In vitro and in vivo rescue of dopaminergic neurons in Parkinson’s disease models after Parkin gene therapy

Preparation of PRKN expression plasmids

All expression plasmids were designed by the authors and were produced at GeneWiz (South Plainfield, NJ, USA). Expression cassettes consist of promoter, intron, human PRKN cDNA (NM_004562.3, transcript variant 1, PRKN-206) (1398 base pairs (bp)), and human growth hormone poly A (477 bp), which was inserted in pUC57. For promoters, chicken β-actin hybrid (CBh, 822 bp) [29], human Elongation Factor 1 alpha (EF1α, 1187 bp) [30], the short, intron-less variant of EF1α (EFS, 232 bp) [31], human phosphoglycerate kinase 1 (PGK1, 511 bp) [32], human synapsin 1 (SYN1, 469 bp) [33], mouse Calcium–calmodulin-dependent protein kinase IIα (CaMKIIα, 1289 bp) [34], human tyrosine hydroxylase (TH, 521 bp) [35], and native human PRKN promoters (339 bp upstream of the ATG start codon) [36] were used. A chimeric intron from chicken β-actin and rabbit β-globin genes (CHMin, 1017 bp, the corresponding sequence of chimeric intron in the pCAGGS vector), human β-globin intron (hBGin, 476 bp) [37], and human chorionic gonadotropin intron (hCGin, 352 bp) [38] were used for intron sequences to improve RNA polymerase II recruitment.

AAV9 vector cloning and production

All PRKN AAV9 vectors were designed by the authors and were produced at the University of Massachusetts Medical School Viral Vector Core, Worcester, MA, USA. All expression cassettes consisted of promoter, intron, human wild-type (WT) or codon-optimized PRKN cDNA, and human growth hormone (hGH) poly A (Table 1 and Table 2). AAV9 vectors were produced using standard triple transient transfection methods. In brief, HEK293 cells were co-transfected with three plasmids: the inverted terminal repeat (ITR) plasmid containing the PRKN transgene, a replication and capsid (RepCap) plasmid, and an adenoviral helper gene plasmid. After 72–96 h, the cells were chemically lysed, and the cell pellet and medium were collected. Full AAV particles were enriched using CsCl sedimentation. The AAV vectors were formulated in phosphate-buffered saline (PBS) with 0.001% Pluronic F-68 (Thermo Fisher Scientific, Waltham, MA, USA) and were subjected to standard characterization, including digital droplet (dd) Polymerase Chain Reaction (PCR) for titrating. Purity assessed by 4–12% SDS-acrylamide gel electrophoresis and silver staining (Thermo Fisher Scientific). The full-to-empty particle ratio was quantified from the transmission electron microscopy (TEM) images using ImageJ.

Design of codon optimization of PRKN cDNA

Ten codon-optimized PRKN (coPRKN) sequences were generated by the GeneGPS^® platform at ATUM (Newark, CA, USA) using a human-specific central nervous system (CNS) codon usage table and codon adaptation index optimized for neurons, with additional annotations and de-risking parameters, including reduction of CpG motifs in WT PRKN coding sequence (from 46 CpG dimers in WT to 5 or 0), restriction enzyme sites were avoided for cloning purpose, removal of potential/cryptic splice sites, removal of cryptic start sites, and removal of nucleotide repeats motifs were included.

Generation of PRKN-null iPSCs

The control 201B7 human iPSC line was a gift from the Center for iPS Cell Research and Application (CiRA), Kyoto, Japan [39]. PRKN homozygous knockout human iPSCs on the 201B7 genetic background were generated by deleting the translation start ATG and inserting stop codons using gene editing tools. Deletion of the ATG codon was confirmed by DNA sequencing and Parkin protein expression by Western blot analysis, with no hParkin expression being detectable (Fig. S1A, B).

