This was an investigator-initiated, open-label, nonrandomized, single-center clinical trial and was registered with the Chinese Clinical Trial Registry (ChiCTR identifier: ChiCTR2200062174). Participants were recruited at the First Affiliated Hospital of Soochow University between September 2022 and May 2024. A total of nine patients diagnosed with advanced nonsyndromic RP were enrolled into two distinct cohorts. Six participants were assigned to the LP cohort, with decimal visual acuity ≤0.02, as measured using the standard logarithmic visual acuity chart, and preserved LP. Three participants were allocated to the NLP cohort, characterized by complete absence of light detection. NLP status was defined based on the standardized testing protocol described in the Endpoints section, consistent with the established clinical definition of total blindness. Comprehensive inclusion and exclusion criteria have been provided in the Supplementary Materials. Participants in the LP cohort were assigned identifiers L001 to L006, while those in the NLP cohort were designated NL001 to NL003. All participants in both cohorts underwent a single intravitreal injection of UGX-201 (1.5 × 10 ¹¹ vector genomes [vg]) in the target eye at baseline and were followed up for a minimum of 52 weeks to assess safety, tolerability, and efficacy.

The trial was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (Approval No.: 2022-048-7) and was conducted in accordance with the principles set forth in the Declaration of Helsinki. Written informed consent was obtained from all participants prior to study enrollment.

For the LP cohort, the eye with inferior best-corrected visual acuity (BCVA, assessed with the standard logarithmic visual acuity chart) was designated as the treated eye. If no significant difference in BCVA was observed between the two eyes, the eye subjectively evaluated as having poorer visual function was selected for treatment. In one case (participant L006), the left eye was selected despite the right eye exhibiting poorer BCVA, as the right eye had previously undergone intraocular gas tamponade for retinal detachment. For the NLP cohort, all three participants lacked LP in both eyes. Since no significant differences in fundus anatomy were identified between the two eyes, the right eye was selected as the treated eye for each participant.

Initially, pupil-dilating drops were administered to the selected eye to facilitate optimal visualization of intraocular structures. After 15-30 minutes, a topical anesthetic (pramecaine hydrochloride eye drops) was then administered to the eye to relieve pain and discomfort during the procedure. A single intravitreal injection of UGX-201, delivering a fixed target dose of approximately 1.5 × 10¹¹ vg in a total volume of 130 to 135 µL or 90 to 95 µL, was carefully administered to the selected eye of each participant. UGX-201 was delivered into the vitreous cavity using a 1 mL sterile tuberculin syringe fitted with a 30-gauge, 0.5-inch needle. The injection was performed slowly and in a controlled manner by an experienced retinal specialist to ensure accurate delivery. Inconsistency in injected drug volume arose from that the drug titer was adjusted after administration to the six participants in the LP cohort, though the injection dose remained the same for both cohorts. The injection was administered under direct visualization through the dilated pupil, with the needle positioned near the macula to ensure optimal drug distribution. Following the procedure, participants remained supine for 30 minutes to promote diffusion of the UGX-201 solution within the retina. To minimize ocular inflammation, participants received 0.5 mg/kg/d of prednisone starting 1 week before treatment, with a maximum dose of no more than 60 mg/d regardless of body weight. Postinjection, the prednisone dose was increased to 1 mg/kg/d during the first week and gradually tapered until Week 8.

Blood samples were collected at baseline, Day 7, Week 2, Week 12, Week 24, and Week 52 (or at premature withdrawal from the study) to evaluate AAV2-7m8 neutralizing antibody (nAb) levels. Additionally, blood samples for AAV vector testing were collected at baseline, Week 12, Week 24, and Week 52 (or end of study) to monitor the leakage of the injected AAV2-7m8 into the systemic circulation.

Serum samples were collected to systematically investigate the humoral immune response to AAV2-7m8 vector. AAV2-7m8 nAb titers were quantified using a cell-based assay, as described in current AAV immunomonitoring protocols.

Prior to transduction into HEK293 cells, serum samples were diluted to different ratios and cocultured with AAV reporter constructs expressing enhanced green fluorescent protein (EGFP). The nAb titer was determined by identifying the serum dilution ratio at which EGFP expression inhibition was closest to or exactly 50%.

To assess the level of systemic vector shedding following intravitreal injection of UGX-201, 8 mL of whole blood was collected from all nine participants at baseline and post-treatment. Genomic DNA was extracted from these samples, and the concentration of the vector was determined using a quantitative polymerase chain reaction assay specifically designed to amplify the unique recombinant junction region between the human MW-opsin and mGluR4 domains within the RCO expression cassette. The lower limit of quantification for this assay was established and validated at 100 genome copies/μL of whole blood. Any amplification signals falling below this threshold were considered unquantifiable background noise and reported as below the limit of quantification.

The primary endpoint of this study was drug safety, including the incidence of adverse events (AEs), treatment-related AEs, procedure-related AEs, serious AEs, and treatment-related serious AEs. The severity of all AEs was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) version 5.0. Safety assessments were performed through continuous monitoring of vital signs, physical examinations, clinical laboratory tests (including hematology, serum chemistry, coagulation profiles, and urinalysis), electrocardiograms, ophthalmic evaluations, and immunogenicity testing for anti-AAV antibodies. Anatomical retinal changes were assessed using color fundus photography (CFP), spectral-domain optical coherence tomography (SD-OCT), OCT angiography, and fundus fluorescence angiography (FFA). The retinal parameters, including central foveal thickness, central macular thickness, ganglion cell complex thickness, and retinal nerve fiber layer thickness, were measured using SD-OCT. Intraocular pressure (IOP) was monitored in both the treated and untreated eyes, and these values were compared to baseline measurements.

