CRISPR for Blindness: Gene Editing Trials for Eye Diseases

Why the Eye Is the Perfect Organ for Gene Therapy

Of all the organs in the human body, the eye may be the single best target for gene therapy and gene editing. This is not a coincidence — it is a convergence of biological properties that make the eye uniquely suited to these interventions, and it is why ophthalmology has led the field in clinical translation.

First, the eye is immune-privileged. The interior of the eye — particularly the subretinal space between the retinal pigment epithelium (RPE) and the photoreceptors — is shielded from the full force of the systemic immune system. This phenomenon, known as ocular immune privilege, is maintained by the blood-retinal barrier, the absence of lymphatic drainage, and the active secretion of immunosuppressive factors like transforming growth factor beta (TGF-beta) and alpha-melanocyte-stimulating hormone. In practical terms, this means that viral vectors and gene editing reagents delivered to the eye are less likely to provoke the kind of inflammatory immune response that has derailed gene therapies targeting other organs, particularly the liver.

Second, the eye is small. The human vitreous cavity holds approximately 4 milliliters of fluid, and the subretinal space much less. This means that therapeutic doses can be delivered in tiny volumes — typically 0.1 to 0.3 milliliters for a subretinal injection. Smaller volumes mean lower total doses of vector or editing reagent, which reduces manufacturing costs, limits systemic exposure, and improves the therapeutic index. A dose of adeno-associated virus (AAV) that would be insufficient for a systemic liver-targeted therapy can be highly effective in the eye.

Third, the eye is surgically accessible. Retinal surgeons routinely perform vitrectomies and subretinal injections, delivering therapeutics directly to the target tissue with remarkable precision. The procedures are well-established, carry manageable risks, and can be performed under local anesthesia in an outpatient setting. Unlike gene therapies that require bone marrow transplantation or systemic infusion, ocular gene therapy is a local procedure.

Fourth, the eye offers built-in outcome measures. Visual acuity, visual fields, optical coherence tomography (OCT), and electroretinography (ERG) provide objective, quantitative ways to measure whether a therapy is working. Clinicians can literally see the retina and track structural and functional changes over time. Few other organs offer this level of real-time monitoring.

Finally, many inherited retinal diseases (IRDs) are caused by mutations in single genes expressed predominantly or exclusively in the retina. There are over 270 genes known to cause IRDs, affecting roughly 1 in 3,000 people worldwide. Many of these conditions are progressive and lead to severe vision loss or complete blindness, with no approved treatments prior to the gene therapy era. The combination of unmet medical need, well-characterized genetics, and a favorable target organ made the eye the natural proving ground for gene therapy.

Luxturna: The Treatment That Started It All

The modern era of ocular gene therapy began with Luxturna (voretigene neparvovec-rzyl), developed by Spark Therapeutics and approved by the FDA in December 2017. Luxturna was the first gene therapy approved in the United States for a genetic disease, and it remains a landmark in the field.

Luxturna treats biallelic RPE65 mutation-associated retinal dystrophy, a condition in which both copies of the RPE65 gene carry loss-of-function mutations. RPE65 encodes an enzyme — retinal pigment epithelium-specific 65 kDa protein — that is essential for the visual cycle. Specifically, RPE65 converts all-trans-retinyl esters to 11-cis-retinol, a critical step in regenerating the visual pigment rhodopsin after it is bleached by light. Without functional RPE65, the visual cycle stalls, photoreceptors cannot regenerate their light-sensing pigment, and patients experience progressive vision loss beginning in childhood. Most are legally blind by their twenties or thirties.

How Luxturna Works

Luxturna uses an AAV2 vector to deliver a functional copy of the RPE65 gene directly to the RPE cells via subretinal injection. The vector does not integrate into the genome; instead, the therapeutic gene persists as an episome in the nucleus of the non-dividing RPE cells. Once the functional RPE65 protein is expressed, the visual cycle resumes, and photoreceptors regain the ability to respond to light.

Each eye is treated in a separate procedure, typically two weeks apart. The subretinal injection is performed during a standard vitrectomy, and patients are monitored closely in the weeks following treatment.

