Successful PSC differentiation into functional cell types has been guided by knowledge of the endogenous developmental signals regulating differentiation of progenitor cells into tissue-specific cell types. Wnt proteins are one such important class of evolutionarily conserved signaling molecules, activating β-catenin signaling to mediate embryonic and tissue stem cell self-renewal and differentiation (Table S4) and Power SYBR Green according to manufacturer's protocol (ThermoFisher) in triplicates. Samples were run and analyzed on a CFX384 Real-Time PCR Detection System (BioRad) using the following thermocycling conditions: 95°C for 10 min, followed by 40 cycles of 95°C for 15 s, 60°C for 1 min and 72°C for 30 s. Mean relative gene expression was analyzed with the ΔΔCt method using CFX Maestro software and normalized to the housekeeping gene PPIB.

For single gene knockout studies, individual gRNAs (see Table S5 for gRNA sequence list) were ligated into Esp3I digested LentiCRISPRv2, a generous gift from Feng Zheng (Addgene, 52961), by following established protocol (Table S5. INDELs and sgRNA cutting efficiency were measured by comparing .ab1 files of targeted amplified genomic regions of the wild-type cells with the knockout cells using the online TIDE software (113631), was used. Synthetic design of pJ151-LMX1A-IRES-EGFP was planned using SnapGene cloning software. Briefly, a IRES-GFP fragment was subcloned into the pJ151 vector downstream of the HDR sites using Gibson Assembly (New England Biolabs). Left and right homology arms (LHA and RHA) from genomic DNA of H1 hESCs were amplified using PCR (Kapa Biosystems), designed for integration immediately after the stop exon of the human LMX1A gene (sequence accessed by GenBank file from NCBI). The homology arms were cloned into the backbone using restriction cloning. The LHA was cleaved using enzymes AscI and BAMHI (New England Biolabs), and the RHA was cleaved using SpeI and ClaI (New England Biolabs), and subcloned into the pJ151-HDR-IRES-GFP vector. The CRISPR nickase strategy was used for targeted insertion (42335). For H1 LMX1A-GFP generation, the pair of nickases were electroporated into H1 hESCs with the pJ151-LMX1A-IRES-GFP vector using the Neon Transfection System (Invitrogen) following default parameters. The cells were selected using puromycin and tested for LMX1A-GFP activity. See Table S5 for the list of sequences for primers and gRNA used for H1 LMX1A-GFP generation.

HEK 293T cells were cultured in DMEM (Gibco) with 10% fetal bovine serum (Gibco) and were transfected at ∼60% confluency with 5 μg of a psPAX2 packaging plasmid and 2 μg of a VSV. G enveloping plasmid and 5 μg of the intended LentiCRISPRv2 gRNA construct in 250 μl of OptiMEM (Gibco) reduced serum media. A 3:1 ratio of polyethylenimine (Sigma-Aldrich) to DNA was prepared separately, diluted in OptiMEM and briefly vortexed and incubated for 20 min, then added dropwise to cell culture plates. Media were changed the following day, and 24 h later those media were harvested and spun down (2000 g for 2 min). Supernatant was filtered with a 0.45 μm filter and Lenti-X Concentrator (Takara) was added to media in a 1:3 volume ratio. Tubes were rotated for 1 h at 4°C then spun at 1500 g for 45 min at 4°C. The viral pellets were resuspended in 250 μl of DMEM/F12, aliquoted and kept at −80°C for long-term storage.

For the monolayer differentiation protocols, the cells were seeded and differentiated on Geltrex-coated round glass cover slips (Fisher Sci). At the timepoint stated, the cells were fixed with 4% paraformaldehyde (PFA) and stored at 4°C until ready for staining. For midbrain organoids, the organoids were fixed in 4% PFA overnight and washed extensively the next day before placing in a 30% sucrose (BioShop) solution in PBS at 4°C overnight, Subsequently, the organoids were then embedded in OCT compound (VWR) for cryosectioning. The frozen organoids were sectioned at 16 μm using a cryostat (Leica) and mounted on Superfrost plus microscope slides (ThermoFisher Scientific). For staining, the slides were permeabilized and blocked in PBS with 0.1% Triton X-100 (Sigma-Aldrich) and 5% donkey serum (Sigma-Aldrich) for 1 h at room temperature. Primary antibody was prepared in blocking solution and added to coverslips for overnight incubation at 4°C, followed by extensive washing. Secondary antibody was added in blocking solution and incubated for 1 h in the dark at room temperature with the cover slips. After washing, the coverslips were mounted onto slides (VWR) using Fluoromount mounting media (Sigma-Aldrich) with DAPI (Sigma-Aldrich). For rat brain imaging, rats were anesthetized with 5% isoflurane before perfusion and then intracardially perfused with 50 ml phosphate-buffered saline (PBS), followed by 50 ml 4% paraformaldehyde in PBS. The brains were removed and stored in 4% paraformaldehyde in PBS overnight at 4°C and then transferred to a 30% sucrose solution for cryoprotection for 48 h at 4°C. Tissue was embedded in the freezing media HistoPrep (Fisher Chemical) and cut in 40 µm coronal sections. Sections were incubated in permeabilizing solution containing 1.2% Triton-X 100 in PBS followed by a blocking solution containing 10% normal goat serum (Thermofisher) in PBS for 1 h to avoid non-specific binding. Sections were then incubated in primary antibody overnight and then briefly washed the next day then incubated secondary antibody in 2% NGS for 1 h and washed. Sections were mounted in Vectashield medium (Vector Laboratories). For the antibody list, see Table S6. All slides were then imaged on a Zeiss LSM 710 confocal microscope. Processing of images was performed with ImageJ.

