An Isogenic Collection of Pluripotent Stem Cell Lines With Elevated α-Synuclein Expression Validated for Neural Induction and Cortical Neuron Differentiation

α-Synuclein (αSyn) is a small, disordered protein that becomes aggregated in Lewy body diseases, such as Parkinson’s disease (PD) and dementia with Lewy bodies (DLB). Human induced pluripotent stem cells (hiPSCs) potentially provide a tractable disease model to monitor early molecular changes associated with PD/DLB. We and others have previously derived hiPSC lines from patients with duplication and triplication of the SNCA gene, encoding for αSyn. It is now recognised that to perform meaningful disease modelling with these hiPSC lines, it is critical to generate isogenic control cell lines that lack the disease causing mutations. In order to complement the existing and emerging hiPSC models for PD/DLB, we have generated an allelic series of αSyn over-expressing hESC lines on the same isogenic background. An unresolved question is whether pluripotent stem cell lines, with elevated levels of αSyn, can undergo efficient differentiation into dopaminergic and cortical neurons to model PD and DLB, respectively. We took advantage of our isogenic collection of hESC lines to determine if increased expression of αSyn affects neural induction and neuronal differentiation. Clonal hESC lines with significantly different levels of αSyn expression proliferated normally and maintained expression of pluripotent markers, such as OCT4. All cell lines efficiently produced PAX6+ neuroectoderm and there was no correlation between αSyn expression and neural induction efficiency. Finally, global transcriptomic analysis of cortical differentiation of hESC lines with low or high levels of αSyn expression demonstrated robust and similar induction of cortical neuronal expression profiles. Gene expression differences observed were unrelated to neural induction and neuronal differentiation. We conclude that elevated expression of αSyn in human pluripotent stem cells does not adversely affect their neuronal differentiation potential and that collections of isogenic cell lines with differing levels of αSyn expression are valid and suitable models to investigate synucleinopathies.


INTRODUCTION
Multiple lines of evidence have implicated αSyn as a major pathological driver in PD (Polymeropoulos et al., 1997;Spillantini et al., 1997;Singleton et al., 2003;Chartier-Harlin et al., 2004). The genetic forms of PD have led to the development of patient-derived and engineered pluripotent stem cell models to gain mechanistic insights into synucleinopathies (Singh Dolt et al., 2017). A key question for the validity of such models is whether mutant or elevated αSyn expression disrupts early neural induction or neuronal differentiation, thereby limiting later phenotypic analysis and disease modelling. Current data in the literature is conflicting, with some studies proposing that high levels of αSyn alter cell fate and differentiation (Schneider et al., 2007;Oliveira et al., 2015), whereas others have found no evidence for impaired neurogenesis (Devine et al., 2011;Prots et al., 2018;Brazdis et al., 2020). Furthermore, αSyn is known to bind DNA and alter expression of other genes, however the exact mechanism of this affect is not known (Surguchev and Surguchov, 2017). In this study, we aimed to generate an isogenic collection of hESC lines with differing levels of αSyn expression and use this to systematically investigate if elevated αSyn expression affects cortical neuron differentiation. If either the efficiency of neural induction or fidelity of neuronal differentiation is affected, then this may be a source of bias in studies comparing cultured neurons with differential αSyn expression. Isogenic cellular models provide the best experimental system to test this directly.
Studies utilising αSyn null mouse models showed no major differences in neuronal development and overall brain structure (Abeliovich et al., 2000;Specht and Schoepfer, 2001;Greten-Harrison et al., 2010). However, subtle differences in dopaminergic neurotransmission were reported and these were age-dependent (Al-Wandi et al., 2010;Anwar et al., 2011;Connor-Robson et al., 2016). Rodent models employing transgenic over-expression of wild-type αSyn have reported impaired adult neurogenesis, but have not determined whether this is a neurodevelopmental defect from birth or a toxic effect of αSyn over-expression (Winner et al., 2012). We have previously shown that hiPSCs with an SNCA triplication mutation have a two-fold increase in αSyn protein levels, and this did not significantly impair dopaminergic neuronal differentiation (Devine et al., 2011). This was also shown by other independent groups using hiPSCs harbouring the SNCA triplication (Byers et al., 2011;Lin et al., 2016) and SNCA duplication mutations (Prots et al., 2018;Brazdis et al., 2020). However, other studies using SNCA triplication mutation-derived hiPSCs reported reduced neuronal differentiation capacity and impaired neurite outgrowth (Flierl et al., 2014;Oliveira et al., 2015). Furthermore, other studies using lentiviral-mediated Doxinducible expression of αSyn, in human iPSC-derived neuronal progenitors, reported altered cell fate and impaired differentiation of neural stem cells into neurons (Schneider et al., 2007;Zasso et al., 2018). The cellular and rodent data described so far report conflicting roles for αSyn in neurogenesis, highlighting the need for work to clarify this point.