Clinical Results

The pivotal Phase 3 trial enrolled 31 patients aged 4 to 44 with confirmed biallelic RPE65 mutations and sufficient viable retinal cells. The primary endpoint was a novel test called the multi-luminance mobility test (MLMT), which measures a patient's ability to navigate an obstacle course at different light levels. Patients treated with Luxturna showed a dramatic improvement in the MLMT score at one year, with a mean change of 1.8 light levels compared to 0.2 in the control group. Many patients who could previously navigate only in bright light were able to navigate in dim or near-dark conditions after treatment.

Longer-term follow-up has shown that the benefits are durable, though not permanent in all patients. Some studies have reported a gradual decline in visual function beginning 3 to 5 years after treatment in a subset of patients, likely because the underlying photoreceptor degeneration continues even after RPE65 function is restored. This observation has fueled interest in combination approaches that address both the metabolic defect and the degenerative process.

Luxturna is priced at $425,000 per eye ($850,000 for bilateral treatment). While expensive, health economic analyses have generally found it to be cost-effective when weighed against a lifetime of blindness-related costs. It is approved for patients of any age with sufficient viable retinal cells, though earlier treatment — before extensive photoreceptor loss — yields better outcomes.

The CEP290 Challenge: Why CRISPR Was Needed

Luxturna demonstrated that gene addition via AAV could work beautifully in the eye. But it also exposed a fundamental limitation: AAV vectors have a packaging capacity of approximately 4.7 kilobases. Any therapeutic gene — including its promoter and regulatory elements — must fit within this constraint. For RPE65, which has a relatively compact coding sequence of about 1.6 kilobases, this was not a problem.

But many genes that cause inherited retinal diseases are far too large for AAV. The most prominent example is CEP290 (centrosomal protein 290), which encodes a 290-kilodalton protein essential for the structure and function of the connecting cilium in photoreceptors. The CEP290 coding sequence alone spans approximately 7.4 kilobases — well beyond the AAV packaging limit even without regulatory elements.

Mutations in CEP290 are the most common cause of Leber congenital amaurosis type 10 (LCA10), one of the most severe forms of inherited blindness. LCA10 typically presents in infancy with severely reduced vision, nystagmus, and an absent or profoundly diminished electroretinogram. The specific mutation responsible for most LCA10 cases is a deep intronic point mutation (c.2991+1655A>G) in intron 26 of CEP290. This mutation creates a cryptic splice donor site that causes inclusion of a pseudoexon containing a premature stop codon, leading to a truncated, non-functional protein.

This single mutation accounts for the majority of LCA10 cases in populations of European descent and represents an estimated 1 to 2 percent of all inherited blindness. But because the full CEP290 gene cannot fit in an AAV vector, traditional gene addition therapy was not feasible. Dual-vector strategies and alternative approaches were explored but came with significant complexity and reduced efficiency.

This is where CRISPR entered the picture. Rather than trying to deliver the entire CEP290 gene, researchers realized they could use CRISPR-Cas9 to simply remove the intronic mutation — excising the region containing the cryptic splice site and restoring normal splicing of the endogenous gene. The target was not a coding sequence but a 280-base-pair segment deep within an intron. By flanking this segment with two guide RNAs and cutting on both sides, the mutation is deleted, the pseudoexon is eliminated, and normal CEP290 protein production resumes from the patient's own gene.

This was a conceptual breakthrough: CRISPR could treat diseases that AAV gene addition could not, by editing the gene in place rather than replacing it entirely.

EDIT-101 and the BRILLIANCE Trial: CRISPR Inside the Eye

Editas Medicine, co-founded by CRISPR pioneer Feng Zhang, developed EDIT-101 — the first CRISPR medicine to be administered directly into the human body (in vivo). EDIT-101 consists of an AAV5 vector carrying the Staphylococcus aureus Cas9 (SaCas9) gene and two guide RNAs targeting the flanking regions of the CEP290 intronic mutation. SaCas9 was chosen because it is smaller than the more commonly used Streptococcus pyogenes Cas9 (SpCas9), allowing it to fit within the AAV packaging constraint along with the guide RNAs and necessary regulatory elements.

