α‐Synuclein antisense transcript SNCA‐AS1 regulates synapses‐ and aging‐related genes suggesting its implication in Parkinson's disease

Abstract SNCA protein product, α‐synuclein, is widely renowned for its role in synaptogenesis and implication in both aging and Parkinson's disease (PD), but research efforts are still needed to elucidate its physiological functions and mechanisms of regulation. In this work, we aim to characterize SNCA‐AS1, antisense transcript to the SNCA gene, and its implications in cellular processes. The overexpression of SNCA‐AS1 upregulates both SNCA and α‐synuclein and, through RNA‐sequencing analysis, we investigated the transcriptomic changes of which both genes are responsible. We highlight how they impact neurites' extension and synapses' biology, through specific molecular signatures. We report a reduced expression of markers associated with synaptic plasticity, and we specifically focus on GABAergic and dopaminergic synapses, for their relevance in aging processes and PD, respectively. A reduction in SNCA‐AS1 expression leads to the opposite effect. As part of this signature is co‐regulated by the two genes, we discriminate between functions elicited by genes specifically altered by SNCA‐AS1 or SNCA's overexpression, observing a relevant role for SNCA‐AS1 in synaptogenesis through a shared molecular signature with SNCA. We also highlight how numerous deregulated pathways are implicated in aging‐related processes, suggesting that SNCA‐AS1 could be a key player in cellular senescence, with implications for aging‐related diseases. Indeed, the upregulation of SNCA‐AS1 leads to alterations in numerous PD‐specific genes, with an impact highly comparable to that of SNCA's upregulation. Our results show that SNCA‐AS1 elicits its cellular functions through the regulation of SNCA, with a specific modulation of synaptogenesis and senescence, presenting implications in PD.


