Where Two Drug Giants Failed, This Startup Just Won

Frontier Signal · Deep Dive · Biotech & Investing · 7 min read · Friday, 2 October 2026

X-linked retinitis pigmentosa slowly blinds boys and men from a single misspelled gene. For a decade the industry's answer has been to deliver a working copy of that gene into the eye inside a harmless virus. Twice, a major company built that drug, ran a pivotal trial, and missed.

Biogen's version missed in 2021. Johnson and Johnson's version missed in 2025. Both had real biology behind them, real dosing data, and real reasons to expect success. Both walked away.

On 21 September, Beacon Therapeutics, a company of about 80 people, said its own version, laru-zova, hit its primary endpoint in a controlled pivotal trial. The more interesting question is why this one worked when the other two did not.

31% AND 24% RESPONDER RATES · ZERO IN THE UNTREATED ARM · 85-PATIENT PIVOTAL TRIAL
Figures are from Beacon's topline release. They are not peer-reviewed, and the FDA has not yet reviewed them for approval.

Beacon announced topline results from VISTA on 21 September 2026: a global, randomised, controlled Phase 2/3 trial in 85 males aged 12 to 48 with a confirmed RPGR mutation, split across a high dose, a low dose and an untreated arm.

The primary endpoint was low-luminance visual acuity: how many extra letters a patient could read on an eye chart in dim light at 12 months. Beacon has not disclosed how many patients sat in each arm. At the high dose, 31.0 percent gained at least 15 letters. At the low dose, 24.1 percent did. In the untreated arm, none did.

The pattern gets less tidy on secondary measures. On a 10-letter version of the same test, the low dose beat the high dose, 58.6 to 48.3 percent. On microperimetry, a retinal-sensitivity test, only the low dose reached significance. Side effects were mostly mild and procedure-related, but more common at the low dose, which the release does not explain.

Beacon will present the full dataset at the American Academy of Ophthalmology meeting in New Orleans on 10 October and plans a rolling FDA submission before year end.

RPGR is one of the hardest genes to deliver as a therapy because the region where nearly all disease-causing mutations sit, called ORF15, is unusually repetitive DNA, which folds into odd shapes and scrambles itself every time it is copied inside the bacteria used to manufacture a gene therapy.

The usual fix is codon optimisation: rewriting the DNA to make the same protein with more stable three-letter codes that avoid the scrambling. Laru-zova uses a codon-optimised, full-length RPGR gene, delivered by an engineered virus aimed at photoreceptor cells.

In plain English: the gene is respelled more stably, packaged into a virus and injected under the retina, the same basic move Biogen's failed drug made. Codon optimisation did not save that trial, so whatever differs about laru-zova is something else: the virus shell, the dose, or a more forgiving finish line.

Delivery, not the edit itself, is often where gene-editing programs actually stall, a problem that showed up differently when immune cells kept blocking a CRISPR delivery vehicle a screen was built to fix.

Biogen's cotoretigene toliparvovec, gained by buying Nightstar Therapeutics for 800 million dollars in 2019, missed its primary endpoint, a retinal-sensitivity map, in 2021. Johnson and Johnson's botaretigene sparoparvovec, built with MeiraGTx, missed its primary endpoint, a dim-light maze test, in 2025 and was handed back to MeiraGTx for 25 million dollars plus royalties.

Both trials still reported encouraging trends on a secondary measure, extra letters read on an eye chart in low light, the same measure Beacon led with. That is either a sign Beacon learned from two expensive failures and picked the more sensitive endpoint, or a sign it picked the easier one. The public record does not settle which.

Beacon is privately held, so there is no direct stock to buy. Syncona Limited (LSE: SYNC), the UK life-sciences trust that helped build Beacon, owns 38.4 percent, valued at 183 million pounds at the end of June. Its shares rose more than 6 percent on the news, and it said the result could support financing, an IPO or an eventual sale. QuotedData analyst Matthew Read said the result materially reduces the program's risk and moves it closer to approval. Regulatory work remains: de-risked, not proven.

The clearest listed competitor is MeiraGTx Holdings (NASDAQ: MGTX), which reacquired Johnson and Johnson's failed RPGR therapy in April 2026 for 25 million dollars upfront plus royalties and is pursuing its own filing on the same low-light reading signal Beacon just hit. If it succeeds, that strengthens the case for this endpoint. If it fails, endpoint choice, not biology, looks likelier to have decided Beacon's result.

Ocugen (NASDAQ: OCGN) takes a different route: its OCU400 resets a broader network of retinal genes whichever is mutated, with two years of sustained benefit in a small trial now advancing to Phase 3. It is an alternative bet, not a direct RPGR competitor.

Beacon is not the only small biotech betting it can out-engineer a well-funded rival's approach: we found a similar dynamic in the attempt to make the body manufacture its own $400,000 cancer therapy.

Atsena Therapeutics and Opus Genetics build AAV gene therapies for other inherited retinal diseases, but neither has a disclosed RPGR program, so they are peers rather than rivals. Nanoscope Therapeutics pursues a light-sensitising approach to retinitis pigmentosa that works whatever the mutation, a different technology from gene replacement.

This is company-announced topline data. It is not peer-reviewed, and the FDA has not reviewed it for approval. Beacon has not disclosed how many patients sat in each of the three arms of an 85-person study meant to support a license filing, a detail regulators will scrutinise.

The dose-response pattern does not fully make sense on the public numbers: the high dose beat the low dose on the primary endpoint, but the low dose beat the high dose on a related secondary measure and had more side effects. Beacon's release addresses neither.

Public interpretation so far is dominated by Beacon, trial investigators and interested parties, not independent ophthalmologists. Whether laru-zova is genuinely more effective than the earlier programs, or its endpoint was simply more sensitive to benefit, is something the topline release cannot settle.

If Beacon had run its trial against the same kinds of primary endpoints used in the earlier Biogen and Johnson and Johnson programs, would it still have met its primary goal?

That matters because the earlier programs used different primary measures yet still showed encouraging low-light reading signals on secondary endpoints. Beacon's result may reflect a better therapy, a better endpoint, or both; the topline release cannot separate them.

On 21 September 2026 it reported that laru-zova met its primary goal in an 85-patient pivotal trial for X-linked retinitis pigmentosa: 31 percent of high-dose and 24 percent of low-dose patients gained meaningful dim-light vision, against none untreated.

Biogen's and Johnson and Johnson's gene therapies both missed their primary endpoints, in 2021 and 2025, though both showed encouraging trends on the low-light reading test Beacon chose as its main measure.

Not necessarily. All three programs used broadly similar AAV gene-replacement strategies, so Beacon's result could reflect a better therapy, a more sensitive endpoint, or both.

No. Most treated patients did not reach the trial's improvement threshold. This is a partial gain in a subset of patients, from a small, company-reported trial regulators have not yet reviewed.

Beacon is private. UK-listed Syncona owns roughly 38 percent and is the closest public read-through, though it holds many other investments.

Trial figures are company-reported topline data, not peer-reviewed. Earlier-trial history comes from the companies' own announcements and independent recaps.

Frontier Signal explains frontier technology in plain English. Company-reported and vendor-reported claims should be independently verified. This is general information, not investment advice.