Loss of Hyperdirect Pathway Cortico-Subthalamic Inputs Following Degeneration of Midbrain Dopamine Neurons

Abstract

The motor symptoms of Parkinson's disease (PD) are linked to abnormally correlated and coherent activity in the cortex and subthalamic nucleus (STN). However, in parkinsonian mice we found that cortico-STN transmission strength had diminished by 50%-75% through loss of axo-dendritic and axo-spinous synapses, was incapable of long-term potentiation, and less effectively patterned STN activity. Optogenetic, chemogenetic, genetic, and pharmacological interrogation suggested that downregulation of cortico-STN transmission in PD mice was triggered by increased striato-pallidal transmission, leading to disinhibition of the STN and increased activation of STN NMDA receptors. Knockdown of STN NMDA receptors, which also suppresses proliferation of GABAergic pallido-STN inputs in PD mice, reduced loss of cortico-STN transmission and patterning and improved motor function. Together, the data suggest that loss of dopamine triggers a maladaptive shift in the balance of synaptic excitation and inhibition in the STN, which contributes to parkinsonian activity and motor dysfunction.

Keywords: NMDA; Parkinson’s disease; basal ganglia; cortex; globus pallidus; glutamate; plasticity; subthalamic nucleus; synapse.

Figure 1. The strength of cortico-STN transmission decreased following loss of dopamine

(A, B) Viral expression of ChR2(H134R)-eYFP (green) in infected cortical neurons at the site of injection in primary motor cortex (M1; A) and at the site of their axon terminal fields in the STN (B). Cortical (A) and STN (B) neurons were immunohistochemically labeled for neuronal nuclear protein (NeuN; red). (C, D) The amplitude of cortico-STN transmission was lower across a range of optogenetic stimulation intensities in 6-OHDA-injected, dopamine-depleted mice relative to vehicle-injected, dopamine-intact mice (C, representative examples; D, population data). (E, F) The amplitude of optogenetically stimulated cortico-STN EPSC2 relative to EPSC1 was similar in 6-OHDA- and vehicle-injected mice across a range of stimulation intensities (E, representative examples; F, population data). (G–I) The frequency (G, H) and amplitude (G, I) of asynchronous cortico-STN transmission optogenetically evoked in the presence of Sr²⁺ was reduced in 6-OHDA- relative to vehicle-injected mice (G, representative examples; H, I, population data). Blue arrow, optogenetic stimulation; *, p < 0.05.

Figure 2. Cortico-STN axon terminals and their post-synaptic targets were eliminated following loss of dopamine

(A-C). The density of vGluT1-immunoreactive cortico-STN axon terminals (red puncta) was lower in 6-OHDA-injected mice compared to vehicle controls (A, B, representative confocal micrographs; C, population data). (D–K) The dendritic morphology of Golgi-impregnated STN neurons was altered in 6-OHDA-treated mice compared to vehicle controls. (D–G) Representative Golgi-impregnated neurons, illustrating loss of dendrites (D–F) and dendritic spines (G; red arrowheads) in 6-OHDA-treated mice. (F) Sholl plots of the neurons in D and E. Concentric shells, which increase in diameter in 10 μm increments, are centered on the soma of each representative neuron. (H) Population data. The number of dendritic intersections at distances greater than 30 μm from the soma was lower in neurons from 6-OHDA-treated mice relative to neurons from vehicle- treated mice. (I–K) Population data for dendritic length (I) and dendritic spines (J, K) in Golgi-impregnated neurons from dopamine-depleted and -intact mice. *, p < 0.05.

Figure 3. The post-synaptic receptor and long-term plasticity properties of cortico-STN transmission were modified following loss of dopamine

(A, B) The sensitivity of optogenetically stimulated cortico-STN transmission evoked at −80 mV to the Ca²⁺ permeable (GluA2-deficient) AMPAR antagonist Naspm is greater in 6-OHDA- relative to vehicle-treated mice (A, representative examples, B, population data). (C–E) Optogenetically stimulated and pharmacologically isolated AMPAR-mediated cortico-STN EPSCs exhibited relatively rapid decay kinetics (C, D) and rectification (C, E) at depolarized voltages in 6-OHDA-injected mice (C, representative examples; D, E, population data). (F–H) Application of an optogenetic stimulation protocol (1 ms light pulses delivered at 50 Hz for 300 ms, repeated 30 times at 0.2 Hz), which induced LTP of cortico-STN transmission in vehicle-treated mice (F), failed to induce LTP in 6-OHDA-treated mice (G). Furthermore, application of 0.5–1.0 μM dopamine failed to rescue LTP in 6-OHDA-treated mice (H). (F–H) Left panel, normalized transmission before and following the induction protocol; right panel, representative cortico-STN EPSCs at time points 1 and 2, prior to and following the induction protocol, respectively. Blue arrow, optogenetic stimulation; *, p < 0.05. See also Figure S1.

