Post-translational proteomics platform identifies neurite outgrowth impairments in Parkinson’s disease GBA-N370S dopamine neurons

Summary

Variants at the GBA locus, encoding glucocerebrosidase, are the strongest common genetic risk factor for Parkinson’s disease (PD). To understand GBA-related disease mechanisms, we use a multi-part-enrichment proteomics and post-translational modification (PTM) workflow, identifying large numbers of dysregulated proteins and PTMs in heterozygous GBA-N370S PD patient induced pluripotent stem cell (iPSC) dopamine neurons. Alterations in glycosylation status show disturbances in the autophagy-lysosomal pathway, which concur with upstream perturbations in mammalian target of rapamycin (mTOR) activation in GBA-PD neurons. Several native and modified proteins encoded by PD-associated genes are dysregulated in GBA-PD neurons. Integrated pathway analysis reveals impaired neuritogenesis in GBA-PD neurons and identify tau as a key pathway mediator. Functional assays confirm neurite outgrowth deficits and identify impaired mitochondrial movement in GBA-PD neurons. Furthermore, pharmacological rescue of glucocerebrosidase activity in GBA-PD neurons improves the neurite outgrowth deficit. Overall, this study demonstrates the potential of PTMomics to elucidate neurodegeneration-associated pathways and potential drug targets in complex disease models.

Graphical abstract.

Introduction

Heterozygous mutations in the glucocerebrosidase gene (GBA) are the strongest common genetic risk factors for Parkinson’s disease (PD) and are present in around 5%–10% of patients with PD,^2,3 resulting in lower age of onset and exacerbating disease progression, including an increased risk of dementia.^4,5 However, the exact mechanisms leading from dysfunction of the enzyme glucocerebrosidase (GCase) encoded by GBA to PD pathogenesis and neurodegeneration remain unclear. GCase is a lysosomal enzyme that degrades glucosylceramide into glucose and ceramide. Homozygous mutations in GBA lead to severe GCase deficiency and glucosylceramide accumulation in lysosomes, preventing their normal function⁶ and causing the lysosomal storage disorder Gaucher’s disease (GD). Heterozygous GBA mutation carriers have a 10%–30% risk of developing PD.^7,8 Over 300 different disease-associated changes in GBA have been identified, of which the N370S mutation is the most common.⁹

Studies of how GBA mutations may lead to PD pathogenesis have generated diverse hypotheses involving either loss-of-function or toxic gain-of-function mechanisms of a number of cellular pathways including the autophagy-lysosome pathway (ALP), endoplasmic reticulum (ER) stress, or lipid dyshomeostasis.^10,11 Decreased GCase activity has been demonstrated in patients with genetic and sporadic PD and also as a consequence of aging, suggesting that loss of enzyme activity may be causal for disease.^12,13 Human cell and animal models of GCase deficiency, obtained by genetic knockout or chemical inhibition, highlighted ALP dysfunction and accumulation of α-synuclein (α-syn).^14–16 This could be the result of reduced ALP function hindering the degradation of α-syn either directly due to a loss of GCase activity or due to a toxic gain of function of mutant GCase.^11,17–20

Induced pluripotent stem cell (iPSC)-derived dopamine neurons from patients with PD have been fundamental in understanding the molecular pathology of PD. Studies using iPSCs derived from patients with GBA mutations show increased dopamine oxidation, decreased lysosomal function, accumulated glycosphingolipids including glucosylceramide, and buildup of α-syn.^17–19 Furthermore, we have previously demonstrated GBA-N370S phenotypes including ALP dysfunction, augmented ER stress response, and increased α-syn release in patient iPSC-derived dopamine neurons.^11,20

Previously, proteomic and single-cell transcriptomic strategies have been used to characterize dysfunction in iPSC-derived neurons from patients with GBA mutations, highlighting the utility of omics approaches.^18,20 However, in order to fully understand the cellular effects of GBA mutations, an unbiased and global characterization of the neuronal proteome is required. Furthermore, given a large proportion of regulation of protein activity, localization, and function is not controlled by changes in protein levels but through post-translational modifications (PTMs),²¹ simultaneous assessment of multiple protein PTMs is essential in understanding complex diseases mechanisms and models.

Numerous PTMs have been associated with cellular dysfunction and PD pathology. Phosphorylation, which is the most studied PTM, has been demonstrated to be involved in regulating activity of major PD-related pathways such as ALP and mitophagy.²² In addition, the phosphorylation states of key proteins involved in neurodegeneration such as α-syn and tau can affect their propensity for aggregation.²³ Another major contributor to PD pathology is oxidative stress, as increased levels of reactive oxidative/nitrosative species cause reversible and irreversible modifications of redox-sensitive cysteine residues, affecting both the structure and function of proteins.^24,25 Additionally, N-linked glycosylation of proteins is a highly relevant PTM for understanding the proteome, with a large number of proteins shown to have N-linked glycosylation.²⁶ The prevalence of this modification and the role sialylated N-linked glycosylation plays in neural development and function, including in processing and transport of GCase and other key lysosomal proteins through the ER and Golgi,^3,27,28 highlights the need to assess the abundance of sialylated N-linked glycosylation in the proteome of disease models. We have previously examined the effect of knockdown of the PD-linked gene PRKN (PARK2) on the proteome of iPSC-derived neurons, focusing on phosphorylation and oxidized cysteine-containing proteins, resulting in identification of a number of phenotypes.^29–31

Applying a multi-part PTMomics enrichment strategy,³² we have isolated and quantified peptides with phosphorylation, reversible cysteine-modification, and sialylated N-linked glycosylation simultaneously. We applied this methodology to iPSC-derived dopamine neurons from patients with GBA-N370S PD and healthy age-matched controls, identifying large numbers of dysregulated native and modified proteins. Quantification of numerous sialylated glycopeptides from lysosomal proteins identified disturbances in the ALP that coincided with upstream perturbations in mammalian target of rapamycin (mTOR) phosphorylation levels. Integrated pathway analysis of all PTMomic datasets revealed an enrichment of dysregulated native/PTM proteins related to neuritogenesis, which we validated in functional assays, showing significant defects in neurite outgrowth in GBA-N370S patient neurons. Knockdown of tau, a predicted upstream regulator of GBA-induced dysfunction, caused a significant decrease in neurite outgrowth, whereas normalizing GCase activity using a small-molecule chaperone significantly improved the neurite outgrowth defects. Together, this study highlights the potential of this workflow for identifying phenotypes and their underlying biology as well as potential drug targets in disease models.

Results

Proteomic analysis reveals the post-translational proteome of human midbrain dopaminergic neurons

To assess the proteome and PTMome in human dopaminergic neuronal cultures, four age- and sex-matched GBA-N370S mutation carriers diagnosed with PD and four control iPSC lines were differentiated into dopaminergic neuronal cultures with comparable efficiency (Figures 1A and S1A–S1D; Table 1). Neurons derived from GBA-N370S mutation carriers exhibited a 50% decrease in GCase activity without significant changes in GCase protein expression (Figures 1B, S1E, and S1F).

Figure 1. Proteomic analysis on GBA and control iPSC-dopamine neurons achieves clear patient stratification.

(A) Immunofluorescence staining for tyrosine hydroxylase (TH; green), β-III-tubulin (TUJ1, yellow), and FOXA2 (red) on differentiation day 35 GBA patient and control neurons. Nuclei stained with DAPI (blue). Scale bar: 50 μm.

(B) Glucocerebrosidase (GCase) enzyme activity, relative to average of controls, of the GBA patient and control iPSC-dopamine neurons included in the proteomic analysis (n = 4 patients with GBA and 4 controls; mean ± SEM). *p % 0.05 (Student’s t test).

(C) Schematic representation of the preparation and enrichment workflow for the proteomic analysis of GBA patient and control iPSC-dopamine neuron cell lysates.

(D) Overview of the identified non-modified proteins and phosphorylated/glycosylated/cysteine-modified peptides. Normalized peptides represent number of individual phosphorylated/glycosylated/cysteine-modified peptides normalized to levels of the corresponding non-modified proteins.

(E) Venn diagram showing the overlap between the resulting differentially abundant non-modified proteins (GBA-PD/control abundance ratio > 1.2, p ≤ 0.05 [t test with Benjamini-Hochberg correction (false discovery rate [FDR] 0.1)]) and modified peptides (GBA-PD/control abundance ratio >1.3, coefficient of variation [CV]% ≤ 30) in the four groups.

(F and G) Principal-component analysis (PCA) plot based on the protein expression data from iPSC-dopamine neurons derived from the four patients with GBA (GBA 2; GBA-PSP) and healthy controls (F) with GBA 2 and (G) without GBA 2 included in the analysis.

Table 1. Information on the iPSC lines included in this study.

Study ID	Donor ID	Clone	Genotype	Age	Gender	Characterization
GBA 1	MK088	07	N370S/WT	46	M	this study
GBA2	SFC834-03	01	N370S/WT	72	M	Fernandes et al.11
GBA 3	SFC871-03	09	N370S/WT	70	F	this study
GBA 4	MK082	31	N370S/WT	51	M	this study
GBA 5	MK071	03	N370S/WT	81	F	Fernandes et al.11
GBA 6	SFC848-03	02	N370S/WT	68	M	this study
Con 1	SFC841-03	02	WT/WT	36	M	Van Wilgenburg et al.79
Con 2	OX2	28	WT/WT	43	M	this study
Con 3	SFC856-03	04	WT/WT	78	F	Haenseler et al.81
Con 4	MK053	06	WT/WT	68	M	Lang et al.20
Con 5	OX1	19	WT/WT	36	M	Van Wilgenburg et al.79
Con 6	NHDF	1	WT/WT	44	F	Hartfield et al.80

WT, wild type; M, male; F, female.

