The Longevity Desk
13 min read

Follistatin-344 Names a Gene, and Every Positive Result I Found Under That Name Came From Delivering One

The protein a cell makes from that gene is 315 amino acids long and carries no manufacturing tag, and when two anti doping laboratories opened seventeen vials sold under the name, eight held no follistatin at all.


What it is

Myostatin tells muscle to stop growing, and follistatin is one of the proteins that mops it up before it reaches its receptor (PMID 11459935). UniProt lists the gene's longer message at 344 amino acids, of which the first 29 are a signal peptide, the address label stripped off as the protein is secreted, leaving a mature chain of 315 and no mature follistatin of 344 anywhere in human biology.

The number names a gene therapy cassette, the packaged stretch of DNA a virus carries into a cell. Rodino-Klapac and colleagues, the Nationwide Children's group behind the published follistatin gene therapy trials, call FS344 an alternatively spliced cDNA, a laboratory copy of the gene's message delivered to muscle by adeno-associated virus, whose transgene product is a peptide of 315 amino acids secreted into the serum (PMID 19208403).

It is also not a peptide. UniProt gives the 344 residue precursor 38,007 daltons and 18 disulfide bonds, the chemical staples that hold a fold in shape, about a hundred times the mass of epithalon at 390 daltons. Catalogue follistatin is bacterial: ProSpec's His tagged version runs 325 residues including a ten residue tag, at 36.0 kilodaltons, research use only.

What the trials found

I searched PubMed for follistatin filtered to the clinical trial publication type and got 93 records, three of which gave anything follistatin derived to a person, and pulled all 79 follistatin studies on the government registry. Go and run both.

Mendell and colleagues injected the virus into both quadriceps of six men with Becker muscular dystrophy, three at 3 x 10^11 vector genomes, a count of gene carrying virus particles, per kilogram of body weight per leg, and three at 6 x 10^11. Six minute walk distance across the six ran from minus 14 to plus 125 metres, measured at one year in the first three and six months in the second three, an 11.5 percent average with two of them not improving at all (PMID 25322757, PMID 27858738). Six patients, no control group.

The same team gave the higher dose to six people with sporadic inclusion body myositis, a muscle wasting disease, and walking distance improved 56.0 metres a year against a decline of 25.8 in eight matched untreated people (PMID 28279643). Every treated participant also followed an exercise programme.

Three boys with Duchenne muscular dystrophy got the same construct, and their posted results, the only ones among the follistatin studies I pulled, show no dose limiting toxicities and zero of three improving on the walk test. Minicircle gave 43 healthy subjects a follistatin plasmid on Roatán, Honduras, and has published nothing since finishing in August 2023.

Whose results these are

Four substitutions, and the sales copy runs all four at once.

The modality is wrong, and that is the load bearing one. Kota and colleagues put 1 x 10^13 vector genomes into the quadriceps of cynomolgus macaques, three per group. Circumference rose 15 percent over each animal's own baseline by 8 weeks under a general purpose switch, and 10 percent by 12 weeks under a switch that only turns on in muscle, a group that stopped being distinguishable from controls by week 16 (PMID 20368179). That paper acquired an erratum on 5 August 2026 which neither I nor the verifier could open. A vector expresses continuously inside a muscle fibre for months. An injection does not.

The isoform is wrong, isoform meaning which version of a protein a single gene makes. A review from the same group describes the delivered gene, FS344, as modified after translation into FS315, the form that circulates in blood (PMID 27858738). FS-288 is the shorter one that sticks to cell surfaces and stays put.

The ACE-083 numbers belong to a different molecule again, a truncated follistatin coupled to an antibody Fc domain and engineered to stay in the muscle it is put into (PMID 31388039). In 58 healthy postmenopausal women, 50 to 200 mg into one muscle produced largest volume increases averaging 14.5 percent and no strength change (PMID 29486514). Volume without strength is the class failure: a 2022 survey reports these drugs failed in trials more often on an inability to improve muscle function than on safety (PMID 34520530).

The myostatin genetics belong elsewhere again: the quadrupling figure in marketing comes from a mouse bred both to carry a follistatin transgene and to have no myostatin (PMID 17726519).

I found two studies that injected the protein into animals, and neither produced muscle growth. Feger and colleagues gave FS-288 by pump under the skin to 72 rats, and the muscles came out smaller after three months of denervation, meaning the nerve supply had been cut (PMID 35747585). Forrester and colleagues injected follistatin under the skin of mice around irradiation, 8 to 9 per group over six months, and measured scarring rather than growth (PMID 28301516).

How it compares

Four things travel under one name, and only the last row is what anybody buys.

What is deliveredHuman evidenceWhat that evidence covers
rAAV1.CMV.FS344A virus carrying the gene into muscle12 patients in two papers, 6 more on the registryWalking distance in Becker muscular dystrophy and inclusion body myositis
Follistatin plasmidA loop of bare DNA, single dose43 healthy subjects, one commercial trial in HondurasNothing yet. No results posted since August 2023
ACE-083A truncated follistatin Fc fusion, into one muscle58 healthy women in phase 1, a phase 2 in muscular dystrophyMuscle volume at 50 to 200 mg, with no strength change
A vial labelled follistatin-344A bacterial protein with a purification tagNone I could find, by any routeNothing

Where the dose came from

The convention that circulates is roughly 100 to 200 micrograms a day under the skin, in cycles of about 10 to 30 days, and it is community convention rather than a tested dose. The dosage guide at peptidewiki.co that carries it concedes that no published human dose-response studies exist.

No derivation can exist. Every dose in the gene therapy record is a count of vector genomes per kilogram of body weight, and those do not convert into micrograms of protein, because how much a muscle the virus has reached puts out was never the dosing variable, and there is no mass to scale up from by body size either.

