Organoids culture and condition

Briefly, the cerebral and ocular organoids were cultured in mTeSR1 medium (Stem Cell Technologies, cat. no. 85850). Human ES cell (Wicell H1 WA01 and H9 WA09) were treated by accutase to generate single cells. Then, 4000 cells were plated in each well of a V-bottom 96-well plate (Sematec Pte Ltd Code: 1009985) with low concentrations of basic fibroblast growth factor (bFGF 4 ng/ml) and 20 µM/ml Rho-associated protein kinase (ROCK) inhibitor (Y27632 Stem Cell). The next day, Embryonic Bodies (EB) were transferred into in low-attachment 96 well U-bottom plate with hESC medium (For 500 ml of medium, combine 400 ml of DMEM-F12, 100 ml of KOSR, 15 ml of ESC-quality FBS, 5 ml of GlutaMAX, 5 ml of MEM-NEAA and 3.5 μl of 2-mercaptoethanol) for cerebral organoid, Differentiation Medium DM (DMEM/F12, 4% knockout serum replacement (KOSR), 4% fetal bovine serum (ESC-quality FBS), 1× non-essential amino acids (NEAA), 1× Glutamax, 1× Pen-Strep. Filter it using a vacuum-driven 0.2-μm filter unit) for ocular organoid. EB were fed every other day for 6 days and then changed into neural induction media for cerebral organoid and into retinal differentiation medium (RDM: DM + 2% B27) for ocular organoid for the next 4 days. After the EB undergone neuro-ectodermal differentiation, they are transferred to Matrigel (Growth factor–reduced Matrigel, Bio-Lab 354230). Making matrigel in 1:1 dilution with cerebral organoid differentiation medium or corneal differentiation medium (CDM). 50 µl of matrigel is added to each well and incubated for 30 min in a 37 °C incubator, followed by adding 100ul cerebral organoid differentiation medium with B27 (−) Vitamin A to each well and cultured for 48 h. After 2–3 days, the aggregates (organoids) were transferred to 6-well clear flat-bottom ultra-low attachment plates. After 4 days of static culture with cerebral organoid differentiation medium with B27 (−) Vitamin A, the embedded organoids were transferred to an orbital shaker at 80 rpm within 37 °C, 5% CO2 incubator for long-term culture with cerebral organoid differentiation medium with B27 (+) Vitamin A.