Dopaminergic neuronal cultures were harvested, and peptides derived by tryptic digestion from neural proteins were labeled to allow multiplexing. The peptide mixture was subsequently subjected to the sequential enrichment strategy, followed by pre-fractionation and subsequent liquid chromatography with tandem mass spectrometry (LC-MS/MS) (Figure 1C). The MS analysis successfully isolated, identified, and quantified more than 5,000 non-modified proteins with high confidence across all 8 samples (Figure 1D). In addition, we obtained the first ever global characterization of several important subtypes of PTMs in human midbrain neurons. In total, 7,988 phosphorylation sites on 3,092 proteins and 2,862 sialylated N-linked glycosites on 1,055 proteins were identified. In addition, 11,148 reversible cysteine modifications on 4,456 proteins were detected (Figure 1D), providing a comprehensive reference dataset of the proteome and PTMome in human iPSC-dopamine neurons.

Proteomic signatures of GBA-PD iPSC-dopamine neurons are specific and distinct from healthy controls

Principal-component analysis (PCA) of all non-modified proteins separated the four controls from the four GBA iPSC-dopamine neurons, with neurons from the GBA patient 2 line clearly separating from the others (Figures 1F and 1G). This was confirmed by hierarchical clustering of the 500 most abundant proteins (Figure S2A; Table S1A). Similarly, PCA on the post-translationally modified proteins separated patients from control iPSC-dopamine neurons and segregated GBA patient 2 from the remaining GBA neurons (Figures 2A, S3A, and S3D; Tables S1C, S1E, and S1G). As previously described elsewhere,²⁰ the patient no longer has the original PD diagnosis but was re-diagnosed with progressive supranuclear palsy (PSP) based on clinical assessment while the proteomic analysis was being performed. As these are two different diseases, which can initially present with similar symptoms, due to nigrostriatal pathology in PSP, and have different pathomechanisms, clinical outcomes, and treatments, we decided to remove the patient with PSP from the PD analysis of the remaining GBA mutation carriers, all of whom had a confirmed PD diagnosis (GBA-PD). This did not noticeably affect the clustering and dimensionality of the proteomic and PTMomic data when comparing GBA-PD with control neurons (Figures 1F, 1G, 2A, 2B, S3A, S3B, S3D, and S3E).

Figure 2. Assessment of glycosylated peptides reveals widespread changes in the lysosomal proteome in GBA-PD iPSC-dopamine neurons.

(A and B) PCA plot based on glycosylated protein levels in iPSC-dopamine neurons derived from the four patients with GBA (GBA 2; GBA-PSP) and healthy controls (A) with GBA 2 and (B) without GBA 2 included in the analysis.

(C) Heatmap of the abundance ratios (GBA-PD/control) of all glycosylated proteins identified by the proteomic analysis showing the three main functional clusters that segregate patients and controls.

(D) Heatmap of the abundance ratios (GBA-PD/control) of non-modified protein and PTMs for all established lysosomal proteins identified by the proteomic analysis. Unidentified proteins/PTMs are marked with gray. An asterisk (*) indicates PTM abundance ratio >1.3 and CV% <30. None of the non-modified proteins were significantly regulated.

(E) Visualization of enriched functional Gene Ontology (GO) term networks based on annotations of significantly regulated glycosylated proteins. The node size indicates the number of proteins connected to the GO term, and the color reflects functionally connected groups of terms. Only terms with an adjusted p value ≤ 0.05 shown (two-sided hypergeometric test with Bonferroni stepdown).

(F) Proteomic data from proteins related to ceramide catabolism with graph showing the levels of glycosylated peptides normalized to control and table showing the abundance ratio (GBA-PD/control) for the non-modified and glycosylated form including the glycosylation site (n = 3 GBA patients and 4 controls; mean ± SEM).

Comparing protein levels between the GBA patient 2 iPSC-dopamine neurons and the four healthy controls revealed levels of 250 proteins to be significantly altered (Table S2A). Pathway analysis of these proteins showed an enrichment of proteins involved in extracellular matrix organization and neurotransmitter release cycle, in addition to regulation of cell biology by calpain proteases, which are known regulators of tau phosphorylation and fragmentation (Figures S2B and S2C).^33,34 GBA patient 2 neurons had, compared with the remaining patients with GBA, significantly increased tau phosphorylation levels at five phosphorylated (phospho)-sites after normalizing to total tau, which was significantly decreased (Figure S2D; Tables S2B and S2C).

Analysis of the proteome in GBA-PD vs. control neurons resulted in the identification of 172 proteins that were significantly differentially abundant with a >1.2-fold change (Figure 1D; Tables S3A and S3B). To understand the contribution of PTMs to cellular dysfunction in GBA-PD, levels of PTM peptides were normalized to the level of the corresponding non-modified protein to avoid protein-level changes affecting the PTM analyses. This analysis resulted in more than 20,000 normalized PTMs (Tables S3C–S3H). Of these, levels of 1,337 reversible cysteine modifications, 701 phosphorylation sites, and 190 sialylated glycosites were altered by more than 1.3-fold in GBA-PD iPSC-dopamine neurons, and these PTMs were included in downstream analyses (Figure 1D; Tables S3C–S3H). In comparison, with GBA patient 2 included in the analysis, only 45 non-modified proteins, 761 reversible cysteine modifications, 322 phosphorylation sites, and 103 sialylated glycosites were identified as differentially abundant due to higher inter-sample variation (Tables S1B, S1D, S1F, and S1H).

When we compared the proteins, which were differentially expressed and/or showed changes in PTM levels, a total of 201 proteins were found to be regulated in at least two of the datasets, signifying increased regulation associated with GBA mutation-related disease mechanisms (Figure 1E; Tables S3B, S3D, S3F, and S3H). Out of the 5,005 identified master proteins, 3.4% showed significantly different protein abundance levels, and 19.3%, 9.6%, and 2.5% had differentially abundant cysteine modification, phosphorylation, and glycosylation levels, respectively (Figure S1G). The overlap between proteins with differentially abundant glycosylation and phosphorylation levels was limited, with only 5 proteins identified showing both (Figure 1E). However, when we examined all the proteins that had both glycosylation and phosphorylation sites identified (n = 130), there was a weak but significant correlation between levels of the two PTMs (Figure S1H).

Overall, these data demonstrate the ability of our comprehensive analysis to differentiate between iPSC-dopamine neurons derived from GBA-N370S carriers and controls and from patients with different forms of neurodegenerative diseases.

Assessment of sialylated glycopeptides reveals widespread changes in the lysosomal proteome of GBA-PD iPSC-dopamine neurons

Glycosylation is an abundant PTM, particularly on membrane-bound and lysosomal proteins, suggesting that their levels may be relevant in GBA-PD.^11,14,35 Hierarchical clustering and pathway analysis of the complete sialylated glycoprotein dataset demonstrated coverage of proteins involved in cell motility, neurite outgrowth, and lysosomal processes, and PCA of these peptides revealed modest separation of GBA-PD and control neurons, in addition to clearly identifying GBA patient 2 (GBA-PSP) (Figures 2A–2C). Using a previously published analysis of the lysosomal proteome,³⁶ we established that 40/72 (56%) of established lysosomal proteins were quantified by examining the sialylated glycoproteome with 33/72 (46%) proteins identified by quantifying non-modified peptides (Figure 2D; Tables S3A and S3C). Examination of the lysosomal proteome revealed a number of significantly altered lysosomal proteins in GBA-PD neurons, including the genome-wide association study (GWAS)-associated lysosomal integral membrane protein 2 (LIMP2)³⁷ and arylsulfatase B (ARSB)³⁸ (Figure 2D). Additionally, the analysis also confirmed our previous findings of increased levels of glycosylated LAMP1, LAMP2, and cathepsin D in GBA-PD iPSC-dopamine neurons¹¹ (Figure 2D).

In addition, pathway analysis of the differentially expressed sialated glycoproteome identified a number of significantly enriched pathways, including ceramide catabolic processes, driven by changes in glycosylation of the PD GWAS-associated genes acid ceramidase (ASAH1),³⁸ galactocerebrosidase (GALC),³⁹ prosaposin (PSAP),⁴⁰ and α-galactosidase A (GLA)⁴¹ (Figures 2E and 2F). Given the role of GCase in the catabolism of glucosylceramide to glucose and ceramide, the dysregulation of several enzymes involved in ceramide catabolism, which are also GWAS hits for PD, is of particular interest. These data indicate a highly relevant role for the sialated glycoproteome in identifying perturbation in lysosomal biology in neurons. In addition to lysosomal function, differentially regulated sialylated glycoproteins were enriched in processes including neuron projection development and axon extension, suggesting a potential phenotype in GBA-PD (Figures 2C and 2E).

PTMomics identifies tau and mTOR as key mediators in GBA-PD pathogenesis

To further understand the biology of GBA-PD dopamine neurons, we examined the significantly differentially abundant proteins (Figure 3A; Table S3B). Among the proteins with the highest abundance increase was the microtubule-associated protein tau (MAPT), a common GWAS-associated genetic risk factor for PD,⁴² in addition to a number of other proteins involved in microtubule dynamics, neurite outgrowth, and axon extension including tubulin β-2B chain (TUBB2B), MAP1B, MAP6, neuronal cell adhesion molecule (NRCAM), neuromodulin (GAP43), copine-1 (CPNE1), and calponin-2 (CNN2) (Figure 3A; Table S3B). Network analysis of the significantly regulated proteins revealed many of these proteins (MAPT, MAP2, MAP6, MAP1B, GAP43) to be part of a larger functional cluster, which also contained a larger number of affected phospho-sites (Figure 3B). Interestingly, a key protein in this functional cluster was apolipoprotein E (APOE), with levels of APOE demonstrated to be 0.77-fold decreased in GBA-PD (p = 0.02) (Figure 3B; Table S3B). Importantly, given the increased prevalence of dementia in GBA-PD, dementia-associated APOE variants cause decreased APOE levels and increased tau and α-syn pathology.^43,44 Also of relevance to tau function were the decreased levels of cysteine modification of the tau regulator glycogen synthase kinase 3β (GSK3β) at C76, a site which when modified has been shown to inhibit GSK3β activity, that we identified (Table S3H).^45,46

Figure 3. Tau and mammalian target of rapamycin (mTOR) are predicted as key regulators of the proteomic changes detected in GBA-PD iPSC-dopamine neurons.

