Subretinal AAV vector delivery restoring functional gene expression in photoreceptor & RPE cells — the Luxturna (voretigene neparvovec) model
Inherited retinal diseases (IRDs) such as Leber congenital amaurosis (LCA) and certain forms of retinitis pigmentosa are caused by mutations in genes essential for photoreceptor and retinal pigment epithelium (RPE) function. When the disease-causing defect maps to a single gene, replacing that gene’s function becomes a tractable therapeutic strategy — rather than treating downstream symptoms, the root molecular cause can be addressed directly.
Most inherited retinal diseases are monogenic — a single defective gene disrupts one specific molecular step, while the rest of the cellular machinery in the photoreceptor or RPE cell remains intact. This is fundamentally different from complex multifactorial disease: if the missing or defective protein can be supplied, the cell’s existing machinery can, in principle, resume normal function.
RPE65 is a clear example: it encodes the isomerase enzyme that converts all-trans-retinyl ester to 11-cis-retinol inside the RPE, a step required to regenerate the visual pigment used by rod and cone photoreceptors (the "visual cycle"). Biallelic loss-of-function mutations in RPE65 block this cycle, starving photoreceptors of usable chromophore and causing severe, early-onset vision loss even though the photoreceptor cells themselves may initially remain anatomically present.
Several anatomical and immunological properties make the eye — and particularly the subretinal space — an unusually favorable site for gene replacement compared to most other organs:
• Immune privilege: the blood-retinal barrier and local immunosuppressive factors limit inflammatory responses to a surgically introduced vector • Compact, enclosed compartment: a small injected volume achieves high local vector concentration without systemic exposure • Direct visualization: the target tissue can be directly visualized and accessed under an operating microscope • Contralateral control: the fellow untreated eye can serve as an internal comparison in clinical evaluation • Post-mitotic target cells: photoreceptors and RPE do not divide, so an episomal (non-integrating) vector genome is not diluted out over successive cell divisions
Because the defect is confined to one well-characterized gene and the target tissue is immunologically privileged and surgically accessible, the retina became one of the first tissues where single-gene replacement therapy achieved regulatory approval in humans.
Adeno-associated virus (AAV) is engineered into a delivery vehicle by removing its own replication genes and packaging a therapeutic expression cassette in their place. The resulting recombinant AAV (rAAV) cannot replicate or cause disease on its own — it functions purely as a shuttle that efficiently transduces non-dividing retinal cells and delivers a functional copy of the gene.
Wild-type AAV is a small, non-enveloped parvovirus with a ~4.7 kb single-stranded DNA genome flanked by inverted terminal repeats (ITRs). To convert it into a therapeutic vector:
• The viral rep (replication) and cap (capsid) genes are removed from the genome and supplied separately, in trans, only during vector manufacturing • The space between the ITRs is replaced with an expression cassette: a promoter, the therapeutic gene’s coding sequence (cDNA), and a polyadenylation signal • The ITRs themselves are retained because they are required in cis for genome packaging into the capsid • Because rep and cap are absent from the final product, the resulting recombinant vector cannot replicate or produce new infectious particles inside the patient
The capsid serotype (e.g. AAV2) determines which cell types the vector preferentially binds and enters — its tropism. AAV2 shows efficient, well-characterized tropism for RPE cells and photoreceptors, which is why it was selected as the delivery vehicle for RPE65 gene replacement.
Inside the capsid, the packaged genome carries a functional cDNA copy of the gene that is defective in the patient — for RPE65-mediated disease, a corrected RPE65 coding sequence under a promoter chosen to drive expression preferentially in RPE cells.
Because AAV’s natural packaging capacity is limited (roughly the size of its own 4.7 kb genome), the therapeutic cassette must fit within that limit — cDNA sequences are used rather than the full genomic gene (with introns) to keep the payload compact while still encoding a complete, functional protein.
The vector particle itself is inert cargo — a protein shell with no replication genes. Its only "activity" is receptor binding, cell entry, and delivering the DNA payload; it cannot spread from cell to cell or produce progeny virus once inside the retina.
Getting the vector to its target requires a precise surgical procedure. Rather than an intravitreal injection into the fluid-filled cavity of the eye, subretinal delivery places the vector suspension directly beneath the neurosensory retina, in immediate proximity to the photoreceptor outer segments and the RPE — the exact cells that need to take up the gene.
The standard subretinal delivery procedure proceeds through several defined steps:
1. Pars plana vitrectomy: the vitreous gel is removed through small surgical ports to give the surgeon clear access and prevent the gel from interfering with the injection 2. Retinotomy: a microscopic opening is made in the retina at a peripheral location, away from the central macula 3. Cannula placement: a fine-gauge subretinal cannula is passed through the retinotomy and advanced just beneath the neurosensory retina 4. Bleb formation: the vector suspension is slowly infused, mechanically separating the neurosensory retina from the underlying RPE and creating a localized fluid "bleb" that bathes both photoreceptor outer segments and RPE cells in vector 5. Bleb resolution: over the following hours to days, the subretinal fluid is naturally resorbed by the RPE, allowing the retina to reappose while the delivered vector remains associated with the target cells
The therapeutic effect depends on the vector physically reaching the correct cell layer. Photoreceptor outer segments and the apical RPE surface face each other across a very thin extracellular space; an intravitreal injection would leave the vector on the wrong side of the retina, with the full thickness of neurosensory retina acting as a diffusion barrier and diluting the effective dose.