Figure 4. The strength of cortico-STN patterning decreased following loss of dopamine

(A–F) Cortico-STN inputs optogenetically stimulated (blue) at 15 Hz less effectively patterned STN neuronal activity in ex vivo brain slices derived from 6-OHDA-injected mice versus vehicle-injected mice. (A, B) Representative examples and (C, D) population peristimulus time histograms derived from vehicle- (A, C) and 6-OHDA-injected (B, D) mice. Cyan dashed line, mean of action potentials preceding optogenetic stimulation. Red dashed line, mean plus 3 standard deviations of action potentials preceding optogenetic stimulation. Blue rectangle, period of optogenetic stimulation. (E) The frequency and regularity (CV) of spontaneous firing were reduced in 6-OHDA-injected mice relative to vehicle-injected controls. (F) The number of cortically driven action potentials (synaptic drive) was lower and the latency of action potentials following each optogenetic stimulus greater in 6-OHDA mice relative to vehicle-injected controls. (G–J) Application of the AMPAR antagonist GYKI 53655 but not the NMDAR antagonist D-APV reduced the effectiveness of optogenetically stimulated cortico-STN patterning both in terms of the number (G-I) and latency (J) of synaptically generated action potentials in both vehicle- (G, I, J) and 6-OHDA-injected mice (H–J) (G, H, representative examples; I, J, population data). *, p < 0.05. ns, not significant.

Figure 5. The strength of cortico-STN transmission decreased following chemogenetic activation of D2-SPNs in dopamine-intact mice

(A, B) Expression of rM3Ds-mCherry in D2-SPNs (A, red arrows) and their axon terminal fields in the GPe (B) in the adora2A-rM3Ds-mCherry mouse. Expression was absent in putative D1-SPNs (A; white arrows) and their axon terminal fields in the SNr (C). (B, C) Expression was also absent in GPe (B; white arrows) and SNr (C; white arrows) neurons. Immunohistochemistry for NeuN (white) was used as a neuronal marker in A–C. (D–F) Chemogenetic activation of rM3Ds in D2-SPNs through subcutaneous (SC) injection of CNO (1 mg/kg) led to inhibition of open field motor activity relative to vehicle injection. (D) Representative examples of open field activity before and after first injection. (E, F) Population data confirming that CNO injection reduced movement traveled in the open field (E, left and right box plots for vehicle and CNO represent movement prior to and following first injection, respectively; F, movement following injection of vehicle or CNO over 3 consecutive days). (G) Simultaneous recordings of the electroencephalogram (EEG) band pass filtered at 0.5–1.5 Hz and 10–100 Hz, and GPe unit activity in a urethane-anesthetized adora2A-rM3Ds-mCherry mouse prior to (control), and 30–45 mins following the SC administration of CNO (1 mg/kg). The rate of GPe unit activity during periods of robust cortical slow-wave activity decreased following the injection of CNO both in each example neuron (G) and across the population sample (H). (I–K) the frequency (I, J) (but not the amplitude; I, K) of mIPSCs in GPe neurons was greater in brain slices treated with CNO (10 μM) ex vivo versus untreated control slices (I, representative examples; J, K, population data). (L–P) Subcutaneous injection of CNO every 12 hours for 2–3 days led to a reduction in the density of vGluT1 expressing cortico-STN axon terminals (L–N) and to a reduction in the amplitude of optogenetically stimulated (blue arrow) cortico-STN transmission (O, P) relative to vehicle-injected control mice (L, M, representative micrographs; N, population data; O, representative traces; P, population data). Blue arrow, optogenetic stimulation; *, p < 0.05. ns, not significant. See also Figure S2.

