Follistatin 344 Guide: Evidence & Sourcing

Follistatin 344 at a glance

What it is

A 38 kDa secreted glycoprotein, not a small peptide. UniProt P19883 annotates the human follistatin precursor as 344 residues with a molecular mass of 38,007, a signal peptide at residues 1-29 and a mature chain at 30-344 — so the secreted species is 315 residues long, and FS-344 names the precursor rather than the circulating protein

Researched for

Myostatin and activin inhibition, and through that skeletal-muscle mass — studied in Becker muscular dystrophy, inclusion body myositis, Duchenne models and facioscapulohumeral dystrophy models

Plasma half-life

No published pharmacokinetic study of injected, unmodified follistatin in humans could be located. This page previously published an animal figure of 'roughly 90 minutes to a few hours' and that claim is RETRACTED — see the half-life section for what replaced it, and why the industry's own systemic candidate had to be re-engineered before it could be dosed at all

Regulatory status

Not an approved medicine in any jurisdiction, in either protein or gene-therapy form. Follistatin is prohibited in sport under section S4 (hormone and metabolic modulators) of the WADA Prohibited List. Engineered follistatin proteins reached human trials but were never approved. Research use only

Educational reference. This page does not publish dose ranges or reconstitution arithmetic.

What it is / mechanism

Follistatin binds activins and other TGF-beta family ligands — myostatin among them — and prevents them from signalling. Myostatin restrains skeletal-muscle growth, so a molecule that sequesters it is expected to increase muscle mass, and in animal models it does. The name is where the confusion starts, and it is worth resolving before anything else. UniProt's human follistatin entry, P19883, describes a 344-residue precursor of 38,007 Da whose first 29 residues are a signal peptide, cleaved during secretion, leaving a mature chain running from residue 30 to 344 — 315 residues. The gene-therapy literature says the same thing in its own vocabulary: the FS344 variant undergoes post-translational modification to FS315. Two independent sources, one conclusion — 344 is the length of the thing before secretion, and 315 is the length of the thing that circulates. A vial labelled 'Follistatin 344' is named after a precursor. The other number worth knowing is 288. FS288 is the isoform that stays bound to cell surfaces through heparan sulfate and carries high activin affinity; FS315 is the serum-based form with roughly ten-fold lower affinity for activin. That difference is not trivia — it is the reason FS344 was the variant selected for human work. Follistatin interacts with the pituitary activin-inhibin axis and suppresses follicle-stimulating hormone secretion, and the lower-affinity circulating isoform was chosen specifically to keep that endocrine effect in check. When a research programme picks an isoform to avoid an endocrine consequence, the consequence is real. One complication belongs here rather than in a footnote. A 2018 analysis in Nature Communications found that in muscle-wasting and atrophying diseases the myostatin pathway is already downregulated — myostatin and activin receptor expression fall while endogenous follistatin rises — and proposed this as an explanation for why several anti-myostatin drugs produced limited efficacy in clinical trials. The simple story, block myostatin and gain muscle, has been tested repeatedly in humans by pharmaceutical programmes with far better molecules than a research vial, and it has mostly underdelivered.

Researched effects

The honest summary is that follistatin's demonstrated functional benefits in humans and in disease models come from gene transfer. For injected protein the human evidence is not absent, as this page previously said, but it is not about the product sold here, and it is negative on function: an engineered follistatin protein made muscles bigger in placebo-controlled trials without making them work better. The clinical landmark is a gene-transfer trial in Becker muscular dystrophy that delivered the FS344 isoform — the gene, packaged in adeno-associated virus, injected directly into muscle — and reported improved ambulation. The choice of isoform in that trial is the same choice described above: FS344 processing to FS315, with its lower activin affinity, to limit interference with the pituitary axis. The animal work points the same way. In a DUX4-driven model of facioscapulohumeral dystrophy, AAV1.Follistatin increased muscle mass and strength even while the toxic transgene was still being expressed. In aged mdx mice, follistatin delivered alongside micro-dystrophin restored force generation and resistance to contraction-induced injury in a way that neither treatment achieved alone. In limb-girdle models, plasmid-delivered human follistatin combined with alpha-sarcoglycan reduced membrane damage. Every one of those results was produced by putting the gene into muscle. What none of them establishes is what happens when you inject the protein. That is not a rhetorical point — it is the design problem the pharmaceutical industry ran into and solved differently. The systemic follistatin candidate developed for muscle wasting is not follistatin; it is FST-delta-HBS-Fc, a follistatin-315 mutant with its heparan-sulfate binding site deliberately disabled, fused to a human IgG4 Fc domain. Two separate pieces of engineering, both aimed at making the molecule survive in circulation long enough to act. The published pharmacokinetic work on that construct in mice and cynomolgus monkeys found clearance still driven by glycosylation, specifically sialic acid content, through the asialoglycoprotein receptor — to the point that the developers had to define sialylation specifications for the manufactured material. A second company took the local route instead. ACE-083 is human follistatin 291 fused to an IgG2 Fc domain, designed to be injected into a single muscle and act there. In a placebo-controlled phase 1 in 58 healthy postmenopausal women it increased the volume of the injected muscles without a significant change in strength (PMID 29486514). In phase 2 trials in facioscapulohumeral dystrophy and Charcot-Marie-Tooth disease, muscle volume again rose against placebo, but walking, stair-climbing, arm function and quality of life did not consistently improve (PMID 35428982, 35545446), and the sponsor terminated the programme because it did not achieve its functional endpoints. That is the closest the record comes to injected follistatin protein in people, and it is a heavily engineered molecule rather than follistatin itself. A research vial of unmodified follistatin-344 is the molecule that engineering was working around.