In this study, we generated an allelic series of clonal isogenic hESC lines expressing a broad range of αSyn, including lines expressing the protein at supraphysiological levels. We then set out to determine whether elevated αSyn expression impairs cortical neuron differentiation of human pluripotent stem cells. We show, using marker analysis during multiple rounds of differentiation and unbiased transcriptomic analysis, that cortical neuron differentiation is not impaired by increased expression of αSyn. The strength of our approach is the use of an isogenic collection of hESC lines with differing and sustained αSyn expression during differentiation, robust cortical neuron differentiation protocols and unbiased analysis of the transcriptome before and after differentiation. This study provides a valuable collection of cell lines for the neuroscience community and provides important evidence for the validity of hESC/iPSC disease models with elevated levels of αSyn expression.
The impact of αSyn on cell proliferation was assessed using an MTS assay. Over a 10-day period, six low αSyn lines (S9, S12, S13, S17, S22, S34) and 3 high αSyn lines (S8, S36, S37) were examined and each showed a similar rate of proliferation (Supplementary Figure S1). There was no significant difference in the proliferation rate of low and high αSyn groups and linear regression analysis showed no correlation between SNCA expression and proliferation rate of the examined cell lines (R 2 = 0.003, p = 0.884).

Elevated αSyn Expression in hESC Lines Is Compatible With Cortical Neuronal Differentiation
Multiple clonal lines were differentiated into cortical neural progenitors using a dual SMAD inhibition protocol ( Figure 2A) (Chambers et al., 2009;Shi et al., 2012). RNA Frontiers in Cell and Developmental Biology | www.frontiersin.org May 2022 | Volume 10 | Article 898560 was isolated and qRT-PCR used to quantify NCAM and MAPT expression at day 11 and compared to SNCA expression at the start of differentiation ( Figure 2B). There was no correlation in day 11 NCAM (R 2 = 0.008, p = 0.761) nor day 11 MAPT (R 2 = 0.019, p = 0.461) levels relative to SNCA expression for the clonal lines ( Figure 2C). Western blotting at day 11 of differentiation also confirmed differentiatial protein expression of αSyn in low and high αSyn cell lines ( Figure 2D). hESC lines with low or high αSyn expression differentiated equally well to form PAX6-positive neuroectoderm by day 12 ( Figure 3A). A direct comparison of Shef4 and selected clonal hESC lines with human iPSC lines from a healthy control (NAS2) and triplication SNCA patient (AST18) showed robust and similar expression of cortical progenitor markers PAX6, OTX1/2, and Vimentin at day 12 of neural induction (Supplementary Figure S2). By day 26 of cortical neuron differentiation the mRNA expression of PAX6 and the cortical marker, TBR1, were similarly induced in the parental Shef4 cell line and two high αSyn clones (S8, S37) ( Figure 3B). Furthermore, by day 45, there was similar protein expression of early cortical markers, TBR1 and CTIP2, in low and high αSyn groups ( Figure 3C). Despite the difference in αSyn levels, both groups formed neuronal networks with the same degree of efficiency based on βIII-Tubulin immunostaining ( Figure 3D). This data provides evidence that αSyn expression levels do not influence neuronal induction and differentiation potential across multiple clonal hESC lines.
To determine if transgenic SNCA expression was maintained during differentiation, total and transgenic SNCA levels were measured by qRT-PCR ( Figure 4A). The primers for total SNCA were designed to target the coding region and primers for transgenic SNCA targeted the IRES region ( Figure 4A). Total and transgenic levels of human SNCA were measured in selfrenewing hESCs (day 0) and day 25 differentiated cortical cells. Day 25 was a suitable time point as the neural induction period has been completed and immature cortical neurons are forming.
