Purpose
To develop and evaluate an rAAV5-based gene therapy vector expressing an optimized human RPGR ORF15 transgene (rAAV5-RPGR) for the treatment of X-linked retinitis pigmentosa caused by RPGR mutations, addressing the challenges of cloning the unstable wild-type ORF15 sequence.
Design
This was a prospective experimental study.
Subjects
This was an animal study.
Methods
An optimized RPGR ORF15 sequence was designed to eliminate problematic secondary structures and cryptic splice sites. In vitro expression was validated in HEK 293T and photoreceptor-like 661 W cells. A complete Rpgr knockout mouse model ( Rpgr -knockout [KO]) was generated and characterized phenotypically. Therapeutic efficacy was assessed in Rpgr -KO mice via subretinal injection of rAAV5-RPGR at low (1 × 10⁹ vg/eye), medium (3 × 10⁹ vg/eye), or high (1 × 10¹⁰ vg/eye) doses. Structural and functional outcomes were evaluated at 12- and 14-month postinjection. Short-term safety was assessed in rabbits 1 month after subretinal injection.
Main Outcome Measures
Level of RPGR protein expression and Protein isoform profile (elimination of truncated isoforms), Cellular localization of transgene expression and Dose-dependence of expression, outer nuclear layer thickness, and electroretinography parameters.
Results
(1) The optimized vector increased RPGR protein expression 3.3-fold in vitro compared to wild-type and eliminated truncated isoforms. (2) Subretinal delivery of rAAV5-RPGR in mice demonstrated dose-dependent transgene expression localized correctly to photoreceptor inner segments. (3) In Rpgr -KO mice, high-dose treatment significantly preserved outer nuclear layer thickness at the injection site (42% greater than controls at 14 months, P <.01) and central retina ( P <.05), reduced aberrant rhodopsin mislocalization ( P <.01), and partially restored retinal function. ERG showed significantly improved scotopic a-wave (≥100 vs <90 µV in controls at 10 cd·s/m²) and photopic b-wave amplitudes (49-66 vs 31-46 µV at 30 cd·s/m²) in treated mice. (4) No vector-related toxicity was observed in rabbits.
Conclusions
rAAV5-RPGR mediated efficiently, targeted expression of optimized RPGR-ORF15, significantly preserved photoreceptor structure and function in a severe X-linked retinitis pigmentosa mouse model, and demonstrated a favorable safety profile. This study provides preclinical proof-of-concept for RPGR-targeted gene replacement therapy.
INTRODUCTION
R etinitis pigmentosa (RP) represents a phenotypic continuum of inherited retinal dystrophies characterized by progressive vision loss. The disorder affects approximately 1 in 4000 to 1 in 3000 individuals. Initial clinical manifestations typically include diminished night vision and peripheral visual field constriction. As the disease advances, central visual acuity and color vision may also become compromised. ,,, The age of symptom onset is variable, although it frequently occurs between the first and third decades of life, and the rate of functional deterioration exhibits significant interindividual heterogeneity. ,
RP arises from mutations in one or several genes critical for retinal homeostasis and function. , Among the monogenic causes of RP, defects in the RP GTPase regulator gene ( RPGR ), located on the X chromosome, constitute the most prevalent form. X-linked RP (XLRP) is recognized as one of the most severe RP subtypes. Mutations in RPGR account for approximately 70% of XLRP cases. ,, The RPGR protein localizes to the connecting cilium (CC) of photoreceptors, where it plays a pivotal role in protein trafficking. , Over 300 distinct RPGR mutations have been identified. , The gene undergoes complex alternative splicing, generating multiple isoforms; notably, the RPGR-ORF15 isoform is predominantly expressed in retinal photoreceptors. RPGR mutations instigate the degeneration of both rod and cone photoreceptors, commencing in early childhood.
As a loss-of-function disorder, RP caused by RPGR mutations presents an ideal candidate for gene replacement therapy. Among current gene delivery vectors, adeno-associated virus (AAV) has emerged as the predominant platform for gene therapy applications due to its favorable safety profile, excellent tissue tropism, capacity for sustained transgene expression, and low immunogenicity, as evidenced by multiple approved gene therapy products. ,,,, However, the development of RPGR -based therapies faces a significant challenge: exon ORF15 contains a highly repetitive purine-rich sequence (predominantly adenine and guanine nucleotides), , which is inherently prone to mutation. This repetitive nature complicates the cloning of the full-length RPGR cDNA into viral vectors. These obstacles can be mitigated through rational codon optimization and sequence engineering strategies. Furthermore, murine models effectively recapitulate key features of human RPGR pathology, ,, providing robust tools for preclinical proof-of-concept studies.