AAV plasmid cloning and virus production

The barcoded eGFP plasmids were constructed by introducing a short sequence TAATAAATCGATCGNNNNNNNN after the eGFP transgene stop codon in the plasmid backbone pZac2.1-CMV-eGFP.rgb, a gift from Luk Vandenberghe. Primers with overhanging barcode were designed for first round PCR to generate barcoded eGFP fragments that terminates at ITR sequences. A second round of nested PCR amplify shorter fragments of barcoded eGFP which are digested with restriction enzyme NheI and BamHI. Digested fragments are ligated with the vector backbone which is digested using the same restriction enzymes. The sequences of the clones were checked by Sanger sequencing. The representing barcodes for each AAV serotype are shown in Table S1. The serotype-specific pAAV-RepCap plasmids were constructed by cloning in the Cap genes from the different serotypes into the pAAV-RepCap backbone using Gibson assembly. The different serotypes Cap genes were ordered as gene blocks (IDT) and cloned into HindIII/PmeI-digested pAAV-RepCap backbone via Gibson assembly to construct the pAAV-RepCap with the different serotypes Cap genes. AAV viruses from different serotypes each bearing its own barcode were produced as per standard protocol [32]. Briefly, AAV were packaged via a triple transfection of 293AAV cell line (Cell Biolabs AAV-100) that were plated in a HYPERFlask ‘M’ (Corning) in growth media consisting of DMEM + glutaMax+pyruvate+10%FBS (Thermo Fisher), supplemented with 1X MEM non-essential amino acids (Gibco). Confluency at transfection was between 70–90%. Media was replaced with fresh pre-warmed growth media before transfection. For each HYPERFlask ‘M’, 200 μg of pHelper (Cell Biolabs), 100 μg of pRepCap [encoding capsid proteins for different serotypes], and 100 μg of pZac-CASI-GFP (barcoded) were mixed in 5 ml of DMEM, and 2 mg of PEI “MAX” (Polysciences) (40 kDa, 1 mg/ml in H2O, pH 7.1) added for PEI: DNA mass ratio of 5:1. The mixture was incubated for 15 min, and transferred drop-wise to the cell media. The day after transfection, media was changed to DMEM + glutamax+pyruvate+2%FBS. Cells were harvested 48–72 h after transfection by scrapping or dissociation with 1×PBS (pH7.2) + 5 mM EDTA, and pelleted at 1500 g for 12 min. Cell pellets were resuspended in 1–5 ml of lysis buffer (Tris HCl pH 7.5 + 2 mM MgCl + 150 mM NaCl), and freeze-thawed 3× between dry-ice-ethanol bath and 37  °C water bath. Cell debris was clarified via 4000 g for 5 min, and the supernatant collected. The collected supernatant was treated with 50 U/ml of Benzonase (Sigma-Aldrich) and 1 U/ml of RNase cocktail (Invitrogen) for 30 min at 37  °C to remove unpackaged nucleic acids. After incubation, the lysate was loaded on top of a discontinuous density gradient consisting of 6 ml each of 15%, 25%, 40%, 60% Optiprep (Sigma-Aldrich) in an 29.9 ml Optiseal polypropylene tube (Beckman-Coulter). The tubes were ultra-centrifuged at 54000 rpm, at 18  °C, for 1.5 hr, on a Type 70 Ti rotor. The 40% fraction was extracted, and dialyzed with 1×PBS (pH 7.2) supplemented with 35 mM NaCl, using Amicon Ultra-15 (100 kDa MWCO) (Millipore). The titer of the purified AAV vector stocks were determined using real-time qPCR with ITR-sequence-specific primers and probe [33], referenced against the ATCC reference standard material 8 (ATCC).

In vitro transduction of organoids

AAV serotypes pool was created by pooling each AAV serotype at 1 × 1010 vg, giving a final viral copy of 9 × 1010 that is used for the transduction of organoids in each well of a 24-well plate. AAV1, 2, 6, 7, 8, 9, rh10, DJ and Anc80 serotypes were used for the pooling. Organoids were transduced for 7–10 days before harvesting for sequencing, fluorescence imaging, and histochemistry.

Immunofluorescence histochemistry

Organoids and marmoset’s corneas were fixed in 4% paraformaldehyde for 4 h at 4 °C followed by washing in PBS three times for 15 min. Organoids and marmoset’s corneas were allowed to sink in 30% sucrose overnight and then embedded in OCT and cryosectioned at 12 µm. Marmoset’s corneas were vertical embedded and sectioned. Sections were permeabilized in 0.2% Triton X-100 in PBS and blocked in block buffer (2% BSA 5% fetal bovine serum) for 1 h at room temperature. Sections were subsequently incubated with the indicated primary antibodies at a 1:100 dilution in block buffer at 4 °C overnight. Secondary antibodies used were donkey Alexa Fluor 488, 568 and 647 conjugates (Invitrogen, 1:1000). After staining with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) in PBS for 5 min, slides were mounted in Vectashield anti-fade reagent (Vector Laboratories). Confocal imaging was performed with Leica TCS SP8 DLS LightSheet microscope. Primary antibodies: PAX6 (rabbit, ab5790), CHX10 (rabbit, ab133636), ZO-1 (mouse, ThermoFisher ZO1-1A12), MAP2 (chicken, ab5392), S100β (rabbit, ab52642), RAX (rabbit, ab23340), CD31 (mouse, ab23340), aSMA (rabbit, ab5694), DAPI (49,6-diamidino-2-phenylindole), ThermoFisher D1306, NeuN (mouse, Sigma-Aldrich MAB377), AAV (rabbit ab45482), Cas9 (Sp)(E7M1H) XP® (rabbit, CST #19526, CK3 (mouse, ab68260).

