Subretinal RS1 gene therapy delivers anatomical rescue in children with X-linked retinoschisis
A subretinal RS1 gene therapy study has shifted the center of gravity in X-linked retinoschisis research. In 12 children treated with scAAV8-hRS1, macular schisis closed by 13 weeks in every treated eye, and outer retinal structure improved in most participants. The anatomical signal was accompanied by an average treated-eye best corrected visual acuity gain of 10.8 ± 4.9 letters at 1 year, compared with 2.4 ± 3.3 letters in untreated fellow eyes.
One sentence stands out: the delivery route changed the biology seen at the bedside.
XLRS is an X-linked recessive retinal disorder caused by pathogenic variants in RS1, which encodes retinoschisin, a secreted octameric protein involved in retinal structural integrity and photoreceptor-bipolar cell signaling. Boys are affected predominantly, and the disease usually appears in childhood with reduced central vision, foveal schisis, and an electronegative full-field electroretinogram. The cystic cavities on optical coherence tomography are only part of the disease burden. The underlying defect reaches deeper, disrupting the adhesive and signaling functions that keep retinal layers organized and responsive.
That biology explains why an intervention can improve anatomy without immediately restoring function.
Why subretinal delivery appears to outperform intravitreal exposure
Earlier intravitreal RS1 trials established the safety challenge without delivering a convincing efficacy signal. Although those studies were generally safe over 1 year, inflammatory ocular adverse events were common and no overall improvement emerged in best corrected visual acuity, visual field, full-field electroretinogram, or cystic-cavity volume relative to fellow eyes. The mechanistic problem is straightforward: an intravitreal vector must cross the vitreous and internal limiting membrane, navigate a structurally abnormal retina, and reach the relevant bipolar and photoreceptor cells. In human eyes, that barrier course appears inefficient.
Subretinal administration avoids much of that obstacle by placing vector directly beneath the retina, closer to the target cells that produce and respond to retinoschisin. In this study, the injection bleb covered the macular region, and the treatment effect was visible early on optical coherence tomography. Treated eyes showed closure of macular schisis cavities and reduced retinal thickness. The absence of grade 3 or higher adverse events and the lack of ocular inflammation over 52 weeks make the safety profile especially notable when contrasted with the inflammatory signal seen after intravitreal dosing.
Visual acuity improved, but electrophysiology did not move
The visual and structural results do not align perfectly, and that mismatch is clinically important. Despite the gain in ETDRS letters, there were no changes in b-wave amplitude or retinal sensitivity. Retinoschisin is not just a scaffold protein; it also supports signaling across the photoreceptor and bipolar-cell layers. If those circuits have already undergone remodeling or irreversible damage, closing schisis cavities may not be enough to normalize function.
In practice, this means OCT improvement cannot be treated as a surrogate for vision restoration. Retinal cysts are visible, quantifiable, and easy to track, but they do not capture the full burden of XLRS. Photoreceptor injury, impaired synaptic transmission, and downstream remodeling can all limit the degree to which structure translates into function. The early visual acuity signal is encouraging, yet it remains only one part of the clinical picture.
The trial’s design supports the signal, but not certainty
The treated cohort was small, open label, and short in duration. Those features matter because XLRS has variable severity and progression, and fellow-eye comparisons can be distorted by eye-specific differences in disease stage, fluctuation in schitic cavities, and uneven natural history. At the same time, bilateral inherited disease does allow each patient to serve as a partial internal control for age-related change and learning effects on visual testing. The comparison is useful, but it does not erase the limitations of a nonrandomized study.
Seven of the 12 treated eyes gained at least 10 letters. That is a meaningful threshold in retinal disease, yet it still needs confirmation in a larger, longer trial that can separate durable treatment effect from short-term fluctuation. Children aged 5 to 11 years were enrolled, which is an age range in which retinal plasticity may remain greater than in later disease. Whether that age window is optimal for preserving structure before function is lost is still unknown.
Genotype and disease stage may determine who benefits most
Most preclinical RS1 work has used knockout models, but more than half of patient alleles are missense variants. Those variants can produce misfolded, retained, or dysfunctional retinoschisin rather than simple absence of the protein. The study cohort included missense variants in the discoidin domain, and one participant had a partial genomic deletion. Different molecular lesions may respond differently to augmentation, especially if residual protein biology shapes how much rescue is possible.
The broader trial field will need genotype-stratified analyses and more sensitive outcome measures. Best corrected visual acuity is likely the most clinically intuitive endpoint, but low luminance visual acuity, microperimetry, and electrophysiology may better capture whether the retina is truly recovering. Natural-history data also matter because XLRS visual acuity often declines slowly, making short-term changes hard to interpret without a clear comparator.
A stronger vector platform may reduce surgical trade-offs
Subretinal injection is not a free solution. In a fragile, schitic macula, the surgery itself can be technically demanding because the retina must be detached at the injection site. Emerging vector engineering may soften that trade-off. A laterally spreading capsid, AAV.SPR, has shown the ability to drive central retinal expression after peripheral subretinal injection in non-human primates, reaching as far as 7.2 mm from the injection edge without requiring a central retinal detachment. That concept is now being tested in XLRS with ATSN-201 in the LIGHTHOUSE trial (NCT05878860).
The near-term question is not whether retinoschisin biology can be repaired. This study suggests it can. The harder task is proving that anatomical rescue is durable, reproducible across genotypes, and matched by lasting functional benefit rather than a transient visual signal in a small pediatric cohort.
- • ClinicalTrials.gov Trial Protocol (NCT05878860): https://clinicaltrials.gov/study/NCT05878860
- • Journal Article DOI 10.1056/NEJMoa2515953: https://doi.org/10.1056/NEJMoa2515953
- • PubMed PMID 42269152: https://pubmed.ncbi.nlm.nih.gov/42269152/
- • Journal Article DOI 10.1089/hum.2015.142: https://doi.org/10.1089/hum.2015.142
- • PubMed PMID 27036983: https://pubmed.ncbi.nlm.nih.gov/27036983/
- • Journal Article DOI 10.1016/j.ymthe.2018.05.025: https://doi.org/10.1016/j.ymthe.2018.05.025
- • PubMed PMID 30196853: https://pubmed.ncbi.nlm.nih.gov/30196853/
- • Journal Article DOI 10.1016/j.oret.2022.06.013: https://doi.org/10.1016/j.oret.2022.06.013
- • PubMed PMID 35781068: https://pubmed.ncbi.nlm.nih.gov/35781068/
- • Journal Article DOI 10.1167/iovs.66.4.37: https://doi.org/10.1167/iovs.66.4.37