Although RPGR gene therapy has advanced into clinical trials (NCT03116113, NCT04671433, NCT04850118), current strategies face two core unresolved issues: (1) the extreme instability of the ORF15 sequence leads to the generation of truncated, nonfunctional isoforms during vector production, reducing the yield and quality of effective vectors; and (2) although overtruncation strategies (such as the deletion of ∼126 amino acids in MeiraGTx’s approach) improve sequence stability, some studies have shown truncation of the highly glutamylated region may impair the protein’s funtion, , which may be a potential reason for the insufficient efficacy observed in its Phase III clinical trials. Therefore, developing a novel vector that both mitigates sequence instability and maximizes the retention of protein functional domains holds significant clinical translational value.
In this study, we engineered an AAV5 vector expressing an newly codon optimized human RPGR ORF15 transgene with limited truncation (rAAV5-RPGR) and validated its expression in vitro using photoreceptor cell lines. We developed a Rpgr knockout mouse model ( Rpgr -KO) and characterized its ocular phenotype. To evaluate the therapeutic potential of our candidate vector, we administered rAAV5-RPGR via subretinal injection (SRI) to Rpgr -KO mice. A comprehensive panel of structural and functional assays was employed to assess preliminary in vivo safety and efficacy.
MATERIALS AND METHODS
plasmid construction
Optimized human RPGR ORF15 CDS and WT human RPGR ORF15 CDS were synthesized by GenScript. The fragment RPGR ORF15 opt1, opt2, and WT were digested with restriction enzyme Nhe I and PspX I, respectively, which allowed the insertion of CDS downstream of a photoreceptor-specific promoter, followed with an intron and upstream of a SV40 polyadenylation site. The expression cassettes above were cloned into an existing AAV plasmid maintained in the lab. The plasmids were designated pAAV-RPGR ORF15 opt1, pAAV-RPGR ORF15 opt2, and pAAV-RPGR ORF15 WT, respectively.
aav production
rAAV5-RPGR virus was provided by Neurophth Inc, which was produced in HEK 293T cells. Briefly, cells were transfected with pAAV-RPGR ORF15 opt1 plasmid/pAAV-noncoding and two helper plasmids by using PEI Pro (Polyplus, 115-010). Forty-eight hours after transfection, cells were collected and disrupted, the cell lysate was digested with Benzonase (Sigma, 30 U/mL, E1014-25KU) at 37°C for 2 hours. Virus particles were then concentrated at 3000 g for 10 minutes. AAV preparations were purified by iodixanol gradient ultracentrifugation. Titers of the virus were determined by real-time PCR using linearized plasmid standards and primers targeting the inverted terminal repeat: Forward: 5′-GGAACCCCTAGTGATGGAGTT-3′, Reverse: 5′-CGGCCTCAGTGAGCGA-3′.
animals
The Rpgr -KO mice model was generated by Cyagen. Briefly, gRNAs targeting exon 1 and exon 18, along with Cas9, were coinjected into C57BL/6 J mouse fertilized eggs to KO the gene via induced double-strand breaks. Rpgr -KO mice, C57BL/6 J mice and New Zealand White rabbits were purchased from Beijing Vital River Laboratories Animal Technology and housed in an AAALAC-accredited standard facility. Animals were randomly assigned to groups based on body weight. All animal procedures and experiments were approved by the Institutional Animal Care and Use Committee (IACUC).
