Alltrna is entering the first human trial of an engineered tRNA therapy — and it almost didn't make it.
The Flagship Pioneering-backed company announced on Monday that it will file an IND this year for AP003, an engineered transfer RNA designed to read through stop codons in genetic disease. If the FDA clears it, 2026 becomes the first year any engineered tRNA is tested in humans. That is a genuine milestone for a field that has been working toward this moment for more than a decade.
The CEO who got it there is gone.
Michelle Werner stepped down as CEO of Alltrna in early March 2026, in the middle of the company's third layoff round — a restructuring that cut 19 people, roughly 34 percent of its staff, according to a WARN notice filing reviewed by Fierce Biotech. She left before the IND filing she spent two years shepherding. The company preserved enough cash to get here. Whether she watches from the outside or someone else files the application is not the story. The story is that it is being filed at all.
What AP003 is designed to do sounds like science fiction. When a mutation sends a "stop" signal where there shouldn't be one, the protein assembly line crashes mid-build. The result is a truncated, useless protein — or none at all. Stop-codon diseases account for roughly 11 percent of all genetic diseases, according to Nature Biotechnology. The Arg-TGA premature termination codon is the most common nonsense mutation — representing 21 to 22 percent of known cases within that class, per Alltrna's own data. Thousands of diseases fall into this category: Duchenne muscular dystrophy, cystic fibrosis, certain forms of hereditary blindness, methylmalonic acidemia, phenylketonuria. Alltrna's bet is that a single chemically modified tRNA can override that stop signal and let the protein finish being built. One drug, many diseases — a platform play dressed up as a single asset.
The preclinical data is what gives the platform argument any credibility. In a mouse model of methylmalonic acidemia, a single dose of AP003 restored protein production to roughly 25 percent of wild-type levels by day 4, well above the 1 to 2 percent rescue threshold the field generally considers clinically meaningful. In a separate mouse model of phenylketonuria, the same candidate restored the PAH enzyme to 7 percent of normal levels within 72 hours and drove phenylalanine down by 76 percent from baseline. Those are real numbers. They are also mouse numbers, which biotech investors have learned to discount heavily.
The mechanism is not novel in principle. Tevard Biosciences has been working with Vertex Pharmaceuticals on a tRNA approach to DMD. HC Bioscience, a tRNA biotech backed by ARCH Venture Partners and Takeda, shut down in March 2025 after its hemophilia A tRNA program failed to meet targets, citing delivery challenges — a useful reminder that the space has claimed real casualties. What makes Alltrna's approach technically distinct is the specific chemical engineering of the tRNA molecule itself and the use of lipid nanoparticles to deliver it to liver tissue. The company is not proposing to edit genes. It is proposing to read through them.
The CEO situation adds a human dimension that is difficult to separate from the science. Werner's son was diagnosed with Duchenne muscular dystrophy in 2020, a fact she disclosed publicly, describing her "deeply personal" motivation for moving into rare disease biotech. DMD is caused by a different kind of mutation than the one Alltrna's current candidate targets, and the company's platform is not presently aimed at DMD. The personal stake is real. Whether it distorts clinical priorities or drives good science — or both — is worth watching as the trial moves forward.
The IND filing is the thing that matters. Everything else is context.
