Systemic delivery of AAV-PHP.eB results in a global-scale expression of hα-syn in the brain

To induce a large-scale hα-syn overexpression in mice brains, we delivered AAV- PHP.eB viral particles overexpressing hα-syn fused to a myc tag to the blood circulation, via intravenous retro-orbital injection, to 3-month-old C57Bl/6 mice. The addition of the myc tag helps with enhancing the distinction between the endogenous and overexpressed α-syn, with minimal impact on the protein’s structure and behavior [35, 36]. As control groups, we injected animals with viral particles overexpressing the fluorescent protein mTurquoise2 (mTurq) or injected with PBS (control). First, we performed post-mortem tissue analysis, 2-weeks post-injection, to evaluate the transduction efficiency and to assess transgene expression levels in different brain regions. Using anti-hα-syn antibody (LB509), we observed that exogenous hα-syn signal was widely detected in all brain regions, with a marked staining in the neuronal soma and neurites in the cortex, the striatum, the hippocampus, the thalamus, the midbrain, and the hypothalamus, as well as the cerebellum (Fig. 1A). Interestingly, we also observed an abundant hα-syn punctate staining, consistent with a presynaptic localization of α-syn in different brain regions [13] (Fig. 1A). Of note, the absence of signal using LB509 antibody in the control group confirmed the specificity of α-syn staining in hα-syn-injected group (Suppl. Figure 1A). Moreover, mTurq signal was also detected in the soma and cell neurites in all brain regions, as previously reported [22] (Fig. 1B). It is important to mention that, although the systemic injection of AAV particles was unilateral, the expression of the two transgenes was detected bilaterally with a similar intensity in the two hemispheres, thus representing an important advantage for the use of the present delivery approach (Suppl. Figure 1B-D).

Fig. 1

AAV-PHP.eB hα-syn transduces several cerebral regions and induces widespread expression of the transgenes in the brain 2-weeks post-AAV delivery. (A-B) Representative confocal microscopy images of sagittal brain slices of mice overexpressing (A) hα-syn or (B) mTurq (scale bar = 1 mm). Numbered boxes represent zoomed images of the different brain regions: [1] cortex, [2] striatum, [3] hippocampus, [4] thalamus, [5] midbrain, [6] hypothalamus, and [7] cerebellum (scale bar = 10 μm). (C-F) Western blot analysis of hα-syn, mα-syn and mTurq protein levels in the soluble protein fraction from the (C) substantia nigra, (D) striatum, (E) motor cortex, and (F) hippocampus. GAPDH was used as a loading control. (G-H) Quantification of the levels of (G) hα-syn and (H) mTurq in different brain regions normalized to GAPDH. (I) Ratio of hα-syn protein levels compared to endogenous mα-syn. The data are expressed as the means ± s.e.m. (n = 4–5 mice per experimental condition). SN: substantia nigra, Str: striatum, Ctx: cortex, Hip: hippocampus. Full Western blots are shown in Suppl. Figure 8

It is important to highlight that despite the well-known strong neuronal tropism of AAV-PHP.eB [22, 37, 38], there was notable variability in the quantification of NeuN+ cells expressing the transgenes, depending on the specific brain region examined. While certain regions showed a lower to moderate percentage of transgene-expressing neurons (e.g., 15% in the striatum, 17% in the hippocampus, and 26% in the cortex), the SN displayed a significantly higher rate of neuronal transduction (48%) at 2-weeks post-injection (Suppl. Figure 2 A). As previously reported [22], we also observed a minor expression of the transgenes in other brain cell types, including astrocytes and microglia (Suppl. Figure 2B and C).

To further assess the transgene expression levels in different brain regions, we extracted the soluble protein fraction from the substantia nigra (SN), the striatum, the cortex and the hippocampus and performed Western blot analysis. Using an anti-α-syn antibody (Syn1/BD Lab), we detected a band at 17 kDa corresponding to the exogenous hα-syn fused to the myc tag (Fig. 1C-F). Analysis of this band intensity revealed a disparity in the transgene expression levels between animals in each condition, probably due to the intravenous retro-orbital injection efficiency (Fig. 1G and H). Moreover, the expression levels of hα-syn and mTurq were also different between the brain regions, (Fig. 1G and H). As anti-α-syn antibody (Syn1/BD Lab) also recognizes endogenous mouse α-syn (mα-syn), we were able to compare exogenous and endogenous α-syn levels, and quantification estimated that the hα-syn levels is representing between 3 and 10% of the mα-syn (Fig. 1I). Finally, we evaluated the stability of the transgenes expression levels and observed a consistent hα-syn and mTurq signals in the different brain regions, 3-months post-injection (Suppl. Figure 3 A and B). Interestingly, the levels of hα-syn, in comparison to its endogenous counterpart, exhibited a progressive increase over time across all brain regions, ultimately reaching approximately 20–30% of the endogenous α-syn level at 3-months post-injection (Suppl. Figure 3 C-I). Collectively, these results demonstrate that systemic delivery of AAV-PHP.eB particles resulted in discrete, but widespread, stable, and bilateral hα-syn overexpression in the mouse brain.