subretinal injection
Mice were anesthetized by intraperitoneal injection of ketamine (80 mg/kg) and xylazine (8 mg/kg). Pupils were dilated with Compound Tropicamide Eye Drops (applied 1-2 minutes preinjection). Injections were performed using a WPI UMP3-Micro2T system with a 10-µL microsyringe and 34 G needles under an ophthalmic surgical microscope (YZ20P5, Suzhou Liuliu). A 32 G needle was inserted at 15° to the optical axis, 1 to 2 mm posterior to the temporal limbus to create a scleral channel (piercing sclera without penetrating RPE). A 34 G beveled needle (NF34BL-2, WPI) was advanced through the channel toward the posterior pole. An amount of 2.0 µL of test article was injected in 2000 nL/min (programmed speed). The needle was retained for 40 to 60 seconds postinjection before withdrawal. After injection, OCT imaging was performed. If a localized hyporeflective area (subretinal fluid) was observed between the retinal neural epithelium and the pigment epithelium, the injection was considered successful. If retinal elevation cannot be observed, the animal was excluded.
pcr analysis
For RPGR mRNA expression analysis, 661 W cells were collected 48 hours after viral infection. Total RNAs were extracted by using the RNAsimple Total RNA kit (Tiangen, DP419), cDNA was synthesized by using the FastKing gDNA Dispelling RT SuperMix (Tiangen, KR118). Gene expression was determined by qRT-PCR on ABI StepOne Plus quantitative PCR system (Thermo Fisher) using TB Green Premix Ex Taq (TAKARA, RR42LR). The sequence of the primers used were as follows: RPGR -F: 5′-TCTCAGCTCGAATGAGACGC-3′, RPGR -R: 5′-CCTCGATGGGAGGAAGTGTG-3′, ACTIN -F: 5′-GGACTTCGAGCAAGAGATGG-3′, ACTIN -R: 5′-AGGAAGGAAGGCTGGAAGAG-3′. Relative gene expression was calculated by the comparative threshold cycle method.
western blot
Virus-infected 661 W or 293T cells were harvested and lysed using RIPA lysis buffer (Biosharp, BS064). Protein concentrations were quantified using a bicinchoninic acid assay. Protein samples were separated by SDS-PAGE, transferred to a PVDF membrane, and blocked with 5% nonfat milk. The membrane was incubated overnight at 4°C with the following primary antibodies: Anti-RPGR (Sigma, HPA001593, 1:1000), anti-ACTIN (Proteintech, 66009-1-Ig, 1:1000). After washing, the membrane was incubated with a goat antirabbit horseradish peroxidase-conjugated secondary antibody (Proteintech, SA00001-2, 1:10,000) or goat antimouse horseradish peroxidase-conjugated secondary antibody (Proteintech, SA00001-1, 1:10,000) for 1 hour at room temperature. Protein bands were visualized by chemiluminescence imaging instrument (Biorad, JY04S-3C) using enhanced chemiluminescence substrate (Boster, AR1191).
tissue preparation and immunostaining
After euthanasia, mouse or rabbit eyes were enucleated, with orientation marked by black dye to ensure consistent analysis between vector- and vehicle-treated eyes. Tissues were fixed in 4% paraformaldehyde (Solarbio, P1112) for 1 hour, cryoprotected in 30% sucrose/PBS overnight, and cryosectioned at 10 µm (Leica HistoCore BIOCUT). Sections were mounted on Superfrost Plus slides and stored at −80°C. Prior to staining, sections were postfixed in 1% formaldehyde/PBS (pH 7.0) for 2 minutes.
Primary antibody incubation was performed using anti-OPSIN (Merck, AB5405). After washing (3 × 0.05% PBST, 1 × PBS), sections were incubated with fluorescent secondary antibodies (Abcam, antimouse ab150080, antirabbit ab150077) for 1 hour at RT in the dark. Following additional washes (3 × PBST, 1 × PBS), nuclei were counterstained with DAPI (Beyotime, C1006). Slides were mounted with antifade medium (Beyotime, P0126), coverslipped, and sealed with nail polish.
h&e staining
Sections were stained in hematoxylin (Servicebio, G1004-500ML), rinsed with tap water, differentiated (1 second in 0.1% HCl/70% ethanol), and blued (Zhuhai Baso, BA4025D). After eosin staining (Zhuhai Baso, BA-4024), slides were dehydrated (70% → 95% → 100% ethanol, xylene) and mounted (Zhuhai Baso, BA7004).
