Gene-Editing Tools Delivered by Extracellular Vesicles Suppress ATXN3 in Spinocerebellar Ataxia Models
Background
Spinocerebellar ataxia type 3 (SCA3) is an autosomal dominant neurodegenerative disease in which an expanded CAG repeat in the ATXN3 gene leads to the production of toxic ataxin-3 protein. The research team packaged gene-editing tools in extracellular vesicles (EVs) and delivered them to patient-derived cells and two mouse models, observing reduced ATXN3 expression and editing at the target site.
To inactivate a disease-causing gene with CRISPR-Cas9, the editing tools must enter the target cells. The possibility that the editing enzyme may persist for a long time also needs to be considered. This study therefore focused on transient delivery of a ribonucleoprotein (RNP) complex consisting of Cas9 protein and single-guide RNA (sgRNA).
Key findings
The team attached a palmitoylation signal that promotes fatty-acid attachment and the photocleavable linker protein PhoCl to SpCas9 to package it into EVs. The design allows Cas9 to be detached from the packaging signal by exposing isolated EVs to 410 nm light before delivery. The light is applied to the isolated EVs, and Cas9 separates from the packaging signal inside the EVs.
The editing target was exon 2 of ATXN3. Rather than restoring the CAG repeat to its normal length, this approach inactivates gene function and does not distinguish between the normal and mutant alleles. To increase cellular uptake, the team used EVs displaying the viral envelope glycoprotein VSV-G on their surface. At 72 hours after treatment of patient-derived induced pluripotent stem cells (iPSCs), mutant ATXN3 protein levels were reduced by a mean of 28.58% and normal ATXN3 by 21.21% compared with the phosphate-buffered saline (PBS) control.
Results were also obtained in two mouse models, but the outcome measures differed. In a striatal model expressing mutant ATXN3 and a luminescent reporter, EVs were injected directly into the brain and normalized luminescence was measured two weeks later. The value was a mean of 34.42% lower than in the control hemisphere injected with EVs carrying Cas9 without sgRNA. In the YAC SCA3 84.2 model, which carries the human ATXN3 gene, EVs and a lentiviral vector expressing green fluorescent protein (GFP) were co-injected into the cerebellum. Four weeks later, up to 15.6% editing at the target site was observed in selected GFP-positive cells. This figure is not the editing rate across the entire cerebellum.
Implications and outlook
These results provide preclinical proof of concept that EVs can transiently deliver editing tools and partially inactivate ATXN3. They cannot be interpreted as demonstrating improved motor symptoms or therapeutic benefit in patients. Normal ATXN3 protein also decreased, and the limited distribution of EVs after direct brain injection remains a challenge.
Transient delivery does not establish safety. In an initial off-target assessment in which SpCas9 and sgRNA plasmids were introduced into cells, one of four predicted sites showed an insertion/deletion level approximately 2.4% above the control background. The researchers explicitly stated that a more comprehensive assessment is needed. This evaluation should be distinguished from a comprehensive safety study in animals treated with EVs.
The study design and results are described in the original paper published in Biomaterials.
Spinocerebellar Ataxia Type 3 (SCA3) is a neurodegenerative disorder caused by an overexpansion of a CAG tract within the ATXN3 gene, leading to toxic properties of the ataxin-3 protein. Genome editing using CRISPR-Cas9 enzymes is a promising strategy to inactivate mutant ATXN3 alleles; however, in vivo delivery remains a challenge. Extracellular vesicles (EVs) are promising vehicles for delivering Cas9 and single guide RNA (sgRNA) ribonucleoproteins, minimizing genomic exposure to highly active endonucleases. In this study, researchers designed SpCas9 with a palmitoylation motif, enabling SpCas9 and sgRNA enrichment into EVs. The introduction of a photocleavable linker, PhoCl, allowed for the photo-inducible release of SpCas9 from the palmitoylation motif in EVs, increasing target engagement to ATXN3 in vitro. EVs loaded with SpCas9 ribonucleoproteins resulted in ATXN3 knockout in SCA3 patient-derived iPSCs and two SCA3 animal models. These findings highlight an innovative route for the transient delivery of gene editing tools, providing a promising therapeutic platform for the treatment of genetic diseases, including SCA3, as reported in Nature Medicine.
In gene editing for brain diseases, a key challenge is not only what to cut, but also where and for how long to deliver the editing tools. This study presents a production and delivery approach in which a Cas9 protein–RNA complex is loaded into EVs and separated from its packaging signal using light before administration. Follow-up studies could use this platform to compare distribution within the brain and routes of administration, and to assess the effects of reducing normal ATXN3 and off-target editing. Even if the editing tools are present only temporarily, the resulting DNA edits may persist. Assessing its value as a therapeutic strategy will require determining whether less invasive administration can reach enough target cells and whether molecular changes lead to improved symptoms.