Alpha‐Synuclein is Involved in DYT1 Dystonia Striatal Synaptic Dysfunction

Abstract Background The neuronal protein alpha‐synuclein (α‐Syn) is crucially involved in Parkinson's disease pathophysiology. Intriguingly, torsinA (TA), the protein causative of DYT1 dystonia, has been found to accumulate in Lewy bodies and to interact with α‐Syn. Both proteins act as molecular chaperones and control synaptic machinery. Despite such evidence, the role of α‐Syn in dystonia has never been investigated. Objective We explored whether α‐Syn and N‐ethylmaleimide sensitive fusion attachment protein receptor proteins (SNAREs), that are known to be modulated by α‐Syn, may be involved in DYT1 dystonia synaptic dysfunction. Methods We used electrophysiological and biochemical techniques to study synaptic alterations in the dorsal striatum of the Tor1a+/Δgag mouse model of DYT1 dystonia. Results In the Tor1a+/Δgag DYT1 mutant mice, we found a significant reduction of α‐Syn levels in whole striata, mainly involving glutamatergic corticostriatal terminals. Strikingly, the striatal levels of the vesicular SNARE VAMP‐2, a direct α‐Syn interactor, and of the transmembrane SNARE synaptosome‐associated protein 23 (SNAP‐23), that promotes glutamate synaptic vesicles release, were markedly decreased in mutant mice. Moreover, we detected an impairment of miniature glutamatergic postsynaptic currents (mEPSCs) recorded from striatal spiny neurons, in parallel with a decreased asynchronous release obtained by measuring quantal EPSCs (qEPSCs), which highlight a robust alteration in release probability. Finally, we also observed a significant reduction of TA striatal expression in α‐Syn null mice. Conclusions Our data demonstrate an unprecedented relationship between TA and α‐Syn, and reveal that α‐Syn and SNAREs alterations characterize the synaptic dysfunction underlying DYT1 dystonia. © 2022 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson Movement Disorder Society.