Two masses in the record point the other way. ACE-083 used 50 to 200 milligrams into a single muscle, 250 to 2,000 times the daily microgram convention. Nguyen and colleagues modelled what a follistatin protein dose would have to be and came out at 3 to 5 milligrams per kilogram weekly of a half life engineered fusion (PMID 32419339), which for an 80 kilogram adult is 240 to 400 milligrams a week against a community convention of about 0.7 to 1.4 milligrams a week.

What could go wrong

Reichel, Gmeiner and Thevis, at anti doping laboratories in Seibersdorf and Cologne, bought seventeen black market products labelled follistatin 344. Nine contained follistatin, and some of the other eight held other growth promoting peptides instead, named as MGF and GHRP-2. All nine carried a His tag, a run of histidine residues added during manufacture, plus oligomers, meaning copies stuck to one another (PMID 31758732). That leaves a tagged, clumped bacterial protein with none of the sugar coating the human version carries, the setup for an immune response against something the body already makes, and I found no measurement of anti follistatin antibodies in anybody injecting one. That paper carries a 2020 erratum neither I nor the verifier has read. How to read a certificate of analysis covers the rest.

Follistatin was isolated from the ovary and suppresses follicle stimulating hormone, the pituitary signal driving egg and sperm production, and the group that took it into humans named that as the reason for their design. Those hormones stayed normal in the six Becker patients. I found no repeat dose toxicology, carcinogenicity or genotoxicity study of injected follistatin protein in any species.

The only FDA document I found naming follistatin is a 2020 warning letter to one compounding pharmacy, where it sits first on a list of bulk substances. That settles that pharmacy's eligibility for an exemption and is not a safety finding, and fda.gov returned a 404 to me and to the verifier independently, so it rests on this project's saved capture. In sport, the doping laboratory paper puts follistatin under chapter S4 of the WADA 2019 list, and I did not check the current one.

What nobody knows

Whether any of it gets in. I found no pharmacokinetic study, meaning a measurement of how much reaches the blood and how fast, of follistatin protein under the skin in any species. The nearest evidence is sheep, where Supersaxo and colleagues injected four compounds under the skin, three animals each, and concluded that molecules above about 16,000 daltons are absorbed mainly by lymph, which is slow and partial (PMID 2137911). Follistatin is about twice that mass.

Whether it even folds. Eighteen disulfide bonds, no bacterial chaperones, the helper proteins that fold a chain into shape, clumped copies in all nine vials that did contain follistatin, and I found nothing published on whether the material people inject binds myostatin at all.

What the gap actually is

The evidence here is not thin. It is strong, replicated across mice, monkeys and people, and it belongs to something else. The gene branch has the opposite problem: a dose of the virus cannot be taken back.

If you have bought a vial labelled follistatin-344, I want to know whether the certificate mentioned a His tag or a mass near 36 kilodaltons, because that is the one line on the paperwork that would tell you what you are holding.

Frequently asked

Is follistatin-344 a peptide?

Not in the usual sense. Follistatin is a folded protein of roughly 35 to 38 kilodaltons with 18 disulfide bonds and two sugar attachment points, about a hundred times the mass of a small peptide like epithalon. The 344 in the name counts a gene sequence, not the protein. The protein a cell makes from that gene is 315 amino acids long, because the first 29 residues are a signal peptide that is cut off as the protein is secreted.

Has follistatin-344 been tested in humans?

The gene has. The protein has not, as far as I could find. Across 93 PubMed clinical trial records and 79 registry studies I found no published study giving follistatin protein to a person as a single agent by any route. Every study I found in which follistatin itself was the thing given delivered it as a gene, either by a virus injected into muscle or by a plasmid, which is a loop of bare DNA, including one commercial trial that gave a follistatin plasmid to 43 healthy subjects in Honduras and has posted no results since completing in August 2023.

Does follistatin actually build muscle?

Delivered as a gene, it built measurable muscle in mice, in macaques and in a small number of muscular dystrophy patients. Given as a protein, the record is different. The only engineered follistatin protein I found in human trials, ACE-083, increased thigh muscle volume by up to 14.5 percent on average in 58 healthy postmenopausal women injected into one muscle, but produced no significant change in strength, and its phase 2 trial hit its muscle volume endpoint without improving function on the tests. A 2022 review of this drug class reports that failures came more often from an inability to improve function than from safety problems.

What is the usual follistatin-344 dose?

I found no derived dose, and the figures that circulate are community convention rather than anything a study produced. One convention that circulates is roughly 100 to 200 micrograms a day under the skin in cycles of about 10 to 30 days, and other pages give 100 to 200 micrograms two or three times a week instead. One of those pages states outright that no published human dose-response studies exist. Doses in the actual research are counts of gene carrying virus particles, which cannot be converted into micrograms of protein.

What is actually inside black market follistatin vials?

Two anti doping laboratories, in Seibersdorf and Cologne, bought seventeen products labelled follistatin 344 and analysed them. Nine contained follistatin, and in some of the other eight they found different growth promoting peptides instead, naming MGF and GHRP-2. All nine that did contain follistatin carried a His tag, which is a run of histidine residues added during manufacture and something the follistatin a human body makes does not have, along with a high proportion of clumped copies.

Is follistatin banned in sport?

It is prohibited. The anti doping laboratory paper that analysed those vials opens by stating that follistatin is prohibited under chapter S4 of the WADA 2019 list of prohibited substances and methods, and that no approved pharmaceutical formulations of follistatin are available. That citation is to the 2019 list, and I did not check the current one, so I am not quoting a section number I have not read.


This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.

ShareEmailBlueskyXReddit

Get the next one in your inbox

Free. One email a week at most. Unsubscribe in one click.

Double opt-in · No spam · Unsubscribe anytime

More in Growth hormone