| INTRODUC TI ON
Aging is the main risk factor for Parkinson's disease (PD), whose prevalence increases by more than 400 times in the elderly population reaching 1% at the age of 60 and 5% in people over 85 (Rodriguez et al., 2015). Numerous neurodegenerative processes occurring in PD are linked to the aging brain, with a great relevance for the degeneration of the nigrostriatal dopaminergic neurons system, a crucial hallmark of PD. Indeed, it has been reported that the most significant difference in PD and aged brains is the number of dopaminergic neurons, suggesting that the etiology of the disease could be overlapping with aging (Rodriguez et al., 2015). The synuclein alpha gene (SNCA) encodes for α-synuclein (α-syn), the small protein mainly known for its implications in PD (Houlden & Singleton, 2012) and principal component of the Lewy Bodies aggregates (Spillantini et al., 1997), the second fundamental hallmark of PD (Poewe et al., 2017).
α-syn physiological functions are still far from clarified. Molecular and transgenic studies involving the protein have not yet completely elucidated its functions, although its main implication seems to be in the synaptic vesicles cycle and dendritic development (Sulzer & Edwards, 2019). In both the aging brain and PD, α-syn promotes the disruption of synapses, with the accumulation of pathogenic proteins leading to an impairment of the ubiquitin-proteasome system, autophagy, and mitochondria, ultimately resulting in dopaminergic degeneration (Bobela et al., 2015;Rodriguez et al., 2015).
It is necessary to eviscerate the possible implication of the SNCA gene locus in α-syn's regulation, as it could represent a common target in PD and aging. Epigenetically, SNCA is highly modulated through twenty-one CpG islands present on the promoter and the first intron of the gene (Guhathakurta et al., 2017). In addition, the SNCA antisense gene (SNCA-AS1), which codes for a long noncoding RNA (lncRNA), has been recently described to localize on the strand opposite to the SNCA gene (Fagerberg et al., 2014). Indeed, the SNCA locus, including SNCA-AS1, has been recently associated with hereditary neurodegenerative diseases and Lewy body dementia (Chia et al., 2021;Zucchelli et al., 2019). Conversely to SNCA's mRNA, SNCA-AS1 is strongly enriched in brain tissue (Fagerberg et al., 2014), and its expression increases during both SH-SY5Y cells and induced Pluripotent Stem Cells (iPSCs) dopaminergic in vitro differentiation (Elkouris et al., 2019). There is currently no evidence reported of what its role in SNCA's regulation and in both aging and PD pathogenesis might be, nor a clear understanding of its function at a cellular level. A wide range of lncRNAs are being implicated in aging and neurodegenerative diseases, through modulation of autophagy, apoptosis, oxidative stress, and even SNCA's expression and subsequent effects on α-syn's aggregation (Lyu et al., 2019;Pereira Fernandes et al., 2018). LncRNAs could even play a role as potential biomarkers, as a significant number of them have been found deregulated in circulating leucocytes or peripheral blood mononuclear cells (PBMCs) from sporadic PD patients (Fan et al., 2019).
It is in this context that there is a crucial need to understand the potential implication for SNCA-AS1 in SNCA's regulation, as it could provide new therapeutic strategies and it could also elucidate new functions underlying cellular senescence and ultimately neurodegenerative diseases. With this work, we aim to elucidate the effect of SNCA-AS1 and SNCA mRNA overexpression in the in vitro SH-SY5Y cell line, a neural model consolidated for its relevance in PD studies (Xicoy et al., 2017). We report how the overexpression of SNCA-AS1 leads to an increase in SNCA's mRNA expression, and through RNA sequencing, we identify a significant number of deregulated genes specifically ascribable to SNCA-AS1 or SNCA's mRNA overexpression. With both bioinformatic approaches and in vitro validations, we describe how these genes influence numerous processes, including neurites extension, synaptogenesis, and cellular senescence, with significant implications in the brain aging process and in PD pathology.