optical coherence tomography
OCT volume scan images were acquired with a spectral domain OCT system (SPECTRALIS, Heidelberg Engineering). Mice were anesthetized, and pupils were dilated as described earlier. Retinal thickness maps were generated by Heidelberg Eye Explorer software. The thickness of the ONL is measured by demarcating its upper and lower boundaries on the image. Twelve equidistant points are selected along the layer on the image for measurement. The final thickness is calculated as the mean of these measurements.
electroretinogram
Mice or rabbits were dark-adapted overnight. Anesthesia and pupil dilation were conducted as described earlier. Recordings were performed using the Celeris Small Animal Visual Electrophysiology System (Diagnosys LLC; Model D430-p- 10). Electrodes were connected to the stimulator. Animals were positioned in a prone posture on the platform, ensuring full body extension and aligned eye height. Corneal electrodes were placed under dim red LED illumination. Corneal ERGs were recorded from both eyes using gold wire loop electrodes with a drop of 2.5% hypromellose ophthalmic demulcent solution. A gold wire loop placed in the mouth was used as reference, and a ground electrode was on the tail. The ERG protocol consisted of recording dark-adapted ERGs using brief flashes of 0.001 to 10 cd.s.m 2/flash. Light-adapted ERGs were recorded for stimulus intensities of 1 to 30 cd.s.m 2/flash. Responses were computer-averaged and recorded at intervals of 3 to 60 seconds depending upon the stimulus intensity.
data analysis
The personnel performing the injections and the data analysts were blinded to the group allocation. Two-tailed paired t test was used to cf outcomes in treated vs control eyes. GraphPad Prism 9 (GraphPad Software) was used for statistical analysis.
RESULTS
design and in vitro expression of optimized vectors
The RPGR ORF15 sequence has represented one of the most formidable challenges in ocular gene therapy development due to its unique structural characteristics and extreme instability during conventional cloning procedures. Multiple studies have documented mutation rates exceeding 80%, , with the majority of clones containing deleterious frameshifts or premature stop codons that render them therapeutically useless. To address these challenges, we developed a multipronged optimization platform combining computational prediction with empirical validation. Computational analysis using RNAfold ( http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi ) and Augustus ( https://bioinf./augustus/submission.php ) identified three instability determinants: (1) A cryptic splice donor site at the exon15-intron15 junction, (2) GC-rich regions forming stable stem-loop structures, (3) an mRNA-destabilizing 40-residue segment (amino acid 865-905). Optimization employed two complementary strategies: (1) Disrupted instability motifs while maintaining 100% amino acid identity, incorporating human codon bias, (2) excised the aa 865 to 905 destabilizing domain. This integrated platform may represent a useful step forward in gene therapy vector design, offering a possible approach for XLRP therapy.
Using the sequence optimization methods described above, we obtained two codon‑optimized nucleic acid sequences of RPGR ORF15 with different codon usage frequencies (designated opt1 and opt2). Compared with the wild‑type (WT) sequence, both optimized sequences encode same amino acid sequences that are partially truncated, which delete about two Glu-Gly repeats ( Figure 1 , A). We then constructed opt1, opt2, and the WT sequence into AAV vectors in the order depicted in Figure 1 , A. The final construct features: Human GRK1 promoter for photoreceptor-specific expression, synthetic intron with cis-acting stability motifs, and optimized RPGR ORF15 sequence followed with SV40 polyadenylate signal. These plasmids were designated as pAAV‑RPGR‑ORF15 opt1, pAAV‑RPGR‑ORF15 opt2, and pAAV‑RPGR‑ORF15 WT, respectively.
Design and in vitro expression of optimized vectors. (A) Schematic map of AAV-RPGR vector design and route of AAV administration. Wild-type RPGRORF15 possesses 11 tandem GEEEG repeats, whereas optimized RPGRORF15 (opt1/opt2) possesses 9 tandem GEEEG repeats; 2 of 11 tandem GEEEG repeats were deleted. (B) Protein extracts from 293T cells transfected with pAAV-RPGR-ORF15 opt1, pAAV-RPGR-ORF15 opt2, and pAAV-RPGR-ORF15 opt1 were subjected to Western blot analysis. Immunoblotting analysis was performed using anti-RPGR antibody, β-actin was used as internal control. The expression level of optimized RPGR (opt1) was 3.3-fold higher than wild-type (WT). (C) 661 W cells were infected by rAAV5-GFP as negative control (NC) and rAAV5-RPGR at different MOIs of 1E4, 3E4, and 1E5, respectively. Forty-eight hours after infection, total RNAs were extracted for expression level analysis. Data are expressed as mean ± 95% CI, N = 3. (D) 661 W cells were infected by rAAV5-GFP as negative control (NC) and rAAV5-RPGR at different MOIs of 1E4, 3E4, and 1E5, respectively. Forty-eight hours after infection, total proteins were extracted for expression level analysis.