A BS TRACT: Background: The neuronal protein alphasynuclein (α-Syn) is crucially involved in Parkinson's disease pathophysiology. Intriguingly, torsinA (TA), the protein causative of DYT1 dystonia, has been found to accumulate in Lewy bodies and to interact with α-Syn. Both proteins act as molecular chaperones and control synaptic machinery. Despite such evidence, the role of α-Syn in dystonia has never been investigated. Objective: We explored whether α-Syn and Nethylmaleimide sensitive fusion attachment protein receptor proteins (SNAREs), that are known to be modulated by α-Syn, may be involved in DYT1 dystonia synaptic dysfunction. Methods: We used electrophysiological and biochemical techniques to study synaptic alterations in the dorsal striatum of the Tor1a + / Δgag mouse model of DYT1 dystonia. Results: In the Tor1a +/Δgag DYT1 mutant mice, we found a significant reduction of α-Syn levels in whole striata, mainly involving glutamatergic corticostriatal terminals. Strikingly, the striatal levels of the vesicular SNARE VAMP-2, a direct α-Syn interactor, and of the transmembrane SNARE synaptosome-associated protein 23 (SNAP-23), that promotes glutamate synaptic vesicles release, were markedly decreased in mutant mice. Moreover, we detected an impairment of miniature glutamatergic postsynaptic currents (mEPSCs) recorded from striatal spiny neurons, in parallel with a decreased asynchronous release obtained by measuring quantal EPSCs (qEPSCs), which highlight a robust alteration in release probability. Finally, we also observed a significant reduction of TA striatal expression in α-Syn null mice. Conclusions: Our data demonstrate an unprecedented relationship between TA and α-Syn, and reveal that α-Syn and SNAREs alterations characterize the synaptic dysfunction underlying DYT1 dystonia. © 2022 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson Movement Disorder Society.
Impairment of the synaptic vesicles machinery and neurotransmission is a characteristic feature of different movement disorders, including Parkinson's disease (PD), dystonia, and parkinsonism with dystonia. 1,2 Alpha-synuclein (α-Syn), a synaptic enriched protein member of the synucleins family, participates in the neuropathophysiology of PD. 3,4 Besides PD, its pathological aggregates characterize a wider group of neurodegenerative disorders defined as synucleinopathies. 3,4 In humans, α-Syn is encoded by the SNCA gene, located on chromosome 4q21. The main SNCA transcript gives rise to the production of a protein of 140 amino acids, which is ubiquitously expressed in the peripheral and central nervous system. 4 Although α-Syn functions are not entirely disclosed, it is known to play a role in maintaining the recycling pool of synaptic vesicles and modulating the assembly of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) complex. 5,6 In particular, α-Syn acts as a chaperone to promote SNARE complex assembly and to limit the trafficking and recycling of synaptic vesicles, thus controlling neurotransmitter release also by direct binding to vesicle-associated membrane protein-2 (VAMP-2/synaptobrevin-2). 6,7 Conversely, the pathological deposition of α-Syn in insoluble aggregates at synaptic terminals affects SNARE proteins (SNAREs) distribution in the brain of PD patients and in experimental synucleinopathy models exhibiting neurotransmitter release failure. 8,9 These findings are of particular interest in the context of PD and Lewy body (LB) dementia, as in the brains of the patients affected by these disorders the deposition of α-Syn aggregates at synaptic sites is several orders of magnitude higher than the amount of the protein composing LB. 10 Despite compelling evidence supporting the existence of possible overlapping mechanisms in PD and dystonia, 11 the role of α-Syn and SNAREs in the latter has never been investigated.
Early-onset generalized torsion DYT1 dystonia (DYT1) is an autosomal dominant movement disorder caused by a GAG deletion in the TOR1A gene coding for torsinA (TA). 12 Loss of the reciprocal modulation between the dopaminergic and cholinergic systems and synaptic plasticity imbalance point to synaptic dysfunction as a major pathophysiological alteration of DYT1 dystonia. [13][14][15] TA is a member of the AAA+ superfamily of ATPases, which typically act as chaperones in the endoplasmic reticulum (ER). 16 However, the interaction between TA and snapin supports the hypothesis that TA may influence synaptic vesicles dynamics in neurons. 17,18 Indeed, ΔE-TA overexpression affects vesicle exocytosis, thus resulting in the accumulation of the calcium (Ca 2+ ) sensor synaptotagmin I (Syt I) on the plasma membrane through a mechanism that involves snapin regulation. 17 In this way, TA acts as a chaperone at the synapse level affecting synaptic vesicles turnover and neurotransmitter release. 19 Intriguingly, TA has been found to accumulate in LB, where it interacts with α-Syn. 20 Moreover, a recent study has shown that dystonia-related genes may converge in common pathways linked to α-Syn and synaptic signaling. 21 Consistently, α-Syn null mice exhibit a decrease in striatal dopamine release as well as in the expression of some synaptic markers in the striatum, such as Syt and the dopamine transporter (DAT). 22 α-Syn and TA can both modulate DAT trafficking [23][24][25] and affect corticostriatal plasticity. 14,26,27 The two proteins are detectable in striatal synapses where α-Syn is mostly localized at glutamatergic terminals and controls the mobilization of glutamate from reserve pools. [28][29][30][31] Finally, TA can affect synaptic vesicle recycling analogously to α-Syn. 17,[32][33][34][35] Together, these findings suggest that both TA and α-Syn play a major role in the control of synaptic homeostasis. [36][37][38] Here, we investigated the possible occurrence of alterations in α-Syn and SNAREs levels in association with functional changes in the striatum of the Tor1a +/Δgag knock-in DYT1 mouse model. Our findings reveal that Tor1a +/Δgag mice exhibit a specific reduction of α-Syn levels in glutamatergic striatal terminals in association with an imbalance of synaptic proteins related to the SNARE complex. In parallel, we observed a remarkable decrease of miniature and quantal excitatory postsynaptic currents (mEPSC and qEPSC, respectively) recorded from striatal spiny projection neurons (SPNs), in the absence of alterations in GABAergic currents, indicating a significant impairment in release probability. These findings suggest that alterations in α-Syn expression and SNAREs may cause vesicle recycling alterations, with an ensuing impact on synaptic activity and plasticity.