(A) Volcano plot showing the fold change and significance level (−log10 [p value]) of all non-modified proteins with the dashed line corresponding to a p = 0.05. The labeled points represent the ten most highly up-/downregulated proteins in patients with GBA-PD relative to control.

(B) STRING network of connected, significantly regulated non-modified proteins with the circle color displaying protein abundance ratios (GBA-PD/control) and significantly regulated phospho-sites marked by surrounding halo, displaying the phosphorylation abundance ratio (GBA-PD/control).

(C) Heatmap of PD-GWAS proteins with altered protein/PTM levels displaying the abundance ratio (GBA-PD/control). Unidentified proteins/PTMs are marked with gray. An asterisk (*) indicates non-modified protein abundance ratio >1.2 (p < 0.05) or PTM abundance ratio >1.3 (CV% < 30).

(D) Pathway analysis identifying the predicted upstream regulators most highly enriched across all datasets. Data are presented as a heatmap of –log(p value), where increasing color intensity corresponds to increasing significance in enrichment, as indicated.

(E–J) Western blotting (E) and quantification of (F) microtubule-associated protein tau (Tau), (G) mTOR, (H) phospho-mTOR, (I) phospho-mTOR/mTOR ratio, and (J) the lipidated form of microtubule-associated protein light chain 3 (LC3II) levels in GBA-PD patient and control neurons from a representative independent differentiation (n = 3–5 iPSC lines). GBA 2; GBA-PSP shown on blots and excluded from quantification. Protein expression levels were normalized to β-actin and are shown relative to control. Mean ± SEM, *p % 0.05 (Student’s t test).

Given the altered levels of PD GWAS targets such as MAPT and ASAH1, we examined changes in protein levels and PTMs for a list of PD-associated genes compiled from the GWASdb SNP-Disease Associations dataset.⁴⁷ Of the 756 PD-associated genes, 138 were quantified, and 75 (54%) of these showed differentially abundant protein or PTM levels in GBA-PD neurons (Figure 3C), corresponding to a significant enrichment of affected proteins encoded by PD-associated genes compared with the general neuronal proteome (Fisher’s exact test, p < 0.0001). Particularly, changes in reversible cysteine modifications were observed on a large number of PD GWAS targets, suggesting differences in the cellular redox environment affecting redox active cysteines in these proteins (Figure 3C; Table S3H).

To understand the biological underpinning of the proteomic changes observed in GBA-PD, we performed a directional network analysis to identify the upstream regulators that cause the observed proteomic dysfunction in GBA-PD neurons (Figure 3D). Integration of this information across the different datasets highlighted MAPT (tau) as a potential regulator of proteomic dysfunction in GBA-PD (integrated p = 0.011), in addition to presenilin 1 (PSEN1), amyloid precursor protein (APP), and mTOR (Figure 3D). This strongly suggests that key proteins implicated in PD and Alzheimer’s disease (AD) could be central in mediating GBA-related proteomic changes.

Given these data, we confirmed the alterations in the levels of tau in the GBA-PD neurons by western blotting in samples from the same differentiation as the proteomics analysis, confirming the validity of the method (Figures S4A–S4B). In subsequent independent differentiations, we found that GBA-PD neurons again tended to have higher tau levels, although this varied between differentiations. (Figures 3E and 3F). The proteomic analysis showed a 1.3-fold increase in the active form of the key autophagy regulator mTOR (phosphorylated at residue S1261) in GBA-PD neurons (Figure 3B; Tables S3E and S3F).⁴⁸ This increase in phospho-mTOR was confirmed by western blot with a significant increase in levels of phospho-mTOR in GBA-PD neurons in both the proteomics differentiation (Figures S4C–S4E) and an independent differentiation (Figures 3E and 3G–3I). Given the role of mTOR in regulating multiple pathways, we examined the mTOR downstream targets p70S6K, 4E-BP1, and PKCα. This analysis revealed a significant increase in the ratio of active phospho-4E-BP1, whereas the other mTOR target showed unaltered phosphorylation levels at relevant mTOR-regulated phospho-sites (Figures S4M–S4Y).⁴⁹ Consistent with previous reports¹¹ and supporting mTOR activation, we found significantly increased levels of the lipidated form of MAP1A/1B-light chain 3 (LC3II), an important marker for autophagosome formation, indicating ALP dysfunction in both the proteomic (Figures S4F and S4G) and an independent differentiation (Figures 3E and 3J). Together, these data suggest that a combined analysis of the proteome and PTMome can accurately predict upstream regulators of proteome changes in disease models. Furthermore, we have observed that mTOR activation contributes toward the increased autophagosome accumulation observed in GBA-PD neurons.

PTMomics identifies cytoskeletal organization and axon extension defects in GBA-PD iPSC-dopamine neurons

To further understand the pathways affected in GBA-PD neurons, we examined the biological processes that were enriched in the differentially expressed proteins/PTM peptides. Among the non-modified proteins, “negative regulation of supramolecular fiber organization” (p = 4.6E–04) and “developmental cell growth” (p = 6.1E–04) were some of the most significantly enriched pathways (Figures 4A and 4B; Table S4A), whereas among the sialylated glycoproteins, “glycosaminoglycan metabolic process” (p = 5.2E−09) and “extracellular matrix organization” (p = 6.5E−19) were enriched (Figure 2E; Table S4B). Among the phospho-proteins, a high enrichment of RNA-related terms was present including “RNA splicing” (p = 8.5E−20) and “regulation of mRNA metabolic process” (p = 3.8E−19) (Figures 4C and S3G; Table S4C). Interestingly, perturbed RNA splicing has recently been suggested as a mediator of PD risk genes.⁵⁰ In the reversible cysteine-modified group, “RNA localization” (p = 6.1E−13) and “ribonucleoprotein complex biogenesis” (p = 3.5E−12) were among the most highly enriched terms (Figure 4D; Table S4D).

Figure 4. Proteomics and PTMomics indicate cytoskeletal organization and axon extension defects in GBA-PD iPSC-dopamine neurons.

Visualization of functional GO term networks based on annotations of significantly regulated (A and B) non-modified, (C and E) phosphorylated, and (D and F) cysteine-modified proteins.

(A, C, and D) The node size indicates the number of proteins connected to the GO term, and the color reflects functionally connected groups of terms.

(B, E, and F) Subnetworks related to axogenesis and cytoskeletal organization with the node size reflecting the enrichment significance of the terms and the leading group term being that of the highest significance. Nodes with similarity in the associated proteins are connected by lines with arrows indicating positive regulation, cropped lines negative regulation, and diamonds undefined regulation. Only terms with adjusted p values ≤0.05 are shown (two-sided hypergeometric test with Bonferroni stepdown).

Together, these pathway enrichments suggest that a diverse set of processes are dysregulated in GBA-PD neurons. Interestingly, terms related to regulation of cytoskeleton organization and neuron projection or axogenesis were common to all four groups (Figures 2E and 4A–4F; Tables S4A–S4D).

PTMomics identifies tau as a modifier of neurite outgrowth defects in GBA patient neurons

Further to this analysis, we performed a pathway analysis on the integrated datasets encompassing significantly dysregulated non-modified and modified proteins. This integrated analysis confirmed “neuritogenesis” (integrated p < 0.001) as the most highly enriched pathway when comparing across all four datasets (Figure 5A).

Figure 5. Functional assay confirms defects in neurite outgrowth predicted by pathway analysis.

(A) Pathway analysis identified the diseases and functions most highly enriched across all datasets. Comparison analysis was performed using Ingenuity Pathway Analysis on the four datasets (non-modified, phosphorylated, glycosylated, and cysteine modified) from the proteomic characterization of GBA-PD iPSC-dopamine neurons. Data are presented as a heatmap of –log(p value), where increasing color intensity corresponds to increasing significance in enrichment.

(B) Representative bright-field images of GBA-PD and control iPSC-dopamine neurons from neurite outgrowth assay day 0, 2, 4, and 6 post-scratch. Scratch areas are marked with blue dotted lines. Scale bar: 100 μm. Representative closeup of processed binary image used for quantification of neuronal processes in the scratch areas from 6 days post-scratch.

(C) Quantification of neurite outgrowth as measured by the percentage of the scratch area covered by processes. Values from day 0 subtracted as background. Data are from 3 independent differentiations (n = 4–5 iPSC lines per group). Mean ± SEM, **p ≤ 0.01 (paired Student’s t test).

(D–F) Analysis of mitochondrial motility and morphology showing (D) representative images of mitochondrial movement visualized with MitoTracker deep red (Red) at 12 time points from T = 0–484. Nuclei stained with NucBlue LiveReady stain (blue). Arrows indicate individual tracked mitochondria. Scale bar: 25 μm. Quantification of (E) mitochondrial movement speed (nm/s) and (F) mitochondrial width-to-length ratio (n = 3 iPSC lines per group). Mean ± SEM, **p ≤ 0.01 (Student’s t test).

