Bigger, but not stronger

Most drugs that fail do nothing. The muscle drugs built on releasing the body’s brake failed a stranger way: they worked. You could see the muscle grow on the scan. The trouble was that the person carrying it was no stronger than before. For roughly fifteen years that gap — between a muscle you can measure and a muscle you can use — was the defining problem of the whole field, and two molecules from one company, ACE-083 and ACE-031, are the cleanest way to understand it.

Both aimed at the same elegant target this site keeps returning to: the myostatin and activin pathway that actively restrains how much muscle you keep (the mechanics are in Born switched off). Release that brake and muscle grows. Both drugs proved it did. Neither turned that growth into function — and the reason they couldn’t is the most useful thing on this page.

ACE-083 was the purest possible test of the idea. It was built from follistatin, the body’s own myostatin antagonist, and engineered to stay where it was put — injected directly into a target muscle so the effect would be local and the rest of the body left alone. Acceleron ran it in facioscapulohumeral muscular dystrophy (FSHD) and Charcot-Marie-Tooth disease, both conditions where specific muscles waste.

On its own terms, it succeeded. In the phase 2 FSHD trial the injected muscles grew by double digits — a robust, statistically significant increase in total muscle volume on MRI (Statland et al., 2022). Then the secondary endpoints came in. The functional tests — the measures of whether patients could actually do more — showed no statistically significant improvement. Bigger muscle, unchanged function. Acceleron stopped the FSHD program.

It is hard to overstate how counterintuitive that result felt at the time. The intuition everyone brings to muscle — bigger is stronger — is right for a healthy body under load. It is not a law of biology. ACE-083 had separated the two variables cleanly, and shown that you can move one without the other.

Its sibling failed from the opposite direction. ACE-031 (ramatercept) was a systemic drug: a soluble decoy receptor, the business end of the activin type II receptor fused to an antibody stalk, floating in the blood to soak up myostatin before it could reach muscle anywhere in the body. Acceleron took it into Duchenne muscular dystrophy, and the early signs were encouraging — trends toward more lean mass, denser bone, less fat, even a hint of held walking distance (Campbell et al., 2017).

Then boys started getting nosebleeds and bleeding from the gums, and small dilated blood vessels — telangiectasias — appeared on the skin. The trial was halted after the second dosing regimen. The cause was baked into the design: that decoy receptor is not selective. It binds myostatin, but it also traps BMP-9 and BMP-10, two related signals the body uses to keep blood vessels sealed and stable. Cut the muscle brake with a wide enough blade and you cut the vascular wiring alongside it.

Here is the turn that makes this a story about science working rather than failing. Acceleron’s ligand-trap platform — the same molecular idea behind ACE-031 — did not die. It was pointed at the parts of the biology that the muscle programs had treated as side effects, and it produced two approved drugs.

Trapping this family of signals nudges red-blood-cell production, so a close relative of ACE-031 became luspatercept (Reblozyl), approved in 2019 for the anemias of myelodysplastic syndrome and beta-thalassemia. Trapping a slightly different member calms the overgrowth of small blood vessels in the lung, so another became sotatercept (Winrevair), approved in 2024 for pulmonary arterial hypertension. The molecules that reached the market were the non-muscle ones. The “off-target” biology that sank the muscle trials was, somewhere else in the body, the entire point.

So why doesn’t added muscle simply work? The answer is the load- bearing idea of this piece: releasing the brake tells a muscle fiber to grow, but a fiber only does useful work when the rest of the system is intact — the motor nerve wired to it, the fiber’s own internal machinery sound, real mechanical demand asking it to contract. Growth supplies mass. It does not supply any of those.

That reframes both failures at once. In FSHD, the muscle ACE-083 inflated was still genetically diseased on the inside; a larger volume of compromised tissue is still compromised. In a body under no particular training load, extra mass has no reason to become extra strength. The drugs weren’t weak. They were answering a question — how do I make this muscle bigger? — that turned out not to be the question that mattered. The one that mattered was: is this muscle able to do work in the first place?

In 2026 a drug on this exact pathway cleared the bar that had stood for fifteen years. Apitegromab, from Scholar Rock, is the sharpened version of the idea: instead of a wide decoy, it binds only the inactive precursor of myostatin, disarming the brake before it is ever switched on, and touching little else (why that selectivity is possible is the subject of One receptor, a whole family). In the phase 3 SAPPHIRE trial in spinal muscular atrophy, it did what ACE-083 could not: it improved a functional motor score, not just a scan (Scholar Rock, 2024/25).

The reason it worked is the whole thesis in one experiment. Apitegromab was added on top of the SMA drugs that repair the underlying fault — the SMN-restoring therapies that get the motor neuron talking to the muscle again. Fix the wiring, then release the brake, and the extra muscle has something to connect to and something to do. The mass finally lands on a system able to use it. Same pathway that failed as ACE-083; this time pointed at a muscle that could answer.

The regulatory path has been its own small drama, and worth stating plainly rather than dressing up: the FDA issued a complete response letter in September 2025 (Scholar Rock, 2025) — but the letter was about a third-party fill-finish facility, with no question raised about the drug’s safety or how well it worked. After a resubmission, the decision date now sits at September 30, 2026. Manufacturing, not mechanism, is the last gate.

This history is not a museum piece. The hottest use of muscle-brake drugs right now is preserving muscle during GLP-1 weight loss — pairing them with semaglutide or tirzepatide so the scale drops fat and spares lean mass. The early readouts are genuinely exciting, and this site has covered them in Muscle loss on GLP-1. But nearly every headline number so far is a body-composition figure — kilograms of lean mass on a DEXA scan. That is a size readout. It is exactly the measurement ACE-083 aced on its way to being shelved.

Here is the honest, and quietly optimistic, distinction. A person losing weight is not a person with dystrophic muscle. Their fibers are healthy, wired, and — if they are moving at all — under load. That is precisely the intact system in which preserved mass has every reason to become preserved strength. The biology is finally lined up in the drug’s favor. What remains is to prove it, by wiring strength, mobility and physical function into the trials rather than trusting the scan to speak for them. Apitegromab showed the pathway can deliver function when the setup is right. The task now is to ask it to — and to keep the difference between bigger and stronger in view every time a new lean-mass number lands.