HEK 293T cells were seeded in 6-well plates, and after 24 hours, each well was transfected with 2 µg either pAAV-RPGR-ORF15 opt1, pAAV-RPGR-ORF15 opt2, or pAAV-RPGR-ORF15 WT (WT) plasmid DNA. After 48 hours, total cellular protein was extracted and subjected to Western blot analysis to assess RPGR protein expression, using untransfected HEK 293T cells as a negative control. As shown in Figure 1 , B, the opt1 vector yielded the highest protein expression level, exhibiting a 3.3-fold increase compared to the wild-type construct. Notably, the wild-type sequence produced truncated isoforms (indicated by the red box), whereas both optimized sequences showed no detectable truncated bands, indicating that the protein expressed from the optimized sequence is more stable compared to the wild-type. Since the opt1 sequence expresses significantly higher protein levels than opt2, which theoretically enables effective treatment with a lower viral load, we therefore selected pAAV-RPGR-ORF15 opt1 as the final vector.
XLRP caused by RPGR mutations primarily affects retinal photoreceptors (rods and cones). As these highly differentiated, specialized cells are difficult to obtain and maintain in vitro, no established human photoreceptor cell lines exist. The 661 W cell line, a murine-derived retinoblastoma cell line, expresses characteristic photoreceptor marker proteins. Crucially, photoreceptor-specific promoters, which are often inactive in common cell lines, can effectively drive downstream gene expression in 661 W cells. Given the requirement for photoreceptor-specific expression in this study, 661 W cells served as a suitable in vitro model to validate transgene expression.
As described above, Opt1 was selected for further development based on its superior protein expression. Therefore, all subsequent experiments (including vector packaging, dose‑finding, and efficacy studies) were performed exclusively with the Opt1‑derived vector. Recombinant AAV5-RPGR-ORF15 opt1 virus (unless otherwise specified, the abbreviation rAAV5-RPGR will be used hereafter) was produced using the triple-plasmid transfection method. A total of 661 W cells were then infected with rAAV5-RPGR at multiplicities of infection of 1 × 10⁴, 3 × 10⁴, and 1 × 10⁵. After 72 hours, mRNA and protein expression were quantified by qPCR and Western blot analysis. Results are presented in Figure 1 , C and D, compared to the negative control, both RPGR mRNA and protein levels were significantly upregulated in rAAV5-RPGR-infected 661 W cells across all tested multiplicities of infection. Furthermore, the expression levels exhibited a clear dose-dependent relationship.
efficient and correct expression of raav5-rpgr in murine retina
Efficient and correctly localized transgene expression in vivo is a critical prerequisite for therapeutic efficacy. To evaluate whether the rAAV5-RPGR vector mediated high-level and appropriate expression in the murine retina, adult C57BL/6 mice received SRIs of rAAV5-RPGR at three different doses: 1 × 10⁹ vector genomes (vg)/eye, 3 × 10⁹ vg/eye, and 1 × 10 10 vg/eye. Four weeks postinjection, retinas were harvested for analysis. Total protein was extracted from retinas and subjected to Western blotting to assess RPGR protein levels, using β-actin as a loading control. As shown in Figure 2 , A, RPGR protein expression was readily detectable in retinas injected with rAAV5-RPGR, in contrast to untreated retinas. Furthermore, expression levels exhibited a clear dose-dependent increase across the low, medium, and high dose groups. Retinal cryosections were prepared and analyzed by immunofluorescence to determine the localization and expression levels of RPGR protein, using rhodopsin (RHO) as a photoreceptor marker. Figure 2 , B demonstrates that robust RPGR expression (red signal) was observed specifically in retinas treated with rAAV5-RPGR, compared to controls. Critically, the RPGR protein was correctly localized to the inner segments (IS) of photoreceptor cells. Consistent with the Western blot results, the intensity of the RPGR immunofluorescence signal increased in a dose-dependent manner.