2 | RE SULTS 2.1 | In silico characterization of the novel SNCA Antisense RNA SNCA-AS1 SNCA antisense RNA (SNCA-AS1) is a recently discovered antisense transcript which localizes on chromosome 4, on the strand opposite to the SNCA gene, partially overlapping its 5′ end with its first exon K E Y W O R D S aging, LncRNAs, Parkinson's disease, RNA-sequencing, SNCA, SNCA-AS1, synapses, synuclein F I G U R E 1 α-synuclein's (α-syn) expression can be modulated by SNCA-AS1 and effects of both SNCA-AS1 and SNCA overexpression on transcription profiles. (a) Genetic localization of SNCA-AS1 gene with respect to the SNCA gene. (b) Evolution conservation as obtained with the AnnoLnc2 database. (c) SNCA-AS1 RNA secondary structure prediction for minimum free energy (MFE) minimization with the AnnoLnc2 database. Binding sites for miRNA are highlighted in blue, and the respective miRNA in red. (d) SNCA-AS1 expression in the brain as obtained with the LncExpDB database. (e) The increase of α-syn's expression in SH-SY5Y-SNCA-AS1 was evaluated by means of realtime PCR and Western blot. GAPDH was used as housekeeping gene for both experiments. For real-time PCR analysis, data are expressed as mean ± SEM of 3 replicate values in 3 independent experiments (n = 9; ***p < 0.001 vs. SH-SY5Y). For Western blot analysis, data are expressed as mean ± SEM of 5 independent experiments (n = 5; **p < 0.01 vs. SH-SY5Y). (f) Cellular localization determined by droplet digital PCR in SH-SY5Y. Data are expressed as mean ± SEM of 3 independent experiments (n = 3; **p < 0.01, ***p < 0.001 vs. Nuclei). (g) Volcano plot showing deregulated genes in SH-SY5Y-SNCA-AS1 vs SH-SY5Y and SH-SY5Y-SNCA vs SH-SY5Y. (h) PCA of differently expressed genes in SH-SY5Y-SNCA-AS1, SH-SY5Y-SNCA, and SH-SY5Y. (i) GSEA representation of MANNO midbrain neurotype as obtained with the cell type signature gene sets database ( Figure 1a). Specifically, its second exon overlaps the first exon of the longest isoform of the SNCA mRNA. With the AnnoLnc2 database, we screened the lncRNA for repeat elements, evolutionary conservation, transcriptional, and miRNA regulation (Ke et al., 2020). The search for repeat elements with the RepeatMasker genomic datasets highlighted no results, suggesting these elements are not present along SNCA-AS1 sequence. PhyloP and PhastCons scoring aimed at assessing base conservation highlighted a tendency for sequence conservation across primates, mammals, and vertebrates ( Figure 1b).
Moreover, phylogenetic analysis was performed by neighbor-joining (NJ) method and the maximum-likelihood (ML) based on 26 aligned predicted sequences of SNCA-AS1 (see Experimental Procedures).
Homo sapiens was used as outgroup. The NJ-based phylogenetic analysis allows to distinguish between four groups, reporting a conservation of the transcripts in mammals. 19 out of 26 are primates, which are mostly grouped together with exclusion of the Cercopithecidae familiae ( Figure S1A). Phylogenetic analysis highlights also similarity between predicted sequences with Camelus, Vicugna, and Mustela ( Figure S1A). The ML-based phylogenetic tree shows similar topology with the NJ-based phylogenetic tree ( Figure   S1B). SNCA-AS1's secondary structure was predicted through conformational bioinformatics studies with the identification of a possible secondary structure specifically based on the minimum free energy (MFE) minimization ( Figure S1C), predicted according to Turner 2004 RNA folding parameters (Lorenz et al., 2011). The MFE reached is −244.30 kcal/mol. We investigated the potential miRNA regulation of SNCA-AS1 through the analysis of miRNA binding sites and we highlighted binding motifs for 17 miRNA (Figure 1c, Table   S1). Transcription factors (TFs) associated with SNCA-AS1 were predicted using both the AnnoLnc2 database through the Gene Transcription Regulation Database database (Table S2) and Ciiider software (Gearing et al., 2019), which highlighted 462 TFs. Among them, 21 out of 462 were neural TFs ( Figure S1D) and, even more interestingly, 9 (e.g., Ahr::Arnt, Arnt, E2f1, Foxj2, Foxo1, Foxo4, Hsf1, Sox5, and Yy1) were associated with the aging process ( Figure   S1E). An expression analysis conducted with the LncExpDB (Li et al., 2021) database highlighted a strong expression of SNCA-AS1 during brain development, with a peak at week 8 ( Figure 1d).

