Fine mapping and eQTL analysis of CTSB variants
To identify the most likely variant driving the PD-association in the CTSB locus, we performed analyses using the summary statistics from the most recent PD GWAS 9 and multiple bioinformatic tools. First, to examine whether there are multiple independent associations in this locus, we used genome-wide complex trait conditional and joint analysis (GCTA-COJO) 26, using default parameters. For downstream analyses, we generated a linkage disequilibrium (LD) matrix for the CTSB locus using PLINK 1.9 27, including all variants within ± 1Mbp from the top variant. Then, we performed fine-mapping of the CTSB locus to nominate the most likely driving variants using FINEMAP 28, with minor allele frequency (MAF) threshold > 0.01. Expression quantitative trait locus (eQTL) analysis was performed using colocalization (COLOC) 29, which examines whether the same variants associated with the trait (PD) are also associated with gene expression. QTLs were tested in a total of 109 tissues and cells (Supplementary Table 1) To further explore the link between genetic variants, QTLs and PD we used Summary-data-based Mendelian Randomization (SMR), which uses Mendelian randomization to suggest potential causality, followed by heterogeneity in dependent instruments (HEIDI) to distinguish between pleiotropy (or causality) and LD 30.
Rare Variant Analysis
Rare variant analysis was performed on 5,801 PD cases and 20,427 controls across six cohorts (Supplementary Table 2). All patients were diagnosed by movement disorder specialists according to the UK brain bank criteria 31 or MDS diagnostic criteria 32. From the AMP-PD and UKBB cohorts we only included participants of European ancestry and excluded any first and second-degree relatives from the analysis. Quality control procedures for AMP-PD and UKBB were performed as previously described in detail 33,34.
In addition, we conducted sequencing on four distinct cohorts at McGill University (McGill cohort, Columbia cohort, Sheba Medical Center cohort and Pavlov and Human Brain Institutes cohort). We performed sequencing of the CTSB gene, including exon-intron boundaries (± 50bps) and the 5’ and 3’ untranslated regions (UTRs) using molecular inversion probes (MIPs) as described previously 35. The full protocol is available at https://github.com/gan-orlab/MIP_protocol. Library sequencing was performed by the Genome Quebec Innovation Centre on the Illumina NovaSeq 6000 SP PE100 platform. We used Genome Analysis Toolkit (GATK, v3.8) for post-alignment quality checking and variant calling 36. We applied standard quality control procedures 37. In brief, only variants with minor allele frequency (MAF) of less than 1% and a minimum quality score (GQ) of 30 were included in the analysis. The average coverage for CTSB in cohorts sequenced at McGill was > 4000X with 95% nucleotides covered at 30x (Supplementary Table 3).
To analyze rare variants, we applied the optimized sequence Kernel association test (SKAT-O, R package) 38 with further meta-analysis between the cohorts using metaSKAT package 39. We performed separate analyses for the whole gene, non-synonymous and functional (nonsynonymous, stop and frameshift variants) and variants with Combined Annotation Dependent Depletion (CADD) scores of ≥ 20 40. We adjusted for sex, age and ethnicity in all analyses. We also analyzed whether rare CTSB variants affected age at onset of PD.
Generation of CTSB-KO and SNCA-KO iPSC
The previously characterized AIW002-02 iPSC cell line 41 was used to generate CTSB and SNCA knockout lines. CRISPR gRNAs were designed using Synthego and the sequences of reagents used are depicted in Supplementary Table 4. The SNCA-KO line was created by using two gRNAs to introduce a 122bp deletion into exon 1 of the gene. The gRNA sequences were cloned into a Cas9/puromycin expression vector PX459 (Addgene, #48139) and transfected into iPSCs using Lipofectamine TM stem reagent (ThermoFisher Scientific). Transfected iPSCs were selected in 0.3µg/mL puromycin for 72h and surviving colonies were manually picked and expanded for PCR screening to confirm deletion of the target region. Colonies confirmed to be knockout by PCR screening were further validated by sanger sequencing, and loss of protein was confirmed by Western blot in differentiated neurons.
The CTSB-KO cell line was created by HDR using a single gRNA and ssDNA repair template to introduce a stop tag in exon 4 of the gene. Cas9 nuclease, gRNA and a ssDNA repair template for HDR were introduced by Lonza Nucleofection. Edited alleles were detected with ddPCR to select edited clones and deletion was verified by PCR screening followed by sanger sequencing, and finally loss of protein was determined by Western blot (Supplementary Fig. 1).