The therapy is delivered via a single subretinal injection during a standard vitrectomy. Once the AAV5 transduces the photoreceptor cells, SaCas9 and the guide RNAs are expressed, the flanking cuts are made, and the intervening sequence containing the mutation is excised. The cell's own DNA repair machinery joins the cut ends, and the corrected gene resumes normal splicing and protein production.

The BRILLIANCE Trial

The BRILLIANCE trial (NCT03872479) was a Phase 1/2, open-label, dose-escalation study that began enrolling patients in 2020. It was a landmark — the first clinical trial to test in vivo CRISPR gene editing in humans. The trial enrolled both adults and children with LCA10 caused by the CEP290 intronic mutation.

The results, presented at multiple medical conferences and published in peer-reviewed form, showed that EDIT-101 was generally safe and well-tolerated. There were no serious treatment-related adverse events and no evidence of significant off-target editing. The most common adverse events were mild and related to the surgical procedure itself.

In terms of efficacy, the results were meaningful if modest. Of 14 treated patients, 11 showed measurable improvement on at least one clinically relevant endpoint. Some patients demonstrated improvements in light sensitivity, as measured by full-field stimulus testing (FST), the most sensitive measure of photoreceptor function in severely affected patients. Several patients reported subjective improvements in vision, including the ability to perceive light, distinguish shapes, or navigate more confidently in dim environments.

The improvements were most notable in the higher-dose cohorts and in younger patients, consistent with the expectation that earlier intervention — before photoreceptor degeneration becomes too advanced — would yield better results. However, the magnitude of improvement was generally smaller than what was seen with Luxturna in RPE65 patients. This is not entirely surprising: LCA10 patients in the trial had more advanced disease, and the editing efficiency achieved by a single subretinal injection of AAV-delivered CRISPR may not have been sufficient to restore CEP290 protein in a large enough proportion of photoreceptors to produce dramatic visual gains.

Despite the modest efficacy, BRILLIANCE established several critical principles. It demonstrated that in vivo CRISPR gene editing is feasible and safe in humans. It showed that AAV-delivered Cas9 and guide RNAs can produce targeted genomic edits in photoreceptor cells. And it confirmed that editing a deep intronic mutation — a strategy with no equivalent in traditional gene addition therapy — could restore some degree of protein function in patients.

Editas Medicine ultimately decided not to advance EDIT-101 further as a standalone program, citing the need for improved editing efficiency and delivery. But the BRILLIANCE trial opened the door for the next generation of ocular CRISPR therapies.

Retinitis Pigmentosa: MCO-010 and Optogenetic Gene Therapy

While CRISPR targets the genetic root cause of retinal disease, another approach takes a fundamentally different strategy: optogenetics. Retinitis pigmentosa (RP) is a group of inherited retinal diseases affecting roughly 1 in 4,000 people worldwide. RP is genetically heterogeneous — over 80 genes can cause it — and is characterized by progressive degeneration of rod photoreceptors followed by cone loss, leading to tunnel vision and eventually blindness.

For patients who have already lost the majority of their photoreceptors, gene replacement for the causative mutation is not useful — there are too few target cells remaining. Nanoscope Therapeutics took a gene-agnostic approach with MCO-010 (sonpiretigene isteparvovec), an intravitreal AAV2-delivered therapy that introduces a multi-characteristic opsin (MCO) gene into the surviving retinal ganglion cells or bipolar cells. These cells do not normally respond to light, but the MCO protein — an ambient-light-activatable channelrhodopsin — converts them into artificial photoreceptors.

In the Phase 2b RESTORE trial, 30 patients with advanced RP received either MCO-010 or sham injection. The treatment group showed statistically significant improvements in visual function, including gains in best-corrected visual acuity and functional vision as measured by the multi-luminance mobility test. The results were encouraging enough that Nanoscope initiated a Phase 3 trial — one of the most advanced optogenetic programs in ophthalmology.

MCO-010 is not CRISPR-based, but it represents an important complementary strategy. Where CRISPR and gene addition aim to prevent or reverse photoreceptor dysfunction, optogenetics provides a rescue pathway for patients who are already past the point of photoreceptor rescue. The combination of preventive gene editing for early-stage disease and optogenetic restoration for advanced disease could eventually cover the full spectrum of RP patients.