| SNCA-AS1 regulates SNCA and αsynuclein's expression
To firstly investigate the possible regulation of SNCA by SNCA-AS1, we analyzed whether this lncRNA could impact SNCA's transcription and subsequent translation. Indeed, SH-SY5Y cells stably transfected with SNCA-AS1 presented an upregulation of SNCA's mRNA expression and a subsequent upregulation of its protein product αsyn ( Figure 1e). As very little is known about SNCA-AS1's expression and function, an analysis of its localization was performed by droplet digital PCR. The results show that SNCA-AS1 and SNCA RNAs are mainly localized in the cytoplasm, both in non-transfected SH-SY5Y ( Figure 1f) and when overexpressed (SH-SY5Y-SNCA and SH-SY5Y-SNCA-AS1) ( Figure S1F).

| Overexpression of SNCA-AS1 and SNCA leads to significantly different RNAs expression profiles
As SNCA-AS1 is a novel uncharacterized lncRNAs, nothing is known about its signal transduction and there is no evidence of specific cellular pathways affected by its expression. Furthermore, the effect of SNCA's mRNA overexpression on cellular transcriptome is yet to be fully characterized and could be helpful in the understanding of α-syn's biology and PD pathogenesis. To this end, a whole transcriptome analysis of wild type SH-SY5Y and SH-SY5Y overexpressing either SNCA-AS1 or SNCA was performed. We detected many differentially expressed coding and non-coding RNAs (DE RNAs) in SH-SY5Y-SNCA-AS1 and SH-SY5Y-SNCA vs. SH-SY5Y, as shown by volcano plot analysis (Figure 1g). PCA analysis of the DE RNAs showed a clear difference in the three expression profiles, suggesting that the two investigated genes deeply affect cellular functions ( Figure 1h). Specifically, a total of 969 transcripts were affected by SNCA-AS1's overexpression (82% of which were coding genes, Table   S3) and 698 by SNCA's overexpression (85% of which were coding genes, Table S4). A validation of 4 selected deregulated transcripts, implicated in pathways of interest, was performed by real-time PCR confirming the RNA-Sequencing (RNA-Seq) evidence ( Figure S2).
Gene set enrichment analysis (GSEA) for the cell type signature gene sets database was performed, and, interestingly, we found that among the top cellular phenotypes impacted by SNCA-AS1, the midbrain neurotype appeared to be profoundly negatively correlated with the lncRNA overexpression, with a high number of genes being downregulated ( Figure 1i).

| Pathway analysis of deregulated transcripts
Transcripts were subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis using GSEA (Table S5), and in Figure S3, the significantly enriched terms obtained with GSEA KEGG pathway analysis are reported. Interestingly, in SNCA overexpression it is possible to notice a reduced expression of terms F I G U R E 2 Pathway and GO analyses for DE RNAs in SH-SY5Y overexpressing SNCA-AS1 and SNCA. Pie chart displaying the % of KEGG processes associated with synapses (a) or aging (b) processes as obtained with EnrichR, g:Profiler and GSEA analyses in SH-SY5Y-SNCA-AS1 or SH-SY5Y-SNCA vs. SH-SY5Y. Only terms with p-value <0.05 were considered. GSEA analysis of top GO terms for CC (c), BP (d), and MF (e) in SH-SY5Y-SNCA-AS1 or SH-SY5Y-SNCA vs SH-SY5Y. Pie chart displaying the % of processes associated with synapses in GO CC (f) BP (g) or MF (h) as obtained with EnrichR, g:Profiler, and GSEA analyses in SH-SNCA-AS1 or SNCA vs SH-SY5Y. Only terms with p-value <0.05 were considered associated with neurodegenerative diseases (Alzheimer's disease and PD) ( Figure S3). The deregulated transcripts with a deregulation ≥1 in terms of |Log 2 FC| were subjected to KEGG pathways analysis through the use of two more tools: the EnrichR web tool (Kuleshov et al., 2016) (Table S6) and g:Profiler (Table S7), where terms were ranked for their importance in terms of fold change (Raudvere et al., 2019). As the only evidence currently present for SNCA-AS1 points to its implication in neural differentiation (Elkouris et al., 2019), along with a synaptic implication being the main physiological function known for SNCA, we extrapolated from the three enrichment analysis datasets all pathways pertaining synaptogenesis ( Figure 2a). All three enrichment methods highlighted that around 1/3 of terms re- In order to obtain more information on SNCA-AS1 involvement in regulating synaptic processes, as it can be relevant for both PD and aging progression, the GO CC, BP, and MF enrichment were analyzed using GSEA (Table S8), EnrichR (Table S9), and g:Profiler (Table   S10). When considering the CC profile, all three enrichment methods reported a strong involvement of both SNCA and SNCA-AS1 in synaptic processes, which appears to be around 50% for most databases analysis (except for EnrichR where it seems to comprise most processes) (Figure 2f). Considering BP, the synaptic involvement ranged from 15% to 25% for g:Profiler and GSEA analysis, reaching 60% with the EnrichR database ( Figure 2g). Lastly, for MF analysis, the synaptic implication ranged from 35%-57% for g:Profiler and GSEA analysis, reaching 89% with the EnrichR database ( Figure 2h).

