Vocal changes in a zebra finch model of Parkinson's disease characterized by alpha-synuclein overexpression in the song-dedicated anterior forebrain pathway

Abstract

Deterioration in the quality of a person's voice and speech is an early marker of Parkinson's disease (PD). In humans, the neural circuit that supports vocal motor control consists of a cortico-basal ganglia-thalamo-cortico loop. The basal ganglia regions, striatum and globus pallidus, in this loop play a role in modulating the acoustic features of vocal behavior such as loudness, pitch, and articulatory rate. In PD, this area is implicated in pathogenesis. In animal models of PD, the accumulation of toxic aggregates containing the neuronal protein alpha-synuclein (αsyn) in the midbrain and striatum result in limb and vocal motor impairments. It has been challenging to study vocal impairments given the lack of well-defined cortico-basal ganglia circuitry for vocalization in rodent models. Furthermore, whether deterioration of voice quality early in PD is a direct result of αsyn-induced neuropathology is not yet known. Here, we take advantage of the well-characterized vocal circuits of the adult male zebra finch songbird to experimentally target a song-dedicated pathway, the anterior forebrain pathway, using an adeno-associated virus expressing the human wild-type αsyn gene, SNCA. We found that overexpression of αsyn in this pathway coincides with higher levels of insoluble, monomeric αsyn compared to control finches. Impairments in song production were also detected along with shorter and poorer quality syllables, which are the most basic unit of song. These vocal changes are similar to the vocal abnormalities observed in individuals with PD.

Fig 1. Comparison of simplified male zebra finch and human vocal motor neuroanatomy.

Human and male zebra finch vocal motor pathways contain both cortico-brainstem circuits (i.e., posterior pathway) involved in vocal production (solid lines) and cortico-basal ganglia-thalamo-cortico circuits (i.e., anterior forebrain pathway, AFP) involved in vocal modulation (dashed lines). Glutamatergic (arrowhead), GABAergic (square), and dopaminergic (circle) input play critical roles within these vocal-dedicated motor pathways. In male zebra finches, however, the vocal-dedicated basal ganglia (Area X, violet) contains multiple neuronal cell types in a single nucleus, whereas in the human basal ganglia these neurons are segregated across distinct regions (light violet). Moreover, similar to humans, the finch basal ganglia receive input from cortical nuclei within the AFP. The cortical output nucleus of this pathway, lateral magnocellular nucleus of the anterior nidopallium (lMAN), receives input from a thalamic region (red, DLM) and outputs onto another cortical region (RA, bottom blue nuclei). The darker colors correspond to the remaining vocally-dedicated areas within each species’ brain. These areas have been extensively reviewed [–35,43]. Cx—cortex. BG—basal ganglia. Th—thalamus. MB—midbrain. HB—hindbrain. Cb—cerebellum. OB—olfactory bulb.

Fig 2. Representative spectrogram of zebra finch song.

Song from an adult, male zebra finch is shown. Finches sing a characteristic song that is comprised of multiple motifs. Each blue box outlines a unique motif in that bird’s song. A red box outlines a repeated motif. Each motif is composed of a unique sequence of syllables. Four broad types of syllables include Harmonic (H), Noisy (N), Mixed (M), and Slide (S). Additionally, Flat Harmonic (H_f) syllables were separately analyzed from general harmonic syllables. Individual acoustic features can be measured from all of these syllables (Table 1). The y axis represents frequency (Hz) and the x axis represents time (ms). Two seconds of song is shown for presentation. The frequency window is between 0 and 13 kHz. The scale bar represents 100ms.

Fig 3. Experimental design for Parkinsonian model.