Total SNCA expression at day 0 was significantly higher in the high αSyn vs. low αSyn hESC group ( Figure 4B). Importantly, at day 25, the elevated expression of SNCA was maintained between the high αSyn vs. low αSyn cortical neuron group (p < 0.01). Total SNCA levels did not significantly increase over each time point, day 0 vs. day 25, for both high αSyn and low αSyn groups. Transgenic SNCA levels in undifferentiated hESCs were significantly higher in the high αSyn vs. low αSyn group ( Figure 4B) (p < 0.01). Day 25 high αSyn vs. low αSyn cortical neurons also had higher transgenic SNCA (p < 0.01). Transgenic SNCA levels did not significantly increase or decrease between day 0 and day 25 for the high αSyn and low αSyn groups. Immunostaining and western blotting confirmed that high αSyn expression was maintained following cortical neuron differentiation ( Figures 4C,D). At day 72 of cortical FIGURE 2 | Level of αSyn does not impair neural induction. (A) Schematic summarising key stages of the cortical neuron differentiation protocol. Cell lifts were carried out at days 12, 17, and 25. Dual SMAD inhibition using SB431542 (SB) and LDN-193189 (LDN) was employed early to drive neural induction and cortical identity, and the growth factors BDNF and GDNF were used to promote neuronal maturation after day 25. (B) Quantitative RT-PCR data measuring total SNCA expression in undifferentiated transgenic Shef4 cell lines, and NCAM and MAPT mRNA levels at day 11 of differentiation. Data shown as relative fold-change to expression in the parental Shef4 line, and error bars represent the standard error of the mean (SEM). (C) Linear regression analysis of SNCA, NCAM, and MAPT expression levels was performed (p = 0.174 and p = 0.191 for NCAM and MAPT, respectively). (D) Western blot for αSyn and ß-actin of whole cell lysates from cortical neuroectoderm (day 11) differentiated from transgenic hESC lines.
Frontiers in Cell and Developmental Biology | www.frontiersin.org May 2022 | Volume 10 | Article 898560 differentiation the hiPSC line, AST18, containing an SNCA triplication mutation exhibited a 3-fold increase in αSyn expression compared to a control hiPSC line, NAS2, derived from a 1st-degree relative (Devine et al., 2011) ( Figure 4D). The S37 transgenic cell line maintained high αSyn expression similar to AST18 neurons, while the S36 hESC line, that expressed the most SNCA at day 0 ( Figure 1B), now had low αSyn expression by day 72 of differentiation, suggesting significant transgene silencing ( Figure 4D). This is not unexpected as transgene expression is contingent on several factors, including site of integration, number of copies of the transgene, as well as transgene silencing over cell passage or differentiation (Liu et al., 2009). The transcriptomic profile of self-renewing hESC lines (day 0) and differentiated cortical neurons (day 25) with high αSyn and low αSyn were investigated using RNA-seq analysis. A total of 17 samples were subjected to UPX 3′-sequencing to measure polyadenylated mRNA transcripts across 4 sample groups 1) low αSyn hESCs, 2) high αSyn hESCs, 3) low αSyn cortical neurons, and 4) high αSyn cortical neurons ( Figure 5A). Principal Component Analysis (PCA) showed a clear segregation between hESCs and cortical neurons, represented by 83% variance on PC1 ( Figure 5B). The PC2 axis was representative of αSyn-related differences between hESC and cortical neuron samples. The hESC samples were intermixed, but cortical neuron samples segregated based on αSyn expression, however, the variance along PC2 was small (5%). This is further corroborated by hierarchical clustering analysis based on all differentially expressed genes across the four sample groups ( Figure 5C). A pairwise comparison of self-renewing high αSyn vs low αSyn hESC lines found only 2 genes (CBR1, CTNNA3) were significantly differentially expressed between the two groups ( Figure 6A), indicating αSyn does not impact on the pluripotency transcriptome. As expected, comparisons of low αSyn cortical neurons and low αSyn hESCs, as well as high αSyn cortical neurons and high αSyn hESCs, identified large numbers of differentially expressed genes (3,429 and 4,188, respectively, Figures 6B,C; Supplementary Table S1). Seven of the top ten most significantly upregulated