| SNCA-AS1 and SNCA transcriptional modifications are implicated in synaptic modulation resembling an aging-related decline
Synaptic dysfunctions are underlying processes in both aging and neurodegenerative diseases, and SNCA-AS1 could prove to be of crucial relevance. We firstly used the Aging Atlas database to identify those genes, among the DE RNAs, which have been implicated in aging processes through transcriptomics, proteomics, epigenomics, and more studies (Consortium, 2021). We highlighted how in both cases more than 85% of genes are correlated with aging, specifically  Figure S5A) and SH-SY5Y-SNCA ( Figure S5B) was performed, highlighting that in both conditions there seems to be a reduced synaptic plasticity, as highlighted by the decrease in c-fos, a gene correlated to increased long-term synaptic plasticity and, intriguingly, implicated in aging (Jaworski et al., 2018). Moreover, in both contexts, there is a decrease in the expression of the gene that codes for vesicular monoamine transporter (VMAT), also decreased in aging neurons (Meyza et al., 2007). A closer look was also given to the impact of both SNCA-AS1 and SNCA on GABAergic synapses ( Figure S6A,B) as these are widely deregulated in the aging process for their relevance in regulating the excitatory/inhibitory balance. Indeed, a significant disruption is observed both in SH-SY5Y-SNCA-AS1 and in SH-SY5Y-SNCA with a decreased uptake of synaptic GABA ( Figure S6). Lastly, focusing on cellular senescence pathways, both in SH-SY5Y-SNCA-AS1 and SH-SY5Y-SNCA ( Figure S7A,B) there was an upregulation in related genes, including the FOXO1 transcription factor, whose prediction site is also present in SNCA-AS1's promoter ( Figure S1E) and which is strongly implicated in aging-related processes (Tia et al., 2018). We When looking at GO BP, the same pattern is maintained: There is a deregulation in synaptic biology, cell junction assembly, and synapse development, but there is also a presence for non-neuronal related functions such as cardiac, renal, and neuromuscular development.
The same is true for KEGG 2021 enrichment analysis.
When analyzing the 453 terms selectively affected in SH-SY5Y-SNCA cells (Figure 4c), the most deregulated cellular components resulted to be the synapses and neuron projections.
The main significantly enriched MF term relates to inorganic cation transmembrane transporter activity. Furthermore, a BP enrichment analysis shows a deregulation in nervous system F I G U R E 4 Analysis of SNCA-AS1 and SNCA-specific functions. ClueGO CC, MF, BP, and dotplot of KEGG top 10 enriched terms in genes deregulated only SH-SY5Y-SNCA-AS1 vs SH-SY5Y (a), in genes shared between SH-SY5Y-SNCA-AS1 and SH-SY5Y-SNCA vs SH-SY5Y (b) and in genes deregulated only in SH-SY5Y-SNCA vs SH-SY5Y (c). Each pie segment refers to the % of terms present per group, p < 0.05 was set as the limit of significance (**p < 0.01 vs. SH-SY5Y). In the dotplots, the y-axis represents the name of the pathway, the x-axis represents the gene ratio, dot size represents the number of DE RNAs and the color indicates the adjusted p-value development and neurogenesis. Moreover, KEGG 2021 analysis shows an implication of variable pathways such as cardiac-and renal-related ones but the dopaminergic synapse is the tenth most dysregulated pathway.
Together, these results highlight how SNCA-AS1 could be of crucial importance in regulating synapses biology through a molecular signature shared with SNCA, and thus, α-syn's involvement in synaptic processes could possibly be due partially to its modulation by SNCA-AS1.