(G) Western blotting confirmed lentiviral-mediated shRNA knockdown (KD) of tau in GBA-PD iPSC-dopamine neurons (GBA 3) and no effect of scrambled shRNA. Expression levels were normalized to β-actin and are shown relative to untreated GBA-PD iPSC-dopamine neurons (n = 3 technical replicates). Mean ± SEM, *p ≤ 0.05 (Student’s t test).

(H) Representative bright-field images of one line of GBA-PD iPSC-dopamine neurons (GBA 3) with lentiviral-mediated shRNA tau KD and untreated from neurite outgrowth assay day 0 and 6. Scale bar: 100 μm.

(I–J) Quantification of neurite outgrowth as measured by the percentage of the scratch area covered by processes over time. Data from (I) GBA-PD (GBA 3) and (J) control (Con 3) iPSC-dopamine neurons with shRNA tau KD or a scrambled shRNA control. Values from day 0 subtracted as background (n = 8–12 wells per group). Mean ± SEM, *p ≤ 0.05, **p ≤ 0.01 (one-way ANOVA, Dunnett’s multiple comparisons).

To investigate the predicted differences in neuritogenesis between GBA-PD and control neurons, we utilized an assay for quantification of neurite outgrowth through the application of standardized scratches to the monolayer of neurons and measured the growth of neurites into the uncovered scratch area over 6 days (Figure 5B). Given that very few cells migrated into the scratch area over the course of the assay, the analysis was able to primarily measure neuronal processes extending into the scratch area. Analysis of neurite outgrowth demonstrated a 31% reduction in the ability to regrow into the scratch area in GBA-PD neurons (p = 0.005), indicating a deficit in neuritogenesis, as predicted by the combined analysis (Figure 5C). The neurite outgrowth defect could be a result of changes in microtubule dynamics, which was the second most highly enriched pathway when comparing across all four datasets. To examine if other phenotypes, potentially arising from microtubule dynamic perturbations,⁵¹ were present, we analyzed mitochondrial motility using MitoTracker. This revealed that the mitochondrial movement speed was significantly decreased in the GBA-PD neurons, whereas the mitochondrial morphology based on the width-to-length ratio was unaltered (Figures 5D–5F).

Given the prominence of tau (MAPT) as an upstream regulator identified in the integrated pathway analysis of GBA-PD neurons (Figure 3D), we explored the role of the tau in the neurite outgrowth phenotype. We knocked down tau during neuronal maturation using short hairpin RNA (shRNA) and performed the scratch assay. shRNA treatment was confirmed to significantly knock down total tau levels to around 50% of untreated, with the scrambled negative control shRNA having no effect (Figure 5G; p < 0.05). We observed that tau knock-down significantly decreased neurite outgrowth in both a GBA-PD patient and a control line, indicating that tau is required for neurite outgrowth in human iPSC-dopamine neurons (Figures 5H–5J).

Chaperoning GCase rescues neurite outgrowth defects in GBA patient neurons

Given that loss of GCase activity is hypothesized to be a major cause of dysfunction in GBA-N370S carriers, we sought to investigate whether rescuing GCase activity could improve the neurite outgrowth phenotype observed in GBA-PD neurons. We therefore applied the non-inhibitory small-molecule GCase chaperone NCGC758^52–54 to day 30 GBA-PD iPSC-dopamine neurons and observed a significant, dose-dependent increase in GCase activity 4 days later (Figure 6A). Treatment of GBA-PD and control neurons with 15 μM NCGC758 for the duration of the scratch assay improved the neurite outgrowth in the GBA-PD neurons without affecting the control neurons (Figures 6B and 6C). These increases in mean neurite outgrowth were independent of alterations of tau, phospho-mTOR, or LC3II levels as measured by western blotting (Figure S5). Importantly, treatment of a GBA-PD line with increasing doses of NCGC758 demonstrated a concentration-dependent rescue of neurite outgrowth. NCGC758 treatment (15 μM) resulted in a significant increase in scratch repair, increasing from 43% to 78% of control levels (p = 0.016; Figures 6C and 6D), demonstrating the importance of the GBA mutation to the observed phenotype and highlighting opportunities for pharmacological rescue.

Figure 6. GCase chaperoning rescues neurite outgrowth defects in GBA-PD iPSC-dopamine neurons.

(A) Representative graph of GCase activity in GBA-PD iPSC-dopamine neurons treated for 48 h, in response to increasing concentrations of the GCase activator NCGC758 (758), relative to healthy control neurons (n = 3–26 wells per condition). Mean ± SEM (non-linear regression).

(B) Quantification of scratch repair as measured by the percentage of the scratch area covered by processes on day 6. Values from day 0 subtracted as background (n = 2–3 iPSC lines per group). Mean ± SEM, *p ≤ 0.05 (one-way ANOVA, Tukey’s multiple comparisons).

(C) Representative bright-field images of control and GBA-PD iPSC-dopamine neurons treated with 5 or 15 μM 758 for 7 days. Scale bar: 100 μm.

(D) Quantification of scratch repair with increasing doses of 758. Data are from one control and one line of GBA-PD iPSC-dopamine neurons (n = 3 independent differentiations). Mean ± SEM, *p ≤ 0.05, **p ≤ 0.01 (one-way ANOVA, Tukey’s multiple comparisons).

Discussion

In the present study, we applied a comprehensive proteomics workflow to profile the post-translational proteome in human neurons. Mutations in the GBA gene are the strongest common genetic risk factor for PD, and studies of the effect of GBA mutations in human dopaminergic neurons are progressing our knowledge of PD pathogenesis.^2,18,35 We differentiated GBA-N370S patient and control iPSC lines using a well-established protocol⁵⁵ to generate high numbers of midbrain-specific dopamine neurons. The simultaneous enrichment and identification of the proteome as well as peptides with phosphorylation, reversible cysteine modifications, and sialylated N-linked glycosylation resulted in an unbiased, global characterization of PTM levels in human dopamine neurons. Moreover, this strategy identified more than 2,000 peptides containing post-translationally modified residues that were dysregulated in GBA-PD neurons, making this dataset a resource for research into GBA-related pathogenesis. Importantly, this approach utilized small amounts of staring material (100 μg/sample), demonstrating applicability to small-scale cultures of PD patient-derived dopaminergic neurons to identify disease phenotypes and mechanistic targets.

The ability of the workflow to assess sialylated N-linked glycosylation was demonstrated to significantly increase the coverage of lysosomal proteins and gave insight into lysosomal dysfunction in GBA-PD, resulting in identification of perturbed levels of several glycosylated lysosomal proteins.¹¹ Sialylated N-linked glycosylation of cathepsin D, LIMP2, LAMP1, and LAMP2 is necessary for their function and transport from the ER via the Golgi to the lysosome, and increased lysosomal content is associated with an increase in the glycosylated forms specifically.^28,56 Building on our previous observations of ALP dysfunction in GBA-PD lines,^11,20 we identified alterations in the levels of a number of proteins that are encoded by PD GWAS-associated risk genes such as PSAP,⁴⁰ ASAH1,³⁸ and GALC,^39,57 observations consistent with the known pathways regulated by GCase.

Interestingly, pathway enrichment analysis identified mTOR, an important autophagy regulator, as an upstream target based on PTMomic changes. Although non-modified mTOR was identified as decreased by the MS analysis, the phospho-proteomics detected increased mTOR phosphorylation. Western blotting confirmed increased levels of phospho-mTOR, suggesting perturbed mTOR signaling, which has been implicated in PD pathogenesis based on post-mortem brain studies and cellular models.⁵⁸ However, the upregulation of autophagy and lysosomal markers in GBA iPSC-dopamine neurons does not suggest mTORC1-mediated autophagy inhibition, and may reflect impaired lysosomal clearance, as previously observed.¹¹ The observation of increased phosphorylation of 4E-BP1, but not p70S6K, suggests increased protein translation via eIF4.⁵⁹ This may explain the over-representation of translation as well as RNA transport and splicing in the pathway analysis across the combined datasets. Further studies are needed to determine the consequences of the observed mTOR signaling changes in GBA-N370S neurons.⁵⁸

The value of large-scale proteomic analysis was also confirmed by the clear separation of patients with PD with the GBA-N370S mutation from controls and the identification of a patient carrying a GBA-N370S mutation but with a final diagnosis of PSP. GBA mutations have been reported in patients with PSP, although they are found infrequently. PSP, which is clinically distinct from PD, also demonstrates nigrostriatal dopaminergic degeneration.⁶⁰ The ability of multi-omics analysis of iPSC-dopamine neurons to stratify patients with clinically overlapping disorders is consistent with our previous work using single-cell RNA sequencing (RNA-seq) profiling to achieve the same separation.²⁰ Further investigation of larger cohorts of patients with GBA-PD, GBA mutation carriers who do not develop PD, and additional PSP, Lewy body dementia (LBD), and AD cohorts will allow the limits of patient and disease stratification of PTMomics to be identified.

Post hoc comparison of the current study with bulk and single-cell RNA-seq data on GBA-N370S patient neurons²⁰ found a similar dysregulation of genes involved in MAPT splicing, microtubule function and formation, neuron projection, and axon development. As the RNA-seq analysis was performed on day in vitro 65 (DIV65) neurons, this suggests that key phenotypes identified by PTMomics such as neurite outgrowth impairment are consistent across techniques and are relevant as neurons further mature. As demonstrated by Lang et al.,²⁰ further analysis of DIV65 neurons would likely identify additional phenotypes in processes such as synaptic function.