Systemic delivery of hα-syn induced progressive motor impairment

To investigate if the large-scale hα-syn overexpression in mouse brains was associated with the manifestation of behavioral phenotype, we performed a battery of tests to evaluate the animals’ motor performance. Three months post-viral delivery, we analyzed animals’ motor coordination, endurance, and balance using the rotarod test. Results revealed that mice overexpressing hα-syn exhibited a significant short latency to fall, compared to the control and mTurq groups [F [2, 22] = 17.07, p < 0.0001; hα-syn vs. control: p < 0.0001; hα-syn vs. mTurq: p = 0.0079], suggesting that hα-syn overexpression affected animals’ motor coordination (Fig. 2A). Of note, the rotarod test analysis overtime revealed that hα-syn overexpression affected animal performance as early as 1-month post-viral delivery and this deleterious effect worsened with time and reached a plateau at 2 months [effect of time: F [3, 66] = 7.711, p = 0.0002; 1 month vs. baseline: p = 0.0299; 2 months vs. baseline: p < 0.0001; 3 months vs. baseline: p < 0.0001] (Fig. 2B). At 3-months post-injection, the mTurq group exhibited a moderated decline in the animals’ performance in the rotarod test, compared to the control group [mTurq vs. control: p = 0.0201] (Fig. 2A). However, this impairment did not reach statistical significance when compared to the animals’ baseline performance [1 month vs. baseline: p = 0.9661; 2 months vs. baseline: p = 0.1493; 3 months vs. baseline: p = 0.1140] (Fig. 2B). This observation suggests that mTurq overexpression had no significant impact on the animals’ overall performance, and the observed effect at 3-months is likely attributed to fluctuations in the performance of the control group. Moreover, analysis of movement coordination, using the gait test, revealed significant deficits in mice overexpressing hα-syn at 3-months post-viral delivery, compared to the control group [F [2, 22] = 12.62, p = 0.0002; hα-syn vs. control: p = 0.0014], whereas no effect was observed in the mTurq group [mTurq vs. control: p = 0.9855] (Fig. 2C). This gait abnormality appears progressively and reaches significant levels after 3-months post-viral delivery [effect of time: F [2, 44] = 2.834, p = 0.0696; 2 months vs. 1 month: p = 0.4724; 3 months vs. 1 month: p = 0.0132] (Fig. 2D). Furthermore, hα-syn overexpression induced impairment of mice locomotor activity at 3-months post-viral delivery, as evaluated by the open field test, compared to the control and mTurq groups [F [2, 21] = 9.845, p = 0.0010; hα-syn vs. control: p = 0.0016; hα-syn vs. mTurq: p = 0.0049] (Fig. 2E). This deficit appears very early after hα-syn overexpression (1-monthpost-viral delivery) [effect of group: F [2, 21] = 7.908, p = 0.0028; hα-syn vs. control: p = 0.0465; hα-syn vs. mTurq: p = 0.0355] and progresses overtime [effect of time: F [3, 63] = 0.3033; p = 0.8229; hα-syn, 1 month vs. baseline: p = 0.1929; 2 months vs. baseline: p = 0.1315; 3 months vs. baseline: p = 0.0193] (Fig. 2F). Finally, no effect was observed after analysis of the cylinder test [F [2, 22] = 0.7333, p = 0.4917] and the grip force test F [2, 22] = 1.620, p = 0.2207] (Suppl. Figure 4 A and B).