iPSC-DA neuronal culture and AAV transduction

Control 201B7 human iPSCs (WT) or isogenic PRKN-null human iPSCs were differentiated into DA neurons based on previously published methods (Fig. S1C) [40, 41]. Briefly, frozen stock WT iPSC or isogenic PRKN-null iPSC-derived midbrain floor plate progenitor cells were cultured for 6 days in ultra-low attachment dishes using the Floor Plate Cell Expansion Kit (Thermo Fisher Scientific) with penicillin/streptomycin (FUJIFILM Wako Chemicals Corporation, Osaka, Japan) in the presence (day 0–1) or absence of 10 μM Y-27632 (day 1–6) while changing the medium once every 2–3 days. Proliferated progenitors were dissociated by Accutase (Innovative Cell Technologies, San Diego, CA, USA), and dissociated cells were seeded in Matrigel-coated 96 well plate (CellCarrier-96 ultra, Perkin Elmer, Waltham, MA, USA) at 20,000 cells/well with total 200 μL medium of Floor Plate Cell Expansion Kit (Thermo Fisher Scientific) with 1× penicillin/streptomycin and 10 μM Y-27632 (FUJIFILM Wako Chemicals) for 1 day. The next day, half of culture medium was replaced with fresh differentiation medium (Neurobasal medium [Thermo Fisher Scientific] with 1× GlutaMax [Thermo Fisher Scientific], 1× B27 supplement [Thermo Fisher Scientific], 1× CulturOne supplement [Thermo Fisher Scientific], 100 μM ascorbic acid [Sigma-Aldrich, St. Louis, MO, USA], 0.5 mM dbc AMP [Sigma-Aldrich], 3 μM PD0325901 [FUJIFILM Wako Chemicals], 20 ng/mL Activin A [R&D Systems, Minneapolis, MN, USA], and 1× penicillin/streptomycin) while replacing half the medium twice a week. Eight days after seeding of single midbrain floor plate progenitor cells, differentiated DA neurons were transduced with AAV. Briefly, 100 μL of culture medium was removed, 100 μL of fresh differentiation medium including AAV were added at 0.03, 0.1, 0.3, and 1 × 10⁵ multiplicity of infection (MOI). Half volume of the culture medium was replaced with fresh differentiation medium twice a week and AAV-transduced iPSC-DA neurons were cultured for additional 13 days.

Parkin protein enzyme-linked immunosorbent assay (ELISA)

The hParkin protein levels were measured by human Parkin ELISA kit (Abcam, Cambridge, UK, ab212159) according to manufacturer’s instruction. DA neurons were lysed with 80 μL of cell extraction buffer PTR at 4 °C for 30 min. Lysates were transferred into 1.5 mL microtubes and were centrifuged at 18,000 g for 20 min at 4 °C. The supernatants of lysates were used in the Parkin ELISA. As a reference of Parkin levels in human brain, whole brain lysate from one human (Novus Biologicals, Centennial, CO, NB820-59177) was applied for the ELISA (Table S1 and Fig. S2).

Mitophagy induction

Carbonyl cyanide m-chlorophenyl hydrazone (CCCP) (Sigma-Aldrich) was dissolved in dimethyl sulfoxide (DMSO) and was used to induce mitophagy in iPSC-DA neurons [42]. Briefly, at 13 days after AAV transduction, half of culture medium was removed and then equal volume of fresh differentiation medium including 1× B27 minus antioxidants (Thermo Fisher Scientific) and CCCP at 20 μM was added. As negative control, neurons were treated with DMSO without AAV transduction. WT DA neurons derived from 201B7 iPSCs were treated with 20 μM CCCP or DMSO as positive control. Twenty-four (24) h after mitophagy induction, culture medium was completely removed, and DA neurons were fixed using 200 μL of cold 4% paraformaldehyde (PFA) for 30 min.

High-content analysis

Chain formation of phosphorylated ubiquitin (poly-pUb) as a readout for Parkin activity and mitochondrial protein TOM20 as a marker for clearance of mitochondria were quantified by immunocytochemistry (ICC). After permeabilization with 0.2% Triton X-100, 4% PFA-fixed DA neurons were blocked with 5% bovine serum albumin (Meso Scale Diagnostics, Rockville, MD, USA, R93BA-1) in PBS buffer with 0.05% Tween 20 (Thermo Fisher Scientific) for 60 min at room temperature. The cells were incubated with anti-phospho-Ubiquitin (Ser65) antibody (Merck, Darmstadt, Germany, ABS1513-I, 1:500), anti-TOM20 antibody (Santa Cruz Biotechnology, Dallas, TX, USA, sc-17764, 1:1000), and anti-MAP2 (Merck, AB5543, 1:1000) overnight at 4 °C. Then, the co-immunostained cells were incubated with Alexa Fluor dye-conjugated secondary antibodies (Thermo Fisher Scientific, A21206, A21203, A21103, 1:400) in combination with Hoechst 33342 (Abcam, 1:5000) to stain nuclei for 2 h at room temperature. Fluorescence of 16 fields (fixed spacing fields) per well was imaged with the IN Cell analyzer 6500HS (Cytiva, Marlborough, MA, USA) and was quantified with IN Cell investigator (Cytiva) using a 40 × objective lens. The number of nuclei was quantified by counting the cell nuclei labeled by Hoechst 33342. The area with positive fluorescence signal (μm²) in 16 fields was expressed as an arbitrary unit for poly-pUb and TOM20 analysis. All pseudocolor images were generated by NIS-Elements (Nikon Corporation, Tokyo, Japan).