Although only around 20% of the regulated non-modified proteins had a role in the cytoskeleton, the proteins showing the highest upregulation were primarily cytoskeletal proteins related to microtubule dynamics and neurite outgrowth. This was also apparent from the pathway analysis, which highlighted axon extension across all four datasets. Interestingly, post-mortem studies of gene expression changes in early PD have also indicated that the axogenesis pathway is perturbed.⁶¹ Our functional analysis confirmed a significant decrease in neurite outgrowth, a phenotype not previously reported in GBA-PD iPSC-dopamine neurons. However, iPSC-derived neurons carrying LRRK2 or PARK2 mutations or with heterozygous GBA knockout also show impaired neurite outgrowth, indicating that this phenotype could be characteristic for PD in general, perhaps even early in the disease development.^30,62,63 Furthermore, it indicates that neurite outgrowth deficits are downstream effects, which can arise from a variety of forms of PD-related cellular stress. The present study does not dissect the sequence of events leading from GCase dysfunction to neurite outgrowth impairment, although the pathway analysis of the proteomic/PTMomic data highlight a number of upstream regulators to potentially be involved. For example, the finding of decreased mitochondrial movement speed could be a contributing factor to neurite outgrowth or another secondary effect of microtubule-associated perturbations.

The top predicted upstream target in the comparison pathway analysis was tau (MAPT). In addition, tau protein levels were increased in the GBA-PD neurons as shown by MS, although the extent was variable across differentiations, perhaps due to subtle differences in the resulting neuronal content. Tau is not only important in AD pathogenesis but has been identified by GWASs as a strong risk factor for PD and is found in Lewy bodies colocalizing with α-syn.^42,64 In addition, GBA mutations are associated with an increased risk of dementia and tau deposition in patients with PD.^4,5 To validate if tau, the top predicted upstream regulator, has a role in the most prominent phenotype identified by PTMomics neurite outgrowth, we used shRNA to knock down tau to around 50% of normal expression levels. This caused a significant inhibition of neurite outgrowth in both GBA-PD and control neurons, indicating that tau is required for neurite outgrowth but that the increased tau levels in GBA-PD neurons are not directly linked to the neurite outgrowth phenotype. However, this confirms the ability of the combined proteomic/PTMomic analysis to predict key regulators of the identified perturbed pathways. Our findings could suggest that the role of tau on PD pathology stems from a loss of function, perhaps through interaction with α-syn.⁶⁴

Interestingly, the neurite outgrowth defect could be rescued using a modulator of GCase activity, NCGC758. This compound is a non-inhibitory GCase chaperone that increases GCase maturation, trafficking, and activity, increases lysosomal GCase activity, and reduces α-syn aggregation.^35,52–54 Chaperone rescue by NCGC758 links mutant GCase processing and activity to the neurite outgrowth phenotype in GBA-PD patient-derived neurons, adding neuritogenesis to other GBA-N370S phenotypes modulated by chaperone treatment.^{11,19,20,35,65} We did not find any clear effect of NCGC758 on the protein/PTM level phenotypes (tau, LC3II, and phospho-mTOR) established by western blotting, which could indicate that these are not directly involved in linking GCase dysfunction to neurite outgrowth. However, detecting potentially subtle effects of a drug treatment on specific protein level changes by western blotting can be highly challenging given the variability between individual patients and controls as well as between differentiations. As evident from the PCA plots and western blots in this study, even the controls show high variation in protein/PTM levels. This highlights the advantage of global analyses such as proteomics/PTMomics, reporting more broadly on pathway level changes, in overcoming inter-individual and differentiation-induced variation when studying multi-factorial diseases such as GBA-PD.

The present proteomic workflow allows for parallel monitoring of three distinct PTMs, in addition to the non-modified peptides observed by traditional proteomics, from a single sample. Importantly, this workflow can be performed on as little as 100 μg protein per sample, allowing the technique to be performed on small-scale samples, such as brain regions or iPSC-derived neurons. This represents a substantial improvement on similar workflows,^30,66,67 and the inclusion of analysis of sialylated glycopeptides resulted in valuable information on alterations in levels of the active/membrane-bound forms of numerous lysosomal proteins. The increased coverage of the functional lysosomal proteome in neurons is valuable to disease biology, with significant lysosomal dysfunction found in PD, frontotemporal degeneration (FTD)/amyotrophic lateral sclerosis (ALS), and AD.^68–70 Indeed, other types of glycosylation including o-linked glycosylation and non-enzymatic glycation are known to have diverse and important roles in neurode-generative diseases⁷¹ and could be added to future iterations of the PTMomics workflow to further expand the PTMs detected.

In conclusion, we have combined GBA-PD patient iPSC-dopamine neurons with a comprehensive MS methodology, allowing the identification of large numbers of dysregulated proteins and PTMs. Integration of the proteomic/PTMomic datasets led to the discovery of a neurite outgrowth phenotype in iPSC-dopamine neurons from patients carrying the GBA-N370S mutation, as well as upstream and downstream modulators of this phenotype, most notably tau. Proteomic/PTMomic analysis of iPSC-derived neurons is thus a powerful method to identify pathways and phenotypes as well as therapeutic targets and compounds relevant to human neurodegenerative disease.

Limitations of the study

This study reports a comprehensive characterization of PTM changes resulting from the GBA mutation. However, there are currently limited methods available for more broad validation of individual PTMs such as cysteine modifications. We have therefore focused on validating phenotypic changes arising from the cellular pathways showing proteomic/PTMomic perturbations. We observe that proteomic signatures or pathways comprising tens to hundreds of proteins are much more robust than comparisons of individual protein targets, which demonstrate more variability. As is evident from the PCA plots and heatmaps in this study, a degree of heterogeneity can be seen between iPSC-derived neurons from different individuals. There are several sources of variability between iPSC lines, such as normal inter-individual heterogeneity among control individuals, patient-to-patient clinical heterogeneity, and technical variability between differentiations. To counter this, validation of the current findings in a larger number of patient and control iPSC lines could be beneficial. For the rescue experiments especially, it would be relevant to gauge in a larger cohort the proportion of patients whose iPSC-derived neurons respond well to GCase chaperone treatment to better assess the potential clinical relevance.

Star⋆Methods

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-tyrosine hydroxylase	Millipore	RRID: AB_390204
Goat anti-FOXA2	R & D Systems	RRID: AB_2294104
Mouse anti-b-III-Tubulin	Biolegend	RRID: AB_10063408
Rabbit anti-LMX1A	Abcam	RRID: AB_2827684
Chicken anti-MAP2	Millipore	RRID: AB_571049
Donkey anti-rabbit AF488	ThermoFisher	RRID: AB_2535792
Donkey anti-goat AF647	ThermoFisher	RRID: AB_2535864
Goat anti-chicken AF647	ThermoFisher	RRID: AB_2535866
Donkey anti-mouse AF555	ThermoFisher	RRID: AB_2536180
Mouse anti-b-Actin-HRP	Abcam	RRID: AB_867494
Rabbit anti-LC3	Sigma-Aldrich	RRID: AB_796155
Mouse anti-Tau	NeoMarkers	#ms-247-P
Mouse anti-mTOR	Cell Signaling	RRID: AB_1904056
Rabbit anti-phospho-mTOR Ser2448	Cell Signaling	RRID: AB_330970
Rabbit anti-p70S6K	Cell Signaling	RRID: AB_390722
Mouse anti-phospho-p70S6K Thr389	Cell Signaling	RRID: AB_2285392
rabbit anti-4E-BP1	Cell Signaling	RRID: AB_659944
rabbit anti-phospho-4E-BP1 Thr37/46	Cell Signaling	RRID: AB_560835
rabbit anti-PKCa	Cell Signaling	RRID: AB_2284227
rabbit anti-phospho-PKCa Ser657	ThermoFisher	RRID: AB_2736408
HRP-conjugated secondary antibody	BioRad	#170-6515
Bacterial and virus strains
One Shot TOP10 Chemically Competent E. coli	ThermoFisher	#C40400
One Shot Stbl3 Chemically Competent E. coli	ThermoFisher	#C737303
Chemicals, peptides, and recombinant proteins
hESC-Qualified Matrigel, LDEV-free	Corning	#354277
Geltrex	ThermoFisher	#A1413302
ROCK inhibitor	Bio-Techne	#1254
LDN193189	Sigma-Aldrich	#SML0559
SB431542	Bio-Techne	#1614
Sonic hedgehoc-C24II	Bio-Techne	#1845-SH
Purmorphamine	Bio-Techne	#4551
FGF8a	Stratech	#16124-HNAE-SIB
CHIR99021	Bio-Techne	#4423
BDNF	Peprotech	#450-02
GDNF	Peprotech	#450-10
TGFb3	Peprotech	#100-36E
DAPT	Abcam	#120633
Ascorbic acid	Sigma-Aldrich	#A4544
(db)-cAMP	Sigma-Aldrich	#D0627
Endoproteinase Lys-C	Wako	#129-02541
Trypsin	Sigma-Aldrich	#T0303
C18 3M EmporeTM disk	Sigma-Aldrich	#66883-U
C8 3M EmporeTM disk	Sigma-Aldrich	#66882-U
Poros 50 R2	Applied Biosystems	#1-1159-06
Oligo R3	Applied Biosystems	#1-1339-03
N-glycosidase F	Biolabs	#P0705S
Sialidase A	Prozyme	#GK80040
PhosSelect IMAC beads	Sigma-Aldrich	#P9740
TSKgel Amide-80 HR, 5 mm column	Tosoh Bioscience	#0021982
ReproSil-Pur C18 AQ 3	Dr. Maisch	#r13.aq.
4-15% precast gel	Biorad	#567-8085
Polyvinylidene difluoride membrane	Biorad	#170-4157
MagicMark XP Protein Standard	ThermoFisher	#LC5602
BLUeye Prestained Protein Ladder	Geneflow	#S6-0024
Conduritol B epoxide	Calbiochem	#234599
Methylumbillifery β-D-glucopyranoside	Sigma-Aldrich	#M3633
NCGC00188758	Sigma-Aldrich	#5316600001
Critical commercial assays
CytoTune-iPS Sendai Reprogramming kit	ThermoFisher	#A16518
AllPrep DNA/RNA Kit	Qiagen	#80204
Human CytoSNP-12v2.1 beadchip array	Illumina	#WG-320-21
OmniExpress-24 v1.4 Kit	Illumina	#2006206
HumanHT-12 v4 Expression BeadChip Kit	Illumina	#BD-103-0204
iTRAQ Reagent 8-plex assay kit	Sciex	#4390811
Bicinchoninic acid assay	Pierce	#23225
Gibson Assembly Master Mix	NEB	#E2611
KAPA HiFi HotStart ReadyMix	KAPA Biosystems	#KK2602
QIAquickGel Extraction Kit	QIAGEN	#28704
PureLink HiPure Plasmid Maxiprep kit	ThermoFisher	#K210006
NucBlue LiveReady	ThermoFisher	#R37605
Mitotracker Deep Red	ThermoFisher	#M22426
Deposited data
Proteomics raw and analyzed data	This paper	PRIDE: PXD026691
Illumina SNP and -HT12v4 expression array data	This paper, GEO Datasets	GSE99142
Code for to prepare images for neurite outgrowth analysis	This paper	https://doi.org/10.5281/zenodo.7606336
Experimental models: Cell lines
MK088-07	University of Oxford	MK088-07
SFC834-03-01	EBiSC	STBCi025-A/B
SFC871-03-09	EBiSC	STBCi084-C
MK082-31	EBiSC	UOXFi002-B
MK071-03	EBiSC	UOXFi001-B
SFC848-03-02	EBiSC	STBCi042-A
SFC841-03-02	EBiSC	STBCi025-C
OX2-28	EBiSC	UOXi006-A
SFC856-03-04	EBiSC	STBCi063-A
MK053-06	EBiSC	UOXFi005-B
OX1-19	EBiSC	UOXFi004-B
NHDF1	Lonza	CC-2511
Oligonucleotides
Primers for Sendai virus-delivered reprogramming gene clearance, see Table S5	This paper	N/A
pMXsAS3200v2: TTATCGTCGACC ACTGTGCTGGCG	This paper	N/A
mNanog forward primer: GCTCCA TAACTTCGGGGAGG	This paper	N/A
Software and algorithms
GenomeStudio	Illumina	N/A
KaryoStudio	Illumina	N/A
Harmony analysis software	Perkin-Elmer	N/A
PluriTest™	Pluritest.org	N/A
Image Lab software	BioRad	N/A
Xcalibur v3.0	ThermoFisher	N/A
Proteome Discoverer 2.1	ThermoFisher	N/A
Ingenuity Pathway Analysis software	Qaigen	N/A
Cytoscape StringApp	Doncheva et al.72	https://apps.cytoscape.org/apps/stringapp
Cytoscape Omics Visualizer app	Legeay et al.73	https://apps.cytoscape.org/apps/omicsvisualizer
Cytoscape ClusterMaker2 app	Morris et al.74	https://apps.cytoscape.org/apps/clustermaker2
Cytoscape ClueGO app	Bindea et al.75	https://apps.cytoscape.org/apps/cluego
The R Project for Statistical Computing	R Core Team76	https://www.r-project.org/
ggplot2	Wickham77	https://ggplot2.tidyverse.org/
Perseus, MaxQuant	Tyanova et al.78	https://maxquant.net/perseus/
GraphPad Prism version 5.0	GraphPad Software	N/A
Other
FACSCalibur	BD Biosciences	N/A
Opera Phenix microscope	Perkin-Elmer	N/A
PHERAStar FSX plate reader	BMG Labtech	N/A
ChemiDoc Touch Imaging system	BioRad	N/A
Agilent 1200 Series HPLC	Agilent	N/A
EASY-nLC system	ThermoFisher	N/A
QExactive HF Mass Spectrometer	ThermoFisher	N/A