Efficient and correct expression of rAAV5-RPGR in murine retina. (A) Detection of rAAV5-RPGR expression in mouse retina by Western Blot. rAAV5-RPGR at different doses (1E9 vg/eye, 3E9 vg/eye, and 1E10 vg/eye) were administered to mouse eyes via subretinal injection. Mice were euthanized for 4 weeks postinjection, and retinas were isolated. Total retinal protein was extracted, and RPGR ORF15 protein expression levels were detected by Western blotting. Retinas from mice injected with vehicle solution served as the negative control group (NC). Band intensity was quantified using ImageJ software. ACTIN served as internal control; the analysis demonstrated that drug expression within the mouse retina was dose-dependent. Three replicates were performed per group. (B) Immunofluorescence detection of tissue localization for rAAV5-RPGR in mouse eye sections. The rAAV5-RPGR at different doses (1E9 vg/eye, 3E9 vg/eye, and 1E10 vg/eye) was administered to mouse eyes via subretinal injection. Four weeks postinjection, mice were euthanized, and eyes were enucleated. Paraffin sections of the eyes were prepared. Expression and tissue localization of the RPGR ORF15 protein were detected by immunofluorescence staining. Eyes from mice injected with the vehicle solution (PBS) served as the control group. Green signal represents opsin RHO, localized in the photoreceptor outer segments (OS); red signal represents RPGR, localized in the photoreceptor inner segments (IS); blue signal represents cell nuclei (DAPI), distinguishing different retinal layers, such as the outer nuclear layer (ONL). Ten replicates were performed per group. Scale bar = 20 µm.
characterization of the rpgr-ko mouse model
To validate the photoreceptor-protective efficacy of the rAAV5-RPGR, we generated a Rpgr KO mouse model ( Rpgr -KO). Given the wide distribution of pathogenic mutations across multiple exons in human patients with RPGR-associated RP, the Rpgr -KO mouse was engineered by deleting sequence from exon 1 to exon 18 of the target gene. RP primarily involves the degeneration of rod photoreceptors initially, with subsequent cone photoreceptor loss, collectively leading to progressive vision impairment. Therefore, demonstrating progressive photoreceptor degeneration in the model is essential to confirm its fidelity in recapitulating the human disease phenotype and underlying pathology. Longitudinal optical coherence tomography (OCT) imaging revealed significant thinning of the outer nuclear layer (ONL) in Rpgr -KO mice compared to WT mice. This ONL degeneration exhibited a pronounced age-dependent progression ( Figure 3 , A). Histological analysis of retinal sections from 14-month-old mice confirmed severe degeneration. The ONL in Rpgr -KO mice displayed a marked reduction in the number of cellular layers compared to age-matched WT mice, indicative of substantial photoreceptor cell loss ( Figure 3 , B). Full-field electroretinogram (ERG) was performed to evaluate photoreceptor function. As shown in Figure 3 , C, 14-month-old Rpgr -KO mice exhibited a significant reduction (at light intensities of 0.1 and 1, the P values were.0034 and.0039, respectively, CI = 95%) in the amplitude of the scotopic (rod-mediated) b-wave compared to WT controls, demonstrating severe impairment of retinal function.