Generation of Prkn knockout (KO) mice

Prkn homozygous KO mice were generated on the C57BL/6 J (CLEA, Fujinomiya, Japan) background by deleting the exon 2 of murine Prkn gene by CRISPR/Cas9 system in the zygote at the Axcelead Drug Discovery Partners (Fujisawa, Japan). Briefly, Cas9, 4 guide RNAs, and single-stranded oligodeoxynucleotide (ssODN) were injected into the zygote to generate F1 heterozygous Prkn KO mice (Fig. S2A, Table S2). Deletion of the exon 2 in F1 mice were confirmed by PCR (forward primer, 5′- CCATTAAGGGTTTGGAGTAATATTCTTTCTT-3’; reverse primer, 5′- AAGATGGACTTTTGCTGTGTGAATTTGAT-3’) (Fig. S2B). Homozygous Prkn KO mice were produced by in vitro fertilization using fresh sperm and oocytes isolated from heterozygous or homozygous Prkn KO mice, followed by a measurement of dosage of exon 2 in Prkn gene by quantitative PCR (forward primer, 5’-TTTGCCGGGAAGGAGCTT-3’; TaqMan probe, 5’-FAM-CGAATCACCTGACGGTT-NFQ-MGB-3’; reverse primer, 5’-CATTAGCTGCAGGGTCTTTGG-3’). The dosage of exon 2 was normalized by a dosage of nerve growth factor (Ngf) gene (forward primer, 5’- TGCATAGCGTAATGTCCATGTTG-3’; TaqMan probe, 5’-VIC- ACGGTTCTGCCTGTACGCCGATCA-NFQ-TAMRA-3’; reverse primer, 5’- TCTCCTTCTGGGACATTGCTATC-3’). No Parkin protein was detectable in the brain of Prkn KO mice by Western blot analysis (Fig. S2C). Prkn KO mice are viable and healthy and show no gross behavior phenotype even at 125 weeks of age, and only a slight reduction of dopamine levels in the striatum compared to that in WT mice at 125 weeks of age (Fig. S2D, E). All mice were housed under laboratory conditions (12 h light/dark cycles) with food (CE-2, CLEA Japan) and water available ad libitum at the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC)- accredited animal facility in Shonan Health Innovation Park (Fujisawa, Japan). The care and use of the animals and the experimental protocols conducted in Takeda were approved by the Institutional Animal Care and Use Committee (IACUC) of Shonan Health Innovation Park.

Characterization of aged Prkn KO mice

Rotarod test was performed on a rotarod treadmill (Muromachi Kikai, Tokyo, Japan, MK-670). Nine male 125-week-old Prkn KO mice and fifteen WT mice were placed on a rotating rod (3 cm diameter), and rotation was accelerated from 3 to 40 rpm over 300 s. Fall latency was recorded three time with more than 30 min interval, and the first fall latency of the third trial was used for analysis. After the rotarod test, striatum of mice was removed, and quickly frozen for stock. Striatal dopamine levels were measured by LC/MS/MS at the Axcelead Drug Discovery Partners (Fujisawa, Japan). Briefly, frozen samples were homogenized with 99 volumes of extraction buffer (acetonitrile/ 500 μg/mL ascorbic acid/ 500 μg/mL EDTA, mixture of equal volume) and zirconia beads. Samples that were spiked with internal standards, centrifuged at 22,500 × g for 5 min at 4 °C and the supernatant was collected and dried. After resuspension in 1 M bicarbonate buffer and 4 mg/ mL dansyl chloride by vortex mixing, followed by sonication, the resuspension was incubated for 15 min at 60 °C, and then treated with 15% formic acid. The mixture was centrifuged at 22,500 × g for 5 min at 4 °C, the supernatant was subjected to analysis by LC/MS/MS.

