In vivo electrophysiology of nigral and thalamic neurons in alpha-synuclein-overexpressing mice highlights differences from toxin-based models of parkinsonism

Abstract

Numerous studies have suggested that alpha-synuclein plays a prominent role in both familial and idiopathic Parkinson's disease (PD). Mice in which human alpha-synuclein is overexpressed (ASO) display progressive motor deficits and many nonmotor features of PD. However, it is unclear what in vivo pathophysiological mechanisms drive these motor deficits. It is also unknown whether previously proposed pathophysiological features (i.e., increased beta oscillations, bursting, and synchronization) described in toxin-based, nigrostriatal dopamine-depletion models are also present in ASO mice. To address these issues, we first confirmed that 5- to 6-mo-old ASO mice have robust motor dysfunction, despite the absence of significant nigrostriatal dopamine degeneration. In the same animals, we then recorded simultaneous single units and local field potentials (LFPs) in the substantia nigra pars reticulata (SNpr), the main basal ganglia output nucleus, and one of its main thalamic targets, the ventromedial nucleus, as well as LFPs in the primary motor cortex in anesthetized ASO mice and their age-matched, wild-type littermates. Neural activity was examined during slow wave activity and desynchronized cortical states, as previously described in 6-hydroxydopamine-lesioned rats. In contrast to toxin-based models, we found a small decrease, rather than an increase, in beta oscillations in the desynchronized state. Similarly, synchronized burst firing of nigral neurons observed in toxin-based models was not observed in ASO mice. Instead, we found more subtle changes in pauses of SNpr firing compared with wild-type control mice. Our results suggest that the pathophysiology underlying motor dysfunction in ASO mice is distinctly different from striatal dopamine-depletion models of parkinsonism.

Keywords: Parkinson's disease; basal ganglia; mouse; recording; thalamus.

Fig. 1.

Motor deficits in human alpha-synuclein-overexpressing (ASO) mice. A–C: cylinder test. The number of steps taken (A), grooming episodes (B) and the number of rears (C) in 3 min were counted after introduction of the mouse into the cylinder. ASO mice were hypokinetic, as they made fewer steps (P < 0.01) and had fewer grooming episodes (P < 0.05) and fewer rears (P < 0.05) than their wild-type (WT) littermates. D and E: challenging beam traversal test. ASO mice took a longer time to traverse the beam (D, P < 0.01) and made more errors during traversal (E, P < 0.001) than WT mice. F–H: pole test. ASO mice took longer to turn around (F, P < 0.001) but, once oriented, descended the pole in a time equal to WT mice (G, P > 0.05). ASO mice fell off the pole more often than WT mice (H, P < 0.05). Mann-Whitney U-tests were performed for all statistical analysis. Means ± SE are shown. *Statistical differences between WT and ASO mice.

Fig. 2.

Alpha-synuclein and tyrosine hydroxylase (TH) staining of WT and ASO mouse brain tissue. A and B: there was a large increase in the overall alpha-synuclein (αSyn) staining in the brain of ASO (B1–B3) mice compared with WT (A1–A3) mice. Higher-magnification pictures of the labeling in substantia nigra pars reticulata (SNpr) of WT mice are shown in A2 and A3, while B2 and B3 illustrate higher-power views of SNpr labeling in ASO mice. Within the SNpr of ASO mice, but not WT mice, several alpha-synuclein-immunoreactive aggregates were observed (A3 vs. B3). C: similar levels of TH immunostaining were observed in the striatum (STR) of WT (C1) and ASO (C2) mice. MGN, medial geniculate nucleus; Hip, hippocampus; AC, anterior commissure. Scale bars: 1 mm (A1, B1, C1, C2), 0.5 mm (A2, B2), or 0.2 mm (A3, B3).

Fig. 3.

Lack of increased beta oscillations in local field potentials (LFPs) recorded in ASO mice. A: representative examples of LFPs recorded in primary motor cortex in WT (top) and ASO (bottom) mice during slow wave activity (SWA; left) and the desynchronized state (DS; right). B: mean power spectra (PSD) during SWA (left) and DS (right) were similar in both WT (black, n = 7) and ASO (red, n = 5) mice. Shaded areas represent 2 × SD for each group. No prominent beta was seen in either state. C: quantification of LFP results. The total power in the beta (15–30 Hz) range was not statistically different between groups during SWA (P > 0.05, WT SWA: 1.9 × 10⁻⁹ ± 3.1 × 10⁻¹⁰, ASO SWA: 9.6 × 10⁻¹⁰ ± 1.9 × 10⁻¹⁰, Mann-Whitney U-test); however, a significant decrease in beta was seen in ASO mice in the DS state (P < 0.05, WT DS: 2.4 × 10⁻⁹ ± 7.5 × 10⁻¹⁰, ASO DS: 8.6 × 10⁻¹⁰ ± 2.5 × 10⁻¹⁰, *P < 0.05, Mann-Whitney U-test).