Research timeline

Dated milestones taken from the primary records: the molecule's discovery, its first registered human study, the trials that shaped the evidence and the regulatory decisions. Each entry links to the record it comes from.

1. Dec 1987A protein isolated from pig ovarian follicular fluid is shown to suppress the release of follicle-stimulating hormone and is named follistatin, to set it apart from the inhibins. PMID 3120188
2. Jul 31, 2001Follistatin is shown to block myostatin from binding its activin receptor, and mice engineered to make extra follistatin in muscle grow far more muscle, on the scale of myostatin knockout mice. PMID 11459935
3. Jan 2015The first human trial delivers the FS344 follistatin gene, not the protein, by viral vector into the thigh muscles of six patients with Becker muscular dystrophy: four walk farther on the six-minute test, two do not change, and no adverse effect is reported. PMID 25322757
4. Apr 5, 2017The same gene approach in six patients with sporadic inclusion body myositis improves walking distance over a year compared with eight matched untreated patients, a comparison group rather than a randomized placebo arm. PMID 28279643
5. Jun 2018ACE-083, an engineered follistatin-based protein rather than follistatin 344, is injected into single leg muscles of 58 healthy postmenopausal women in a placebo-controlled study: muscle volume rises, mean strength does not change significantly. PMID 29486514
6. Nov 2019An anti-doping laboratory tests 17 black-market products sold mostly as follistatin 344: only nine contain follistatin, all nine carry a His-tag and clumped oligomers, and some of the others contain other growth-promoting peptides instead. PMID 31758732
7. Nov 2020A case series from one institution describes 11 male bodybuilders who developed central serous chorioretinopathy, fluid under the retina, after injecting follistatin-344 under the skin; it recurred in the three who had injected more than once. PMID 32671599
8. Jun 6, 2022A randomized phase 2 trial of ACE-083 in 63 adults with Charcot-Marie-Tooth disease increases the volume of the treated shin muscle and ankle-lifting strength over placebo, but the other functional measures do not improve. PMID 35545446
9. Jul 2022A randomized phase 2 trial of ACE-083 in facioscapulohumeral muscular dystrophy reports larger treated muscles than with placebo but no consistent improvement in function or in what patients reported, after up to 12 months. PMID 35428982

Records checked Oct 4, 2026

Claims at a glance

The claims most often made about this compound, each with the strongest kind of evidence we found for it. The grade says what was studied, not whether the claim is true; the line beside it says what that study showed.

• Builds muscleHuman, randomised trial. An engineered follistatin drug grew injected muscles against placebo in a randomised trial (16.4% upper arm, 9.5% shin); no study has measured muscle after injecting unmodified follistatin. PMID 35428982
• Increases strengthHuman, randomised trial. In a randomised placebo-controlled study in 58 healthy women, injected muscles grew but mean strength did not change significantly. PMID 29486514
• Improves walking in muscular dystrophyHuman, non-randomised. In a gene-therapy trial in six men with Becker dystrophy, four walked further, with no control group; in the randomised engineered-drug trials, walking did not consistently improve. PMID 25322757
• Extends lifespanNo study found. The only published lifespan result, a 2022 mouse gene-therapy paper, was retracted by its journal in August 2025; no other study reports one. PMID 40789040
• Prevents hair lossNo study found. The hair result came from the same retracted mouse paper; no other study of follistatin and hair loss was found. PMID 40789040
• Free of serious side effectsHuman, non-randomised. The only published account of people who injected the product sold as follistatin-344 describes eleven bodybuilders with a retinal disorder (central serous chorioretinopathy). PMID 32671599

The reasoning behind each grade, study by study, is in this guide's Benefit & Outcome Review.

