Why this study matters
Leber congenital amaurosis type 2 (LCA2) is an inherited retinal disease caused by changes in RPE65, a gene that is essential for the visual cycle—the process that allows the eye to respond to light. RPE65 is active in the retinal pigment epithelium (RPE), a thin layer of supportive cells at the outer edge of the retina. When RPE cells cannot make enough functional RPE65 protein, vision can be severely affected early in life.
Gene therapy for LCA2 aims to deliver a working copy of RPE65 to RPE cells. Reaching those cells efficiently, however, presents a major technical challenge. The established approach for directly targeting the RPE is subretinal injection, in which treatment is placed beneath the retina. This can provide strong RPE exposure, but the procedure is invasive and intentionally creates a temporary retinal detachment. The new study explored whether a less invasive approach—injecting a gene therapy vector into the vitreous gel in the center of the eye—could reach the RPE successfully.
Searching for a vector that can reach the RPE
The research team focused on adeno-associated virus (AAV), a widely used delivery vehicle for retinal gene therapy. Specifically, they created and screened an extraordinarily large library of 1.28 billion AAV2 capsids. A capsid is the outer protein shell of a virus; changing its surface can alter which cells it can enter.
Each AAV2 capsid in the library contained a different short, seven–amino acid peptide insertion. The goal was to identify a version of AAV2 that could travel from the vitreous to the outer retina and enter RPE cells after an intravitreal injection.
This screening identified a candidate called AAV2.Var3, or variant 3. In mouse eyes, the researchers found that AAV2.Var3 could deliver a green fluorescent protein reporter gene to RPE cells. They tested both a broadly active genetic control element, CAG, and a control element from the Best1 gene, which is associated with RPE cells. These experiments showed that the new capsid could support RPE expression after intravitreal delivery.
How did the new capsid compare with an existing vector?
The investigators compared AAV2.Var3 with 7m8, another AAV2-based capsid that has reached the clinical-stage development pathway. The abstract notes that 7m8 effectively transduces the outer retina but has had limited success targeting RPE cells through intravitreal injection.
This distinction is important. Retinal diseases affect different cell types, and a vector that reaches photoreceptors or other outer-retinal cells may not adequately treat a condition in which the key therapeutic target is the RPE. AAV2.Var3 was developed to address this specific delivery gap: reaching RPE cells from inside the eye without placing the injection directly under the retina.
Testing an RPE65 treatment in an LCA2 model
To evaluate whether AAV2.Var3 could do more than deliver a reporter gene, the researchers packaged it with human RPE65 DNA. They then injected the resulting vector, AAV2.Var3-hRPE65, into the vitreous of rd12 mice. These mice are a model of LCA2 because they have deficient RPE65 function.
Treated mouse groups showed increased RPE65 protein expression and improved RPE65-related function. They also demonstrated improved visual responses to light flashes. Together, these findings indicate that intravitreal delivery using AAV2.Var3 was able to produce biologically meaningful effects in this disease model.
The results were encouraging but not uniform: the authors note that therapeutic gene delivery produced functional results with efficiency limitations. In other words, the approach showed proof of therapeutic activity, while also underscoring that further work is needed to improve the consistency and extent of delivery.
Early evidence in non-human primates
The study also examined whether AAV2.Var3 could reach RPE cells in macaques, an important step between mouse research and potential human studies. After intravitreal administration of AAV2.Var3 carrying Best1-GFP, the researchers observed green fluorescent protein expression in peripheral RPE cells.
This result supports the translatability of the capsid for intravitreal RPE delivery, while also highlighting an important question for future studies: how broadly and efficiently can the vector reach RPE cells across the retina, including the central regions most relevant to vision?
What this could mean for future LCA2 treatment
This publication advances a promising idea in retinal gene therapy: tailoring viral capsids to reach the right retinal cell type through a less invasive route. For LCA2, an intravitreal RPE-targeting vector could eventually complement existing subretinal strategies and may reduce some procedural limitations associated with subretinal delivery.
AAV2.Var3 remains a preclinical research vector, but its ability to restore RPE65-associated function in an LCA2 mouse model and target peripheral RPE in macaques makes it a notable addition to the retinal gene therapy landscape. Future research will need to refine delivery efficiency, characterize retinal coverage, and build on these findings as researchers seek safer, more precise ways to bring genetic treatments to the cells that need them most.