A) Experimental timeline for characterizing song changes. Two hours of undirected (male singing alone, 2HRUD) song was recorded for at least three days during each time point over a four-month window (0, 1, 2, and 3 months, black ticks). 0 month was treated as baseline (pre-injection time point). 1, 2, and 3-month time points are post-injection. Viral injection (red tick) occurred immediately after the last morning of baseline song collection (i.e., 0 month). The brain was collected immediately following the last day of the 3-month song collection period for further tissue processing (red arrow). Area X (outlined by hot pink dashes) was biopsied for molecular validation. B) Schematic of recombinant adeno-associated viral construct bilaterally injected into Area X. SNCA virus did not contain a GFP reporter. Both viruses were under the control of the broad promoter, chicken beta-actin (CBA), driving transfection of various cell types. Either rAAV5-CBA-SNCA (bottom) or rAAV5-CBA-eGFP (top) were therefore bilaterally injected into Area X to monitor effects on song. C) Representative confocal images of αsyn levels in the zebra finch basal ganglia and cortex relative to control. A single label of αsyn protein (red) with counterstaining using a fluorescent nissl stain (green) to outline cytoarchitecture was conducted on a 2HRUD singing ASYN bird (top). A single label of αsyn protein (red) was similarly conducted on a 2HRUD singing GFP bird (bottom). Representative images demonstrate elevated αsyn protein in the ASYN group relative to control GFP finches (top relative to bottom) and highlight successful viral-mediated expression of gfp (green) in Area X but not within lMAN (bottom). Area X is demarcated by a dashed line forming the characteristic teardrop structure, while lMAN is demarcated by dashed lines forming an ellipse structure dorsal and lateral to Area X (top panel). White, dashed lines extending from or around Area X outline striatal-nidopallidal border. Top part of the ASYN and GFP images is dorsal and bottom is ventral. ASYN images are of right hemisphere, in which the left-side of the image is lateral and the right-side is medial. In contrast, GFP images are of the left hemisphere, where the left-side is medial and the right-side is lateral. Scale bar represents 200 μm. Images were taken on a Zeiss 880 Inverted Confocal Microscope.

Fig 4. Asyn overexpression in zebra finch AFP.

A) Representative images of an αsyn (green signal) and PanNeuronal (purple signal) double label taken from an ASYN (top) or GFP (bottom) bird following 2HRUD singing highlight overexpression of αsyn within neuronal processes in Area X. B) Schematic representation highlighting Area X in a coronal slice of zebra finch brain. C) Representative images of an αsyn and PanNeuronal double label taken from an ASYN (top) or GFP (bottom) bird highlight overexpression of αsyn within neuronal cell bodies and processes in lMAN (white signal). D) Schematic representation highlighting lMAN in a coronal slice of zebra finch brain. Tissue was collected from a cohort of both ASYN or GFP birds collected at 3 mpi. Images were taken near center of target region on a Leica DMI 6000B wide field fluorescence microscope with a DFC 450 color camera at 40x magnification. Scale bar represents100μm.

Fig 5. Asyn expression levels by molecular weight in Area X.

A) A representative Western blot loaded with low salt (LS) or urea (U) soluble fractions obtained from Area X of birds injected with either AAV5-CBA-eGFP or AAV5-CBA-ASYN following 2HRUD singing is shown on the top. A blot with non-surgical (NS) birds is a supplementary figure (S4 Fig). Western blots were labelled with an αsyn antibody for quantification of protein levels in Area X relative to GAPDH from the LS lane of the same sample. B) A representative coronal slice of zebra finch brain highlighting punched Area X, the ventral striato-pallidum (VSP) ventral and medial to Area X, and nidopallium (NP). Tissue biopsies were taken of song-dedicated nucleus Area X and neighboring non-song related motor area VSP prior to thionin staining to confirm punch accuracy. Image from Miller et al. [56] C) Quantification of blots. For quantification, αsyn expression levels for each fraction are grouped by molecular weight within each group. Western blots indicate that higher levels of monomeric (15kD) urea-soluble αsyn are expressed in the ASYN group compared to the NS group and for the GFP group. Levels of trimeric (~45-50kD) αsyn protein are lower in ASYN and GFP groups compared to NS. Interestingly, total levels of urea soluble αsyn are also lower in ASYN compared to NS. Summary statistics provided in Table 2. Subscripts under experimental condition on representative blot denote different birds (i.e., sample). The representative blot contains raw data from birds 1 and 2 of both ASYN and GFP control groups (all raw blot images can be found in Supporting Information). Horizontal dashed lines indicate where blot was cut for incubating blot with either ASYN (~15kD, 40–250+kD) or GAPDH (~35kD) antibody. Names on left side of ladder indicate primary antibody used for immuno-labeling that section of the blot. For boxplots, vertical lines are 95% confidence intervals. Bottom end of the box is first quartile (25 percentile), middle horizontal line is second quartile (50 percentile), and top end of the box is the third quartile (75 percentile). The circle points represent the mean value for each group. Statistical comparisons were made using a Welch’s test. * indicate p < 0.05. # indicates 0.05 < p < 0.1.