Resource Availability

Lead contact

Further information and requests for resources and reagents must be directed to and will be fulfilled by the lead contact, Brent J. Ryan (brent.ryan@dpag.ox.ac.uk).

Materials availability

All iPSC lines generated in this study are available from the European Bank for induced pluripotent Stem Cells (EBiSC) or from the lead contact upon request with a completed Materials Transfer Agreement.

Experimental Model and Subject Details

Participant recruitment and GBA-N370S mutation screening

Participants gave signed informed consent to mutation screening and derivation of iPSC lines from skin biopsies (Ethics committee: National Health Service, Health Research Authority, NRES Committee South Central, Berkshire, UK, REC 10/H0505/71).

PD patients and controls from the Discovery clinical cohort established by the Oxford Parkinson’s Disease Center were screened for heterozygous GBA-N370S mutation as earlier described.¹¹ All patients fulfilled the UK Brain Bank diagnostic criteria for clinically probable PD at presentation. Fibroblasts from six GBA-N370S PD patients and six age and gender-matched healthy controls were reprogrammed to generate iPSC lines (Table 1). Two GBA and five control iPSC lines have earlier been characterized.^{11,20,79–81}

Reprogramming of primary fibroblasts

Skin punch biopsies were obtained from participants and low passage fibroblast cultures established from these were used for re-programming either by retroviral delivery or using the CytoTune-iPS Sendai Reprogramming kit (ThermoFisher) as previously described.^11,80

iPSC QC analyses

Detailed characterization of the previously unpublished iPSC lines was performed as detailed in Fernandes et al. (2016).¹¹ High-resolution Illumina SNP-based karyotyping was generated using the Human CytoSNP-12v2.1 beadchip array or OmniExpress24 array (Illumina) on genomic DNA generated using the All-Prep kit (Qiagen) and analysis with GenomeStudio and Karyostudio software (Illumina). The SNP-based karyotyping confirmed their genomic integrity, demonstrating no gross chromosomal amplifications, deletions or loss of hetero-zygosity (Data S1). Additionally, SNPs datasets also enabled confirmation that the iPSC lines derived from the expected fibroblast line. Flow cytometry for pluripotency markers was performed with the following antibodies and appropriate isotype control, at the same concentration, from the same supplier (clone, isotype control, supplier): TRA-1-60 (B119983, IgM-488, Biolegend) and Nanog (2985S, IgG-647, Cell Signaling) on a FACSCalibur (BD Biosciences). Silencing of retroviral-delivered reprogramming genes was assessed by quantitative real time-PCR (qRT-PCR) using primers published by Takahashi et al.,⁸² with modifications of the pMXsAS3200v2 (TTATCGTCGACCACTGTGCTGGCG) and mNanog forward primer (GCTCCATAACTTCGGGGAGG).

RT-PCR to assess clearance of Cytotune Sendai virus-delivered reprogramming genes was performed with the primers listed in Table S5.

‘Pluritest’ Analysis was performed on RNA from iPSC lines subjected to Illumina HT12v4 transcriptome array with image data files uploaded to www.pluritest.org and scored for pluripotency, as previously described.⁸³ All lines tested had acceptable pluripotency scores (Figure S6).

iPSC propagation and differentiation

Post-thawing iPSCs were propagated as a monolayer on Matrigel-coated (Corning) 6-well plates (Corning) in mTeSR1 medium (Stem Cell Tech.) with 1% penicillin-streptomycin (pen-strep, ThermoFisher), passaged 1:2-3 when confluent using Tryple (ThermoFisher). When thawing or passaging the iPSCs, ROCK inhibitor (Y27632, Bio-Techne) was added to promote survival. iPSCs were differentiated using a modified dual-SMAD inhibition protocol as previously described.^20,84 At D20 cells were replated at 3 x 10⁵ cells/cm² and maintained until assays performed at D35.

Method Details

Immunofluorescence

Differentiation day 35 cells cultured in full or ½ area 96-well plates (Greiner Bio-One) were fixed for 15 min at room temperature (RT) in 4% (w/v) paraformaldehyde (PFA, Sigma-Aldrich) in distilled phosphate buffered saline (dPBS, ThermoFisher), pH 7.4, and rinsed with dPBS. Cells were permeabilised and unspecific binding blocked with dPBS with calcium and magnesium (dPBS⁺⁺/0.1% Triton (Sigma-Aldrich)/10% donkey serum (Sigma-Aldrich) for 1 h at RT. Primary antibodies were diluted in dPBS⁺⁺/0.1% (v/v) Triton/1% (v/v) donkey serum and incubated overnight (ON) at 4°C. The following primary antibody dilutions were applied: rabbit anti-tyrosine hydroxylase (TH, Millipore #152) 1:500, goat anti-FOXA2 (R&D #AF2400) 1:250, mouse anti-β-III-Tubulin (TUJ1, Biolegend #801202) 1:500, rabbit anti-LMX1A (Abcam #ab139726) 1:250, chicken anti-MAP2 (Millipore #ab5543) 1:2000.

Cultures were rinsed in PBS and incubated with donkey anti-rabbit AF488 (ThermoFisher #A21206) 1:1000, donkey anti-goat AF647 (ThermoFisher #A21447) 1:500, goat anti-chicken AF647 (Invitrogen A21449) 1:500 and donkey anti-mouse AF555 (Invitrogen #A31570) 1:1000 in dPBS⁺⁺/0.1% Triton/1% donkey serum for 1 h at RT. Cell nuclei were counterstained with 10 μM 4″,6-diamidino-2-phenylindole dihydrochloride (DAPI, Sigma-Aldrich) in dPBS⁺⁺. Fluorescence images were acquired on the Opera Phenix microscope (PerkinElmer) and analyzed using Harmony analysis software (PerkinElmer).