In the LP cohort, the secondary endpoints included changes in BCVA over time, as well as evaluations of retinal sensitivity and vision-related quality of life. BCVA was assessed in all participants after refractive correction. Standardized low-vision protocols were used for participants with ultralow off-chart acuity (counting fingers [CF], hand movement [HM], LP, NLP), including fixed-distance CF/HM testing and validated light projection assessment for LP. All tests were conducted under uniform conditions by a single trained examiner to ensure consistency and reduce interexaminer variability. To enhance the accuracy and reliability of visual function assessments, two complementary visual acuity measurement tools were utilized in this study. The first was the Freiburg Visual Acuity and Contrast Test (FrACT, Version 3.10.5; http://michaelbach.de/fract/ ), which was performed on a display with an approximate screen illuminance of 250 cd/m². During FrACT testing, two investigators were present: one conducted the visual acuity assessment, while the second monitored the distance between the participant’s eye and the display screen using a fixed 50-cm measuring ruler, and participants were instructed to maintain a stable head position to ensure a consistent testing distance throughout the examination. Furthermore, repeated measurements were implemented to minimize variability. The second was the standard logarithmic visual acuity chart (GB11533-2011), commonly referred to as the tumbling “E” chart. Retinal sensitivity changes were assessed by full-field stimulus threshold (FST) testing using white and chromatic stimuli (blue, red, and green) with a reference luminance of 0.01 cd·s/m². ^, Testing was performed using the Espion system (Diagnosys LLC; software version V6.64.14), and device calibration was conducted using the manufacturer’s calibration software prior to testing. Lower FST values indicated improved retinal sensitivity to light. A clinically significant change in retinal sensitivity was defined as a change of at least 0.6 log (cd·s/m²), equivalent to 6 dB, based on previous studies of CEP290-associated and RPE65-associated inherited retinal degenerations. ^,, Additionally, differences between blue and red lights sensitivities were analyzed to determine types of visual response, which included rod-mediated (blue-red difference >19.3 dB), cone-mediated (blue-red difference <3.6 dB), or mixed (blue-red difference between 3.6 and 19.3 dB) types. ^, The 25-item National Eye Institute Visual Function Questionnaire-25 (NEI VFQ-25), a 25-item questionnaire with 47 questions, was used to measure improvements in quality of life affected by vision. Higher scores indicated a better quality of life and less dependency on others for assistance. For the NLP cohort, the secondary endpoints primarily included the assessment of the proportion of participants who regained LP in the treated eye following treatment, as well as changes in their visual function, including retinal sensitivity detected by FST testing. According to the clinical definition, NLP refers to the inability to detect light regardless of its intensity or distance. In this study, LP was defined as the accurate detection of a white light stimulus (30 lux) presented in front of the eye in a darkened environment. Participants who failed to detect the light stimulus under these conditions were classified as nonlight perceivers.

In addition, several exploratory endpoints were assessed, including: (1) Visual field (VF) assessments were conducted using the Octopus perimeter to evaluate both central (30°) and peripheral VFs. (2) Full-field electroretinogram (ff-ERG) was conducted to assess the retinal function and progression of RP.

Sample size calculations were based on the efficacy endpoint of visual acuity, as measured by the FrACT. Previous studies have reported a standard deviation of approximately 0.15 logMAR for changes in visual acuity (ΔVA) from baseline in individuals with low vision. For this study, an expected ΔVA of −0.2 logMAR in the treated eye was assumed, corresponding to approximately 25% of participants achieving a clinically meaningful change of 0.3 logMAR which is indicative of an improvement from LP to HM or from NLP to LP. To ensure that the 95% CI for ΔVA did not include 0 with a probability of 85%, a sample size of five participants was required. Accounting for a 10% dropout rate, the total number of participants needed was set at six.

Demographic data were summarized using descriptive statistics. Baseline characteristics were presented for the overall study population. Continuous variables were described using the mean, standard error (SE) of mean, quartiles, and minimum/maximum values, while categorical variables were summarized as frequencies and percentages. Safety analyses were conducted for all participants and included ophthalmological examinations, laboratory tests, and reported ocular and systemic AEs. AEs were categorized and summarized by type and frequency. For efficacy assessments, quantitative indicators were summarized at each follow-up visit, and trends from baseline were calculated as mean with SE of mean. Regarding missing data, only one participant missed the 6-week follow-up visit due to COVID-19-related disruptions. No data imputation was performed for missing data. As this was an exploratory study, no formal hypothesis testing was conducted. Statistical analyses were performed using SAS (version 9.4) and R (version 4.4.1).

After obtaining informed consent, nine legally blind participants with nonsyndromic RP met the eligibility criteria and were successfully enrolled. Their baseline characteristics and genetic mutations are shown in Table 1 . All nine participants are of Han ethnicity. The LP cohort comprised two males and four females, with a mean age of 44.67 (8.87) years. The NLP cohort included three participants, two males and one female, with a mean age of 53.00 (6.56) years.

In the LP cohort, baseline BCVA in the treated eyes ranged from HM at 10 centimeters (HM/10 cm) to CF at 40 centimeters (CF/40 cm), while that in untreated eyes ranged from LP to 0.01 measured using a standard logarithmic visual acuity chart. In the NLP cohort, both eyes of the three participants were NLP. VF testing showed severe defects in all nine participants, consistent with the natural progression patterns of late-stage disease. No participants withdrew from the study throughout the 52-week follow-up period. Additionally, L001 in LP cohort missed the 6-week postoperative follow-up due to COVID-19 infection. However, the missing data from this visit were not believed to have influenced the main conclusions of the study.