LHON: Mitochondrial Disease Meets Gene Therapy

Leber hereditary optic neuropathy (LHON) is a maternally inherited mitochondrial disease that causes sudden, painless vision loss in young adults, predominantly males. It is caused by mutations in mitochondrial DNA — most commonly m.11778G>A in the ND4 gene — that impair Complex I of the mitochondrial electron transport chain, leading to selective death of retinal ganglion cells and optic nerve atrophy.

LHON presents a unique challenge for gene therapy because the causative gene resides in the mitochondrial genome, not the nuclear genome. Standard AAV gene therapy delivers DNA to the nucleus. To circumvent this, GenSight Biologics developed Lumevoq (lenadogene nolparvovec), which uses an AAV2 vector to deliver a nuclear-encoded, mitochondrially-targeted version of the ND4 gene. The gene is engineered with a mitochondrial targeting sequence so that after translation in the cytoplasm, the protein is imported into mitochondria, where it integrates into Complex I and restores electron transport.

Lumevoq is injected intravitreally — a simpler procedure than subretinal injection — and targets retinal ganglion cells, which are the affected cell type in LHON. In the REFLECT Phase 3 trial, patients with ND4-LHON received Lumevoq in one eye and sham injection in the other. The treated eyes showed significant improvement in best-corrected visual acuity at 1.5 years compared to baseline. Unexpectedly, the sham-treated contralateral eyes also improved substantially, likely due to transfer of the therapeutic vector from the treated eye via the optic chiasm — a phenomenon called the contralateral effect.

Lumevoq received marketing authorization in the European Union in 2024 but has faced a more complex regulatory path in the United States. The FDA issued a complete response letter, requesting additional data. GenSight has continued to work toward U.S. approval, and the LHON gene therapy story highlights both the promise and the regulatory complexity of treating mitochondrial diseases.

Age-Related Macular Degeneration: The Biggest Prize

Inherited retinal diseases collectively affect roughly 2 million people worldwide. Age-related macular degeneration (AMD) affects over 200 million. AMD is the leading cause of irreversible vision loss in people over 50 in the developed world, and the wet (neovascular) form — in which abnormal blood vessels grow beneath the retina, leaking fluid and blood — is responsible for the most severe and rapid vision loss.

The current standard of care for wet AMD is repeated intravitreal injections of anti-VEGF (vascular endothelial growth factor) drugs: ranibizumab (Lucentis), aflibercept (Eylea), and the newer faricimab (Vabysmo). These injections must be given every 1 to 3 months indefinitely, imposing an enormous burden on patients and healthcare systems. Many patients receive fewer injections than optimal due to access barriers, leading to preventable vision loss.

Gene therapy for AMD aims to turn the eye into its own anti-VEGF factory, producing therapeutic protein continuously from a single treatment. Several programs are in clinical development:

Ixoberogene soroparvovec (Adverum Biotechnologies) uses an AAV.7m8 vector delivered intravitreally to express aflibercept. The LUNA Phase 2 trial showed sustained aflibercept protein levels in the aqueous humor and a significant reduction in injection burden, though some cases of inflammation required careful dose optimization.

4D-150 (4D Molecular Therapeutics) uses an engineered intravitreal AAV vector (4D-R100) to co-express aflibercept and an anti-VEGF-C microRNA, targeting both VEGF-A and VEGF-C pathways simultaneously. The Phase 2 PRISM trial reported promising results, with a majority of patients requiring no supplemental anti-VEGF injections through one year.

RGX-314 (REGENXBIO) uses an AAV8 vector to express ranibizumab. It is being tested via both subretinal injection (ATMOSPHERE trial) and a novel suprachoroidal injection route using the Clearside Biomedical SCS Microinjector, which delivers the vector to the suprachoroidal space — the potential space between the sclera and the choroid. Suprachoroidal delivery is an office procedure that avoids vitrectomy, potentially making treatment more accessible.

These are gene addition therapies, not gene editing. But CRISPR approaches to AMD are also being explored. Researchers have investigated using CRISPR to knock out VEGF-A or its receptor directly in the retinal pigment epithelium, and preclinical studies in mouse models of choroidal neovascularization have shown promising results. The challenge is achieving sufficient editing efficiency in the relevant cell population and ensuring long-term safety. Given that AMD is a common disease of aging — not a rare monogenic disorder — the regulatory bar for safety is correspondingly higher.

