AAV gene replacement restores Parkin pathway activity in preclinical tests

An adeno-associated virus carrying functional human PRKN restored a key molecular signal associated with mitochondrial cleanup in Parkin-deficient cells and restored the same signal in fibroblasts from patients with PRKN-associated early-onset Parkinson’s disease. In wild-type rodents, direct injection of the therapy into the substantia nigra produced dose-dependent Parkin expression, including in dopaminergic neurons, but the highest tested dose was associated with glial activation and changes in striatal dopamine. The study therefore supports further preclinical development while showing that careful control of vector dose and promoter activity will be essential.

The work, reported in Frontiers in Molecular Neuroscience on 21 October 2026, focuses on PRKN-related early-onset Parkinson’s disease, a monogenic form in which damaging variants affect both copies of the gene encoding Parkin, an E3 ubiquitin ligase involved in mitochondrial quality control. Under normal conditions, the PINK1–Parkin pathway helps detect damaged mitochondria and initiate their selective removal through mitophagy. The researchers used phosphorylated ubiquitin at serine 65, or pUbSer65, as a marker of pathway initiation because PINK1 produces this signal and Parkin normally amplifies it. The experiments addressed whether delivering healthy PRKN with an AAV vector could reinstate this early signaling response.

First, the team compared normal and PRKN-knockout SH-SY5Y neuroblastoma cells. Under basal conditions, the two cell types showed no differences in the mitophagy and ubiquitin-proteasome endpoints examined. The researchers then applied the mitochondrial uncoupler FCCP to collapse membrane potential and impose a stress that can initiate both Parkin-dependent and Parkin-independent mitophagy. In normal cells, pUbSer65 became detectable after two hours and peaked at four hours, with statistically significant differences from vehicle-treated controls at two, six and eight hours. Parkin-knockout cells showed no significant response to FCCP at any tested time, indicating that PINK1 can initiate the pathway but functional Parkin is needed for robust signal amplification.

Transduction with AAV1-Ef1a-Parkin produced a multiplicity-of-infection-dependent increase in Parkin protein in the knockout cells. After mitochondrial stress, the treated cells regained a significant pUbSer65 response, while vehicle-treated cells did not. Immunofluorescence showed that the restored signal co-localized with mitochondria. Parkin protein levels also fell significantly after FCCP exposure relative to vehicle, a pattern consistent with the delivered enzyme binding dysfunctional mitochondria and being removed with mitochondrial cargo. The authors were careful about the limits of this observation: the decrease in Parkin is an indirect indication of cargo processing and does not directly demonstrate completed mitophagic flux. Dynamic flux reporters and measurements of mitochondrial turnover would be needed to establish the full kinetics of mitophagy.

The researchers next examined primary fibroblasts carrying diverse biallelic PRKN mutations, a setting intended to test whether gene replacement could overcome different mutant backgrounds. They compared eight or nine patient-derived lines with four healthy control lines and used valinomycin to impose mitochondrial stress. Basal pUbSer65 levels did not differ significantly between patients and controls, but stressed patient fibroblasts produced substantially less of the signal, with a difference of p less than 0.0001. The response also failed to rise significantly after valinomycin in the patient cultures, for which the reported p value was 0.0665. This residual signal is consistent with the established sequence in which PINK1 generates phosphorylated ubiquitin and Parkin subsequently amplifies it.

After screening capsids, the investigators selected AAV2 for efficient delivery to the primary fibroblasts. Transduction with AAV2-CAG-Parkin restored valinomycin-induced pUbSer65 amplification in every patient line examined, regardless of the underlying PRKN mutation. Post-transduction signal in the patient cells was comparable to the upregulation seen in healthy controls, with a reported significance of p less than 0.0001. The result provides a cell-based proof of principle for functional gene replacement across a range of disease-associated variants. It does not, however, establish downstream metabolic recovery: the study did not test whether treatment restored ATP production, mitochondrial respiration, or clearance of reactive oxygen species.