genes were the same for these two pairwise comparisons, including SOX5, NPAS3, MALAT1, MAP2, QKI, PCDH9, and AC10729.1 ( Figure 6D). Most of these genes have a role in neurogenesis (Izant and McIntosh, 1980;Hardy et al., 1996;Wunderle et al., 1996;Strehl et al., 1998;Brunskill et al., 1999;Chen et al., 2016). Similarly, five of the top ten significantly downregulated genes were the same in the hESCs vs cortical neuron comparisons for high and low αSyn cell lines, including DPPA4, L1TD1, RBM47, XACT and DNMT3B Frontiers in Cell and Developmental Biology | www.frontiersin.org May 2022 | Volume 10 | Article 898560 9 ( Figure 6D). Most of these top downregulated genes have roles in pluripotency (Madan et al., 2009;Hu et al., 2012;Vallot et al., 2013;Emani et al., 2015;Radine et al., 2020). There was a large overlap in the genes that were significantly upregulated and downregulated in cortical neuron vs hESC groups with high or low αSyn ( Figure 7A). There was also a substantial overlap of gene ontology (GO) terms significantly upregulated and downregulated in the high or low αSyn comparisons ( Figure 7B). The most significantly enriched KEGG pathway in both high and low αSyn cortical neuron vs. hESC comparisons for upregulated genes was "axon guidance" (KEGG:04360). Five other top ten KEGG pathways linked with upregulated genes were the same for both the high αSyn and low αSyn cortical neuron vs hESC comparisons ( Figure 7C). The "ribosome" term (KEGG:03,010) was the most significant downregulated pathway common to both comparisons. Interestingly, the Parkinson's disease KEGG pathway was downregulated in high αSyn cortical neurons vs high αSyn hESCs, due to the Frontiers in Cell and Developmental Biology | www.frontiersin.org May 2022 | Volume 10 | Article 898560 10 downregulation of several mitochondrial genes and cytochrome c oxidase genes, including NDUFB9, MT-CO2, MT-CO3, MT-ATP6, MT-ATP8, MT-CYB, COX7C, and CYCS. This may reflect a mitochondrial phenotype caused by elevated αSyn expression in cortical neurons. The data so far suggests that the process of cortical neuron differentiation induced a large number of significant gene expression changes, and that the level of αSyn in these cells does not impair this process.
The key group comparison was that of high αSyn cortical neurons vs. low αSyn cortical neurons. The volcano plot, in comparison to the cortical neurons vs. hESC plots, identified a total of 47 differentially expressed genes ( Figure 8A; Supplementary Table S1). KEGG pathway analysis did not show any relevant neurogenesis-related pathways enriched in high αSyn vs low αSyn cortical neurons ( Figure 8B). Selected gene expression analysis revealed that the pluripotency markers POU5F1 (OCT4) and NANOG were significantly downregulated and neurogenesis markers, ASCL1, MYT1L, and POU3F2, were significantly upregulated in cortical neuron samples relative to hESC samples for both high αSyn and low αSyn groups ( Figure 8C) (Nichols et al., 1998;Chambers et al., 2003;Vierbuchen et al., 2010). Analysis of genes implicated in axon guidance, including MAP2, SLIT2, EPHB1 and NCAM1 showed no difference in cortical neurons with low or high αSyn, except for NTN1 (p < 0.01) ( Figure 8D) (Kennedy et al., 1994;Izant and McIntosh, 1980;Borrell et al., 2012;Enriquez-Barreto et al., 2012). There were no significant differences in genes associated with synaptic development, in low αSyn and high αSyn cortical neurons, including CADM1, NLGN1, SNAP25, SYP, and DLG4 ( Figure 8E) (Wiedenmann and Franke, 1985;Oyler et al., 1989;Cho et al., 1992;Ichtchenko et al., 1995;Stagi et al., 2010). Other markers were explored to ascertain telencephalic development. FOXG1 and PAX6, cortical progenitor markers, and CTIP2 and TLE4, markers of deep cortical layers, showed no significant differences in low and high αSyn cortical neuron groups ( Figure 8F) (Raciti et al., 2013). The superficial layer marker SATB2, was not significantly different between the low and high αSyn cortical neurons, which is in keeping with its involvement in the later stages of cortical development (Mariani et al., 2012). Expression of the neuronal migration marker, DCX, also presented no significant variation between low or high αSyn cortical neurons ( Figure 8F) (Gleeson et al., 1999).