| Implication for SNCA-AS1 and SNCA in PD
With the aim to correlate this gene expression evidence with Lewy bodies' pathology, we focused our attention on the identification of sense and antisense regulation of PD-associated genes.  Table S13, and MAPT is also present in this case as the most PDcharacterized gene.
With the aim to identify whether it was SNCA-AS1-specific genes, shared genes, or SNCA-specific genes which were more likely to lead to the development of PD pathogenesis, the three classes of terms were analyzed with the ClinVAR database. This database allows for the identification of all possible diseases associated with a certain gene set. Remarkably, the most significant correlation with PD was observed in the shared terms dataset, where out of all the possible disease associated a correlation was obtained with "Parkinson-dementia Syndrome" (3,12% terms per group) (Figure 5c). This is interesting to report as there could be a deregulation in the overlapping SNCA-AS1 and SNCA regulated pathways in PD, with a specific focus on the processes which are affected by their synergic regulation. To validate the implication for SNCA-AS1 in PD, we assessed its expression in PD-related datasets (Dumitriu et al., 2016;Nido et al., 2020;Schulze et al., 2018) and we found that in all analyzed conditions (PD brain frontal cortex, PD iPSCs, and PD mature neurons) SNCA-AS1 was upregulated (Figure 5d). To provide initial evidence for this claim in PD patients, the expression levels of SNCA-AS1 and SNCA were analyzed in PBMCs of PD-affected  (Table S14).

| SNCA-AS1 downregulation leads to a concordant downregulation of SNCA and an impact on synapses-related genes
Our results have so far shown that the overexpression of SNCA-AS1 upregulates SNCA mRNA and protein. Moreover, our results reported a reduced expression of markers associated with synaptic plasticity. To validate whether these results were specifically due to SNCA-AS1 overexpression, we inhibited SNCA-AS1 in SH-SY5Y differentiated for 7 days with retinoic acid (Figure 6). SNCA-AS1 was efficiently knocked-down and SNCA expression was concordantly downregulated (Figure 6a). α-syn expression was also significantly downregulated after SNCA-AS1 inhibition (Figure 6b). We also validated the impact of SNCA-AS1 downregulation on neurite length and synaptogenesis, and we found a reduced neurites extension ( Figure 6c). Lastly, we found that the expression trends of genes im-

| DISCUSS ION
More and more research is now establishing how PD is an evolution of the aging process, with numerous overlapping mechanisms (Bobela et al., 2015;Rodriguez et al., 2015). Indeed, α-syn, crucial for PD pathogenesis, can promote the accumulation of pathogenic proteins and lead to an impairment of the ubiquitin-proteasome system, ultimately resulting in dopaminergic degeneration in both PD and the aging brain (Rodriguez et al., 2015). Knowledge on α-syn's physiolog- synapses and aging as terms involved in this process were amongst the top enriched. GO enrichment analysis further implicated SNCA in synapses and SNCA-AS1 to a lesser extent, as in this case there is again also a strong involvement of cellular interaction with the surrounding environment. In both SH-SY5Y-SNCA-AS1 and SH-SY5Y-SNCA, there is a reduction in genes associated with synaptic plasticity, such as c-fos, implicated in the aging process and longterm synaptic plasticity (Jaworski et al., 2018). Moreover, in both contexts, there is a decrease in the expression of the gene that codes for VMAT. Interestingly, multiple studies report how this transporter is dysfunctional in PD and implicated in aging (Meyza et al., 2007).
A significant disruption of GABAergic synapse was observed both in SH-SY5Y-SNCA-AS1 and in SH-SY5Y-SNCA, and the downregulations in these specific genes have been reported also in the aging brain (Rozycka & Liguz-Lecznar, 2017). The cellular senescence pathway was also found deregulated, and SNCA-AS1 could thus be implicated in aging-related processes. As we demonstrated that SNCA's expression is increased when SNCA-AS1 is overexpressed, it was not surprising to identify several shared terms between the two deregulated gene sets. A global analysis allowed us to perform a primary characterization of the potential genes' involvement, but it did not discriminate between those pathways, if any, selectively due to SNCA-AS1 overexpression, to SNCA mRNA overexpression, or to a mechanism of action shared between the two and due to the overlapping deregulated genes. When discerning the two datasets, we found that SNCA-AS1 involvement in synapses biology was predominant when analyzing the terms shared among the two datasets, while "SNCA-only" genes also independently affect synapses biology. This highlights that SNCA-AS1 could be a player in synaptic modulation through its modulation of SNCA and α-syn. It is thus also possible to speculate that α-syn's involvement in synaptic modulation could be partially due to a common pathway initiated or shared with SNCA-AS1. The presence of PD-related genes in both datasets also shows that SNCA-AS1 could be a direct player in the pathogenesis of the diseases, and the evidence reported in PBMCs of PD patients shows that the SNCA/SNCA-AS1 ratio could be what is disrupted in the disease and could prove to be a relevant biomarker for the pathology.
Even if further research is needed, the findings here reported demonstrate a strong impact for SNCA-AS1 in the aging process and ultimately in PD, suggesting that this lncRNA may be even more crucial than we thought.