By contrast, subretinal injection delivers a small volume at very high local concentration directly to the interface between the two target cell populations, maximizing the fraction of cells exposed to a transducing dose of vector while minimizing the total amount of vector required — and therefore minimizing the risk of an inflammatory or immune response to the capsid.
The bleb itself is transient: as subretinal fluid resorbs over hours to days, the retina settles back into apposition with the RPE, but the vector genome delivered during that brief window has already begun entering target cells.
Once positioned in the subretinal bleb, individual AAV particles must bind, enter, and successfully deliver their genome into the nucleus of target photoreceptor and RPE cells. Only after the cell’s own transcription and translation machinery reads out the delivered cDNA does the therapeutic protein actually appear — restoring the biochemical function that was missing.
Successful transduction of a target cell involves a defined molecular sequence:
1. Cell-surface binding: the AAV capsid engages receptor and co-receptor molecules on the apical membrane of the RPE or photoreceptor outer segment 2. Endocytosis: the bound particle is internalized inside an endosomal vesicle 3. Endosomal escape and nuclear trafficking: the capsid escapes the vesicle, traffics through the cytoplasm, and enters the nucleus through a nuclear pore 4. Uncoating: the capsid releases its single-stranded DNA genome inside the nucleus 5. Second-strand synthesis: host-cell machinery converts the single strand into double-stranded DNA — a rate-limiting step for the onset of expression 6. Episome formation: the double-stranded genome, still flanked by its ITRs, forms a stable circular episome that persists in the nucleus without integrating into the host chromosome 7. Transcription and translation: the promoter drives transcription of the therapeutic cDNA; the resulting mRNA is translated into the functional protein the patient’s own gene could not produce
For RPE65 gene replacement specifically, once functional RPE65 enzyme is expressed inside transduced RPE cells, the visual cycle can resume: all-trans-retinyl ester is again isomerized to 11-cis-retinol, replenishing the chromophore supply that photoreceptors need to regenerate light-sensitive visual pigment.
The proportion of cells that are successfully transduced — the transduction efficiency — is a key determinant of how much functional protein is produced across the treated retinal area, and therefore how much of the underlying biochemical deficit is corrected.
Gene expression does not appear instantaneously — second-strand synthesis and episome formation mean functional protein typically becomes measurable over days to a few weeks, after which expression is expected to plateau at a level reflecting the fraction of cells successfully transduced.
A defining feature of subretinal AAV gene therapy is durability. Photoreceptors and RPE cells are largely post-mitotic — they do not divide and therefore do not dilute or lose the episomal vector genome the way a rapidly dividing cell population would. This allows gene expression established shortly after treatment to persist for an extended period, potentially providing lasting visual function improvement from a single administered dose.
In rapidly renewing tissues, an episomal (non-integrating) vector genome is progressively lost as cells divide, because the episome is not reliably replicated and partitioned to daughter cells the way chromosomal DNA is. Retinal photoreceptors and RPE cells largely escape this problem: they are terminally differentiated and essentially do not divide across a normal lifespan.
As a result, once a cell is successfully transduced and its episome is established, that cell and its descendants (there being effectively none) continue to carry and express the therapeutic gene indefinitely, without the dilution effect seen in proliferative tissue.
Clinical experience with subretinal AAV gene therapy for RPE65-mediated inherited retinal disease has shown that functional improvements — including gains on validated functional mobility and light-sensitivity testing — established in the months after treatment have been maintained over years of follow-up in treated patients, consistent with sustained transgene expression in the non-dividing target cell population.
Because treatment is administered as a single subretinal procedure per eye rather than a repeated dosing schedule, the durability of the initial transduction event is central to the therapy’s overall clinical value — the treatment is designed to be a one-time intervention rather than an ongoing course.
The single-administration model has practical implications:
• Pre-existing or treatment-induced immunity to the AAV capsid can complicate re-dosing of the same eye, reinforcing the importance of achieving adequate transduction on the first attempt • Baseline disease severity and the amount of viable retinal tissue at the time of treatment influence how much functional benefit is achievable, since gene therapy restores function in surviving cells rather than regenerating lost ones • Long-term monitoring remains important to confirm that expression and functional benefit remain stable over time, even though the underlying biological rationale predicts durability in a non-dividing cell population
This informational illustration is not intended to represent guaranteed real-world clinical outcomes for any specific patient — actual durability and degree of visual improvement vary with baseline retinal health, transduction efficiency achieved, and individual disease course.