Follistatin 344

01 · What it is

The human evidence for follistatin-344 divides cleanly in two: a gene therapy programme that produced real results and does not apply to an injected protein, and an ophthalmology case series that does apply and reports harm.

Follistatin-344 is not a drug name and was not coined by anyone selling it. FST-344 is the 344-residue precursor produced by one of two alternative splice forms of the follistatin gene, documented in rat by a group at Prince Henry’s Institute in Melbourne in 1990 ¹ and followed the next year by an ontogeny study across tissues ². The precursor carries a 29-residue signal peptide and is processed to the 315-residue form that circulates. The alternative splice product, FST-317, gives rise to the shorter cell-surface-associated form. The number in the name is a residue count from a 1990 molecular biology paper, not a product code.

Follistatin’s biology is well established and is not in dispute. It binds and neutralises members of the TGF-β superfamily — activin A and, importantly for the use this compound is put to, myostatin (GDF-8), the negative regulator of skeletal muscle mass. Sequestering myostatin prevents it engaging the type II activin receptor and shuts down the Smad2/3 signalling that restrains muscle growth. That this works is demonstrable at the whole-animal level: transgenic Duroc pigs expressing human follistatin specifically in muscle showed reduced Smad2 phosphorylation, increased Akt Ser473 phosphorylation, myofibre hypertrophy, and lean meat percentage of 72.95 ± 1.0 % against 69.18 ± 0.97 % in wild-type animals (n = 16, P < 0.05), with no cardiac hypertrophy or reproductive abnormality observed ⁷.

The trouble begins when that biology is used to support an injected recombinant protein, because the human work was not done that way.

The human trial evidence is for gene transfer. A follistatin isoform expressed from an adeno-associated virus serotype 1 vector — AAV1-FS344 — was injected into the quadriceps of cynomolgus macaques and produced pronounced and durable increases in muscle size and strength with long-term transgene expression ³. That work moved into people: follistatin gene therapy improved ambulation in Becker muscular dystrophy ⁶. A viral vector that installs a gene in muscle fibres, producing continuous local expression for months or years, and a subcutaneous injection of recombinant protein into abdominal fat are different interventions with different exposure profiles, different distribution and different durations. The trial result does not transfer to the injection, and this guide does not let it.

The only indexed human experience with injected follistatin-344 reports harm. A retrospective series from a single ophthalmology department described 11 bodybuilding athletes who presented with reduced visual acuity and optical coherence tomography findings consistent with central serous chorioretinopathy after subcutaneous injection of complete 1 mg vials into the abdomen ¹⁰. The details are in Section 4 and Section 6.

And most of what is sold under the name is not follistatin. Of 17 black-market products tested by a doping control laboratory, nine contained follistatin. All nine contained His-tagged FS344 — a recombinant protein carrying a purification tag, made for biochemical work — together with a high degree of oligomers. Some of the remaining eight contained other growth-promoting peptides instead, including MGF and GHRP-2 ⁸.

Follistatin is prohibited under chapter S4 of the World Anti-Doping Agency’s List, as stated in that paper for the 2019 edition ⁸.

02 · Evidence at a glance

Evidence grade

Preclinical — the human trial evidence is for AAV gene transfer, a different intervention; the only human data on injected protein is an adverse-event series

What FS-344 is

The 344-residue precursor from one of two FST splice forms ¹

Processing

To the 315-residue circulating form; FST-317 gives the cell-surface form

Molecular targets

Activin A and myostatin (GDF-8), neutralised by direct binding

Downstream effect confirmed in vivo

Reduced Smad2 phosphorylation, increased Akt Ser473 phosphorylation in transgenic pig muscle ⁷

Whole-animal effect

Lean meat 72.95 ± 1.0 % vs 69.18 ± 0.97 % wild-type, n = 16, P < 0.05 ⁷

Human trials of injected FS-344 protein

None

Human trials of AAV1-FS344 gene transfer

Yes — improved ambulation in Becker muscular dystrophy ⁶

Nonhuman primate gene transfer

Pronounced, durable increases in muscle size and strength ³

Only human data on the injected protein

11 cases of central serous chorioretinopathy after 1 mg subcutaneous injections ¹⁰

Product identity

9 of 17 black-market products contained follistatin; all nine were His-tagged and oligomerised ⁸

WADA status

Prohibited under chapter S4 ⁸

Approval status

None, any jurisdiction

03 · Mechanism of action

Ligand trapping, not receptor antagonism

Follistatin does not act at a receptor. It is a soluble binding protein that intercepts ligands before they reach one. It binds activin A with very high affinity and also binds myostatin, and in both cases the bound ligand is prevented from engaging the type II activin receptor at the cell surface. The signalling that would follow — receptor complex assembly, phosphorylation of Smad2 and Smad3, nuclear translocation and transcriptional repression of muscle growth programmes — does not occur.

