Held up by the hands, most five-month-old babies just dangle. One in Michigan pulled himself sideways into an iron cross, the ring position competitive gymnasts spend years failing to hold.

His mother told the Associated Press that this was roughly the moment the family worked out something unusual was going on. By eight months he could do a pull-up. By nine he was managing stairs. Specialists eventually landed on a diagnosis, as reported by ABC News: myostatin-related muscle hypertrophy. Tabloids preferred “the Hercules gene”, which is admittedly better copy.

Muscle biology found the brake instead of the pedal

For decades the hunt was on for whatever it is that makes muscle grow. In 1997 a team at Johns Hopkins found the thing that stops it instead.

Alexandra McPherron, Ann Lawler and Se-Jin Lee, writing in Nature, described a previously unknown protein they called GDF-8, produced almost exclusively in skeletal muscle. Then they disabled the gene for it in mice. Individual muscles in the resulting animals weighed two to three times what they should have, and the mice were otherwise unremarkable.

GDF-8 was later renamed myostatin, and its job was restraint. Nearly every vertebrate carries a version of it, quietly telling muscle cells to stop well short of what they could manage.

“It’s one of the reasons why I don’t look like Arnold Schwarzenegger,” Dominic Wells of Imperial College London told ABC News.

The boy in Berlin who could hold dumbbells out sideways

Could the same defect exist in a person rather than a mouse? Seven years after the Nature paper, Markus Schuelke and his team at the Charité in Berlin found out, publishing their findings in the New England Journal of Medicine. The boy they described was born to a healthy woman who had been a professional athlete. Ultrasound at six days old showed close to double the expected muscle mass. At four and a half he could hold two three-kilogram dumbbells out horizontally with his arms extended, which is not a thing four-year-olds do.

Family history is the part that sticks. Schuelke’s team noted relatives reputed to be freakishly strong, including one who worked construction and shifted curbstones by hand.

Underneath all of it sat a single letter change at a splice site, enough to stop functional myostatin being made at all.

One copy fast, two copies cramping

Whippet breeders had spotted their own version of this long before anyone sequenced it. Litters occasionally threw up a puppy built like a bodybuilder, known in the sport as a “bully” whippet. Dana Mosher, Elaine Ostrander and colleagues at the National Human Genome Research Institute traced the trait to a two-base deletion in the canine gene, writing in PLOS Genetics.

Dogs carrying two copies were grossly overmuscled, prone to cramping in the shoulders and thighs, and rarely competitive. Dogs carrying one copy were more muscular than average and significantly faster on the track. Per the NHGRI announcement, that made it the first myostatin mutation quantitatively linked to athletic performance.

Same gene, doing the same thing in mice, cattle, sheep, racing dogs and one extraordinarily strong boy in Germany.

Human cases stay vanishingly rare. MedlinePlus Genetics lists the prevalence as unknown, notes no medical problems known to follow from it, and describes an inheritance pattern where two altered copies produce dramatic muscle and one produces something milder. Lee told ABC News in 2009 that the 2004 case was the only clear-cut documented human mutation he was aware of.

Twenty years of drugs that did not work

So why, after two decades of trying, is there still no myostatin drug sitting on a pharmacy shelf?

First serious attempt was MYO-029, an antibody built to neutralise myostatin and tested in adults with three forms of muscular dystrophy. Kathryn Wagner and her co-authors reported in Annals of Neurology in 2008 that it was reasonably safe, aside from skin reactions at higher doses, and produced no improvement in strength or function at any dose. Later candidates in muscle-wasting disease went much the same way, one after another, for the better part of two decades.

Myostatin blockers find a use in weight loss

Then GLP-1 drugs happened.

People shedding serious weight on semaglutide or tirzepatide do not shed fat exclusively. A decent slice of what goes is lean tissue, which matters a great deal in older patients and anyone already close to the sarcopenia line. Suddenly a muscle-sparing add-on had a market.

In June 2026, Richard Pratley’s group published the phase 2 EMBRAZE trial in Nature Medicine. 102 adults with overweight or obesity took tirzepatide alongside either apitegromab, an antibody that blocks myostatin activation, or a placebo. Total weight loss came out much the same in both arms. Lean mass loss did not: the apitegromab group lost 1.9 kilograms less of it, a 54.9 per cent retention relative to placebo. Lean tissue made up roughly 30 per cent of everything lost on placebo, and closer to 15 per cent with the drug.

That is one trial, 102 people, 24 weeks, with no measured gain in physical function. Proof of concept, which is a long way short of proof.

Still, there is something almost funny about where the Hercules gene has ended up. A mutation famous for producing babies who could do gymnastics before they could crawl is being developed as a defensive drug, hired to protect the muscle of people who are voluntarily getting smaller.