The cells were harvested using TrypLE (Gibco) for single cell dissociation before staining. For the H1 hESC LMX1A-GFP reporter line, the cells were incubated with the viability dye eFluor 780 (Invitrogen) for 30 min and washed extensively before analysis. For FZD antibody staining, the differentiated cells were harvested and blocked for 1 h with 3% BSA on ice. The cells were then stained with the selective FZD IgG antibody (https://www.r-project.org/) using default parameters (https://satijalab.org/seurat/) in R software v4.1. Seurat objects were created with the cell-gene expression files generated from the CellRanger pipeline for each treatment. The Seurat objects generated from each timepoint (i.e. day 11 or day 30) were merged. Cells were filtered based on a 200 cut-off for minimum number of genes, 7500 for maximum of genes and less than 5% of mitochondrial expression. The gene expression of each cell was normalized using the global-scaling normalization method to normalize RNA expression measurements for each cell by the total expression, which was then multiplied by a scale factor of 10,000. Highly variable genes by cell were computed and the top 2000 features from each dataset were used for subsequent downstream clustering. An elbow plot was used to estimate the number of principle components required for UMAP embedding. This distance matrix was then reduced to low dimensionality using UMAP with the first 15 principal components dimensions and clusters identified using a resolution of 0.5. For differential gene expression between clusters, the FindMarkers function was used with a comparison of the indicated clusters. GSEA was performed with GSEA software (Broad Institute) using a ranked gene expression list based on Log2Fold change for day 11 cluster 0 versus clusters 1, 2 and 5. The GO:0045746 (Notch), GO:0000165 (MAPK), M5923 (PI3K_AKT) gene sets were used for GSEA extracted from MSigDB. For all subsequent analyses, meta, filtered and normalized data were exported from Seurat for integration in various R packages. The CellCycleScoring was performed using the list of signature genes from (https://github.com/carmonalab/UCell. The list of gene markers used for UCell signature scoring is listed in Table S3. The scRNA datasets were prepared for the ShinyApp using the ShinyCell package (https://andyydh.shinyapps.io/scRNA_VM_progenitor/.

HiDDEN clustering analysis was performed on the 3990 DA⁺ neurons from the combined CHIR and F5L6 populations (https://github.com/tudaga/LabelCorrection. Dimensionality reduction was performed using principal component analysis. HiDDEN Continuous Scores were derived as the fitted conditional probability of label 1 given the gene expression profile for each cell computed by training a logistic regression. The optimal number of principal components (P=55) to use in the logistic regression was determined using the Kolmogorov–Smirnov heuristic (Fig. S6I). GO term enrichment analysis of the 394 genes enriched in F5L6_L1 versus L0 (combined CHIR_L0 and F5L6_L0) was performed using the web tool PANTHER Overrepresentation Test (Released in 20221013) to identify significant GO biological process terms using Fisher's Exact test and a Bonferroni multiple testing correction with adjusted P< 0.05. Dot plots in Fig. S6J were produced in scanpy based on the HiDDEN-refined binary labels and the log-normalized scaled gene expression.

All statistical methods are explained in the figure legends. Data were analyzed using GraphPad Prism (version 8) software in at least three independent experiments. Unless stated otherwise, values are shown as mean±s.e.m and asterisks in figures indicate significance according to a two-tailed Student's t-test between two groups (*P≤0.05, **P≤0.01).

We thank Kin Chan at the Network Biology Collaborative Centre and Troy Keleta at Princess Margaret Genomic Center for next-generation sequencing service. The center for Pharmaceutical Oncology provided the support and instruments that were used in these experiments. We also thank the Temerty Faculty of Medicine flow cytometry facility for assisting us with cell sorting experiments. We would also like to thank all members of the Angers labs for helpful discussion throughout the course of this study, particularly Dr Graham Macleod for providing feedback on the manuscript.

This article has an associated ‘The people behind the papers’ interview with some of the authors.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.202545.reviewer-comments.pdf