This mechanism is unusual among the compounds in this project and worth stating precisely: the molecule’s effect is entirely subtractive. It does not stimulate anything. It removes a brake.

The evidence that the brake is actually removed

Mechanistic claims about ligand traps are easy to make and hard to demonstrate in a living animal. The transgenic pig work does demonstrate it. Muscle-specific follistatin overexpression produced a significantly reduced level of Smad2 phosphorylation — the direct readout of blocked myostatin signalling — and a significantly increased level of Akt Ser473 phosphorylation, the readout of the anabolic pathway that myostatin normally restrains. Myofibre hypertrophy in the longissimus dorsi accompanied both ⁷. Independent work using a different genome-editing approach to integrate a myostatin inhibitor at a defined locus in pigs reached the same phenotype ¹¹.

This is a rare instance in this project of a proposed mechanism being confirmed by measurement of the specific phosphorylation events it predicts, in the target tissue, in a large mammal.

Why gene transfer and injection are not interchangeable

AAV1-FS344 delivers a gene to muscle fibres, which then express follistatin continuously and locally. Kota and colleagues reported long-term transgene expression in macaque quadriceps ³, and the downstream translational programme co-delivered follistatin with micro-dystrophin in an aged Duchenne model ⁵, with the broader gene therapy strategy set out by the same centre ⁴.

A subcutaneous injection of recombinant FS-344 protein into abdominal tissue produces a transient systemic concentration that is cleared. The exposure differs in route, in locality, in magnitude and in duration. Follistatin also binds activin A, which has roles well outside muscle — in reproductive endocrinology, inflammation and vascular biology — so a systemic exposure engages targets that a muscle-local transgene largely does not.

That distinction is not a technicality. It is the reason the ambulation result in Becker muscular dystrophy ⁶ cannot be cited in support of an injected product, and the reason this guide grades the compound Preclinical despite the existence of a published human trial with the string “FS344” in it.

Aggregation

The one analytical study of commercial material found not only a histidine tag on every follistatin- containing product but a high degree of oligomers ⁸. Protein oligomerisation is a formulation failure with immunological consequences: aggregated protein is substantially more immunogenic than correctly folded monomer. Nothing in the mechanistic literature describes what oligomerised His-tagged FS344 does, because that material was never meant to be administered to anything.

04 · Key research findings

Gene transfer in nonhuman primates. AAV1-FS344 injected into cynomolgus macaque quadriceps induced pronounced and durable increases in muscle size and strength, with long-term expression of the transgene ³. This is the strongest positive result anywhere in Family D, and it is a result about a viral vector.

A gene delivered to muscle produced a large, sustained, measurable effect. Nothing in this file shows that a protein injected under the skin does the same thing.

Gene therapy in Becker muscular dystrophy. Follistatin gene therapy improved ambulation, reported by the group that ran the preclinical primate work, with a review of the evidence supporting follistatin as a genetic enhancer of cellular performance ⁶. The broader programme at that centre placed follistatin gene transfer alongside exon skipping and mutation suppression as tools for molecular treatment of the dystrophies ⁴, and co-delivery of micro-dystrophin with follistatin restored muscle function in an aged Duchenne model ⁵.

A real clinical programme, in a real disease, with a real result — for a modality that no one is selling as a peptide.

Transgenic livestock. Human follistatin-344 expressed specifically in the muscle of transgenic Duroc pigs increased the proportion of skeletal muscle and reduced body fat, with lean meat percentage 72.95 ± 1.0 % against 69.18 ± 0.97 % in wild-type animals (n = 16, P < 0.05), myofibre hypertrophy in longissimus dorsi, reduced Smad2 phosphorylation, increased Akt Ser473 phosphorylation, and no cardiac hypertrophy or reproductive abnormality observed ⁷. A separate group achieved a comparable phenotype by site-specific integration of a myostatin inhibitor using homology-mediated end joining ¹¹.

Two independent groups, one species, the same phenotype, with mechanism confirmed at the phosphorylation level. This is the best-supported biology in the family — and it is the biology of lifelong genetic overexpression, not of an injection.