Fig 6. Song production affected by αsyn overexpression in the AFP.

A) Total amount sung affected by αsyn overexpression. ASYN birds sing less at 2 mpi compared to pre-injection. No within-group differences were detected across time points for GFP birds (N = 6). For boxplots, vertical lines are 95% confidence intervals. Bottom end of the box is first quartile (25 percentile), middle horizontal line is second quartile (50 percentile), and top end of the box is the third quartile (75 percentile). The circle points represent the mean value for each group. B) Singing during the first 30 minutes of a two-hour singing session is affected by αsyn overexpression. Amount sung during the first 30 and last 30 minutes of the two hour singing period during 0, 1, 2, and 3 months-post injection plotted for both ASYN and GFP groups. GFP birds tend to sing less during the last 30 minutes (Motif.last30) compared to the first 30 minutes (Motif.first30) at the pre-injection time point (0 mpi). The effect persists in the GFP birds during one and three mpi time point while they sang less during the last 30 minutes at 2 mpi compared to first 30, but the effect is not significant. ASYN birds similarly sing significantly less during the last 30 minutes compared to first 30 minutes at the pre-injection time point (0 mpi) but this ends at one mpi. Summary statistics provided in Tables 3 and 4. The raw values are plotted. For boxplots, horizontal lines are 95% confidence intervals. Left end of the box is first quartile (25 percentile), middle vertical line is second quartile (50 percentile, median), and right end of the box is the third quartile (75 percentile). The circle points represent the mean value for each group. Statistical comparisons were made using a Welch’s test. * indicates p < 0.05. # indicates 0.05 < p < 0.1.

Fig 7. Asyn overexpression differentially affects select acoustic features depending on syllable type.

The adjusted mean value of select individual acoustic features is plotted for all syllables (top) sung by αsyn and gfp expressing groups (top). Overall, the ASYN group’s syllables were shorter (syllable.duration) at two and three mpi compared to the GFP control group (top left). A trend was detected at one mpi between the ASYN group and GFP control group. Mean amplitude modulation (mean.AM.2) was higher in the ASYN group compared to GFP control group at 3 mpi (top right). The adjusted mean value of select individual acoustic features is plotted for Harmonic (middle left), Slide (middle right), and Mixed (bottom) syllables sung by asyn and gfp expressing groups. The amplitude modulation (mean.AM.2) of harmonic syllables sung by the ASYN group was higher than the GFP control group at 3 mpi. The pitch goodness of slide syllables was lower in the ASYN group compared to GFP control at 2 mpi. The duration of mixed syllables was shorter in the ASYN group compared to GFP control at 1 and 2 mpi with a statistical trend at 3 mpi (bottom left). Additionally, the amplitude modulation of mixed syllables was higher for the ASYN group compared to GFP control at 3 mpi (bottom right). Lastly, for noisy syllables, no differences were found in the ASYN group (N = 7) compared to GFP control group (N = 7, S7 Fig). Remaining acoustic features that were not affected by αsyn overexpression are reported in S6 and S7. The adjusted values are plotted for between group comparisons. Summary statistics provided in Table 5. For boxplots, vertical lines are 95% confidence intervals. Bottom end of the box is first quartile (25 percentile), middle horizontal line is second quartile (50 percentile), and top end of the box is the third quartile (75 percentile). The circle points represent the mean value for each group. Purple boxplots represent data from ASYN birds, whereas green boxplots represent data from GFP birds. Statistical comparisons were made using a Wilcoxon Rank Sum Test. * indicates p < 0.05. # indicates 0.05 < p < 0.1.