CRISPR Advantages Over AAV Gene Addition for Eye Diseases

The BRILLIANCE trial and subsequent research have clarified several advantages that CRISPR-based editing holds over traditional AAV gene addition for certain retinal diseases:

Overcoming the packaging limit. As the CEP290 example demonstrates, CRISPR can treat diseases caused by mutations in genes too large for AAV. Other large retinal disease genes include ABCA4 (Stargardt disease, coding sequence ~6.8 kilobases), USH2A (Usher syndrome type 2A, ~15.6 kilobases), and MYO7A (Usher syndrome type 1B, ~6.6 kilobases). For these diseases, no conventional single-vector AAV gene addition therapy is possible, and dual-vector strategies have struggled with efficiency. CRISPR editing of specific mutations — whether by excision, correction, or base/prime editing — offers a viable path.

Preserving endogenous regulation. When CRISPR corrects a mutation in the patient's own gene, the repaired gene remains under its native regulatory control — the original promoter, enhancers, and splicing signals. This means the protein is expressed at physiological levels, in the right cells, at the right time. AAV gene addition, by contrast, uses an exogenous promoter that may not precisely replicate the native expression pattern, potentially leading to overexpression or expression in the wrong cell types.

Permanent correction in dividing cells. In non-dividing cells like photoreceptors, AAV episomes can persist for years. But in any context where cell division occurs — during retinal development in pediatric patients, or in RPE cells that occasionally divide — episomal AAV DNA is gradually diluted and lost. CRISPR edits, by contrast, are inscribed directly into the genome and are permanent, inherited by all daughter cells.

Allele-specific precision. CRISPR can be designed to target a specific dominant-negative mutation on one allele while leaving the healthy allele intact — something gene addition cannot do. This is particularly relevant for dominant retinal diseases like some forms of retinitis pigmentosa caused by gain-of-function mutations in rhodopsin (RHO).

Expanding the therapeutic toolbox. Beyond simple gene knockout or excision, the CRISPR platform encompasses base editing (which can correct point mutations without double-strand breaks), prime editing (which can make precise insertions, deletions, and substitutions), and epigenetic editing (which can modulate gene expression without altering the DNA sequence). Each of these modalities has potential applications in retinal disease, and several are in preclinical development for ocular indications.

Next-Generation Approaches and the Future Outlook

The first generation of ocular gene editing — exemplified by EDIT-101 — proved the concept but also revealed the limitations: editing efficiency was modest, delivery was limited to the region of retina directly exposed to the subretinal injection, and the AAV-based delivery system constrained the size and design of editing components.

The next generation is addressing each of these limitations:

Improved delivery. Engineered AAV capsids with enhanced retinal tropism after intravitreal injection are in development. Intravitreal delivery is a simple office procedure that can reach the entire retina, unlike subretinal injection, which typically treats only a localized area. Companies including 4D Molecular Therapeutics and Adverum Biotechnologies have developed capsid variants (4D-R100, AAV.7m8) that can transduce photoreceptors and RPE cells from the vitreous side, though achieving sufficient efficiency in the outer retina remains a challenge.

Non-viral delivery. Lipid nanoparticles (LNPs) and virus-like particles (VLPs) are being explored as alternatives to AAV for delivering CRISPR components to the retina. LNPs carrying Cas9 mRNA and guide RNA offer the advantage of transient expression — the editing machinery is present only briefly, reducing the risk of prolonged Cas9 activity and off-target editing. A single transient pulse of Cas9 may be sufficient to achieve therapeutic editing, and the absence of a viral capsid could reduce immunogenicity. Preclinical studies have demonstrated that LNPs can deliver CRISPR reagents to the retina via subretinal injection, though the efficiency and cell-type specificity of LNP-based retinal delivery are still being optimized.

Base and prime editing. For point mutations that cause retinal disease — and the majority of inherited retinal disease mutations are point mutations or small insertions/deletions — base editing and prime editing offer the precision to correct the exact nucleotide change without creating double-strand breaks. This eliminates the risk of unintended insertions, deletions, or chromosomal rearrangements at the cut site. Beam Therapeutics, Prime Medicine, and academic groups are developing base and prime editors for retinal applications, with several programs in preclinical stages.