Intra-SN AAV9 administration

AAV9 vectors were administered to the SN of Prkn KO mice. Mice were randomly allocated to each group. Four-month-old or six-month-old male Prkn KO mice were anesthetized with the mixture of 3 anesthetic agents (medetomidine 0.3 mg/kg, midazolam 4 mg/kg, and butorphanol tartrate 5 mg/kg). Anesthetized mice were placed in a stereotaxic frame (Kopf Instruments). In case of the 6-OHDA lesion model four-month-old mice were injected with five (5) μL to allow a broader distribution towards the MFB, the injection site of the 6-OHDA. In case of untreated or αSyn PFF model 2 μL of AAV was injected, with the αSyn PFF being injected into the striatum. The same dose of 1 × 10⁸ vector genome (VG) was delivered at 0.5 μL/min unilaterally into the left SN (AnteroPosterior (AP) = −2.8 mm, MedioLateral (ML) = 1.3 mm from the bregma, DorsoVentral (DV) = −4.4 mm from the dura surface) [22] using a 28-gauge cannula (Eicom, Kyoto, Japan, SMI-3T) with an infusion pump (KD scientific, Tokyo, Japan), and the cannula was removed 5 min after the end of the administration. This study was not blinded.

Tissue Collection

For hParkin protein expression analysis, mouse brains were dissected three weeks after AAV injection and washed with chilled saline. The brains were sliced using a brain matrix, and the midbrain including SN was separated from other regions of the brain slice.

Quantification of hParkin protein levels and VG in the midbrain of AAV9-PRKN injected Prkn KO mice

The midbrain was homogenized in Radio-Immunoprecipitation assay (RIPA) buffer with protease inhibitor cocktail (Roche, Basel, Switzerland) and phosphatase inhibitor cocktail (Roche) using a bead beating tissue homogenizer (Bertin Technologies, Tokyo, Japan). Lysates were centrifuged at 20,000 rpm for 20 min at 4 °C. The hParkin protein levels in the supernatant were measured by human Parkin ELISA (Abcam, ab212159) after diluting the supernatant with Cell extraction buffer PTR according to manufacturer’s instruction. In supernatants from WT mouse brains, we could not detect hParkin, indicating that the human Parkin ELISA does not cross-react with the mouse Parkin protein (Fig. S3).

DNA was extracted from tissue RIPA homogenate utilizing the KingFisher™ Flex system (MagMAX™ DNA Multi-Sample Ultra 2.0 Kit, adapted for RNase treatment, Thermo Fisher Scientific). Quantitative PCR assays were conducted using QuantStudio7 Flex Real-Time PCR system (Thermo Fisher Scientific). The specific primer/probe set used for detecting the AAV expression cassette specific sequence for VG was designed as follows: forward primer, 5′-CCCCGTGTGAACGATTGGT-3’; reverse primer, 5′-CGTATTTCCCGTTTAGGCTTTCG-3’; and TaqMan probe, 5′-FAM-AACCCGGTGTCCTGTGAG-NFQ and MGB-3’. The PCR protocol was 40 cycles (95 °C for 15 s, 60 °C for 1 min) following initial incubations at 50 °C for 2 min and 95 °C for 10 min, conducted in a 20 μL volume containing 2× TaqPath ProAmp Master Mix (Thermo Fisher Scientific), primers (150 nmol/L), probe (250 nmol/L), and DNase-free and RNase-free water. All reactions were carried out in triplicate for each sample.