All lines were subject to quality control as previously described 41 and included verification of pluripotency by immunofluorescent staining for pluripotency markers (Nanog, Tra1-60, SSEA4 and OCT3/4), verification of normal karyotype and verification of normal profile on genome stability test (Supplementary Fig. 1).
iPSC culture and dopaminergic neuron differentiation
All cell culture reagents used, and media compositions are depicted in Supplemental Table 5. Midbrain neuronal precursor cells (NPCs) and dopaminergic neurons were generated following previously established protocols 42,43. Briefly, iPSCs were dissociated with Gentle Cell dissociation reagent and transferred to uncoated flasks in NPC Induction Media to allow for embryoid bodies (EB’s) to form. EBs were cultured for 7 days and then transferred to polyornithine/laminin coated flasks and grown for another 7 days in NPC induction media. To expand NPCs the EBs were then dissociated into small colonies by trituration in Gentle Cell dissociation media and replated as a monolayer on polyornithine/laminin coated flasks. After reaching confluence, NPCs were harvested and frozen in FBS with 10% DMSO and stored in liquid nitrogen.
To differentiate neurons, NPCs were thawed in NPC Maintenance Media with Y-27632 (ROCK inhibitor) and plated on polyornithine/laminin. NPCs were grown for 5–9 days until confluent. For final differentiation into dopaminergic neurons, NPCs were dissociated using Accutase and plated on polyornithine/laminin in Dopaminergic Differentiation Media. After 5 days, media was supplemented with mitomycin C to remove proliferative cells. Dopaminergic neurons were maintained by exchanging 1/3 of the culture volume for fresh dopaminergic differentiation media every 5–7 days. Neurons from every batch were assessed by immunofluorescence for expression of Map2 and TH (Supplementary Fig. 2A,B), and only batches achieving at least 50% Map2/TH positivity after 4 weeks of differentiation were used for the experiments included in this manuscript.
For high-content imaging experiments neurons were plated on 96-well plates at a density of 15,000 cells per well. For protein and RNA isolation experiments neurons were plated on 6-well plates at a density of 750,000 cells per well. For live-imaging experiments, neurons were plated on 4-chamber imaging dishes at a density of 100,000 cells per well.
Organoid culture, treatment, and imaging
The patient derived iPSCs with SNCA triplication mutation (3xSNCA) and corresponding SNCA knockout (SNCA-KO) were previously described 44 and provided by Dr. Tilo Kunath. These cells were used to generate midbrain organoids following a protocol previously established in our labs 45,46. Three months after organoid induction they were treated with either DMSO (vehicle) or 1 µM CA074me and treatment was maintained for 60 days. All organoids (12 per group) were fixed, cryo-sectioned, and prepared for immunofluorescence using antibodies against Map2, TH, α-syn and pSyn-129 as previously described 45. Verification of expression of TH and absence of α-syn in SNCA-KO in the organoids used in this study is depicted in Supplementary Fig. 3. Images were acquired using the Leica TCS SP8 confocal microscope and image analysis was performed with an in-house developed script for quantification of immunofluorescent signal in organoids (https://github.com/neuroeddu/OrgQ).
α-synuclein preformed fibril (PFF) generation and characterization
PFFs were generated from recombinant α-synuclein monomers as previously described 47,48. All PFFs underwent quality control assessment by electron microscopy (Supplementary Fig. 2C,D).
RPE1 CRISPRa and CRISPRi Cell Line Generation
Human retinal pigment epithelial-1 cells (RPE1) were grown in Dulbecco’s Modified Eagle Medium (Wisent) supplemented with 10% fetal bovine serum (Wisent). To generate CRISPRa and CRISPRi parental cell lines, lentivirus was used to stably transduced RPE1 cells with either pLX_311-KRAB-dCas9 49 (Addgene #96918, henceforth referred to as CRISPRi) or EF1a-FLAG-dCas9-VPR 50 (Addgene #114195, henceforth referred to as CRISPRa) and single clones were selected and characterized to generate monoclonal parental lines stably expressing the CRISPRa and CRISPRi machinery. The gRNA sequences targeting our genes of interest were selected from previously published CRISPRa/i libraries 51 (Supplemental Table 6), synthesized by IDT and cloned into pCRISPRi/a-v2 51 (Addgene #84832). Lentivirus was used to stably transduce parental CRISPRa and CRISPRi cell lines which then underwent puromycin selection to generate polyclonal cell lines expressing the gRNA of interest. For each target, several gRNAs were tested and the best performing sequences were selected by assessing target modulation by RT-qPCR analysis of gene expression.
