

A and B, data from a representative neuron that was recorded in the perforated patch configuration. Aa, top, dynamic clamp waveform mimicking GABAergic synaptic transmission at 50 Hz under control conditions. Bottom, resetting of neuronal activity by the control waveform (50 superimposed trials). Ab, raster plot from the neuron in Aa. Ac, phase plot of an autonomously generated action potential before (black) and immediately after inhibition (blue). Inset illustrates the lowering of action potential threshold (dots) by inhibition. B, effects of a dynamic clamp waveform mimicking GABAergic synaptic transmission at 50 Hz in the presence of 10 μm dopamine (arrangement of panels as for A). The modulated waveform less powerfully reduces the threshold of action potentials compared to the control waveform. C, differences in the latency (a), variability (b; s.d. of latency) and threshold (c; APth) of action potentials following the control and modulated waveform. The latency and the threshold of the first action potential following the modulated waveform were significantly shorter and greater than for the control waveform, respectively. *P < 0.05.

Figure 10. Presynaptic dopaminergic modulation reduces the capability of GABAergic transmission to evoke rebound activity

A–D, data from a representative neuron that was recorded in the perforated patch configuration. Aa, top, dynamic clamp waveform mimicking GABAergic synaptic transmission at 50 Hz under control conditions. Bottom, a rebound burst induced by the control waveform. Ab, instantaneous frequency (Inst freq) of discharge associated with the trial in Aa. Dashed line = basal firing rate + 3 s.d.Ba, top, dynamic clamp waveform mimicking GABAergic synaptic transmission at 50 Hz in dopamine (10 μm). Bottom, a rebound burst induced by the modulated waveform. Bb, instantaneous frequency of discharge associated with the trial in Ba. Note that the maximum frequency of discharge is lower than for A. Ca and Da, expanded view of rebound bursts evoked by control (Ca) and modulated (Da) waveforms. Raster plots (Cb and Db) and peristimulus time histograms (Cc and Dc) for the control (Cb and c) and modulated (Db and c) waveforms. E, results for seven neurons (each represented by a distinct symbol). The maximum instantaneous rebound burst frequency (a), rebound burst duration (b) and number of action potentials (APs) per rebound burst (c) were significantly lower for the modulated waveform. Black horizontal bars represent mean values. *P < 0.05.