Results
Striatal Levels of α-Syn are Reduced in The loss of the reciprocal modulation between the dopaminergic and cholinergic systems and corticostriatal plasticity imbalance suggest that synaptic dysfunction contributes to the pathophysiology of DYT1 dystonia. 13,14 Western Blot studies and confocal imaging were performed to analyze the levels of α-Syn in the dorsal striatum of Tor1a +/Δgag DYT1 mice. In striatal lysates from mutant mice, we observed significantly reduced levels of α-Syn protein compared to control samples ( Fig. 1A; **P < 0.01). In addition, confocal analysis showed a significant reduction of α-Synpositive signal in mutant mice striatum compared to control ( Fig. 1B; *P < 0.05). Interestingly, this downregulation was peculiar for DYT1, since the striatal levels of α-Syn were unchanged in a different dystonia model, the GNAL (DYT25) rat model ( Fig. 1C; P > 0.05). DYT1 mice exhibit a decrease of TA levels of approximately 50% with respect to wild-type littermates, suggesting that the Δgag is a loss-of-function mutation. 38,39 To evaluate whether α-Syn could affect TA protein levels, we also assessed striatal TA levels in α-Syn null mice. Surprisingly, we found a significant reduction (47%) of striatal TA level in α-Syn null mice compared to controls ( Fig. 1D; *P < 0.05), suggesting the existence of a reciprocal modulatory interaction between these two proteins.

Impaired Protein Expression of SNAREs Complex in Tor1a +/Δgag Mice
The SNARE complex mediates the fusion between synaptic vesicles and the presynaptic terminals. It consists of a number of proteins including the vesicle-associated SNAREs (v-SNAREs) VAMP-2 and the target cell-associated SNAREs (t-SNAREs) syntaxin I and synaptosomeassociated protein 25 kD (SNAP-25) or homologs. SNARE complex formation is maintained by canonical chaperones but also by non-classical chaperones such as α-Syn. 40 Since α-Syn may promote SNARE complex assembly through direct binding to VAMP-2 5,7 we quantified the expression of syntaxin-1, VAMP-2, SNAP-25, and its ubiquitously expressed homolog SNAP-23 in striatal lysates. We found that the levels of VAMP-2 were significantly reduced in the lysates from Tor1a +/Δgag mice when compared to Tor1a +/+ samples ( Fig. 2; **P < 0.01). In addition, in the lysates from Tor1a +/Δgag mice we also observed a significant reduction of SNAP-23 ( Fig. 2; **P < 0.01) when compared to Tor1a +/+ samples, although the (t-SNAREs) syntaxin-1 and SNAP-25 were unchanged ( Fig. 2; P > 0.05).

α-Syn Co-localizes with VGLUT-1 in Corticostriatal Glutamatergic Terminals
In the striatum, α-Syn is most abundant in excitatory when compared to inhibitory synapses and co-localizes mainly with vesicular glutamate transporter-1 (VGLUT-1), and, to a lesser extent, with VGLUT-2. 29,41,42 Thus, in order to evaluate possible changes of striatal dopaminergic, glutamatergic, and GABAergic synaptic terminals, we measured the areas immunopositive for the specific markers DAT, VGLUT-1, and vesicular GABA transporter (VGAT), respectively. (Fig. 3A). We found a significant reduction of DAT-and VGLUT-1-immunopositive areas (Fig. 3B; *P < 0.05 and **P < 0.01, respectively), while the VGAT-positive area was unchanged ( Fig. 3B; P > 0.05). Then, we performed a co-localization analysis of the areas positive for both α-Syn and the specific immunolabeling for the different synaptic markers (Fig. 3A). In particular, we assessed the co-localization rate, which represents the α-Syn immunopositive signal (in pixels) overlapping with the immunopositive signal of each of the specific markers (DAT, VGLUT-1, and VGAT). This co-localization rate was then normalized versus the overall area of immunopositivity of each of the assessed markers in order to estimate the amount of α-Syn localized in DAT-, VGLUT-1-, or VGAT-positive terminals. Interestingly, the co-localization analysis showed a significant decrease only in the amount of α-Syn localizing within VGLUT-1-immunopositive corticostriatal terminals in Tor1a +/Δgag mice when compared to Tor1a +/+ animals ( Fig. 3C; *P < 0.05), while no changes were detected in the α-Syn within VGAT-and DAT-positive terminals. This supports the observation that mutant TA-associated striatal α-Syn decrease mainly involves glutamatergic terminals.