Amplicon barcode sequencing and analysis

AAV serotypes pool were checked for percentage distribution by NGS library preparation and the correction factor was used for re-adjusting the pooling of the AAV serotypes for subsequent experiment. Custom primers were designed for a first round of 20-cycle PCR of the target site containing the AAV barcodes as shown in Table S1. Target bands are extracted using gel extraction and a second round of 15-cycle PCR were used for adding P5 and P7 adapter sequences to the enriched fragments and the final libraries were cleaned up by gel extraction. Primers used for library construction are shown in Table S2. Library concentrations were determined using a Qubit dsDNA HS kit (Agilent). NGS sequencing were carried out on the MiSeq using 2 × 75 bp PE run with 20% PhiX spike-in. An in-house python script was utilized to search for the 8 unique nucleotide barcode sequences representing each serotype within the MiSeq FASTQs generated from the MiSeq run of the amplicon libraries and the total count was tabulated for each barcode sequence for each sample (Refer to PythonScriptforBulkAnalysis package). Transduced organoid samples were harvested as single cells and processed through the 10X chromium machine for cell barcoding of the transcripts. The total cDNA was purified via the 10X workflow and 5ul was aliquoted for custom bulk-sequencing as described above. The rest of the cDNA were used to proceed with the remaining 10X workflow for single-cell sequencing.

Single cell sequencing and RNA transcriptomic analysis

Samples were prepared as indicated in the 10X Genomics Single Cell 3′ v2 Reagent Kit user guide. The single-cell libraries were prepared by following the manufacturers’ protocol followed by sequencing on an Illumina HiSeq4000 flow cell. The sequencing data were processed by the standard Cell Ranger pipeline using the modified gtf and genome manifest files. Briefly, the samples were washed twice in PBS (Life Technologies) + 0.04% BSA (Sigma) and re-suspended in the same solution. Sample viability was assessed using Trypan Blue (Thermo Fisher) under a light microscope. Following viability counting, the appropriate volume for each sample was calculated for a target capture of 10,000 cells and loaded onto the 10x Genomics single-cell-A chip along with other reagents and barcoded beads by following the protocol guide. The chip is then loaded onto a 10X Chromium machine for droplet generation and samples were transferred onto a pre-chilled strip tube (Eppendorf), and reverse transcription was performed using a 96-well thermal cycler (Thermo Fisher). After the reverse transcription, cDNA was recovered using Recovery Agent provided by 10X Genomics, followed by Silane DynaBead clean-up (10X Genomics). Purified cDNA was amplified for 12 cycles before being cleaned up using SPRI-select beads (Beckman). Samples were diluted 4 times in water and ran on a Bioanalyzer (Agilent Technologies) to determine cDNA concentration. cDNA libraries were then prepared following the Single Cell 3′ Reagent Kits v2 user guide with appropriate PCR cycles based on the cDNA concentration as determined by the bioanalyzer. The molarity of the single cell libraries was calculated based on their library sizes as measured using a bioanalyzer (Agilent Technologies) and using the KAPA qPCR quantification (KAPA) method on a qPCR cycler (Roche). Samples were normalized to 10 nM before sequencing. Each organoid sample was sequenced on a full lane on a HiSeq 4000 with the following run parameters: Read 1–26 cycles, read 2–98 cycles, index 1–8 cycles. Using the FASTQ files from each sample, the standard Cell Ranger Count command pipeline was performed for transcripts read alignment, UMI counting, and clustering (Amazon Web Services via the Ronin cloud platform). Raw data were processed using standard Cell Ranger transcriptomics command, while using modified genome reference file and the modified gtf file. For command lines for the modification of gtf file and genome reference file to include the barcoded GFP sequences, refer to Supplementary Data S1. Command lines for cell count using the modified files are also shown in Supplementary Data S1. Finally, upon successful cellranger count run, the output files can be used for single cells visualization and analysis on the Seurat software (v4.0.1).