Drug and PFF treatments
CA074me (Selleckchem) and PADK (Bachem) treatments were performed at the indicated final concentrations with DMSO as vehicle. For PFF experiments in Figs. 2 and 3, a single drug treatment was performed simultaneous with PFF administration, after which media was refreshed every 5–7 days. For lysosomal assays in Fig. 3, drug was administered 5 days prior to the assay unless otherwise specified.
For PFF treatments on RPE1 cells, 50,000 cells were plated on 12-well plates. After 24 hours, PFF was added and cells were allowed to continue growing for 48 hours before cells were washed with PBS and dissociated with trypsin to remove non-internalized PFFs before being lysed in RIPA buffer.
For PFF high-content imaging assays with PFF, neurons were treated with PFF after 2 weeks of differentiation in 96-well plates. After treatment, media was refreshed every 5–7 days normally. At completion of treatment, cells were washed with PBS and fixed with 4% paraformaldehyde.
High-content imaging - immunofluorescence
Cells were permeabilized for 10 minutes with 0.3% saponin (lysosome immunostaining) or 0.2% triton X-100 (pS129- α-syn assay and TFEB assay) in PBS and blocked with 1% BSA, 4% goat-serum and 0.02% triton X-100 in PBS. Antibodies used are described in Supplemental Table 7. High content imaging was performed on an Opera Phenix high-content confocal microscope (Perkin Elmer) and image analysis was performed using Columbus (Perkin Elmer). Data processing was then conducted using R studio as previously described 52. Briefly, nuclei were first identified by the Hoechst channel, and surrounding soma was identified as Map2-positive region. Relevant secondary stains were then quantified within this Map2-defined region. Single-cell data were then processed using a custom script in R studio to filter objects based on nuclear size, nuclear shape and Map2 staining intensity to identify only the neuronal cells for inclusion in subsequent analysis.
High content imaging - live cell assays
PFB-FDGlu GCase activity assay: Cells in 96-well plates were pre-loaded for 30 minutes with lysotracker deep red (1:20,000, Invitrogen). Media was then exchanged for FluoroBrite imaging media (Thermo) containing 25uM of PFB-FDGlu (Invitrogen) and cells were then imaged on the Opera Phenix every 15 minutes for 2 hours to monitor GCase activity. Using the Columbus software, lysotracker signal was used to identify cells for quantification of GCase substrate fluorescence, which is depicted as the mean fluorescence per cell.
DQ-BSA: Cells were pre-loaded with DQ-BSA (Invitrogen) for the indicated duration in standard culture media. Cells were then stained with lysotracker deep red (1:20,000, Invitrogen) for 30 minutes and media was exchanged for FluoroBrite and imaging was conducted on the Opera Phenix. Using the Columbus software, lysotracker signal was used to identify cells and DQ-BSA fluorescence intensity was measured.
Whole-cell proteomics mass spectrometry
For proteomics experiments on iPSC-derived DA neurons, 750,000 neurons were plated on 6-well plates and differentiated for 3 weeks. After 3 weeks, neurons were treated with CA074me and/or 300 nM of PFF and then maintained normally for 3 weeks. No additional drug or PFF were added during the maintenance period. Sample processing, mass-spectrometry and data analysis was performed by The Proteomics and Molecular Analysis Platform at the Research Institute of the McGill University Health Centre (RI-MUHC). Samples were processed for TMT labelling according to the manufacturer recommendations (ThermoFisher TMT-16plex reagents). Labelled peptides were fractionated using Pierce™ High pH Reversed-Phase Peptide Fractionation Kit into 8 fractions. Each fraction was re-solubilized in 0.1% aqueous formic acid and 2 micrograms of each was loaded onto a Thermo Acclaim Pepmap (Thermo, 75uM ID X 2cm C18 3uM beads) precolumn and then onto an Acclaim Pepmap Easyspray (Thermo, 75uM X 15cm with 2uM C18 beads) analytical column separation using a Dionex Ultimate 3000 uHPLC at 250 nl/min with a gradient of 2–35% organic (0.1% formic acid in acetonitrile) over three hours running the default settings for MS3-level SPS TMT quantitation 53 on an Orbitrap Fusion instrument (ThermoFisher Scientific) operated in DDA-MS3 mode.
To translate .raw files into protein identifications and TMT reporter ion intensities, Proteome Discoverer 2.2 (ThermoFisher Scientific) was used with the built-in TMT Reporter ion quantification workflows. Default settings were applied, with Trypsin as enzyme specificity. Spectra were matched against the human protein fasta database obtained from Uniprot(2022). Dynamic modifications were set as Oxidation (M), and Acetylation on protein N-termini. Cysteine carbamidomethyl was set as a static modification, together with the TMT tag on both peptide N-termini and K residues. All results were filtered to a 1% FDR.