Fig. 2

Overexpression of hα-syn induces motor and cognitive impairment. (A) Evaluation of motor impairment using the rotarod test 3-months post-AAV delivery. (B) Progression of animals’ performances on the rotarod test, evaluated at 1-month, 2-months, and 3-months post-AAV delivery. (C) Evaluation of the animals’ motor coordination using the gait test 3-months post-AAV delivery. (D) Progression of animals’ deficits on the gait test, evaluated at 1-month, 2-months, and 3-months post-AAV-delivery. (E) Evaluation of animals’ performance in the open field at 3- months post-AAV-delivery. (F) Progression of animals’ performances in the open field, evaluated at 1-month, 2-months, and 3-months post-AAV delivery. (G) Evaluation of animals’ performance in the training phase of the MWM test at 3-months post-AAV-delivery. (H) Evaluation of animals’ performance in the probe trial of the MWM test at 3-months post-AAV-delivery. The data are presented as the means ± s.e.m. (n = 6–10 mice per experimental condition). One-way ANOVA followed by Tukey’s multiple comparisons test; (A, C, E & H) * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.01 and **** p ≤ 0.0001. Two-way repeated measures ANOVA followed by Tukey’s multiple comparisons test; (B, D & F) $ p ≤ 0.05, $$ p ≤ 0.01, $$$ p ≤ 0.001 and $$$$ p ≤ 0.0001 versus control group; £ p ≤ 0.05 and ££ p ≤ 0.01 versus mTurq group; # p ≤ 0.05 versus baseline for each group; (G) * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.01 and **** p ≤ 0.0001. MWM: Morris water maze

Given the fact that non-motor symptoms are also a major characteristic of PD, we investigated the impact of hα-syn overexpression on mice cognitive performance, 3- months post-viral delivery. Our data revealed deficits in learning abilities in mice overexpressing hα-syn, compared to the control or mTurq groups during the training of the Morris water maze (MWM) test (Fig. 2G). This learning deficit was significant with the probe trial [F [2, 23] = 7.468, p = 0.0032; hα-syn vs. control: p = 0.0028; hα-syn vs. mTurq: p = 0.0342] (Fig. 2H). Finally, no effect of hα-syn overexpression was observed on the animals’ spatial working memory and anxiety as assessed using the Y-maze [F [2, 24] = 1.338, p = 0.2813] and elevated-plus maze (EPM) [F [2, 26] = 0.3687, p = 0.6952] tests, respectively (Suppl. Figure 4 C and D). Collectively, our results demonstrate that widespread hα-syn overexpression in the brain was associated with early onset and progressive motor impairment and learning deficits.

Widespread hα-syn overexpression in the mouse brain induced selective dopaminergic neuronal loss in the midbrain

We next evaluated if the wide-scale overexpression of hα-syn could induce neuronal degeneration in the brain. As exogenous transgenes were equally expressed in both brain hemispheres (Suppl. Figure 1B-D), we post-fixed one hemisphere in paraformaldehyde (PFA) for the immunohistochemistry and the other hemisphere was used for the collection of soluble and insoluble protein fractions for biochemical analysis.

First, we evaluated the neuronal density (NeuN+ cells/mm3) in different brain regions, 3-months post-viral delivery. Analysis showed that neither hα-syn nor mTurq overexpression affected the total number of neuronal cells in the striatum [F [2, 20] = 0.1798, p = 0.8368], the cortex [F [2, 21] = 0.4501, p = 0.6436],the hippocampus [F [2, 22] = 2.313, p = 0.1225] or the SN [F [2, 21] = 3.393, p = 0.0529] (Fig. 3A-D). Then, we focused our analysis on the dopaminergic neurons, the principal neuronal population of the SN affected in PD [39,40,41,42,43]. Using a stereological approach for the quantification of the total number of TH+ neurons in this region, we observed a significant reduction in the number of dopaminergic neurons in the hα-syn group, when compared to the control or mTurq groups [F [2, 20] = 5.447; p = 0.0129; hα-syn vs. control: p = 0.0477; hα-syn vs. mTurq: p = 0.0172] (Fig. 3E and F). Moreover, the reduction of the number of Nissl+ cells in the SN in the hα-syn group confirmed that TH+ neuronal loss was indeed attributed to neurodegeneration rather than a loss of the dopaminergic phenotype [F [2, 20] = 4.805, p = 0.0198; hα-syn vs. control: p = 0.0420; hα-syn vs. mTurq: p = 0.0363] (Fig. 3G). Interestingly, stereological quantification of the TH+ neurons in the VTA showed that the dopaminergic neurons were preserved in this region, despite hα-syn overexpression [F [2, 20] = 0.1125, p = 0.8942; hα-syn vs. control: p = 0.8846; hα-syn vs. mTurq: p = 0.9672] (Fig. 3H), thus confirming the selective vulnerability of the dopaminergic neurons of the SN and the resistance of VTA dopaminergic neurons to α-syn-induced toxicity [44,45,46]. Finally, analysis of the integrity of the nigrostriatal pathway showed a significant decrease in the dopaminergic fiber density in the striatum in the hα-syn overexpression group, compared to the control and mTurq groups [F [2, 21] = 13.63, p = 0.0002; hα-syn vs. control: p = 0.0006; hα-syn vs. mTurq: p = 0.0005], confirming the degeneration of the dopaminergic nigral terminals associated with the neuronal loss in the SN (Fig. 3I and J). Of note, degeneration of the nigrostriatal pathway was associated to a synaptic dysfunction specifically in the striatum, as reflected by the significant decrease of synaptic markers (drebrin [F [2, 11] = 6.354, p = 0.0146; hα-syn vs. control: p = 0.0338; hα-syn vs. mTurq: p = 0.0223], PSD95 [F [2, 11] = 4.023, p = 0.0488; hα-syn vs. control: p = 0.0489; hα-syn vs. mTurq: p = 0.1567] and synaptophysin [F [2, 11] = 10.22, p = 0.0031; hα-syn vs. control: p = 0.0129; hα-syn vs. mTurq: p = 0.0040]) in the hα-syn overexpressing mice, compared to the control and mTurq groups (Suppl. Figure 5), thus mimicking the early neuropathological events observed in the brain of PD patients [47,48,49,50]. Altogether, these observations demonstrate that, despite the widespread overexpression in the brain, hα-syn accumulation induced a selective loss of the dopaminergic neurons in the midbrain, mimicking the neurodegeneration observed in PD. Moreover, compared to current α-syn-based animal models, our model presents a suitable tool to investigate selective dopaminergic neuronal vulnerability in PD.