Fig. 4.

SNpr neurons in WT and ASO mice. A: representative 1-s traces from histologically confirmed SNpr neurons from a WT (top) and an ASO (bottom) mouse. Both neurons were recorded under urethane-ketamine-xylazine (uKZ) anesthesia. The WT neuron fired at 20 Hz with a coefficient of variation (CV) of 0.36. The ASO neuron fired at 29.6 Hz with a CV of 0.57. There was no difference in firing rate across groups (P > 0.05). B: there was a significant main effect of anesthesia [uKZ vs. urethane only (U)] on CV (*P < 0.05) on SNpr neurons. C: burst rate was not significantly different across groups (P > 0.05). D: a significant effect of genotype (**P < 0.01) on pause rate was seen.

Fig. 5.

Phase preference of SNpr neuron firing to the cortical field potential in WT and ASO mice pooled for anesthesia. A: example phase raster for a SNpr neuron recorded under urethane-only anesthesia. A 6-s low-pass-filtered depth cortical field potential is shown in blue. To construct a phase raster, the instantaneous Hilbert phase was calculated (green, range ± π) from a low-pass-filtered LFP from primary motor cortex during SWA. UP states in the cortex are depth negative and occur at ±π. Each spike recorded from a SNpr neuron is shown in red and is plotted on the instantaneous Hilbert phase at that spike time. The distribution of instantaneous phases for this cell is shown in a circular plot in B. Since the mean resultant length (red) is small, a dashed line indicating the preferred direction is depicted. This neuron shows a weak but significant phase preference to fire during the cortical UP state (P < 0.05, Rayleigh test). A stronger, significant phase preference is seen for the onset of bursts (P < 0.05, Rayleigh test) C: second-order 1-sample circular plots showing the preferred directions for all WT (top) and ASO (bottom) SNpr neurons (all spikes included). Black and green circles represent neurons recorded in U anesthesia and uKZ anesthesia, respectively. The circle has a radius of 0.25. A larger vector length represents an increased phase preference for that angle. Unlike ASO neurons, WT SNpr neurons had a significant overall mean direction (P < 0.05; Moore's test; if one exists, a second-order grand mean is shown in red). D and E: second-order 1-sample circular plots showing the preferred direction (if any) for burst (D) and pause (E) onsets in WT (top) and ASO (bottom) SNpr neurons. All circles have a radius of 1.0. Burst onsets for both WT and ASO SNpr typically occurred during the cortical UP state (P < 0.05, Moore's test); however, pause onsets for both WT and ASO SNpr neurons did not show a significant phase preference (P > 0.05, Moore's test).

Fig. 6.

Periodic oscillations in spiking from an ASO SNpr neuron pair. A: representative cross-correlogram recorded from 2 simultaneously recorded SNpr neurons in an ASO mouse. The time to the first subpeak after 0 is ∼25 ms. Confidence intervals (calculated by 2 × SD of 50 time-shifted spike trains) are also shown. B: auto-correlograms of SNpr spiking for the 2 cells shown in A.

Fig. 7.

VM neurons in WT and ASO mice. A: representative 10-s trace from a WT VM neuron recorded under uKZ anesthesia. A second smaller unit can be seen in the background. The simultaneously recorded cortical LFP is shown above the trace. This VM neuron fired in phase with the cortex, firing either low-threshold spike (LTS)-type bursts or single spikes. Inset: an example LTS burst of 3 spikes (scale bar: 2 ms, 5 μV). B: circular plot constructed from LTS burst spikes for the cell shown in A. C: there was a significant effect of anesthesia on interburst rate (***P < 0.001, 2-way ANOVA) but no significant effect of genotype or anesthesia × genotype interaction (P > 0.05, 2-way ANOVA). D: phase preference for the onset of LTS-type bursts to the cortical field potential during SWA. Units are shown on a standard unit circle with a radius of 1. The numbers next to the group header indicate the number of cells with a significant phase preference out of the total group. The sample mean is shown with a gray bar for each group with a significant second-order phase preference (Moore's test).