Fig 8. Asyn overexpression differentially affects across-rendition variability in select acoustic features depending on syllable type.

The adjusted value of across rendition variation (i.e., CV, rendition-to-rendition variability) in select acoustic features is plotted for all individual syllables sung by ASYN (purple) and GFP (green) groups. Overall, the CV of amplitude (mean.amplitude.adjusted) for syllables sung by the ASYN group was lower than the GFP control group at 2 mpi (top left). Additionally, the CV of pitch goodness (mean.pitch.goodness) was higher in the ASYN group than in the GFP group at 3 mpi (top middle). The adjusted value of across rendition variation in select acoustic features is also plotted for Harmonic (middle left), Noisy (middle center), Mixed (middle right), and Slide (bottom) syllables. For harmonic syllables, the CV of entropy (mean.entropy) was significantly higher in the ASYN group compared to GFP control at 1, 2, and 3 mpi. For noisy syllables, the CV of amplitude (mean.amplitude.adjusted) was lower in the ASYN group than in the GFP control at 2 mpi. For mixed syllables, the CV of pitch goodness (mean.pitch.goodness) was higher in the ASYN group compared to the GFP control at 3 mpi For slide syllables, the CV of amplitude (mean.amplitude.adjusted) was lower in the ASYN group compared to GFP control at 2 and 3 mpi (bottom left). The CV of pitch (mean.pitch) was higher for the slide syllables in the ASYN group compared to GFP control at 3 mpi with a trend observed at 1 mpi (bottom middle). The CV of frequency modulation (mean.FM) for slide syllables was lower in the ASYN group compared to GFP control at 1 mpi (bottom right). Remaining acoustic features whose CV scores were not affected by αsyn overexpression are reported in S9-11. Summary statistics provided in Table 6. Reference Fig 7‘s legend for explanation of boxplots. Statistical comparisons were made using a Wilcoxon Rank Sum Test. * indicates p < 0.05. # indicates 0.05 < p < 0.1.

Fig 9. Asyn overexpression differentially affects within-rendition variability in select acoustic features depending on syllable type.

The adjusted value of within rendition variation (i.e., var) in acoustic features is plotted for all individual syllables sung by ASYN (purple) and GFP (green) expressing groups (top left). Only the variance in amplitude modulation (var.AM) was statistically greater in the ASYN group compared to GFP control at 3 mpi. The adjusted value of within rendition variation (i.e., var) in acoustic features is plotted for Harmonic (top right), Mixed (bottom left), and Slide (bottom right) syllables. For harmonic syllables, the variance of amplitude modulation (var.AM) was greater in the ASYN group compared to GFP control at 3 mpi. For mixed syllables, the variance of amplitude modulation was also greater in the ASYN group compared to GFP control at 3 mpi with a trend observed at 2 mpi. For slide syllables, the variance of frequency (var.mean.freq) was greater in the ASYN group than in the GFP control at 1 mpi with a trend detected at 3 mpi. Lastly, for noisy syllables, no differences were found in the ASYN group (N = 7) compared to GFP control group (N = 7, S12). Remaining acoustic features whose variance scores were not affected by αsyn overexpression are reported in S12. Summary statistics provided in Table 7. Reference Fig 7‘s legend for explanation of boxplots. Statistical comparisons were made using a Wilcoxon Rank Sum Test. * indicates p < 0.05. # indicates 0.05 < p < 0.1.