Frequency selectivity and dopamine-dependence of plasticity at glutamatergic synapses in the subthalamic nucleus

Abstract

In Parkinson's disease, subthalamic nucleus (STN) neurons burst fire with increased periodicity and synchrony. This may entail abnormal release of glutamate, the major source of which in STN is cortical afferents. Indeed, the cortico-subthalamic pathway is implicated in the emergence of excessive oscillations, which are reduced, as are symptoms, by dopamine-replacement therapy or deep brain stimulation (DBS) targeted to STN. Here we hypothesize that glutamatergic synapses in the STN may be differentially modulated by low-frequency stimulation (LFS) and high-frequency stimulation (HFS), the latter mimicking deep brain stimulation. Recordings of evoked and spontaneous excitatory post synaptic currents (EPSCs) were made from STN neurons in brain slices obtained from dopamine-intact and chronically dopamine-depleted adult rats. HFS had no significant effect on evoked (e) EPSC amplitude in dopamine-intact slices (104.4±8.0%) but depressed eEPSCs in dopamine-depleted slices (67.8±6.2%). Conversely, LFS potentiated eEPSCs in dopamine-intact slices (126.4±8.1%) but not in dopamine-depleted slices (106.7±10.0%). Analyses of paired-pulse ratio, coefficient of variation, and spontaneous EPSCs suggest that the depression and potentiation have a presynaptic locus of expression. These results indicate that the synaptic efficacy in dopamine-intact tissue is enhanced by LFS. Furthermore, the synaptic efficacy in dopamine-depleted tissue is depressed by HFS. Therefore the therapeutic effects of DBS in Parkinson's disease appear mediated, in part, by glutamatergic cortico-subthalamic synaptic depression and implicate dopamine-dependent increases in the weight of glutamate synapses, which would facilitate the transfer of pathological oscillations from the cortex.

Fig. 1. No effect of high-frequency stimulation on glutamatergic EPSCs in STN neurons in dopamine-intact slices.

(A) Normalized evoked EPSC amplitude (P1) 10 min before and 30 min after application of HFS (n=8 neurons). Each data point represents the mean (± SEM) of six evoked responses per neuron that were averaged across all neurons. Initiation of HFS (100 pulses at 100 Hz) is at time zero. (B) Paired evoked EPSCs (P1 and P2, average of 10 consecutive traces) separated by an interstimulus interval of 50 ms before (black) and 20 min after (red) HFS. (C) Paired pulse ratio (P2/P1) 10 min before and 30 min after application of HFS (n=8). (D) Shows the relative changes in normalized P1 EPSC amplitude plotted against the relative changes in CV-2 induced by HFS for each neuron recorded (white circles) and their average (black). (E) Raw traces showing spontaneous EPSCs before (black) and 20 min after (red) HFS (five consecutive sweeps). (F) Cumulative probability plots of interevent interval (IEI) before (black) and after (red) HFS (P>0.05).

Fig. 2. High-frequency stimulation depresses glutamatergic EPSCs in STN neurons in dopamine-depleted slices.

(A) Normalized eEPSC (P1) amplitude 10 min before and 30 min after application of HFS (n=6 neurons). (B) Paired evoked EPSCs (P1 and P2) before (black) and 20 min after (red) HFS. (C) Paired pulse ratio 10 min before and 30 min after application of HFS (n=6). (D) Shows the relative changes in normalized P1 EPSC amplitude plotted against the relative changes in CV⁻² induced by HFS for each recording (white circles) and their average (black). (E) Raw traces showing sEPSCs before (black) and 20 min after (red) HFS. (F) Cumulative probability plots of IEIs before (black) and after (red) HFS (P<0.005).

Fig. 3. Low-frequency stimulation potentiates glutamatergic EPSCs in STN neurons in dopamine-intact slices.

(A) Normalized evoked EPSC (P1) amplitude 10 min before and 30 min after application of LFS (n=7 neurons). Duration of LFS (10 pulses at 40 Hz delivered every second for 5 min) is indicated by blue bar. Each data point represents the mean (± SEM) of six evoked responses per neuron that were averaged across all neurons. (B) Typical paired EPSCs (P1 and P2, average of 10 consecutive traces in one neuron) evoked before (black) and 20 min after (red) LFS. (C) Paired pulse ratio (P2/P1) 10 min before and 30 min after application of LFS (n=7). (D) Shows the relative changes in normalized P1 EPSC amplitude plotted against the relative changes in CV⁻² induced by LFS for each neuron recorded (white) and their average (black). (E) Raw traces showing sEPSCs before (black) and 20 min after (red) LFS (five consecutive sweeps) (F) Cumulative probability plots of interevent interval (IEI) before (black) and after (red) LFS (P<0.001).

Fig. 4. No effect of low-frequency stimulation on glutamatergic EPSCs in STN neurons in dopamine-depleted slices.

(A) Normalized eEPSC (P1) amplitude 10 min before and 30 min after application of HFS (n=6). (B) Paired evoked EPSCs (P1 and P2) before (black) and 30 min after (red) LFS. (C) Paired pulse ratio (P2/P1) 10 min before and 30 min after application of LFS (n=6). (D) Shows the relative changes in normalized P1 EPSC amplitude plotted against the relative changes in CV⁻² induced by LFS. (E) Raw traces showing sEPSCs before (black) and 20 min after (red) HFS. (F) Cumulative probability plots of interevent interval (IEI) before (black) and after (red) LFS (P>0.05).

Fig. 5. Summary scheme of frequency selectivity and dopamine dependence of plasticity at cortico-subthalamic synapses.

This figure illustrates the potential increase in synaptic weight of cortico-subthalamic input promoted by the depletion of dopamine, the definitive pathology in Parkinson’s disease. High-frequency stimulation (HFS) of glutamatergic inputs to STN, which mimics deep brain stimulation (DBS), has no effect in dopamine-intact slices but elicits synaptic depression in dopamine-depleted slices. In contrast, low-frequency stimulation (LFS), which mimics pathological bursting, has no effect in dopamine-depleted but elicits synaptic potentiation in the dopamine-intact slices.