The human safety series. Eleven bodybuilding athletes presented to a single ophthalmology department with decreased visual acuity after high-dose subcutaneous follistatin-344 injection. All were male; mean age 36.8 ± 8.1 years. All had injected complete 1 mg vials into the abdomen. Eight had a history of a single previous injection; three had injected multiple times. Ten had unilateral findings, one bilateral. In all eight with a single exposure, subretinal fluid disappeared completely over a mean 2.3 ± 0.7 months and symptoms regressed. All three with repeated exposure developed recurrent disease. The authors concluded that follistatin-344 injection can be considered a risk factor for central serous chorioretinopathy ¹⁰.

Eleven cases is a small series and a retrospective one, and it establishes association rather than causation. It is also the entire indexed human record for this route of administration, and it is negative. A dose-response pattern within it — recurrence confined to those who injected repeatedly — is the kind of internal structure that makes a case series harder to dismiss.

Product composition. Seventeen black-market products labelled follistatin were analysed by a doping control laboratory. Nine contained follistatin. All nine contained His-tagged FS344 and a high degree of its oligomers. Growth-promoting peptides including MGF and GHRP-2 were found in some of the others. The presence of the histidine tag allowed unambiguous differentiation from endogenous follistatin, with a urine detection limit around 0.1 ng/mL for a 10 mL sample and a serum limit of 5 ng/mL for 100 µL ^8,9.

Under half of the products contained what the label said. Every one that did contained a laboratory reagent rather than a pharmaceutical preparation. This is the clearest identity failure documented for any compound in this project.

Contemporary appraisal. A 2026 narrative review of approved and unapproved peptides in sports medicine included FS-344 among the compounds it examined and made the general finding for the class: favourable tissue repair and metabolic outcomes in animal models, scarce rigorous human safety data, and potential for serious harm ¹².

05 · Evidence overview

06 · Safety profile

Animal data. The transgenic pig work reported no cardiac hypertrophy and no reproductive abnormality ⁷, which is a meaningful negative finding given that activin and myostatin signalling operate in both systems. It is also a finding about muscle-restricted transgenic expression from conception, not about systemic exposure in an adult. The gene therapy programme reported safety and efficacy of intramuscular follistatin gene delivery in its preclinical work ⁶, again for local transgene expression.

Human data. The only indexed human data for injected follistatin-344 protein is the chorioretinopathy series: 11 males, mean age 36.8 ± 8.1 years, all having injected complete 1 mg vials subcutaneously, presenting with reduced visual acuity and subretinal fluid on OCT; complete resolution over a mean 2.3 ± 0.7 months in the eight with single exposure, recurrent disease in all three with repeated exposure ¹⁰. Central serous chorioretinopathy is a serous detachment of the neurosensory retina; it is often self-limiting, as it was in most of this series, and recurrent disease carries a risk of permanent visual loss.

Human data from the gene therapy route exists and is separate: an ambulation benefit in Becker muscular dystrophy ⁶, with the safety profile of an AAV vector rather than of a peptide injection.

What is genuinely unknown. The pharmacokinetics of injected recombinant FS-344 in a human — no study reports a concentration-time curve by any route. The mechanism linking systemic follistatin exposure to chorioretinopathy, which the case series raises and does not explain ¹⁰. Whether neutralising circulating activin A systemically, as distinct from blocking myostatin locally in muscle, has endocrine, immunological or vascular consequences — activin A is not a muscle-specific ligand and the question has not been asked for this exposure. Immunogenicity, unassessed, and of particular concern because the material analysed carried a histidine tag and was substantially oligomerised ⁸, both strong drivers of antibody response, and because antibodies raised against recombinant follistatin would be expected to cross-react with the endogenous protein. Repeat-dose toxicology, genotoxicity, reproductive toxicology and carcinogenicity: none performed for the injected protein. And whether any given vial contains follistatin at all — in the one systematic analysis, eight of seventeen did not ⁸.

07 · US regulatory status

Current as of 6 September 2026. Follistatin-344 is not approved as a drug in the United States or any other jurisdiction and is not a controlled substance. No recombinant follistatin protein holds a marketing authorisation anywhere; the analytical paper states plainly that no approved pharmaceutical formulations of follistatin are available ⁸.

Follistatin gene therapy is a separate regulatory object. AAV1-FS344 has been studied under investigational programmes in muscular dystrophy ^4,6; those are investigational new drug applications for a biologic vector, not approvals, and they confer nothing on a protein sold under the same number.

Under the World Anti-Doping Code, follistatin is prohibited under chapter S4 — hormone and metabolic modulators — as stated for the 2019 List in the analytical literature ⁸. Detection methods based on immunomagnetic purification followed by SDS-PAGE and immunoblotting have been published, and they exploit the histidine tag present in black-market material to distinguish administered from endogenous protein ^8,9. Competitors should consult the current Prohibited List directly rather than rely on secondary summaries, including this one.

