A September 2026 study grew dramatically bigger muscles in diseased mice — and the muscles didn't get any stronger. The field's own takeaway: judge muscle drugs by contractile performance, not bulk.
Blocking myostatin pharmacologically rebuilt muscle mass in a disease model — and the rebuilt muscle didn't get stronger. The contractile defect was intrinsic to the disease, not a mass problem.
The model: mice with dynamin-2 (DNM2)-related centronuclear myopathy, a genetic disease where muscle fibers waste and weaken. Dynamin 2 organizes the transverse tubules — the membrane invaginations that carry electrical signals deep into the fiber. When that machinery is broken, the excitation–contraction coupling is broken with it. No amount of fiber bulk repairs a broken signal line.
The pharmacologic arm: four weeks of soluble ActRIIB-Fc (sActRIIB-Fc), a decoy receptor that traps circulating myostatin and activin A before they can bind their cell-surface receptor. Result: muscle mass normalized to — and in some muscles exceeded — wild-type levels. Maximal force and specific force (force per unit of muscle): zero improvement. The underlying pathology — protein-degradation signaling (E3 ligases), autophagic flux, and excitation–contraction genes like RyR1 and Casq1 — was untouched. And when the drug stopped, all of the gained mass was gone within five weeks.
The genetic arm: mice with lifelong myostatin knockout told a different story. Mass and strength were restored to wild-type — grip strength and rotarod performance included. The mechanism was developmental reprogramming: more satellite cells, better autophagy, corrected protein homeostasis. Permanent myostatin inactivation reshapes how the muscle is built; transient blockade just inflates what's there.
The comparison between the two arms is inherently unfair: a lifetime of genetic reprogramming vs. one month of a circulating decoy. The paper shows that timing and duration matter enormously — it does not show that no drug could ever work.
This is the limitation the paper's harshest critics would name first, and it's worth naming yourself: the genetic knockout had the animal's entire development to rebuild the muscle correctly. The drug got four weeks in an already-diseased adult animal. Of course the outcomes differ.
But the unfairness is also the point. Every myostatin drug a human could ever take is the short-arm version: transient, adult, pharmacologic. Nobody gets lifelong genetic reprogramming. So the fair question isn't "could a drug match the knockout" — it's whether any achievable version of this biology produces muscle that performs. On this paper's evidence, four weeks of broad ligand trapping produced muscle that looked the part and couldn't do the job.
"Myostatin inhibition builds useless muscle" is not what this paper proves, and the human record contradicts it as a blanket claim. Here's the class as it actually stands.
The functional win: apitegromab (Isembyld) — a selective anti-myostatin antibody — became the first FDA-approved myostatin inhibitor in September 2026 (spinal muscular atrophy, ages 2+). Its Phase 3 SAPPHIRE trial (n=188) improved a functional endpoint: Hammersmith Functional Motor Scale–Expanded scores rose +2.2 points vs placebo. Mass with function, in humans, in a registrational trial. That single data point kills the lazy generalization.
The counterweight on the same drug: the FDA label carries a bone-fragility signal — fractures in 9% of the 10 mg/kg arm vs 2% on placebo — plus fetal-harm and reproductive-function warnings. The class is not "clean." Nobody gets to cherry-pick the efficacy and ignore the label.
The graveyard: Acceleron's ACE-083 grew muscle locally without improving function — the program was discontinued. That remains the class anchor for this exact caution: the regulators and sponsors have already lived the mass-≠-function verdict once.
The obesity split (the GLP-1 lean-mass lane): trevogrumab (Regeneron, anti-myostatin) + semaglutide halved lean-mass loss vs semaglutide alone in the Phase 2 COURAGE trial, with the 75 mg dose preserving >70% of thigh-muscle loss on MRI at 52 weeks — the strongest positive read in the space. Apitegromab's obesity Phase 2 showed +55% lean-mass retention vs tirzepatide alone. On the other side: Roche's emugrobart failed a Phase 2 obesity interim analysis and was handed back to Chugai. The hit rate is genuinely mixed — and every one of these trials reported body composition, not strength. The mass-vs-function question this paper raises is precisely the endpoint the human obesity trials haven't measured.
The pipeline watch: iBio's IBIO-610 — a first-in-class Activin E antibody — reported 8-week monkey data where lean mass was only 6% of weight lost vs 22% on semaglutide alone. Activin E is a different ligand node from myostatin/GDF8, which is exactly the direction the JCSM authors pointed. Evidence level: preclinical, company-reported. Watch pile, not action pile — first-in-human expected H1 2027.
No human trial has tested whether a myostatin-class drug makes a healthy lifter stronger. DXA gains are not performance gains — and this paper is the cleanest demonstration of why that distinction matters.
Here's the uncomfortable version for the enhancement-minded: the physique world treats "lean mass gained" as the scoreboard. This paper — and ACE-083's corpse — say the scoreboard is wrong. If a future myostatin drug reaches healthy users, the question won't be "how many pounds of lean mass did the DXA show." It will be whether those pounds contract, produce force, and stick around after the drug stops. On all three, the best available evidence is either negative (this paper), mixed (the obesity trials, which didn't measure force at all), or positive-but-diseased (SAPPHIRE, in SMA patients).
The paper's own conclusion is the standard to hold: judge future therapies by durable contractile performance, not bulk. Apply that standard to every myostatin headline from here on — including the exciting ones.
The bottom line: bigger is not better if it doesn't work. A 2026 mouse study made the distinction brutally clear — mass restored, force untouched, gains gone in five weeks off-drug — while the human class record says the story isn't that simple: apitegromab improved real function in SMA, ACE-083 died on this exact hill, and the obesity trials still haven't measured strength at all. The field's new rule is the right one: don't count the muscle. Test it.