Glucocerebrocidase (GCase) activity assay

GCase activity was measured by cleavage of 4-methylumbelliferyl-β-D-glucopyranoside (4-MUG) to 4-methylumbelliferone, as previously described.³¹ Briefly, cell pellets were sonicated at 10 amp for 10 s in citrate phosphate buffer pH 5.4 consisting of 0.1 M citric acid (Sigma-Aldrich) and 0.2 M dibasic sodium phosphate (Sigma-Aldrich) with 0.25% (v/v) Triton X- and 0.25% (w/v) taurocholic acid (Sigma-Aldrich). Samples were centrifuged at 800 g for 5 min at 4°C and supernatant collected. Following protein determination equal amounts of protein from each sample were diluted in citrate phosphate buffer in quadruplicates. One replicate of each sample was treated with 1 mM conduritol B epoxide (CBE, Calbiochem) for 10 min before all samples were incubated for 1 h with 2.5 mM of the fluorescent GCase substrate methylumbillifery β-D-glucopyranoside (4MUG, Sigma-Aldrich) at 37°C in the dark. The reaction was quenched with 1 M glycine buffer (Sigma-Aldrich) pH 10.8 and the fluorescent levels analyzed on the PHERAStar FSX plate reader (BMG Labtech). Values from CBE-treated wells were subtracted as background.

Western blotting procedure

Cell pellets were lysed in RIPA buffer (50 mM Tris-hydrochloride (Sigma-Aldrich), 150 mM sodium chloride (Sigma-Aldrich), 1% Tergitol-type NP-40 (Sigma-Aldrich), 0.5% sodium cholate hydrate (Sigma-Aldrich), and 0.1% SDS (Sigma-Aldrich), pH 8) containing phosphatase (PhosSTOP tablets, Roche) and protease inhibitor (Complete Tablets, Roche) and sonicated for 10 s at 10 amplitude microns on ice. Protein concentrations were determined with bicinchoninic acid assay (BCA, Pierce) and equal amounts of protein were separated by SDS-PAGE. Membranes were probed overnight with primary antibodies as follows: mouse anti-β-Actin-HRP (Abcam #49900) 1:50,000, rabbit anti-LC3 (Sigma-Aldrich #L7543) 1:1000, mouse anti-Tau (Neomarkers #ms-247-P) 1:1000, mouse anti-mTOR (Cell Sig. #45175) 1:1000, rabbit anti-phospho-mTOR (Ser2448) (Cell Sig. #2971) 1:100, rabbit p70S6K (Cell Sig. #2708) 1:1000, mouse anti-phospho-p70S6K (Thr389) (Cell Sig. #9206) 1:100, rabbit anti-4E-BP1 (Cell Sig. #4923) 1:1000, rabbit anti-phospho-4E-BP1 (Thr37/46) (Cell Sig. #2855) 1:100, rabbit anti-PKCα (Cell Sig. #2056) 1:1000, and rabbit anti-phospho-PKCa (Ser657) (ThermoFisher #PA5-78124) 1:500. Representative blots of all antibodies shown in Figure S7.

Membranes were subsequently probed with horseradish peroxidase (HRP)-conjugated goat anti-mouse or -rabbit IgG (Bio-Rad), diluted 1:5000 in 5% (w/v) skim milk in TBST, for 1 h and visualised using ECL (Millipore #P90720) on a ChemiDoc Touch Imaging system (Bio-Rad). The optical density of each band was quantified using Image Lab software (Bio-Rad).

Mass spectrometry sample collection and protein isolation

Differentiation day 35 neurons from four GBA and four control iPSC lines were collected on ice in phosphate-buffered saline (PBS, ThermoFisher) with protease- (Complete tablets, Roche) and phosphatase inhibitors (PhosSTOP tablets, Roche). Samples were sonicated for two times 10 s at 50% amplitude on ice and incubated for 30 min at RT in lysis buffer consisting of 6 M urea (Sigma-Aldrich), 2 M thiourea (Sigma-Aldrich), 20 mg/mL sodium dodecyl sulfate (SDS, GE Healthcare), 40 nM N-ethylmaleimide (NEM, for alkylation of free cysteines, Sigma-Aldrich) and protease inhibitor.

Reduction and enzymatic digestion

Following methanol-chloroform precipitation (Sigma-Aldrich), the proteins were denatured and reduced in 6 M urea, 2 M thiourea and 10 mM TCEP (ThermoFisher) at RT. After vortexing, the samples were incubated at RT for 2 h with 1 μL endoproteinase Lys-C (Wako). The samples were diluted 10 times in 20 mM TCEP in 20 mM triethylammonium bicarbonate (TEAB, Sigma-Aldrich) buffer, pH 7.5 and sonicated for two times 10 s at 50% amplitude on ice. Digestion with 1 μg trypsin (Sigma-Aldrich) per 50 μg peptide was done ON at RT.

Desalting and iTRAQ labeling

The samples were acidified with 0.1% trifluoroacidic acid (TFA, Sigma-Aldrich) and desalted using two self-made P200-tip-based columns per sample. A small plug of C₁₈ material from a 3M Empore disk (Sigma-Aldrich) was inserted in the constricted end of a P200 pipette tip and 1.5 cm of the tip was packed with reversed-phase resin material consisting of a 1:1 mix of Poros 50 R2 (Applied Biosystems) and Oligo R3 (Applied Biosystems) dissolved in 100% acetonitrile (ACN, Sigma-Aldrich). The acidified samples were loaded onto the first micro-columns and washed with 0.1% TFA. The peptides were eluted using 60% ACN/0.1% TFA. This process was repeated with the second column before the samples were dried by speed vacuum centrifugation.

100 μg of each sample were labeled with Isobaric Tags for Relative and Absolute Quantification (iTRAQ) Eight-plex Isobaric label Reagents (Sciex) according to the manufacturer’s instructions. Efficient labeling was confirmed by MALDI, the ratios adjusted, the labeled peptides mixed 1:1:1:1:1:1 and dried.

Enrichment of phosphorylated, sialylated glyco- and cysteine-modified peptides

Phospho-peptide enrichment and fractionation were essentially performed as earlier described.⁸⁵ Peptides were dissolved in 80% ACN/5%TFA with 1 M glycolic acid (Sigma-Aldrich) and incubated with 0.6 mg titanium dioxide (TiO₂) beads (Titansphere 10 μm, GL Sciences) per 100 μg peptide for 30 min at RT with vigorous shaking. The beads were centrifuged briefly and the supernatant transferred to a new tube with 0.3 mg TiO₂ beads per 100 μg peptide. After 15 min incubation at RT with vigorous shaking and a brief centrifugation the supernatant was collected. The beads were subsequently washed with 80% ACN/1% TFA and 10% ACN/0.1% TFA. The supernatant with the unbound TiO₂ fraction and the washing fractions, both containing the non-phosphorylated peptides, were combined and dried for further processing (see below). The phosphorylated peptides were eluted from the beads by incubation with 1.5% ammonium hydroxide solution (Sigma-Aldrich), pH 11.3, at RT with vigorous shaking. The beads were spun down and the supernatant passed through C₈ material from a 3M Empore disk (Sigma-Aldrich). Any remaining peptides were eluted from the filter with 30% ACN and all peptide samples were dried.

The dried phosphorylated peptides were dissolved in 20 mM TEAB, reduced with 1 M dithiothreitol (DTT, Sigma-Aldrich) for 30 min at RT and alkylated with 1 M NEM for 30 min at RT in the dark. To separate out sialylated glyco-peptides that also bind to the TiO₂ beads, the sample was deglycosylated with N-glycosidase F (Biolabs) and Sialidase A (Prozyme) at 37°C ON. The sample was dried, resuspended in SIMAC loading buffer (50% ACN/0.1% TFA) and incubated with PhosSelect IMAC beads (Sigma #P9740), pre-equil-ibrated in SIMAC loading buffer, for 30 min at RT with gentle shaking. The beads were washed in SIMAC loading buffer before mono-phosphorylated and deglycosylated peptides were eluted with 20% ACN/1% TFA, combined with the washing fraction and dried. The multi-phosphorylated peptides were then eluted with 1.5% ammonium hydroxide solution, pH 11.3, and dried. The mono-phosphorylated and deglycosylated peptides were separated by adjusting the sample to 70% ACN/2% TFA and repeating the TiO₂ bead enrichment as described above (washing with 50% ACN/0.1%TFA). Deglycosylated peptides were collected in the supernatant.

The dried non-phosphorylated sample was desalted on R2/R3 columns as described above and dissolved in 50 mM TEAB and 20 mM TCEP, pH 7, for 1 h. The sample was incubated with 10 mM cysteine-specific phosphonate adaptable tag (CysPAT), synthesized as previously described,⁸⁶ for 1 hr at RT in the dark with gentle shaking. The CysPAT reacts with reversibly modified cysteines and the phosphonate group allows for the subsequent isolation of CysPAT-labeled peptides from non-modified peptides using TiO₂ bead enrichment as earlier detailed.

All the eluates were dried and desalted on homemade columns as described previously, using only R3 material for the mono- and multi-phosphorylated peptides and R2/R3 material for deglycosylated, cysteine- and non-modified peptides.

Hydrophobic interaction liquid chromatography (HILIC) and high pH fractionation

Mono-phosphorylated-, deglycosylated-, CysPAT-labeled- and non-modified peptides were fractionated to reduce sample complexity using HILIC as described previously.⁸⁵ The multi-phosphorylated sample was not fractionated, but run directly on nanoLC-ESI-MS/MS. The non-modified sample, expected to contain the highest concentration of peptides, was first diluted in 0.1% TFA and approximately 50 μg peptide was fractioned. All mono-phosphorylated-, deglycosylated- and CysPAT samples were fractionated. The samples were dissolved in 90% ACN, 0.1% TFA (solvent B) by adding 10% TFA followed by water and finally ACN slowly in order to prevent peptide precipitation. Samples were loaded onto an in-house packed TSKgel Amide-80 column, 5 μm (Tosoh Bioscience) using an Agilent 1200 Series HPLC (Agilent). Peptides were separated using a gradient from 100–60% solvent B (A = 0.1% TFA) running for 30 min at a flow-rate of 6 μl/min. Fractions were collected every 1 min and combined into 8-12 final fractions based on the UV chromatogram and subsequently dried by vacuum centrifugation.