Figure 6. Knockdown of STN NMDARs prevented loss of cortico-STN transmission and patterning in dopamine-depleted mice

(A–B) Viral expression of eGFP (A) and cre-eGFP (B) centered on the STN (ic, internal capsule; parasagittal plane). (C, D, G) The density of vGluT1-immunoreactive axon terminals in the eGFP expressing, NMDAR-intact STN was lower in 6-OHDA-injected (D, G) versus vehicle-injected Grin1^lox/lox mice (C, G). (C–G) The densities of vGluT1-immunoreactive axon terminals in the cre-eGFP expressing, NMDAR knockdown STN (E–G) were elevated relative to the eGFP expressing, NMDAR-intact STN (C, D, G) in both dopamine-depleted (D, F, G) and dopamine-intact (C, E, G) Grin1^lox/lox mice. (A–F) Representative confocal micrographs. (G) Population data. (H–J) The amplitude and frequency of optogenetically evoked (blue arrow), asynchronous, miniature cortico-STN transmission in eGFP expressing, NMDAR-intact STN neurons were lower in 6-OHDA-injected versus vehicle-injected Grin1^lox/lox mice. (H, I) The frequency of optogenetically evoked asynchronous cortico-STN transmission in cre-eGFP expressing, NMDAR knockdown STN neurons was significantly elevated in both vehicle- and 6-OHDA-injected Grin1^lox/lox mice relative to their eGFP expressing, NMDAR-intact counterparts. (H, J) The amplitude of asynchronous cortico-STN EPSCs in STN neurons in cre-eGFP expressing, NMDAR-knockdown neurons in 6-OHDA-injected Grin1^lox/lox mice was lower than in eGFP expressing, NMDAR-intact neurons from vehicle-injected Grin1^lox/lox mice. (H) Representative traces. (I–J) Population data. (K–M) In eGFP-expressing NMDAR-intact neurons cortico-STN inputs optogenetically stimulated at 15 Hz less effectively patterned neuronal activity in slices from 6-OHDA- versus vehicle-injected Grin1^lox/lox mice. In slices from 6-OHDA-injected Grin1^lox/lox mice loss of cortico-STN patterning was largely prevented in cre-eGFP-expressing neurons with knockdown of STN NMDARs. (K) Population peristimulus time histograms. Cyan dashed line, mean of action potentials preceding optogenetic stimulation. Magenta dashed line, mean plus 3 standard deviations of action potentials preceding optogenetic stimulation. Blue rectangle, period of optogenetic stimulation. (L, M) Population data for synaptic drive and latency of action potentials in response to optogenetic stimulation. *, p < 0.05. ns, not significant. See also Tables S1–S3.

Figure 7. Activation of STN NMDARs ex vivo downregulates cortico-STN transmission in brain slices from dopamine-intact but not dopamine-depleted mice

(A–F) Activation of STN NMDARs ex vivo led to a significant reduction in the density of vGluT1-immunoreactive cortico-STN axon terminals in brain slices derived from vehicle (A–C; A, B, examples; C, population data) but not 6-OHDA-treated (D–F; D, E, examples; F, population data) mice. (G–I) Activation of STN NMDARs ex vivo led to a significant reduction in the amplitude of optogenetically stimulated cortico-STN transmission in brain slices derived from vehicle (G–I; G, H, examples; I, population data) but not 6-OHDA-treated (J–L; J, K, examples; L, population data) mice. *, p < 0.05. ns, not significant. See also Figure S3.

Figure 8. Knockdown of STN NMDARs reduced motor dysfunction in dopamine-depleted mice but had no discernable impact on motor function in dopamine-intact mice

(A–D) Ipsilateral forelimb use (A), ipsiversive rotational behavior (B) and distance traveled in the open field (C, D) were compared in vehicle-injected dopamine-intact and 6-OHDA-injected dopamine-depleted Grin1^lox/lox mice. eGFP-expressing NMDAR-intact, 6-OHDA-injected dopamine-depleted mice exhibited significantly increased ipsilateral forelimb use (A), ipsiversive rotational behavior (B) and significantly reduced travel in the open field (C, D) compared to eGFP-expressing NMDAR-intact, vehicle-injected dopamine-intact mice. cre-eGFP-expressing NMDAR-knockdown, 6-OHDA-injected dopamine-depleted mice exhibited reduced ipsilateral forelimb use (A) and increased distance traveled in the open field (C, D) relative to eGFP-expressing STN NMDAR-intact, 6-OHDA-injected dopamine-depleted mice. (A–D) The motor behavior of eGFP-expressing STN NMDAR-intact, vehicle-injected dopamine-intact mice and cre-eGFP-expressing STN NMDAR-knockdown, vehicle-injected dopamine-intact mice were not significantly different. (A–C) Population data. (D) Examples of open field activity. *, p < 0.05. ns, not significant. See also Figures S4 and S5 and Tables S1–3.