Born switched off

Myostatin is the body’s governor on skeletal muscle - the signal that holds growth below its genetic ceiling. Knock it out and the ceiling lifts: the famously muscular whippets, cattle, and the rare children born without working myostatin all tell the same story. A signal that powerful is dangerous to leave lying around switched on. So the body does something telling with it.

It ships the brake disabled. Most of the myostatin in your blood is not active protein at all - it circulates as a latent complex, the finished signal folded shut inside a wrapper made from its own front half, inert and unable to touch its receptor. That is why its measured half-life reads as long: what is circulating is the safed, stored form, not the live one. The active molecule is manufactured, then immediately holstered.

Turning latent myostatin into a live signal takes two cuts, by two different enzymes, in a fixed order. Neither one alone is enough - the sequence is the safety catch.

The first cut is made by furin, a proprotein convertase, which snips the chain at a four-residue RXXR motif to separate the front prodomain from the business end, the mature growth factor. But separation is not release: the prodomain stays clamped over the mature dimer by non-covalent grip, shielding it from the receptor. This is the latent complex (genetic analysis of myostatin proteolysis, PLOS One).

The second cut is the one that actually arms it. A protease from the BMP-1/tolloid family cleaves the prodomain itself, at a single site just ahead of aspartate-76. That nick destabilizes the clamp, the wrapper falls away, and the mature myostatin dimer is finally free to bind its receptor and signal (Wolfman et al., PNAS 2003). Same molecule the whole time; what changed was permission.

Here is where the biology becomes a strategy. If the live signal is only the last step of a cascade, then every step before it is a place to intervene - and the earlier you cut in, the more selective you can be. The muscle-preservation programs map exactly onto the states of this pathway:

• Catch it latent - apitegromab. Scholar Rock's antibody (SRK-015) binds the pro and latent forms and blocks that second, arming cut - so myostatin never gets released in the first place. Because only myostatin is stored this way, going after the latent form is the most selective move available.
• Trap it once free - follistatin. The body's own antagonist waits at the next altitude down, binding the mature ligand after release and neutralizing it. Raising follistatin is the endogenous way to mop up whatever does get armed.
• Block the dock - bimagrumab. At the bottom of the pathway sits the receptor itself; an antibody there shuts out myostatin and its relatives together. Broadest effect, lowest selectivity.

The elegance of the latent-form approach is that it exploits a feature the body built for its own reasons. A structural study of apitegromab’s parent antibody showed it works by gripping the prodomain-shrouded precursor and jamming the activation step (SRK-015 structural study, JBC) - a drug designed around a switch that was already there. For the human data on what this does during weight loss, see keeping the muscle on GLP-1.

Myostatin did not invent this trick. Its whole family runs on it: classic TGF-β is itself stored latent, wrapped by its prodomain and tethered in the tissue until a local signal frees it. Storing the finished molecule inert, then arming it on the spot with a proteolytic cut, lets the body pre-position a powerful signal and spend the energy of making it long before it needs the effect (myostatin pathway review, JCI).

It is a recurring theme on this frontier: the body regulates its loudest signals not by how much it makes but by when it lets them speak. Native GLP-1 is controlled by how fast it is destroyed; myostatin, by how deliberately it is switched on. Read the control system and the drug targets fall out of it - which is the whole reason the pro-form, not the active one, became the most selective way in.