Characterization of the Rpgr -knockout mouse model. (A) Measurement of ONL thickness in mice. The ONL thickness was measured in WT (wild-type) and Rpgr -KO mice at two time points (12 and 14 months of age). Retinas were scanned using the circular scanning mode of an OCT (optical coherence tomography) instrument. ONL thickness was statistically analyzed across different scanning angles (45°, 90°, 135°, 225°, 275°, 315°). Data are expressed as mean ± 95% CI, N = 10. ** P <.01. (B) H&E staining of mouse eye sections. Paraffin sections were prepared from enucleated eyes of 14-month-old WT and Rpgr -KO mice. Sections were stained using H&E staining to observe the number of cell nuclei and layer thickness in the ONL. Significant differences were observed between the two groups. Scale bar = 20 µm. (C) Full-field ERG was performed on 14-month-old mice. The amplitude of dark-adapted b wave at stimulus intensity of 0.01, 0.1, and 1 cd·s/m² were measured. Fourteen-month-old Rpgr -KO mice exhibited a significant reduction compared to WT controls. Data are expressed as mean ± 95% CI, N = 10. ** P <.01. (D) Immunostaining and fluorescence quantification in mouse eye sections. Paraffin sections were prepared from enucleated eyes of 14-month-old mice. Sections were subjected to immunofluorescence staining, with red fluorescence representing opsin RHO and blue fluorescence representing cell nuclei (DAPI). Observation of RHO distribution across retinal layers revealed disorganized opsin localization in gene-edited mice, exhibiting ectopic leakage into the outer nuclear layer (ONL). Mean fluorescence intensity analysis of red fluorescence within the ONL was performed using ImageJ software. The opsin signal distributed in the ONL of RPGR-KO retinas was significantly higher than in wild-type mice. *** P <.001; data are expressed as mean ± 95% CI, N = 10. Scale bar = 20 µm.
RPGR localizes to the CC, a critical structure bridging the photoreceptor IS and outer segment (OS). The CC serves as the exclusive conduit for the vectorial transport of photo-transduction proteins (eg, RHO), synthesized and modified in the IS, to the OS. RPGR, within a multiprotein complex, functions as a key regulator at this ciliary gate. Mutations in RPGR disrupt CC function, impairing the trafficking and proper localization of opsins. Immunofluorescence analysis of retinal sections from 14-month-old mice ( Figure 3 , D) demonstrated the critical consequence of RPGR loss. In WT retinas, RHO immunolabeling was predominantly and correctly localized within the OS. In contrast, Rpgr -KO retinas exhibited significant RHO mislocalization, characterized by aberrant accumulation within the ONL. Quantitative analysis of fluorescence signal intensity within the ONL confirmed a significant increase ( P =.0003, CI = 95%) in RHO levels in Rpgr -KO mice compared to WT controls, confirming the trafficking defect.
therapeutic efficacy of raav5-rpgr in rpgr-ko mice
To evaluate the in vivo efficacy of the rAAV5-RPGR, Rpgr -KO mice (2 months old) received SRIs of either rAAV5-RPGR at low (1 × 10⁹ vg/eye), medium (3 × 10⁹ vg/eye), or high (1 × 10¹⁰ vg/eye) doses, or a control vector (rAAV5-noncoding) at equivalent doses. Given the protracted disease progression in this model, requiring extended time for a therapeutic window to manifest, structural and functional endpoints were assessed at 12 and 14 months of age.
OCT revealed that treated mice exhibited greater ONL thickness compared to rAAV5-noncoding-injected controls. Before injection, we measured the baseline ONL thickness values in both treatment and control groups across all doses, and the results showed no significant differences between the treatment and control groups (Figure S1). As shown by Figure 4 , A, in the low dose group: At 12 and 14 months, ONL thickness in both the injection site and central retina region were slightly increased compared to controls, but these differences did not reach statistical significance (at 12 months, the P values were.8636 and.1230, respectively. At 14 months, the P values were.3286 and.0843, respectively, CI = 95%). In the medium dose group: At 12 months, ONL thickness showed a nonsignificant increase over controls in both regions (the P values were.0590 and.3129, respectively, CI = 95%). By 14 months, ONL thickness at the injection site was significantly greater than controls ( P =.0140, CI = 95%), while the difference in the central retina region remained nonsignificant ( P =.1947, CI = 95%). In the high dose group: At both 12 and 14 months, ONL thickness was significantly greater than controls in both the injection site and the central retina region (at 12 months, the P values were.0215 and.0011, respectively. At 14 months, the P values were.0186 and.0452, respectively, CI = 95%). These data demonstrate that high-dose rAAV5-RPGR administration mitigated photoreceptor degeneration caused by RPGR deficiency. Immunofluorescence analysis of retinal cryosections from 14-month-old high-dose mice, stained with DAPI and anti-RHO antibody, revealed a significant reduction in the aberrant RHO signal (red) leaking into the ONL compared to rAAV5-noncoding controls ( Figure 4 , C). Quantitative analysis confirmed significantly attenuated RHO mislocalization in treated retinas ( P =.0057, CI = 95%), indicating restoration of normal RHO trafficking to the OS.