Fig. 3

Widespread overexpression of hα-syn induces selective dopaminergic neuronal loss in substantia nigra 3-months post-AAV delivery. (A-D) Stereological quantification of NeuN+ cells in the (A) striatum, (B) motor cortex, (C) hippocampus, and (D) SN. (E) Representative bright-field microscopy images of coronal brain slices illustrating TH+ neurons in the midbrain. Tissue was counterstained with Nissl stain (scale bar = 200 μm). (F-G) Stereological quantification of (F) TH+ dopaminergic neurons and (G) Nissl+ cells) in the SN. (H) Stereological quantification of TH+ dopaminergic neurons in the VTA. (I) Representative images of coronal brain slices illustrating the TH staining in the striatum (scale bar = 1 mm). (J) Quantification of TH staining optical density values in the striatum. The data are expressed as the means ± s.e.m. (n = 6–10 mice per experimental condition). One-way ANOVA followed by Tukey’s multiple comparisons test; * p ≤ 0.05, *** p ≤ 0.001. SN: substantia nigra; TH: tyrosine hydroxylase; VTA: ventral tegmental area

Dopaminergic neuronal loss in the substantia nigra is associated with the accumulation of pathological forms of α-syn

To investigate if the widespread overexpression of hα-syn and the dopaminergic neuronal loss at 3-months post-viral delivery are associated to abnormal accumulation of pathological α-syn, first, we verified the presence of the disease-associated phosphorylated form of α-syn at the residue Ser129 (pS129) in different brain regions. Unexpectedly, immunohistochemistry did not reveal any accumulation of pS129 α-syn nor the formation of dystrophic Lewy neurites in the cortex, the striatum, the hippocampus, or the midbrain (Suppl. Figure 6 A). This observation suggests that ⍺-syn pathology is either absent or undetectable due to the low levels of the exogenous overexpressed protein.

Then, we focused our analysis on the SN, and performed dot blot analysis of protein extracts from the insoluble fraction from the midbrain, using a battery of antibodies raised against monomeric α-syn, namely Syn1/BD Lab, α-syn-FL140 and α-syn LB509. Significant levels of α-syn expression were observed using three different antibodies in the protein extracts from mouse brains overexpressing hα-syn, as well as in the positive control samples of human recombinant α-syn (Mono and PFF α-syn) (Syn1/BD Lab [F [4, 10] = 81.41, p < 0.0001], α-syn-FL140 [F [4, 10] = 58.40, p < 0.0001] and α-syn LB509 [F [4, 10] = 96.08, p < 0.0001]). However, no α-syn signal was detected in the control or the mTurq samples, indicating that only the exogenous hα-syn, and not the endogenous mα-syn, was present in the insoluble protein fraction (Fig. 4A-C). We also examined pS129 levels and we observed a weak signal in hα-syn mice protein extracts (pS129-Wako [F [4,