The KOSDAQ-listed HLB Therapeutics said its experimental eye drop RGN-259 did not beat placebo at healing the cornea in a European Phase 3 trial, a miss the company blamed on a control arm that recovered better than anyone expected.
The drug is a thymosin beta-4 ↗ peptide, a 43-amino-acid piece of a natural repair protein that helps skin and surface cells crawl into a wound and rebuild it. On paper that is a clean fit for the disease it was tested against. Neurotrophic keratitis is a rare condition where damage to the trigeminal nerve robs the cornea of sensation, and a cornea that cannot feel stops healing itself. Small scrapes turn into persistent open sores, and in the worst cases the eye perforates and sight is lost. A peptide that pushes corneal cells to close the wound is exactly the kind of thing you would want.
It did not separate from placebo. In the trial, called SEER-3, patients put either a 0.1 percent RGN-259 solution or a placebo drop in the eye five times a day for four weeks. The primary endpoint was complete corneal healing at four weeks, scored by staining the eye with fluorescein dye and checking whether the defect had closed. As Ophthalmology Times reported ↗, the drug arm did not show a statistically significant advantage over the placebo arm across the roughly 70 patients treated at sites in Spain, Italy, Poland, and Germany. The company's explanation was that the placebo group healed better than the trial was designed to expect, which left no room for the drug to prove a difference.
That is a real and recurring problem in corneal trials, not a dodge. Plain saline and the careful eye care that comes with being in a study can let some defects close on their own, and the smaller the trial, the easier it is for an unusually good placebo run to swamp the signal. It is also the kind of result that is hard to argue with after the fact. The endpoint was missed, whatever the reason.
RGN-259 was meant to be the more convenient option in a category that has exactly one approved drug. That drug is Italy's Dompe and its Oxervate (cenegermin), a lab-made copy of human nerve growth factor that has to be kept cold and dosed every two hours for eight weeks. RGN-259, licensed from RegeneRx Biopharmaceuticals, was pitched as a room-temperature drop given five times a day for four weeks at a lower price. Convenience does not matter if the efficacy box stays unchecked, and for now it does.
There is a second reason this result is worth sitting with. Thymosin beta-4 is one of the most heavily sold ingredients in the unregulated injectable-peptide market, where it goes by the name TB-500 and is marketed for healing tendons, muscles, and skin with confident claims and almost no controlled human data. We wrote in June about how the field of injectable repair peptides has drifted into a regulatory gray zone ↗ on exactly that kind of mechanistic plausibility. SEER-3 is one of the few places the same molecule was put through a controlled test with a hard endpoint, and on the cornea it could not separate from a drop of placebo. That does not condemn every use of the peptide, but it is a useful data point for anyone buying the gray-market version on the strength of its repair story.
HLB is not done. A second, identical Phase 3 in the United States, SEER-2, is still running, with topline results expected in the second half of 2026. Two matched trials were started together to compress the timeline, and the plan now is to lean on the US readout. If SEER-2 also leans on a strong placebo arm, the thymosin beta-4 healing story runs out of trials to prove itself in.
The peptide lives on our platform as a tissue-repair ↗ candidate, an entry that sequesters actin to help cells move. The biology is genuine. SEER-3 is a reminder that genuine biology and a drug that beats placebo are two different things.