To increase the coverage, high pH fractionation was also performed using approximately 50 μg peptide of the non-modified peptide sample. Briefly, the sample was dissolved in 1% ammonium hydroxide (NH₃, Sigma-Aldrich), pH 11, and loaded on a R2/R3 column equilibrated with 0.1% NH₃. The peptides were eluted in a stepwise fashion using a gradient of 5%-60% ACN/0.1% NH₃. All fractions were dried by vacuum centrifugation and stored at -20°C.

Reversed-phase nanoLC-ESI-MS/MS

The samples were resuspended in 0.1% formic acid (FA) and loaded onto an EASY-nLC system (Thermo Scientific). The samples were loaded onto a two-column system containing a 3 cm pre-column and a 17 cm column both consisting of fused silica capillary (75 μm inner diameter) packed with ReproSil–Pur C18 AQ 3 μm reversed-phase material (Dr. Maisch). The peptides were eluted with an organic solvent gradient from 100% phase A (0.1% FA) to 34% phase B (95% ACN, 0.1% FA) at a constant flowrate of 250 nL/min. Depending on the samples based on the HILIC, the gradient was from 1 to 30% solvent B in 60 min or 90 min, 30% to 50% solvent B in 10min, 50%-100% solvent B in 5 min and 8 min at 100% solvent B.

The nLC was online connected to a QExactive HF Mass Spectrometer (ThermoFisher) operated at positive ion mode with data-dependent acquisition. The Orbitrap acquired the full MS scan with an automatic gain control (AGC) target value of 3x10⁶ ions and a maximum fill time of 100ms. Each MS scan was acquired at high-resolution (120,000 full width half maximum (FWHM)) at m/z 200 in the Orbitrap with a mass range of 400-1400 Da. The 12 most abundant peptide ions were selected from the MS for higher energy collision-induced dissociation (HCD) fragmentation (collision energy: 34V). Fragmentation was performed at high resolution (60,000 FWHM) for a target of 1x10⁵ and a maximum injection time of 60ms using an isolation window of 1.2 m/z and a dynamic exclusion. All raw data were viewed in Thermo Xcalibur v3.0.

Mass spectrometry data analysis

The raw data were processes using Proteome Discoverer (v2.1, ThermoFisher) and searched against the Swissprot human database using an in-house Mascot server (v2.3, Matrix Science Ltd.) and the Sequest HT search engine.

Database searches were performed with the following parameters: precursor mass tolerance of 10 ppm, fragment mass tolerance of 0.02 Da (HCD fragmentation), TMT 6-plex (Lys and N-terminal) as fixed modifications and a maximum of 2 missed cleavages for trypsin. Variable modifications were NEM on Cys and N-terminal acetylation along with phosphorylation of Ser/Thr/Tyr, deamidation of Asn and N-Succinimidyl iodoacetate (SIA) on Cys for the phosphorylated, deglycosylated and CysPAT-modified groups, respectively. Only peptides with up to a q-value of 0.01 (Percolator), Mascot rank 1 and cut-off value of Mascot score > 15 were considered for further analysis. Only proteins with two or more unique peptides were considered for further analysis in the non-modified group. Subsequently, proteins which were at least 1.2-fold increased or decreased in the GBA patient neurons were selected and Student’s t-test with Benjamini-Hochbergs correction for multiple testing (FDR 0.1) was applied with a p-value cut-off of 0.05.

Peptides with PTMs which 1. could be normalized to the level of the corresponding non-modified protein, 2. were at least 1.3-fold increased or decreased in the GBA patient neurons and 3. had a coefficient of variation (CV) of ≤ 30% or less were selected for further analysis. Furthermore, peptides with N-linked glycosylation (NxS/T/C motif) were manually sorted based on information from UniProt⁸⁷ on known glycosylation and cellular localisation (Golgi/endosome/lysosome/membrane/extracellular) to exclude spontaneous deamidations.

Pathway and enrichment analyses

The pathway analysis was performed using the Ingenuity Pathway Analysis software (IPA, QIAGEN) on the four datasets; the differentially regulated non-modified-, phosphorylated-, glycosylated- and cysteine-modified proteins. Direct and indirect relationships, either experimentally observed or highly predicted based on data from CNS tissue or cell lines, were included. The IPA comparison analysis was used to combine results from all four datasets and rank the most significantly enriched pathways according to their IPA calculated p-values. An integrated p-value was calculating based on the mean of the p-values from each dataset.

The Cytoscape StringApp was used to visualize networks of differentially expressed proteins and PTMs using the Omics Visualizer app with a confidence score cut-off of 0.7.^72,73 Furthermore, enrichment analysis was performed focusing on Reactome pathways in the StringApp. Hierarchical clustering of the 500 most abundant proteins was computed using the ClusterMaker app analyzing the pairwise average linkage with Euclidean distance as distance metric and results shown as a heatmap.⁷⁴ The ClueGO Cytoscape plug-in was applied to visualize the functionally grouped Gene Ontology (GO) terms in the category “Biological process”, which the differentially expressed proteins in the four datasets were annotated to.⁷⁵ Enrichment was determined by two-sided hypergeometric test with Bonferroni step-down. The Homo sapiens (9606) marker set was applied with the following settings: GO tree levels of minimum 3 and maximum 8 and a Kappa score threshold of 0.4. For GO term selection a minimum of 3 genes and 4% of genes were used for the non-modified, whereas a minimum of 4/5/10 genes and 5/8/15% were used for the larger lists of glycosylated proteins, phosphorylated and cysteine-modified proteins, respectively. Visualisations of GO term enrichment data was performed in R using ggplot2.^76,77

PCA plots and heatmaps for PTM datasets were generated using Perseus.⁷⁸ Heatmaps were generated from z-score median normalized data and were clustered using Euclidean distance and k-means clustering. Subsequent categorical enrichment was performed using g:Profiler⁸⁸ and significant GO molecular function/cellular component/biological process and KEGG terms aggregated to provide an overview of cluster function/compartment.

Enrichment of PD-associated genes was calculated using Fisher’s exact test and based on the number of proteins identified in one or more of the datasets (Tables S3A, S3C, S3E, and S3G) and the number of regulated proteins showing protein and/or PTM level changes (Tables S3B, S3D, S3F, and S3H) compared to the number of identified and regulated PD-associated proteins.

Mitochondrial analysis

iPSC-derived dopaminergic neuronal cultures were cultured on 96-well half area plates (Greiner). On DIV 35 cells were incubated in neuronal maturation medium with NucBlue LiveReady dye (ThermoFisher) for 10 min and then 500 nM Mitotracker Deep Red was spiked in for 60s before cells were washed and incubated in pre-warmed neuronal maturation medium. Cells were imaged at 44s intervals for 12 cycles (total time 484s), NucBlue was imaged using ex375nm/em435-550nm, mitotracker Deep Red was imaged using ex640/em650-760nm, in confocal mode with 1x pixel binning, using an Opera Phenix microscope (PerkinElmer). Images from duplicate wells were analyzed using Harmony analysis software (PerkinElmer) with an average of 8391 mitochondria per well quantified for morphology and speed of motility.

Neurite outgrowth assay

Scratches were applied to differentiation day 30 neurons across the centre of each well in a 96-well plate using a 10 μl pipette tip followed by replacement of the neuronal maturation medium. Cells were treated with NCGC00188758 (NCGC758, Sigma-Aldrich #5316600001) after the scratch was applied and media +/- NCGC758 was replaced every 48h. The scratch area was imaged daily for a week using the Opera Phenix microscope (Perkin-Elmer). Images were analyzed using ImageJ by sharpening and converting images to binary, marking the scratch area and for each day calculating the area of the scratch covered by processes (particle size: 3-1000, circularity: 0.0-0.8).⁸⁹ Values from day 30 were subtracted as background.

Lenti-virus construction and production of shRNA targeting MAPT exons 12-13 (constitutive exons for targeting total MAPT) and no known RefSeq transcript (scrambled, non-targeting shRNA) was performed as previously described.⁹⁰ IPSC-derived neuronal precursors were transduced with lentiviral particles encoding shRNAs targeting MAPT or scrambled on differentiation day 20 following replating.

Quantification and Statistical Analysis

Analysis was performed in GraphPad Prism version 5.0 (GraphPad Software) using two-tailed paired or unpaired Student’s t-tests, one sample t-test, Fisher’s exact test and one-way ANOVA with Dunnett’s multiple or Tukey’s multiple comparisons, where appropriate. Results are expressed as mean ± SEM, and p-values ≤ 0,05 were considered statistically significant. Statistical details for each experiment can be found in the figure legend.

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2023.112180.

Highlights.

• Proteomic/PTMomic analysis of GBA patient iPSC neurons identify disease phenotypes

• PTM changes are highly abundant on lysosomal proteins and PD GWAS-related proteins

• mTOR activation and mitochondrial movement perturbations are identified

• GBA patient neurons show neuritogenesis defects that are rescued by a GCase chaperone

Data and code availability

• The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE¹ partner repository. Illumina SNP datasets and Illumina HT12v4 expression array datasets have been deposited in GEO. Data are publicly available as of the date of publication. Accession numbers are listed in the key resources table.

• All original code has been deposited at Github and is publicly available as of the date of publication. DOI is listed in the key resources table.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.