DISCUSSION
In this study, we generated and investigated 29 SNCA transgenic hESC lines and found no correlation between αSyn expression level and neural differentiation potential. We further addressed the hypothesis that increased αSyn expression does not impair cortical neuronal differentiation of human pluripotent stem cells. Across multiple clonal lines, multiple rounds of cortical neuron differentiation using two robust differentiation protocols and unbiased transcriptomic analysis, we show data to support this hypothesis. The level of SNCA expression, whether normal or increased, did not have a significant impact on the efficiency of cortical neuron differentiation. These findings resolve conflicting data in the field, and are highly relevant to studies utilising human pluripotent stem cells (hESCs or iPSCs) with differential αSyn expression to model synucleinopathies, such as PD, DLB and multiple system atrophy (MSA).
As expected, significant differential gene expression was observed when cortical neurons were compared to undifferentiated hESCs for both high αSyn and low αSyn hESCs. Gene ontology (GO) and KEGG pathway analysis identified significant overlap in the group comparisons with high αSyn or low αSyn hESCs and their differentiated counterparts. Furthermore, most of the top differentially expressed genes were the same in these comparisons, indicating neurogenesis proceeded in a similar manner despite the level of αSyn expression. The gene expression differences that were unique to the high αSyn hESCs or to the low αSyn hESCs could be due to a number of factors, including clonal variation, experiment-to-experiment variation, or due to a phenotype of elevated αSyn expression, such as mitochondrial dysfunction. When high αSyn and low αSyn cortical neurons were directly compared to each other, only 47 genes were differentially expressed, and pathway analysis did not reveal anything relevant to cortical neuron induction or differentiation. Furthermore, read count analysis for well-characterised genes with known roles in neurogenesis and the development of cortical identity confirmed cortical differentiation occurred similarly between the high αSyn and low αSyn groups.
Research groups, including ourselves, have used hiPSCs derived from patients with SNCA multiplication mutations to model synucleinopathies (Byers et al., 2011;Devine et al., 2011;Flierl et al., 2014;Oliveira et al., 2015;Lin et al., 2016;Prots et al., 2018;Chen et al., 2019;Brazdis et al., 2020). We previously showed, using eight clonal hiPSC lines, harbouring the SNCA triplication mutation and six hiPSC lines from a non-affected first-degree relative, that the main sources of variation in dopaminergic differentiation efficiency were due to differences between clonal lines, reprogramming efficiency, and the process of neuronal differentiation itself (Devine et al., 2011). Whilst cellular reprogramming is a powerful method for generating human disease models, reprogramming can be incomplete and can introduce coding mutations or large chromosomal abnormalities that may lead to altered differentiation potential of different clones (Mayshar et al., 2010;Boulting et al., 2011;Hussein et al., 2011). Furthermore, in these patient-derived iPSCs the size of the multiplicated region is variable and other adjacent genes such as MMRN1 may be incorporated (Ross et al., 2008). Overexpression of other coding genes in these iPSC lines presents another potential confounding factor when investigating and interpreting neuronal differentiation potential. A number of these caveats apply to studies differentiating hiPSCs with a SNCA triplication into neuronal progenitors or dopaminergic neurons (Flierl et al., 2014;Oliveira et al., 2015). In particular, limited clonal lines were examined and siRNA knock-down of SNCA "rescue" of dopaminergic differentiation was partial (Oliveira et al., 2015).
