There is a gene sitting in your muscle cells that would poison them if it ever switched on. In most people it never does.

The gene is called DUX4. It does its job in early embryonic development, and in healthy adult muscle it is packed away and kept silent. In facioscapulohumeral muscular dystrophy, FSHD for short, that silencing fails. DUX4 switches on in muscle tissue where it has no business being, and the protein it makes is toxic to the very cells producing it. The muscles of the face, shoulder blades and upper arms weaken first, which is what the awkward name is actually describing. There is no approved treatment.

On August 11, Epicrispr Biotechnologies closed a $90 million Series C round co-led by Octagon Capital and Janus Henderson, with Sanofi’s venture arm among the other backers. The money arrives sixteen months after a $68 million Series B, and follows an early look at the company’s Phase 1/2 trial: three patients given a single intravenous infusion of EPI-321 gained between 0.5 and 1.3 pounds of lean muscle mass at six months.

That difference shapes the whole strategy. In most genetic diseases the instruction itself is wrong, so fixing it means rewriting the text. In FSHD the text is fine. What has failed is the machinery that stops it from being read.

So EPI-321 never touches the DNA. It uses a viral vector to carry an epigenetic editor into muscle cells, a protein that finds the DUX4 region and puts back the chemical marks that keep it switched off. If DNA is sheet music, epigenetics is the instruction about whether to play a given passage. Epicrispr leaves every note alone and simply puts the rest symbol back over a bar the orchestra has started playing by mistake.

Conventional gene editing makes a permanent change, and if it lands in the wrong place you get a permanent mistake. Silencing is, in principle, reversible: stop supplying the editor and the marks fade. There is no cut in the DNA, which is where much of gene editing’s off-target risk comes from.

The trade-off is durability. A permanent edit stays fixed by definition. A chemical mark might need topping up, and nobody yet knows how long a single dose of EPI-321 holds. Six months is just how far the data goes right now, so we can’t say if that is close to the limit or far from it.

Three patients is three patients, and half a pound of lean muscle is a small effect in a very small group. But muscle mass actually going up is an unusual direction in muscular dystrophy, where the honest goal is usually to slow the decline. FSHD now has four different mechanisms in clinical trials: Epicrispr’s silencer, the Phase 3 antibody-oligonucleotide conjugate Novartis picked up in its $12 billion purchase of Avidity, an siRNA from Sarepta, and a Scholar Rock candidate in Phase 2. For a disease that until recently had nothing, four different shots at it is the real news.

Sources: Epicrispr Biotechnologies Series C announcement (Businesswire, 10 August 2026); ClinicalTrials.gov NCT06907875.