Western blot analysis

The supernatant of cell extraction buffer PTR lysate or midbrain RIPA lysate was mixed with LDS sample buffer and sample reducing buffer (Thermo Fisher Scientific) and denatured at 70 °C for 10 min. Equal volume of protein samples was separated by 4–12% NuPAGE Bis-Tris gel and transferred to polyvinylidene fluoride (PVDF) membranes by iBlot (Thermo Fisher Scientific). Blots were blocked with Starting Block blocking buffer (Thermo Fisher Scientific) for 60 min at room temperature. The blots were incubated with anti-Parkin antibody (Thermo Fisher Scientific 702785, 1:200 or BioLegend, San Diego, CA, USA, 865602, 1:200) and anti-β actin antibody (Sigma Aldrich, A5441, 1:10000) overnight at 4 °C. The membranes were washed three times with TBST buffer (Takara Bio, Shiga, Japan) for 10 min, then incubated with secondary antibody (CST, 1:2000). All antibodies were diluted with solutions of Can Get Signal (Toyobo, Osaka, Japan). Protein bands on the membrane were detected using ECL Prime Western Blotting Detection Reagent and an ImageQuant LAS 4000 imager (Cytiva).

6-OHDA injection into medial forebrain bundle (MFB)

A 6-OHDA hydrochloride (Sigma-Aldrich, H4381, lot # MKNL2977) was dissolved with 0.01% (W/V) ascorbic acid in saline to 0.25 μg (calculated as free base of 6-OHDA)/0.8 μL immediately prior to injection. The dose of 6-OHDA was determined in a dose-finding study to mimic the early phase of PD, resulting in approximately 30% loss of DA neurons (Fig. S4). The 6-OHDA solution was kept on ice and protected from light to minimize oxidation. Three weeks after AAV injection into the SN of Prkn KO mice (4-month old), 0.8 μL of 0.25 μg 6-OHDA was delivered at 0.2 μL/min unilaterally into the left MFB (AP = −1.2 mm, ML = 1.2 mm from the bregma, DV = −4.75 mm from dura surface) through a 28-gauge cannula (Eicom, SMI-3T) using an infusion pump (KD scientific) (n = 8/group), and the cannula was removed 5 min after end of the injection. 0.01% (W/V) ascorbic acid in saline was administered for the control group (n = 6).

α-Syn PFFs injection

Mouse α-Syn PFFs (2 mg/mL in PBS, lot. 2018.11.21) were provided by Dr. Hasegawa, Tokyo Metropolitan Institute of Medical Science[28, 43, 44]. α-Syn PFFs were sonicated with an ultrasonic homogenizer (Misonex, XL2020) for 5 s at 10-s intervals four times prior to the injection. Three weeks following AAV injection into the SN of Prkn KO mice (6-months old), 5 μL of 10 μg α-Syn PFFs were unilaterally injected into the left striatum at 0.25 μL/min (AP = 0.2 mm, ML = 2.0 mm from the bregma, DV = −2.6 mm from dura surface) via a 28-gauge cannula (Eicom, SMI-3T) with an infusion pump (KD scientific) (n = 7–8/group). The cannula was removed 10 min after end of the injection. PBS (Thermo Fisher Scientific, 10010023) was administered in the control group (n = 7).

Immunohistochemistry (IHC)

Parkin immunohistochemistry

Antigen retrieval on brain sections was performed at 90 °C for 15 min with pH 6.0 citrate buffer (Agilent Dako, S2369, Santa Clara, CA, USA). Treated sections were incubated with mouse Ig blocking reagent (Vector Laboratories, MKB-2213, Newark, CA, USA) and then blocked with 2.5% normal horse serum. After blocking non-specific signal, anti-Parkin (Cell Signaling Technology, Danvers, MA, USA, Prk8, 4211S, 1:200) and anti-tyrosine hydroxylase (Abcam, ab137869, 1:500) antibodies were applied to slides and incubated overnight at 4 °C. Subsequently, slides were incubated with secondary antibodies (Thermo Fisher Scientific, A11001, A11037, 1:500) in combination with Hoechst 33342 (Abcam, 1:2000). Immunostained slides were coverslipped and scanned using a NanoZoomer to obtain the images and representative images were generated by HALO.

Statistical analysis

The data are presented as the mean ± standard error of the mean (SEM). In in vitro studies, which were done in triplicate, statistical analyses were performed using GraphPad Prism version 9.2.0 or 10.4.1 (Dotmatics, GraphPad Software, Woburn, MA, USA). In PRKN-null DA neurons, comparisons between the groups were performed using Dunnett’s post-hoc test after one-way ANOVA test. The effect of the AAV treatment in the presence of CCCP was analyzed using Dunnett’s test compared with a control group (no AAV and no CCCP). In WT iPSC-DA neurons, the effect of CCCP was analyzed by T-test compared with a control group in WT neurons (no AAV and no CCCP). The effect of the co PRKN on Parkin protein expression in vitro was analyzed using Dunnett’s test compared with a control group (WT PRKN sequence) at each MOI. P values less than 0.05 were considered statistically significant.