To further highlight the point regarding clonal variation and conflicting reports in the literature, recent studies have shown, using hiPSCs from patients with a SNCA duplication mutation, that cortical neurons (Prots et al., 2018) and dopaminergic neurons can be efficiently and comparably generated using these hiPSC lines (Brazdis et al., 2020). Other independent groups have also used hiPSCs with a SNCA triplication mutation to show functional dopaminergic neuron generation comparable to control hiPSC lines (Byers et al., 2011;Lin et al., 2016). We recently showed that reducing SNCA alleles in isogenic hESC lines also does not affect dopaminergic neuronal differentiation; wild type, SNCA +/− and SNCA −/− hESC lines showed no differences in differentiation into FOXA2:TH double-positive dopaminergic neurons (Chen et al., 2019). This study, and the triplication and duplication SNCA hiPSC studies (Devine et al., 2011;Brazdis et al., 2020) strongly predict that the isogenic collection of hESC lines generated here will have similar efficiencies of midbrain dopaminergic differentiation. In the studies reporting no impairment of dopaminergic neuron differentiation with a SNCA multiplications, PD-related phenotypes, including reduced synchronous firing on microelectrode recordings and increased susceptibility to oxidative stress were observed (Lin et al., 2016;Brazdis et al., 2020). While there may be αSyn-related phenotypic differences or inherent vulnerabilities due to elevated αSyn expression in mature neurons, this does not imply that the process of neuronal differentiation itself is impaired.
When αSyn was over-expressed using a lentiviral system in hESC-derived neuroectoderm impaired neuronal patterning and acute toxicity were reported (Schneider et al., 2007). Dopaminergic and GABAergic neuron populations were affected by αSyn overexpression, but this was only quantified in a single hESC line (H9), and the toxicity shown to be occurring could result in selective neuronal loss, making the interpretation of differentiation marker analysis challenging (Schneider et al., 2007). Similar caveats apply to work performed using AF22 neural stem cells with doxycyclineinducible αSyn expression. Exposure to doxycycline may cause pleotropic effects, and αSyn expression was not sustained throughout differentiation (Zasso et al., 2018).
It is important to consider that neuronal differentiation protocols often require optimisation for each clonal line and small variations in ligand concentrations, specifically CHIR99021 in the midbrain dopaminergic differentiation protocol, can impair neurogenesis. For example, in the study by Oliveira et al., 2015 the dopaminergic differentiation protocol yielded less than eight percent TH-positive neurons in the control lines. Whilst our study does not yield insight into the impact of elevated αSyn expression on dopaminergic differentiation, a subject for future studies, the cortical neuron protocols used were robust with the vast majority of cells producing cortical neurons.
The next steps for these cell lines are to investigate differential susceptibility to know triggers of synucleinopathy, such as environmental risk factors that affect autophagy and lysosomal biology, as well as compounds that inhibit mitochondrial function. We have previously shown that reduction of SNCA alleles in hESC-derived neurons protects against synucleinopathy (Chen et al., 2019), and it is predicted that increasing αSyn expression will make neurons more susceptible to conditions that promote Parkinson's and other Lewy body conditions.
In summary, this work has produced an allelic series of isogenic clonal hESC lines with differing levels of αSyn expression, and we have demonstrated that elevated αSyn expression does not impair cortical neurogenesis. This supports the validity of using human pluripotent stem cells, such as iPSCs with SNCA multiplications and transgenic hESC lines, to model synucleinopathies. Since this collection of cell lines are isogenic and the only genetic variable is the level of αSyn expression, their utility for disease modelling and research into novel therapeutics is very significant. However, close matching of the differentiation stage and maturity of neurons from the different clones is still required for accurate interpretation of results. The practical application of these cell lines is most promising for investigating treatments that are targeted against αSyn itself or downstream effects of its over-expression. per manufacturer's protocol. Purified and digested DNA was ligated to PCR products at 16°C overnight in a final volume of 20 μl using 2 U of T4 DNA ligase and 1x ligation buffer (Roche, 10481220001). Plasmid DNA was transformed into TOP10 chemically competent cells (Invitrogen, c4040-10) by the heatshock method. Plasmid DNA was extracted using QIAprep ® -Spin Miniprep kit (Qiagen,27,104) or Maxiprep (Qiagen, 12,662) systems per manufacturer's protocol, then desalted using Millipore centrifugal filter units (Millipore, UFC503024 24PK). The final pCAG-SNCA-IRES-Venus construct contained human SNCA, internal ribosome entry site (IRES) and Venus expression cassette under the constitutive pCAG promoter, as well as a puromycin resistance gene (Puro r ) driven by the PGK promoter. The control construct, pCAG-IRES-Venus, contained the same elements except for the SNCA gene ( Figure 1A).