Glutamatergic mEPSC are Altered in
Tor1a +/Δgag Mice α-Syn limits the trafficking and recycling of synaptic vesicles attenuating neurotransmitter release by its interaction with VAMP-2. 7,43,44 To explore potential differences in neurotransmitter release induced by a reduced expression of α-Syn, VAMP-2, and SNAP-23, we performed whole-cell patch-clamp recording experiments to analyze spontaneous inhibitory (GABA-mediated) and excitatory (glutamate-mediated) postsynaptic currents (sIPSCs and sEPSCs, respectively) in SPNs from both Tor1a +/+ and Tor1a +/Δgag mice. Then, we recorded the frequency and amplitude of miniature currents (mIPSC and mEPSC) to evaluate presynaptic vesicle release. GABAergic sIPSCs and mIPSCs were unchanged in Tor1a +/Δgag with respect to Tor1a +/+ littermates (Fig. 4A,B; P > 0.05) in line with our recent observations. 45 In addition, glutamatergic sEPSCs did not differ between genotypes ( Fig. 4C; P > 0.05), as previously demonstrated. 14 However, we found a significant decrease in the amplitude and frequency of mEPSCs recorded from Tor1a +/Δgag mice when compared to wildtype animals ( Fig. 4D; *P < 0.05). No changes in kinetic properties were observed between genotypes (data not shown; decay time constant: Tor1a +/+ 8.18 AE 0.86 ms; Tor1a +/Δgag 11.10 AE 1.52 ms, rise time: Tor1a +/+ 3.11 AE 0.19 ms; Tor1a +/Δgag 3.50 AE 0.26 ms; P > 0.05). The reduction of mEPSC reflects an impairment of the vesicular glutamate content that is reminiscent of that observed in α-Syn knockout (KO) mice. 31 Of note, western blot analysis showed a down-regulation of SNAP-23 which, unlike SNAP-25, is important for the functional regulation of glutamate receptors. 46 Remarkably, these are pivotal for glutamatergic transmission that is significantly reduced in neurons from VGLUT-1 KO mice where the loss of glutamate presynaptic loading and release impacts on synaptic vesicle cargoes turnover. 47,48 Our electrophysiological results also appear consistent with our present data showing a reduced α-Syn expression specifically in the VGLUT-1 terminals. Collectively, these data are supportive of the possible occurrence of an altered vesicle turnover, indicative of a dysfunctional presynaptic glutamatergic transmission in DYT1 mice.