In in vivo studies, statistical analyses were performed using Exsus 10.1.6 (EP Croit). The sample size for in vivo studies was n = 6–8, except for AAV9 transduction study with coPRKN and 6-OHDA dose-finding study with a sample size of n = 4. For phenotype analysis in 125-week-old Prkn KO mice, comparisons between WT (n = 15) and Prkn KO (n = 9) mice were performed using Student’s T-test. P-values less than 0.05 were considered significant.

For in vivo VG analysis, the data are presented as the mean ± standard deviation (SD).

In 6-OHDA study, comparisons between the groups were performed using Dunnett’s test after one-way ANOVA test. The effect of the vector AAV9-PK041 was analyzed using Dunnett’s test compared with vehicle control in 6-OHDA lesioned mice. P-values less than 0.05 were considered statistically significant. In the α-Syn PFFs study, the effect of α-Syn PFFs was analyzed by Student’s T-test. P-values below 0.05 were considered significant. The effect of the AAV9-PK041 was analyzed using Williams’ multiple comparison test compared with vehicle control in α-Syn PFF-injected mice. P values less than 0.025 were considered statistically significant.

Parkin expression, polyubiquitin formation, and mitophagy induction after AAV-PRKN transduction of PRKN-null iPSC-DA neurons

We have compared hParkin protein expression in PRKN-null iPSC-DA neurons using AAV9-PRKN vector driven by different promoters (four ubiquitous, three neuronal, and one endogenous PRKN promoter) at MOI of 3 × 10³, 1 × 10⁴, 3 × 10⁴, and 1 × 10⁵ (Table 1 and Table S3). In control (buffer-treated) PRKN-null iPSC-DA neurons, hParkin protein levels were below the limit of quantification and in WT iPSC-DA neurons at 526.1 ± 6.3 pg/mg total protein (Fig. S5). In all AAV-PRKN groups, we observed a dose-dependent, exponential increase in hParkin expression levels in PRKN-null iPSC-DA neurons, with no substantial effect on cell viability or total cell number for most MOIs tested (Figs. 1, 2A). The expression cassette PK031 (chicken β-actin hybrid, CBh promoter) showed the strongest expression even at the lowest MOI, always above endogenous WT levels (10-fold above WT at the lowest MOI) and reached the plateau at 3 × 10⁴ and 1 × 10⁵. At this highest MOI, there was a clear reduction in cell number (Fig. 2A). In contrast, in the group treated with the expression cassette PK026, using an endogenous PRKN promoter, only very low levels of hParkin protein below the endogenous WT levels were observed, at the 2 highest MOIs (3 × 10⁴ and 1 × 10⁵) with no detectable expression at the two lowest MOIs (3 × 10³ and 1 × 10⁴). AAV9 vectors driven by ubiquitous or neuron-specific but not the PRKN endogenous promoter (PK026) achieved hParkin protein expression levels in PRKN-null neurons which exceeded endogenous hParkin levels of WT iPSC-DA neurons using MOIs above 3 × 10⁴. These results show that seven AAV9 vectors can fully restore hParkin expression in PRKN-null iPSC-DA neurons at the MOI 3 × 10⁴ to the equivalent or higher expression level of WT iPSC-DA neurons. The promoter rank order for hParkin protein expression levels is the following: CBh > PGK1 = TH > EFS > EF1α = SYN1 > mCaMKIIα » PRKN promoter.