| RNA secondary structure prediction
RNA secondary structure was predicted using the Geneious software (Geneious version 2020.1 created by Biomatters. Available from https://www.genei ous.com) based on Vienna RNA Fold (Lorenz et al., 2011) with default settings.

| Transcription factors' prediction
Transcription factors binding sites were predicted through Ciiider software. Ciiider analysis was performed using the human GRCh38 genome and the 2020 JASPAR core non-redundant vertebrate matrices (Fornes et al., 2020). All promoter regions were defined as spanning −1500 bases to +500 bases relative to the transcription start site.

| Phylogenetic analysis
Phylogenetic analysis was performed using Geneious software (Geneious version 2020.1 created by Biomatters. Available from https://www.genei ous.com). The SNCA-AS1 sequence was used as query to search the sequences with high similarity in databases using Megablast (Chen et al., 2015). Sequences with high pairwise identity were chosen and used as input for multiple alignment. Multiple alignment was performed using Clustal Omega (Sievers & Higgins, 2014).
The final sequence alignment was used to perform phylogenetic analysis employing the distance-based NJ method and the ML implemented in the PHYML program (Guindon et al., 2010). The genetic distance for NJ method was calculated through the Tamura-Nei model.
Cytoplasmic and nuclear RNA fractions were extracted using the Cytoplasmic & Nuclear RNA Purification Kit (Norgen Biotek Corp) according to the manufacturer's instructions.

| Real-Time PCR
Total RNA (1 μg) was reverse transcribed using iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's instructions. Using gene sequences available from NCBI for target genes (http://www.ncbi.nlm. nih.gov/nucle otide), PCR oligonucleotide primers were selected and are reported in Table S15. This was done with the NCBI's Primer-BLAST tool. Real-time PCR was performed with StepOnePlus TM Real-Time PCR System (Invitrogen) using SSOSYBR Green Supermix (Bio-Rad).
Genes were quantified in triplicates; GAPDH was used as housekeeping gene. Gene expression was calculated using the 2 −ΔΔCt method.

| Droplet digital PCR
Total RNA (200 ng) was reverse transcribed using iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's instruc-

| Pathway analysis and Gene Ontology
Gene set enrichment analysis was performed on clusterProfiler R package (Yu et al., 2012). Gene set from Molecular Signature databases such as curated gene set (C2) and ontology gene sets (C5) and a p-value cutoff <0.05 were considered for this analysis (Yu et al., 2012). Moreover, functional enrichment analysis was performed on coding DE RNAs using the EnrichR webtool (Kuleshov et al., 2016) and g:Profiler, ranking terms according to their absolute fold change and using a Bonferroni-Hochberg FDR of 0.05 as threshold. The R software was used to generate Volcano plots , Dotplot graphs (ggplot2 library), and Pathview graphs (Pathview library [Luo & Brouwer, 2013]). All other representations of functional enrichment were generated using the Cytoscape software (Shannon et al., 2003) and the DisGeNET plugin (Piñero et al., 2017).