Downregulated Asynchronous Release in
Tor1a +/Δgag Mice The synaptic membrane-fusion machinery is controlled by Syt I, which acts as a calcium (Ca 2+ ) sensor to regulate exocytosis during synchronous and asynchronous release. 49,50 Remarkably, while the synchronous release relies on the immediately releasable vesicles pool, asynchronous release can also involve recycling and reserve pools, which are regulated by α-Syn. 33,34 In glutamatergic neurons, synchronous release requires SNAP-25, while SNAP-23 only supports asynchronous release. 51 In addition, glutamatergic transmission is reduced in neurons from VGLUT-1 KO mice, specifically in quantal size. 46  Therefore, in order to corroborate our electrophysiological and biochemical data we investigated quantal-like events (qEPSCs) evoked after corticostriatal stimulation in SPNs from Tor1a +/+ and Tor1a +/Δgag mice following the replacement of extracellular Ca 2+ with strontium (Sr 2+ ). When Sr 2+ -induced asynchronous release was recorded in Tor1a +/ΔGAG SPNs, both the frequency and the amplitude of qEPSCs were significantly decreased compared to Tor1a +/+ neurons (Fig. 5A,B; ****P < 0.0001). Interestingly, our confocal analysis in the dorsal striatum from mutant mice showed a significant increase of the immunostaining of Syt I ( Fig. S1A; **P < 0.01). Accordingly, the quantification Syt I levels revealed a significant increase in the Tor1a +/Δgag striatum compared to Tor1a +/+ mice ( Fig. S1B; *P < 0.05). The observation that Syt I governs the synaptic vesicle endocytosis time-course by delaying the kinetics of vesicle retrieval in response to increasing Ca 2+ levels 52 supports the hypothesis that the increase of Syt I plays a role in the onset of asynchronous release deficits by affecting synaptic vesicles turnover. Indeed, the increase of Syt I may induce the formation of Syt I oligomers, which control asynchronous neurotransmitter release. 52 Our findings are also consistent with previous evidence in DYT1 cell models showing that mutant TA overexpression promotes Syt I accumulation on the plasma membrane through the reduction of synaptic vesicle turnover. 17,18 Collectively, these data suggest that DΥΤ1 mutant mice exhibit a deficit in asynchronous glutamate release, which reflects a robust dysfunction of synaptic vesicle turnover.

Rodent Models and Experimental Design
Studies were carried out in adult (P60-P90) mice and rats: knock-in Tor1a +/Δgag mice heterozygous for ΔE-torsinA, 53 in rats heterozygous for GNAL 54 and α-Syn null mice, carrying a spontaneous deletion of α-Syn gene (Harlan Olac, Bicester, UK) and their respective wild-type littermates (C57BL/6 for mice; Sprague Dawley for rats). Animal breeding and handling were performed in accordance with the guidelines for the use of animals in biomedical research provided by the European Union's directives and Italian laws (2010/63EU, D.lgs. 26/2014;86/609/CEE, D.Lgs 116/1992). Genotyping was performed as previously described. 55,56 Each observation was obtained from an independent biological sample. For electrophysiology, each cell was recorded from a different brain slice. All data were obtained from at least three animals in independent experiments. Biological replicates are represented with 'N' for number of animals and 'n' for number of cells.

Image Analysis of Striatal Immunopositive Area
The acquisition parameters during confocal imaging were maintained constant for all the images acquired. The optical density of the striatal positive area from digitized images acquired by confocal microscopy were examined by a researcher blind to the experimental conditions using FIJI Software. Five sections from each mouse were analyzed by examining an average of 10 fields per section. The threshold setup for FIJI was fixed between 30 and 150. The area of co-localization between α-Syn-immunolabeling and DAT-, VGAT-, or VGLUT-1-positive signal was quantified using Zen software (Carl Zeiss). The co-localization rate was then normalized on the total DAT-, VGAT-, or VGLUT-1-positive area for each field, respectively, in order to estimate the percentage amount of α-Synimmunoreactivity within each specific synaptic terminal.

Quantification and Statistical Analysis
Data analysis was performed with MiniAnalysis 6.0, ImageJ (NIH), and Prism5.3 (GraphPad). Data are reported as mean AE SEM. Statistical significance was evaluated as indicated in the text, and two-tailed unpaired or paired Student's test (t-test) was used for two-sample comparison. Normality tests were used to assess Gaussian distribution. Statistical tests were twotailed, the confidence interval was 95%, and the alphalevel used to determine significance was set at P < 0.05.