To validate the functionality of hParkin protein expressed from the AAV-PRKN vectors, we assessed Parkin-mediated poly-pUb by detecting pUb and evaluated the clearance of damaged mitochondria by measuring levels of the mitochondrial outer membrane protein TOM20 in PRKN-null iPSC-DA neurons following CCCP treatment. As expected, CCCP treatment did not induce pUb chain formation in untreated PRKN-null iPSC-DA neurons, consistent with the requirement of Parkin for pUb accumulation [10, 11] (Fig. 2B, S6). All eight AAV9 vectors increased poly-pUb levels in CCCP-treated PRKN-null iPSC-DA neurons in a dose-dependent manner (Fig. 2B). The promoter rank order measured by their efficiency on poly-pUb levels is the following: CBh > PGK1 > TH = EFS > EF1α > SYN1 > mCaMKIIα » PRKN. For PK031 (CBh promoter) which showed the highest hParkin protein expression in Fig. 1, poly-pUb levels were above that in CCCP-treated WT neurons at all MOIs and reached a plateau (2-fold of poly-pUb levels in CCCP-treated WT neurons) at the lowest dose. These findings suggest a saturation point for Parkin-mediated poly-pUb accumulation, beyond which further increases in hParkin expression of 10-fold or more do not result in additional poly-pUb induction. PK012 (PGK1), PK025 (TH), and PK030 (EFS) yielded 1- to 2-fold hParkin protein expression over endogenous levels at the MOI 3 × 10³ (Fig. 1), and poly-pUb levels above of CCCP-treated WT levels at the same MOI, indicating that 1 to 2-fold hParkin protein expression over endogenous levels is sufficient to induce poly-pUb accumulation at levels of WT cells. In the case of PK026 (PRKN promoter), only a marginal increase in poly-pUb was detectable at the highest MOI, and still about 50% below the WT level, consistent with the very low hParkin protein expression (Fig. 1). As a second parameter for Parkin function, we induced mitophagy using CCCP in iPSC-DA neurons from WT and PRKN-null iPSCs and assessed mitophagy activation by immunostaining for TOM20. However, we observed considerable variability in the results of the TOM20 assay, particularly across different experimental runs. In summary, PRKN-null iPSC-DA neurons have a higher TOM20 signal, and do not respond to CCCP treatment except in the case of the PK031 (Fig. 2C). High AAV9 doses for all vectors tested could restore TOM20 levels to WT levels, indicating functional mitophagy. These data support that the Parkin delivered by AAV is functional for all AAV vectors and mitigates mitophagy deficits in PRKN-null iPSC-DA neurons (Fig. 2C). Integrated in vitro analyses of hParkin expression, poly-pUb formation, TOM20 reduction (mitophagy), and cell viability collectively, nominate several promoter constructs for further exploration: the ubiquitous promoters PGK1 and EFS, along with the neuron-specific promoters SYN1 and TH.

Parkin expression by AAV9-codon-optimized (co) PRKN in PRKN-null iPSC-DA neurons and Prkn knockout mice

Codon optimization, especially the removal of immunostimulatory sequences, such as CpG motifs, confers a clear advantage in the context of gene therapy by reducing innate immune activation [45]. Codon optimization of the PRKN cDNA sequence was performed to minimize CpG dinucleotide content while preserving adequate expression of Parkin protein. The WT PRKN cDNA contains 46 CpGs and we were able to reduce the CpGs to five (CpG5-2, CpG5-4) or zero (CpG0-3) applying a codon optimization algorithm optimized for neuronal expression (Table 2 and Table S4). Three codon-optimized sequences were cloned into a cassette driven by the neuron-specific human SYN1 promoter and packaged into AAV9 for in vitro and in vivo analyses. The SYN1 promoter was selected as it achieves moderate hParkin protein expression (Fig. 1) and allows to test the effect of Parkin expression specifically in neurons.

PRKN-null human iPSC-DA neurons were transduced with AAV9-CpG5-2 (PK041), CpG5-4 (PK042), or CpG0-3 (PK043) at MOIs of 3 × 10³, 1 × 10⁴, 3 × 10⁴, and 1 × 10⁵. All three vectors induced MOI-dependent hParkin protein expression, with expression above WT vectors (not codon optimized) at all MOIs (Fig. 3A). Significantly higher hParkin protein expression was observed in all coPRKN vector groups compared to WT vector group in human iPSC-DA neurons at the highest MOI of 1 × 10⁵ (1.6 folds for CpG5-2 and CpG5-4, and 2.4 folds for CpG0-3).