| Western blot
Cell protein extracts were obtained by means of RIPA lysis buffer.
Equal amounts of solubilized proteins were heated in Laemmli sample buffer (Bio-Rad) containing 2βmercaptoethanol (70 mM, Sigma), separated by SDS-PAGE gel 10% and electroblotted onto a nitrocellulose membrane (GE Healthcare, Amersham™). Membranes were fixed for 30 min in 0.4% paraformaldehyde (Lee & Kamitani, 2011). Membranes were blocked in 5% slim milk (diluted in TBS with 0.05% Tween-20) and probed with the appropriate primary antibody against α-syn (AB138501; abcam) and GAPDH (2118;Cell Signaling) overnight at 4°C. The membrane was then incubated with specific secondary antibody Peroxidase AffiniPure Goat Anti-Rabbit/Mouse IgG (Jackson). Proteins were visualized by means of an enhanced chemiluminescence detection system (ECL™; Amersham). After acquisition by a GelDoc™ image capture system (Kodak), proteins were quantified using ImageJ software.

| Immunofluorescence and confocal microscopy
SH-SY5Y cells were seeded on ethanol-washed glass coverslips, maintained in the appropriate culture medium, and processed for immunocytochemistry following an already described protocol (Marfia et al., 2011). Briefly, cells were fixed with 4% paraformal-

| Neurite length analysis
For immunofluorescence images, TUJ was used to mark the cytoskeleton and neurite extension, and the Simple Neurite Tracer plugin of Fiji software was used (Longair et al., 2011).

| Study subjects
Sixteen PD patients and sixteen age-and sex-matched healthy controls were recruited after obtaining written informed consent.
PD patients underwent clinical and neurologic examination at IRCCS National Neurological Institute "C. Mondino" (Pavia, Italy).
All patients were diagnosed with PD as defined by Movement Disorder Society clinical diagnostic criteria (Postuma et al., 2015).
The control subjects were recruited at the Transfusional Service and Centre of Transplantation Immunology, Foundation San Matteo, IRCCS (Pavia, Italy). The study protocol was approved by the Ethical Committee of the National Neurological Institute "C. Mondino", IRCCS (Pavia, Italy). Before being enrolled, the subjects participating in the study signed an informed consent form (Protocol n° p-20170001758).

| Isolation of human PBMCs
Peripheral blood mononuclear cells were isolated by centrifugation on a Ficoll-Histopaque layer (Sigma-Aldrich), and cells were used for subsequent RNA extraction.

| Statistical analysis
Data are expressed as mean ± SEM. The statistical analysis was performed with Student's t test when two datasets were considered, and one-way ANOVA with Bonferroni's post-test when three datasets were considered. The Prism 7 software (GraphPad Software Inc.) was used assuming a p-value <than 0.05 as the limit of significance.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
FR involved in design, conception, data generation, data acquisition, data interpretation, and manuscript writing. CP involved in design, data generation, data acquisition, data interpretation, and data analysis. LM involved in bioinformatic data analysis and interpretation. RL involved in data generation and data acquisition.
BB involved in data generation, data acquisition, and discussion.
RZ involved in patient selection and clinical data generation. MTR involved in supervision of the work and manuscript revision. SG involved in transcriptomic data interpretation. CC involved in supervision of the work, draft and manuscript revision and financial support. GVZ involved in supervision of the work, manuscript revision, and financial support. SC involved in conception and design, supervision of the work, data interpretation, manuscript writing and revision.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data obtained from the RNA-Seq analysis are deposited on the Gene Expression Omnibus repository: GSE183410 for SH-SY5Y and GSE186255 for SH-SY5Y-SNCA and SH-SY5Y-SNCA-AS1.