Grades checked Oct 6, 2026

Half-life & pharmacokinetics

This section previously stated that follistatin 344 has 'a short half-life, with animal-model estimates ranging from roughly 90 minutes to a few hours', and that tissue binding extends the effect window. That claim is RETRACTED. It was published without a citation, and re-checking the literature did not produce a primary study supporting it for unmodified follistatin. Here is what the record does contain. No published pharmacokinetic study of injected, unmodified follistatin protein in humans could be located. The nearest published pharmacokinetics belong to FST-delta-HBS-Fc, and that molecule is follistatin-315 with its heparan-sulfate binding site disabled by mutation and an IgG4 Fc domain fused on — two modifications whose entire purpose is to change how long the molecule persists. Its clearance in mice and cynomolgus monkeys was driven by glycosylation, specifically sialic acid content, through the asialoglycoprotein receptor, to the degree that the developers set sialylation specifications on the manufactured material to control it. The inference that survives is qualitative rather than numeric, and it is the useful one: unmodified follistatin is not a molecule that persists well in circulation, which is why making a systemic follistatin drug required removing a binding site and adding an Fc domain. A specific half-life in minutes for the unmodified protein is not something this page can source, and it will not print one it cannot. If you see a figure quoted elsewhere, the useful question is which study measured it, in which species, on which construct.

Storage & handling

• Lyophilized (powder): store sealed and refrigerated at 2-8°C, protected from light and moisture; long-term storage is commonly at -20°C or below. Do not use past intended research shelf life.
• Reconstituted (liquid): refrigerate at 2-8°C, protect from light, and avoid freeze-thaw cycles and vigorous shaking. This matters more here than for a short peptide — follistatin is a 38 kDa glycoprotein, and larger glycoproteins are the sort of molecule that loses structure to repeated freezing and to foaming.
• Aliquot if a vial will be entered repeatedly. Bacteriostatic water's benzyl alcohol is a preservative for multi-dose entry, but it does not protect a protein from mechanical or thermal stress.
• Label the vial with the reconstitution date and the resulting concentration in mcg/mL. At 10-30 unit draws the concentration is doing all the work in the arithmetic, and it is the number nobody remembers later.

Evidence File

The Follistatin-344 Evidence File: Gene Transfer, an Engineered Protein That Failed on Function, and the First Human Harm Report

People have received follistatin in three ways: as a gene injected into muscle, as an engineered follistatin protein in company trials, and as research-market vials injected at home. This file sets out what each route produced — including a Duchenne result that exists only on the trial registry, a programme stopped because bigger muscles did not work better, and an eye disorder reported in eleven bodybuilders.

• 15 enrolled in the Becker and myositis gene-transfer trial; the papers report 12
• 0 of 3 Duchenne boys walked further, a result posted only on the registry
• 0 published trials of unmodified follistatin protein in people
• 11 bodybuilders with a retinal disorder after injecting follistatin-344

Sourcing File

The Follistatin-344 Sourcing File: A 315-Residue Glycoprotein, 18 Disulfide Bonds, and Black-Market Vials That Were Often Something Else

Follistatin is not a peptide in the sense this library usually means: it is a 35-kilodalton glycoprotein held together by 18 disulfide bonds, which only a living cell can make. This file sets out what the identity record says, what a correctly made protein should weigh and look like, what doping laboratories found when they tested black-market follistatin-344 — and why a peptide-style certificate cannot answer the questions that matter here.

• 9 of 17 black-market products contained follistatin at all
• 18 disulfide bonds a correctly folded mature chain must form
• 34,718 Da: the mature chain's protein mass before any sugar
• 0 PubChem compound records for follistatin; 1 unrelated one named ACE-083

Benefit & Outcome Review

The Follistatin-344 Benefit & Outcome Review: Bigger Muscles, Rarely Better Function, and None of It From a Vial

Follistatin is sold for muscle growth, strength, fat loss, longevity and hair. This review grades each claim against the human and animal record and finds a consistent pattern: muscle size goes up — in mice, monkeys and people — while strength and function mostly do not, every human result came from a gene or an engineered drug rather than the protein in a vial, and the lifespan headline was retracted.

• 0 human outcomes from injected unmodified follistatin
• 16.4% more muscle volume than placebo, with no consistent functional gain
• 4 of 6 Becker patients walked further, in a trial with no control group
• 1 published lifespan result for follistatin, retracted in 2025

Guide FAQ

Quick answers about guide scope, access, and educational use context.