Pathway analysis of differentially abundant proteins was conducted using STRING 54.
Western Blot
Cultured cells were washed with PBS and collected in RIPA lysis buffer with protease inhibitors. Protein concentration was determined using the Pierce™ BCA Protein Assay Kit (Thermo Scientific™) and proteins were prepared at the desired concentration in 6X Laemmli buffer and heated at 95°C for 5 minutes. 20 µg of protein were loaded on polyacrylamide gels, run with SDS running buffer and transferred onto nitrocellulose membranes and blocked for 30 mins in 5% skim milk made in 1X PBS with 0.1% Tween. For alpha-synuclein blots membranes were fixed using 4% PFA and 0.1% glutaraldehyde for 30 minutes before blocking. Blocked membranes were incubated with primary antibody (Supplemental Table 7) at 4°C overnight followed by HRP-conjugated secondary antibodies for 90 mins at room temperature. Protein detection was performed by chemiluminescence using Clarity Western ECL Substrate (Biorad) and Western blots were quantified using ImageJ.
RNA Extraction and qRT-PCR
RNA isolation was performed using the RNeasy Mini Kit (Qiagen) and cDNA was generated by RT-PCR using the MMLV Reverse Transcriptase kit (Thermo) with random hexamer primers. Real-time quantitative PCR was performed using SSoAdvanced SYBR Green Master Mix (Biorad). Primers for specified target genes were designed using NCBI PrimerBlast (Supplemental Table 8).
Live cell confocal imaging and analysis
Neurons were plated on CELLview™ 4-chamber imaging dishes (Greiner) at 100k cells per well. After 3 weeks of differentiation neurons were treated with alexa633 labelled PFF and/or CA074me. After 72 hours neurons were washed and incubated with 50nM of Lysotracker™ Green DND-26 (Invitrogen) for 30m at 37°C in standard culture media. The dye solution was exchanged for FluoroBrite™ DMEM (Gibco) and plates were immediately imaged. Images were acquired on a custom Andor spinning disc confocal microscope at 100X magnification. Single frames were acquired for cell bodies (488nM Lysotracker, 647 PFF). For neuronal trafficking movies, frames were acquired every 1. 5s for a total of 61 frames.
For the analysis of somatic lysotracker and PFF colocalization cell bodies were masked manually using FIJI ImageJ. For each image, Lysotracker and PFF signal were binarized using Otsu automatic thresholding, and binarized co-cluster signal was obtained using the “Image Calculation > > AND” function. Somatic slice densities of Lysotracker, PFF, or co-clusters were calculated via “Analyze Particles”. Finally, percentage of PFF in lysosomes was obtained by normalizing co-cluster particle density to Lysotracker density, and percentage of lysosomes containing PFF was similarly obtained by normalizing co-cluster particle density to PFF particle density.
Lysosomal motility was analyzed using the FIJI ImageJ TrackMate plugin 55,56. The generated tracks were then filtered by max track speed and then analyzed using Python.
Electron Microscopy
RPE1 cells grown in Lab-Tek chambers (Nunc) were rinsed in 0.1M Na Cacodylate buffer and fixed with 2.5% glutaraldehyde in 0.1M Na Cacodylate for 24 hours at 4°C. Cells were then post-fixed with 1% aqueous osmium tetroxide (Mecalab) for 1 h at 4°C, and stained with 4% uranyl acetate (EMS) in 70% ethanol for 45min at 4°C. After dehydrations in ascending alcohols, cells were embedded in Epon resin (Mecalab), and cut at 75 nm thickness in the ultra microtome. Sections were collected in 200 Mesh cooper grids (EMS) and stained with 4% uranyl acetate for 5min following by Reynold’s lead citrate for 2 min. Sections were visualized using a transmission electron microscope (Tecnai G2 Spirit Twin 120 kV Cryo-TEM) coupled to a camera (Gatan Ultrascan 4000 4 k × 4 k CCD Camera model 895). The identification of cellular elements was based on standard descriptions 57.
Statistical Analysis
Statistical analysis was conducted in GraphPad Prism9 software. For experiments with iPSC-derived neurons biological replicates were defined as experiments conducted at different times from the same batch of banked NPCs, or as experiments conducted in parallel from different batches of NPCs. A minimum of 3 distinct batches of NPCs were used for each experiment. Statistical comparisons were performed using t-tests (only 2 conditions), Bonferroni-corrected t-tests (more than 2 conditions compared). Significance levels are depicted in figure legends.