Next, we evaluated the impact of codon-optimized PRKN (coPRKN) on hParkin protein expression in vivo by administrating 1 × 10⁸ VG of each AAV9 vector containing coPRKN (PK041, PK042, or PK043) into the SN of Prkn knockout (KO) mice. AAV9 transduction efficiency was assessed in mice by quantifying by VG copies in the midbrain three weeks post AAV administration. In the PK041 (CpG5-2) treated group, VGs were significantly higher (2.6-fold) than that of WT sequence (PK013), whereas VGs in PK042 (CpG5-4) and PK043 (CpG0-3) treated groups were only slightly higher (Fig. 3B). All three AAV vectors achieved hParkin protein expression in the midbrain of Prkn KO mice with PK042 (CpG5-4) and PK043 (CpG0-3) reaching similar levels to WT, and only PK041 (CpG5-2) showing increased expression (Fig. 3C). Parkin protein expression in the midbrain and striatum after intra-SN injection of each AAV vector was also confirmed by Western blot in an independent study by administrating 1 × 10⁷ VG of each AAV9 vector (Fig. S7). We have selected PK041 (CpG5-2) with the highest hParkin expression (1.7 folds over WT) and highest transduction efficiency (VG) for further evaluation in mouse PD models.

Neuroprotection of DA neurons by AAV9-PK041 (SYN1-coPRKN CpG5-2) in PD mouse models

We have evaluated AAV9-PK041 vector encoding coPRKN CpG5-2 driven by the SYN1 promoter in two PD mouse models, the 6-OHDA lesion model and α-Syn PFFs-lesion model on the Prkn KO background. In the 6-OHDA PD mouse model, oxidative stress-mediated apoptotic cell death is a key driver of subsequent DA neuronal degeneration, consistent with observations in PRKN-PD patients [46, 47]. Additionally, 6-OHDA induces mitophagy deficits in neurons in vivo [48]. In our study, 6-OHDA was injected three weeks after intra-SN AAV injection and brains were analyzed two weeks after 6-OHDA lesioning. In the ipsilateral (Ipsi) side, we could detect a significant (p = 0.003) protection of tyrosine hydroxylase (TH)-positive neurons, the DA neurons in the SNpc, after AAV9-PK041 injection compared to vehicle control with 6-OHDA stimuli (Table 3, Fig. 4A, B). In the contralateral (Contra) side, the TH-positive neurons were not protected by AAV9-PK041, are similar to the vehicle control and serve as intra-animal control, indicating that the AAV does not spread to the contralateral side. The ratio of TH-positive neurons on the ipsilateral side to the contralateral side is significantly (63%, p = 0.023) increased in the AAV9-PK041-treated group compared to the 6-OHDA-lesioned vehicle group. In the AAV9-PK041 treatment group, hParkin is restored in the SN of ipsilateral side and colocalizes with TH, with almost all TH-positive cells in the SN at the injection site expressing hParkin (Fig. 4C, D). These results indicate that expression of Parkin exerts a neuroprotective effect, and that targeted delivery via AAV vectors under the control of the neuron-specific SYN1 promoter is an effective strategy for achieving cell-type restricted expression in vivo.

As a second model, we used the α-Syn PFF injection mouse model, which is relevant for studying the progressive pathology of sporadic PD as the injected PFFs seed the misfolding and aggregation of endogenous α-Syn, resulting in the formation of Lewy body-like inclusions that progressively spread throughout anatomically connected brain regions. This model recapitulates key pathological features of human disease, including phosphorylated α-Syn (pS129) accumulation, neuroinflammation, and dopaminergic neuronal loss. We evaluated the effect of Parkin protein on neuroprotection in the context of α-Syn propagation [49]. About 35% DA neuronal loss in the SNpc was detectable one month after α-Syn PFFs injection (Table 4, Fig. 5A, B), consistent with previously published data [50]. Pretreatment with AAV9-PK041 prior to α-Syn PFFs injection resulted in a dose-dependent preservation of TH-positive neuronal cell bodies in the SNpc, with the highest dose (1 × 10⁸ VG) restoring TH-positive cell counts to levels comparable to those in the non–PFF injected control group (Fig. 5A, B, p = 0.038 in 0.3 × 10⁸ VG group; p = 0.006 in 1 × 10⁸ VG group). The findings from both, the 6-OHDA and α-Syn PFFs-lesion models illustrate the neuroprotective effects of hParkin protein. As α-Syn-induced DA neuron loss is a hallmark of idiopathic PD [49], Parkin supplementation with its neuroprotective activity shows potential for the treatment of idiopathic PD in addition to PRKN-PD.