Generation of Transgenic hESC Lines by Nucleofection
The Neon Transfection System (Invitrogen, MPK5000) was used for nucleofection of both constructs into Shef4 hESCs as per the manufacturer's protocol. 1 μg/ml puromycin was added to the culture media for at least two weeks to isolate clones and colonies, manually picked for expansion and cryopreservation. Clones were screened for SNCA over-expression by qRT-PCR.

Quantitative RT-PCR
The MasterPure ™ Complete DNA and RNA Purification kit (Epicentre, MC85200) or the RNeasy kit (Qiagen, 74,104) was used for RNA extraction. Genomic DNA was removed using DNase I (Promega, M6101). cDNA was synthesised from 500 ng

MTS Assay
Cell proliferation was assessed using a colorimetric assay, CellTiter96 ® AQ ueous One Solution Cell Proliferation Assay or Figure S1).

Transcriptomic Analysis
Total RNA was isolated with RNeasy kit (QIAGEN, 74104). RNA integrity (RINe ≥ 7) was confirmed using Tapestation 4200 (Agilent). The median RINe score was 9.2. Samples were collected at day 0 and day 25, across three sets of cortical neuron differentiations ( Figure 5A). The samples were processed by Qiagen Genomic Services using their QIAseq UPX 3′ Transcriptome kit and libraries were sequenced on a NextSeq500 instrument. The total number of polyadenylated 3′ transcript reads and the mean number of reads per unique molecular identifier were counted. The raw transcript counts were analysed using DESeq2 (Love et al., 2014) differential expression analysis in R studio. Pairwise analysis was used to compare hESC vs. cortical neurons with high or low αSyn. A padj cut off value of 0.05 and a log 2 fold-change cut off value of 1.2, were used. KEGG pathway analysis was performed using the g:GOSt function in the CRAN gprofiler2 package (Raudvere et al., 2019). RNA-seq data has been deposited on the Gene Expression Omnibus (GEO Accession number: GSE195877).

Immunocytochemistry
Immunostaining was performed on cells cultured on 13-mm glass coverslips or in Ibidi 8-well plates. Spent medium was removed and cells fixed with 4% PFA for 15 min. Following 3 PBS (ThermoFisher Scientific) washes, the cells were permeabilised and blocked with 2% goat (or 2% donkey) serum (Sigma) in 0.1% Triton-X-100 (Fisher) in PBS for 45 min prior to overnight incubation at 4°C with primary antibodies. The primary antibodies used were ß-III tubulin (1:1,000,

Image Capture, Processing and Quantification
The Eclipse Ti (Nikon) and/or the Axio Observer (Zeiss) microscopes were used to acquire the images presented. Huygens Software (Scientific Volume Imaging) was used for deconvolution of Z stack images (at least 10 images/stack and maximum intensity pixel projection used). Fiji software was used for image analysis and quantification, with identical brightness and contrast values used for each image channel and for each experiment. Macro scripts were used to split the image channels, threshold and binarize for quantification.
The Pierce ™ ECL Western Blotting Substrate was again used prior to imaging the membrane.

Statistics
Statistical tests were performed using SPSS v23 and include the Student's t-test or t-test with Welch's correction, linear regression analysis and the Mann-Whitney U test for non-parametric data. Each Figure legend details which test was used for each statistical comparison. Significance level cut off was p < 0.05.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
TK conceived the project. TK, AN, RY designed the experiments. AN and RY performed the experiments and analysed the data. RB performed bioinformatic analysis. AN and TK wrote the manuscript, and all authors reviewed and approved the manuscript.

ACKNOWLEDGMENTS
The authors with to thank the Wellcome Trust and the MRC for funding this project. We are very grateful to Prof Steven Pollard for critical comments on the manuscript, and Dr James Ashmore for bioinformatics guidance. We also sincerely thank Prof John Hardy and Dr William Hamilton for pcDNA3.1-SNCA and PGK-Puro-pCAGS plasmids, respectively, and Prof David Hay for Shef4 hESCs.