To assess delivery in animals, the team stereotaxically injected AAV1-Ef1a-Parkin into the substantia nigra of two-month-old rats at doses of 3 × 10^8, 3 × 10^9, or 3 × 10^10 vector genomes per substantia nigra. Human Parkin expression increased with dose and co-localized with the tyrosine hydroxylase marker used to identify dopaminergic neurons. At the lowest dose, measured Parkin ranged from 35% to 260% of endogenous rat Parkin, with a mean of 135.25% in four animals. This placed the average above the study’s desired 50%–100% replacement range, but the wide variation showed that stereotaxic delivery into a small, deep brain region does not produce uniform physiological replacement. The intermediate and highest doses produced expression above physiological levels.

Body weights remained comparable among groups in the rat study, and no significant change in striatal dopamine or dopamine turnover was detected at the two lower doses. In contrast, the 3 × 10^10 vector-genome dose was associated with increased Iba1, a microglial marker, and GFAP, an astrocytic marker. It also reduced striatal dopamine and impaired dopamine turnover. The highest-dose cohort did not include a dose-matched empty-vector control, so the investigators could not conclusively distinguish effects of viral load from those of excessive Parkin expression. Within that limitation, the results indicate that the 3 × 10^10 vector-genome dose was not tolerated as well over the study’s short observation period, while the lower doses appeared well tolerated. The study therefore supports a preliminary dose-finding range for further development rather than establishing a therapeutic dose.

A complementary experiment in wild-type mice compared promoters in the proprietary Spark100 capsid after a single intraparenchymal dose of 1 × 10^10 vector genomes per substantia nigra. Eight weeks later, the ubiquitous CAG promoter produced the greatest Parkin expression, followed by Ef1a, the full-length tyrosine hydroxylase promoter and human Synapsin1, while the shortened TH promoter produced little detectable expression. Despite the high expression driven by CAG, comparison with control animals did not show a significant reduction in substantia nigra TH levels, although striatal TH was lower than in the Synapsin1 group. The full-length TH construct also produced a visual trend toward reduced TH immunoreactivity, although no inflammatory marker increased with it. Ef1a and Synapsin1 were identified as the most favorable candidates for reaching relevant neuronal populations without detectable glial activation.

The mouse comparison also clarified the promoter-related dose finding. Other Spark100 constructs delivered at the same vector dose did not significantly increase Iba1 or GFAP, while the high-expression CAG construct did, supporting the interpretation that excessive Parkin expression rather than capsid dose alone was associated with gliosis in the promoter study. Together, the experiments establish foundational feasibility, not clinical efficacy. The researchers did not test a PRKN-deficient animal that reproduced the overt nigrostriatal degeneration and motor abnormalities of patients, and the available knockout models do not fully capture the range of human PRKN mutations. Long-term safety, metabolic benefit, neurological protection, and efficacy in more representative disease models remain open questions requiring larger and longer studies before a clinical approach could be evaluated.

Subject of Research: Preclinical AAV gene replacement for PRKN-associated early-onset Parkinson’s disease

Article Title: Restoring Parkin function: an AAV gene therapy approach for early-onset Parkinson’s disease

Article References: Basu, S., Demarest, T. G., Gattone, N. J., Gilsrud, A. J., Wicks, B., Khatiwada, A., Nayal, M., Gentzel, R., Cohen, D. M., Kostuk, E. W., Narendra, D. P., Gonzalez-Alegre, P., Biferi, M.-G., & Ramsburg, E. A. (2026). Restoring Parkin function: an AAV gene therapy approach for early-onset Parkinson’s disease. Frontiers in Molecular Neuroscience, 19, Article 1939065. https://doi.org/10.3389/fnmol.2026.1939065

Image Credits: AI Generated

DOI: 10.3389/fnmol.2026.1939065

Keywords: AAV gene therapy, Parkin, PRKN mutations, early-onset Parkinson’s disease, mitophagy, substantia nigra, dopaminergic neurons, mitochondrial stress, promoter optimization, Restoring, function, gene