A gene-augmentation framework for selected early-stage autosomal recessive retinitis pigmentosa genotypes

Technical Summary: A Gene-Augmentation Framework for Selected Early-Stage Autosomal Recessive Retinitis Pigmentosa Genotypes

Problem Statement
Retinitis pigmentosa (RP) encompasses over 100 associated genes with diverse inheritance patterns, creating a significant translational challenge for gene therapy. While gene augmentation is theoretically applicable to many recessive forms of RP, only one therapy (voretigene neparvovec for RPE65) has received regulatory approval. The diversity of genotypes leads to small patient populations for each specific mutation, making conventional, independent drug development programs economically and logistically difficult to sustain. Furthermore, the packaging capacity of adeno-associated virus (AAV) vectors limits the size of transgenes that can be delivered. There is a need for a shared platform strategy that utilizes common capsids, regulatory elements, and manufacturing processes to address multiple genotypes, provided that specific transgenes are validated for safety and efficacy.

Methodology
The authors developed a multi-stage preclinical framework to identify and validate candidate genes and vector components:

1. Gene Prioritization Screen: The team screened over 100 RP-associated genes listed in RetNet. Selection criteria included:

   • Inheritance: Focusing on autosomal recessive patterns.
   • Expression: Using single-cell RNA-sequencing datasets to identify genes predominantly expressed in rods, retinal pigment epithelium (RPE), or Müller glia.
   • Packaging: Ensuring the coding sequence fits within the AAV genome (~4.7 kb limit).
   • Promoter Availability: Verifying the existence of publicly described, cell-type-specific promoters.
     This process yielded 24 potentially tractable genes, which were further narrowed to 19 based on promoter compatibility in primates.

2. Promoter Validation in Non-Human Primates: To ensure translational relevance, four human regulatory sequences (RHO for rods, BEST1 for RPE, GFAP for Müller glia, and RLBP1 for RPE/Müller) were tested in rhesus monkey eyes via subretinal injection of AAV8-GFP vectors. Expression was assessed via fundus imaging, optical coherence tomography (OCT), and histology at days 28 and 56.

3. Proof-of-Concept Vector Design: Based on the screen, a shared cassette architecture was designed using AAV8 capsids, the human RHO or BEST1 promoters, WPRE3/SV40 polyadenylation signals (W3SL) for compactness, and a TLR9 inhibitory sequence (TLR9io2) to reduce innate immune sensing. PDE6B was selected as the proof-of-concept transgene.

4. Dose-Ranging and Efficacy Studies:

   • Dose-Ranging: AAV8-RHO-PDE6B was administered to wild-type mice (neonatal) and rhesus monkeys (adult) at varying doses to identify a safe exposure window.
   • Efficacy in Mouse Models: The selected lower dose (5 × 10⁸ vg/eye) was tested in two Pde6b-deficient mouse models (rd1 and rd10) with different degeneration kinetics. Assessments included Outer Nuclear Layer (ONL) thickness via histology and OCT, electroretinography (ERG) for retinal function, and visually guided behavioral assays (light-dark discrimination and optomotor response).

Key Contributions

• Identification of 19 Candidate Genes: The study defines a list of 19 autosomal recessive RP genes (15 rod-specific, 3 RPE-specific, and 1 dual-expressed) that are compatible with a shared AAV8-based gene augmentation framework.
• Primate-Validated Promoters: The study demonstrates that human RHO and BEST1 promoters support rod- and RPE-enriched expression, respectively, in rhesus monkey eyes. Conversely, it highlights that GFAP and RLBP1 promoters failed to provide suitable, non-toxic expression profiles in primates, excluding five Müller glia-associated genes from the current platform.
• Safety and Efficacy of AAV8-RHO-PDE6B: The proof-of-concept vector demonstrated that lower doses (5 × 10⁸ vg/eye in mice; 1 × 10¹¹ vg/eye in monkeys) preserve retinal structure and function without significant toxicity, whereas higher doses were associated with retinal abnormalities.

Results

• Promoter Performance: In rhesus monkeys, AAV8-RHO-GFP and AAV8-BEST1-GFP showed cell-type-specific expression in the photoreceptor layer and RPE, respectively. AAV8-GFAP-GFP showed no detectable expression, and AAV8-RLBP1-GFP resulted in heterogeneous expression accompanied by RPE abnormalities.
• Dose-Dependent Toxicity: In wild-type mice and monkeys, higher doses of AAV8-RHO-PDE6B caused focal waviness, protrusions in the ONL, and RPE abnormalities. Lower doses were well-tolerated with no significant structural defects observed through day 56 in monkeys.
• Structural Preservation: In rd1 and rd10 mice treated neonatally, the lower dose of AAV8-RHO-PDE6B preserved ONL thickness at P20, P90, and up to 6 months post-treatment, significantly outperforming controls.
• Functional Rescue: Treated mice exhibited increased scotopic and photopic ERG amplitudes, indicating preserved rod and cone pathway function. Visually guided behaviors, including light avoidance and optomotor responses, were also preserved in treated animals compared to untreated degenerating controls.
• Timing Sensitivity: Treatment at postnatal day 2 (P2) was effective, whereas treatment at P0/P1 did not yield detectable benefits, suggesting a critical window related to rod genesis and degeneration onset.

Significance and Claims
The authors posit that this work defines a "preclinical experimental approach" for a shared gene-augmentation framework. The study claims to provide a prioritized starting point for developing therapies for selected early-stage autosomal recessive RP genotypes by leveraging shared vector components to reduce duplicated development work.

However, the paper maintains a modest tone regarding clinical translation. The authors explicitly state that:

• The findings support further evaluation of the feasibility of reusing vector components but do not establish clinical safety or efficacy for PDE6B or the other 18 genes.
• Each candidate transgene still requires individual validation for expression, potency, biodistribution, immunogenicity, and toxicology.
• The current data are limited by small non-human primate cohorts, short follow-up durations (56 days in monkeys), and the use of neonatal mouse models, which may not fully reflect the therapeutic window for adult human patients.
• The proposed "staged access model" for patient enrollment is a development concept, not a clinical protocol, and requires further regulatory and operational establishment.

In summary, the paper provides a biological basis and a structured pipeline for targeting a specific subset of recessive RP genes but emphasizes that platform efficiency cannot substitute for product-specific validation.