Discussion
The present results support the existence of an interplay between TA and α-Syn in synaptic homeostasis that is particularly relevant for glutamate neurotransmission. In particular, our findings show that α-Syn is downregulated in the striatum of mutant Tor1a +/Δgag mice but not in a distinct dystonia model, the DYT25 GNAL rat model. A further clue to the α-Syn-TA relationship is provided by our biochemical experiments on α-Syn null mice, indicating a significant down-regulation of TA. Our data also suggest that TA loss of function might alter synaptic machinery stability by inducing an increase in the Ca 2+ sensor Syt I, and a decrease in α-Syn and SNAREs thereby affecting the glutamate release process.
It has been shown that TA down-regulation induces persistence of Syt I on plasma membrane suggesting that the DYT1 mutation compromises synaptic vesicle recycling. 17,18 Many studies point to a crucial role for Syt I in promoting the synchronous release coupling Ca 2+ to SNARE-mediated fusion mechanism, but also in suppressing the asynchronous release, especially upon oligomer formation. 58,59 We found high levels of Syt I in striatum from mutant mice, which may be supportive of Syt I oligomers, whose formation is relevant for the kinetics of synaptic vesicle recycling during asynchronous release, 52 and that also appears in line with the increase of plasma membrane Syt I observed in DYT1 cell models. 17 While Syt I deletion induces an increase in asynchronous neurotransmitter release, its increase can negatively impact on this process. 60 Consistently, we found that in Tor1a +/Δgag mice, Sr 2+ , which normally stimulates the asynchronous release with a consequent increase in events, 61 reduced qEPSCs, thus confirming an impairment of asynchronous synaptic release. This finding, in parallel to the increase in Syt I, suggests that TA mutation impairs glutamatergic synaptic vesicles turnover, which in turn affects asynchronous glutamate release, a process governing the recovery of neuronal excitability following post-spike hyperpolarization. 62 We showed a significant reduction of the v-SNAREs member VAMP-2, which is the direct interactor of α-Syn. 7 Moreover, we also found a significant decrease of SNAP-23, which, unlike SNAP-25, is more important for the functional regulation of the glutamate receptors and in modulating asynchronous release. 46,51 Our present findings are in agreement with a proposed modulatory role of α-Syn on glutamatergic synaptic activity, in line with evidence supporting it regulating presynaptic mobilization of reserve pools of vesicles at glutamatergic terminals. 31,63 Consistently, we found that in the Tor1a +/Δgag mice α-Syn was reduced at VGLUT-1-positive terminals, whereas GABAergic transmission was normal.
Our confocal imaging data demonstrate a diffuse reduction of VGLUT-1 and DAT signals, in accordance with previous evidence on DYT1 experimental models. [64][65][66] Indeed, Ip and co-workers, showed that Tor1a AE mice exhibit a reduction of striatal DAT level as well as DAT binding decrease after sciatic nerve crush. 67 This appears consistent with the fact that DAT is an α-Syn interactor and that TA can affect DAT expression. 25,67 Of note, VGLUT-1 plays a key role in controlling cargo protein recovery, including VAMP-2, but not Syt I, and is essential for ensuring the quantal efficiency of glutamatergic transmission. 48,68 Therefore, the VGLUT-1 reduction in DYT1 mice does not appear to contradict the observed Syt I accumulation. Interestingly, the relevance of VGLUT-1 in cargo protein recovery may suggest that its decrease could also underlie the reduction of both α-Syn and SNAREs. This notwithstanding, it has been shown that unlike other synaptic vesicle-associated proteins such as SNAREs or synapsin, which rapidly recluster synaptic terminals colocalizing with VGLUT-1 in the post-depolarization recovery phase, α-Syn dissociates from synaptic vesicle membranes after their fusion and exhibits a different and slower recovery. 71 Therefore, it appears unlikely that a VGLUT-1 reduction-associated lowering of cargo protein recovery could be the basis of the observed α-Syn decrease at glutamatergic terminals, though we do not exclude the possibility that by affecting quantal synaptic efficiency 47 it could blunt qEPSC.
Overall, our results highlight the existence of a strong relationship between TA, α-Syn, and SNAREs in the control of glutamate release. In particular, they indicate that TA mutations affect striatal glutamatergic transmission mainly by impinging on asynchronous release. This phenomenon may very well be driven by the reduction of α-Syn and SNAREs occurring in parallel to Syt I increase (Fig. 5C). Our findings further support the notion that different pathways may converge to cause basal ganglia synaptic abnormalities as a main determinant in the pathophysiology of dystonia. 21,70,71 Furthermore, this evidence envisages a pivotal involvement of alterations of α-Syn, SNAREs, and related synaptic vesicle-associated protein in the molecular underpinnings of synaptic imbalance in DYT1 dystonia warranting further investigation.

Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.