RNA editing. An entirely different approach targets RNA rather than DNA. ADAR-based RNA editing can correct G-to-A mutations at the transcript level, leaving the genome untouched. Because RNA edits are reversible and do not alter the permanent genetic code, they may offer a favorable safety profile. Shape Therapeutics (acquired by Roche in 2024) has developed engineered ADAR-recruiting guide RNAs for retinal diseases, and the approach is advancing toward clinical testing.

Combination strategies. The future of retinal gene therapy likely involves combining approaches. A patient with early-stage RP caused by a known mutation might receive CRISPR-based correction to halt degeneration in remaining photoreceptors, while a patient with advanced RP and few surviving photoreceptors might receive optogenetic therapy. Patients with AMD might receive a one-time gene therapy that produces anti-VEGF protein continuously, eliminating the need for repeated injections. And for diseases like Stargardt that involve large genes, dual-AAV or CRISPR-based approaches might be used in tandem.

The pace of progress is remarkable. In 2017, Luxturna was the only approved ocular gene therapy. By 2026, the clinical pipeline includes more than 40 active gene therapy and gene editing trials for eye diseases. The eye has served as the test case for the entire gene therapy field — and the lessons learned here are informing the development of therapies for diseases throughout the body.

For the millions of people living with inherited blindness and the hundreds of millions affected by AMD, the question is no longer whether gene editing can treat eye disease. The question is how quickly these treatments can be made safe, effective, and accessible enough to reach everyone who needs them.

---

Sources

1. Russell, S., et al. "Efficacy and safety of voretigene neparvovec (AAV2-hRPE65v2) in patients with RPE65-mediated inherited retinal dystrophy: a randomised, controlled, open-label, phase 3 trial." The Lancet 390.10097 (2017): 849-860.

2. Editas Medicine. "Editas Medicine Announces Clinical Data from Ongoing BRILLIANCE Trial of EDIT-101 for LCA10." Press releases and conference presentations, 2022-2024.

3. Maeder, M.L., et al. "Development of a gene-editing approach to restore vision loss in Leber congenital amaurosis type 10." Nature Medicine 25.2 (2019): 229-233.

4. Pendse, N.D., et al. "In vivo assessment of potential therapeutic approaches for CEP290-associated Leber congenital amaurosis." Investigative Ophthalmology & Visual Science 60.7 (2019).

5. Nanoscope Therapeutics. "Nanoscope Therapeutics Announces Positive Topline Results from Phase 2b RESTORE Trial of MCO-010." Press release, 2023.

6. GenSight Biologics. "Lumevoq (lenadogene nolparvovec) REFLECT Phase 3 Trial Results." The Lancet (2023).

7. Adverum Biotechnologies. "LUNA Phase 2 Trial of Ixoberogene Soroparvovec for Wet AMD." Clinical data presentations, 2023-2025.

8. 4D Molecular Therapeutics. "PRISM Phase 2 Trial of 4D-150 for Wet AMD." Clinical data presentations, 2024-2025.

9. REGENXBIO. "ATMOSPHERE Trial of RGX-314 for Wet AMD." Clinical trial updates, 2023-2025.

10. Cideciyan, A.V., et al. "Effect of an intravitreal antisense oligonucleotide on vision in Leber congenital amaurosis due to a photoreceptor cilium defect." Nature Medicine 25.2 (2019): 225-228.

11. Pierce, E.A., and Bennett, J. "The Status of RPE65 Gene Therapy Trials: Safety and Efficacy." Cold Spring Harbor Perspectives in Medicine 5.9 (2015).

12. Bainbridge, J.W., et al. "Long-term effect of gene therapy on Leber's congenital amaurosis." New England Journal of Medicine 372.20 (2015): 1887-1897.

13. Sahel, J.A., and Roska, B. "Gene therapy for blindness." Annual Review of Neuroscience 36 (2013): 467-488.

14. RetNet: Retinal Information Network. "Genes and Mapped Loci Causing Retinal Diseases." Available at: https://web.sph.uth.edu/RetNet/