Single cell AAV tropism analysis

For the purpose of parallel sequencing of the AAV barcodes in single cells along with the RNA transcripts, the human genome reference file and the genome transcript file (gtf) were modified (Supplementary Data S1). Briefly, the names and barcodes of each AAV serotypes are manually included into both files that will be used for the execution of the Cell Ranger Count command pipeline in order to include the AAV barcode transcripts into the read alignment, UMI counting, and clustering. To include the AAV barcode representation in the genome reference file, the command line “>GFP1 TAAATCGATCGNNNNNNNN” is included for each barcode, where the 8Ns represent a unique 8 nucleotide barcode sequence. The command line “GFP me exon 1 19 - + - gene_id “GFP1”; transcript_id “GFP1”” was included in the genome transcript file for each AAV barcode representation added to the genome reference file (Supplementary Data S1). After processing by Cell Ranger to obtain the gene count matrices, further computations were carried out in R (version 4.0.4). Quality control, normalization, PCA, clustering and downstream analysis were performed with Seurat (version 4.0.1). Briefly, cells with less than 200 RNA features or more than 10% mitochondrial reads were removed from analysis. The dimensions of the final data matrices are 17,022 features across 5741 cells (Ocular dataset) and 20281 features across 15,335 cells (Cerebral dataset). After normalization, PCA was performed with the top 2000 most variable features, followed by Louvain clustering with a resolution of 0.3–0.4 to achieve reasonable distinction of cell type clusters, and dimensional reduction using tSNE. Plots and figures were assembled using Seurat and the ggpubr (version 0.4.0) package. To determine the transduction specificity of a specific viral vector against a specific cell niche relative to other cell niches, we calculated the frequencies with which the presence of that specific viral vector is detected in the cells of the specific cell niche, against the frequencies with which the presence of the same specific viral vector is detected in the cells of other specific cell niches.

Marmoset animal work

A total of 2 marmosets were used in this study. The animals were sedated with 10–15 mg/kg Ketamine and 4–5% isoflurane, and then intubated with an appropriate endotracheal tube by an experienced veterinarian. The eyes were cleaned with 2.5% povidone-iodine and draped, before a small-sized eyelid speculum was inserted. A single injection of 30 μL of AAV6-SpCas9 (Total 1.9 × 1011 vgs) suspension was injected into the anterior chamber with a 30-gauge needle without touching the lens, iris, or other ocular tissue. After the injection, a Weck-Cel eye sponge was applied to the incision site for 1 minute to prevent bleeding and AAV suspension efflux. The anterior chamber was left formed with no aqueous leak. All the eyes received topical tobramycin ointment (Alcon, Fort Worth, TX, USA) once at the end of the procedure. Three months after injection, the animals were euthanized under general anesthesia with intravenous injection of pentobarbital (60–100 mg/kg). After enucleation of the eyeballs, the corneas were harvested. A small limbal incision was made with a diamond knife, and the cornea was excised 360° alongside the limbus with corneal scissors. There was no lens-cornea touch, damage of iris and lens tissue, or excessive corneal curvature deformation during the harvesting. The excised cornea was kept moist with normal saline. Subsequently, DM peeling was performed manually. The corneal tissue was placed onto an anatomically shaped bowel, of a Coronet corneal trephine vacuum punch (Network Medical Products, North Yorkshire, UK). During the DM peeling, the corneal tissue was submerged in normal saline solution to minimize the surface tension on the tissue and allow the DM to settle back onto the corneal stroma. The DM was then grasped with two fine non-toothed Kelman forceps and was slowly and gently stripped completely away from the stroma from the edge towards the center. The remaining tissue (with epithelial & stroma) is then cut into 4 quarters for different analytical procedures. One quarter was used for DNA extraction, one quarter for RNA extraction and the remaining two quarters for immunohistochemistry. No adverse event was noted.

qPCR quantification for marmoset tissues

Total DNA were extracted from a quarter of the harvested marmoset cornea using QuickExtract (Lucigen) by following manufacturer’s protocol. The titer of the purified AAV vector stocks were determined using real-time qPCR with ITR-sequence-specific primers and probe (Table S5), referenced against the ATCC reference standard material 8 (ATCC). Total RNA were extracted from a quarter of harvested marmoset cornea using the RNeasy universal plus mini kit (Qiagen) by following manufacturer’s protocol. cDNA was synthesized using Superscript III (Thermo Fisher) and polyT primer, and subsequently quantified using real-time qPCR with Cas9-sequence-specific primers and probe (Table S5). For normalization, GAPDH was quantified using real-time qPCR with marmoset GAPDH-sequence-specific primers and probe (Table S5).