08 · Limitations of the evidence

1. The human trial evidence is for a different intervention, and this is the single most important caveat in this guide. AAV1-FS344 gene transfer produced durable muscle gains in macaques ³ and an ambulation benefit in Becker muscular dystrophy ⁶. It delivers a gene to muscle fibres for sustained local expression. Injecting a recombinant protein subcutaneously does not do that. Any claim that follistatin-344 “has been through human trials” rests on this conflation.
2. The only human data on the injected protein describes harm. Eleven cases of central serous chorioretinopathy, with recurrence confined to those who injected more than once ¹⁰. It is a retrospective series from one centre and cannot establish causation, and it is nonetheless the entire evidence base for this route.
3. The livestock evidence is genetic, not pharmacological. Transgenic pigs express follistatin in muscle from conception ^7,11. Lifelong tissue-restricted transgenic expression and intermittent systemic injection in an adult are not the same exposure, and no bridging study exists.
4. Animal-to-human translation is untested in the relevant direction. Everything positive comes from mice, macaques and pigs ^3,5,7,11, and in each case by gene transfer or transgenesis. There is no animal study of injected recombinant FS-344 protein at all — which means the standard preclinical package that would normally precede human exposure is not merely incomplete, it is absent.
5. No randomised controlled trial of any follistatin preparation was identified.
6. Systemic activin neutralisation is unstudied. Follistatin’s highest-affinity ligand is activin A, not myostatin. Muscle-restricted expression limits the consequences; a systemic injection does not. The reproductive, inflammatory and vascular effects of transient systemic activin neutralisation in an adult human have not been characterised.
7. Material identity fails more often than it succeeds. Eight of seventeen products contained no follistatin; the nine that did contained His-tagged, oligomerised recombinant protein ⁸. No published analysis has found a follistatin product made to any pharmaceutical specification.
8. Publication bias is likely in the gene therapy literature. The primate, mouse and clinical follistatin results share authors and a single translational centre ^3,4,5,6. That is a normal feature of a focused programme and it does mean independent replication of the human ambulation result is not present in the indexed record.
9. This guide asserts absence from the indexed record. No human trial of injected follistatin-344 protein appears in PubMed. Unindexed or unpublished work cannot be excluded and cannot be cited.

09 · References

1. Michel U, Albiston A, Findlay JK. Rat follistatin: gonadal and extragonadal expression and evidence for alternative splicing. Biochem Biophys Res Commun. 1990 Nov 30;173(1):401–407.

   PMID 1979488 ↗

2. Michel U, et al. Rat follistatin: ontogeny of steady-state mRNA levels in different tissues predicts organ-specific functions. Biochem Biophys Res Commun. 1991.

   PMID 1718275 ↗

3. Kota J, Handy CR, Haidet AM, et al. Follistatin gene delivery enhances muscle growth and strength in nonhuman primates. Sci Transl Med. 2009 Nov 11;1(6):6ra15.

   PMID 20368179 ↗

4. Mendell JR, Rodino-Klapac L, Sahenk Z, et al. Gene therapy for muscular dystrophy: lessons learned and path forward. Neurosci Lett. 2012 Oct 11;527(2):90–99.

   PMID 22609847 ↗

5. Rodino-Klapac LR, Janssen PM, Shontz KM, et al. Micro-dystrophin and follistatin co-delivery restores muscle function in aged DMD model. Hum Mol Genet. 2013 Dec 15;22(24):4929–4937.

   PMID 23863459 ↗

6. Al-Zaidy SA, Sahenk Z, Rodino-Klapac LR, Kaspar B, Mendell JR. Follistatin gene therapy improves ambulation in Becker muscular dystrophy. J Neuromuscul Dis. 2015 Sep 2;2(3):185–192.

   PMID 27858738 ↗

7. Chang F, et al. The transgenic expression of human follistatin-344 increases skeletal muscle mass in pigs. Transgenic Res. 2017 Feb.

   PMID 27787698 ↗

8. Reichel C, Gmeiner G, Thevis M. Detection of black market follistatin 344. Drug Test Anal. 2020.

   PMID 31758732 ↗

9. Reichel C, Gmeiner G, Thevis M. Erratum: Detection of black market follistatin 344. Drug Test Anal. 2020 Oct;12(10):1522–1533.

   PMID 33460286 ↗

10. Dağ U, et al. Central serous chorioretinopathy associated with high-dose follistatin-344: a retrospective case series. Int Ophthalmol. 2020 Nov;40(11):3155–3161.

    PMID 32671599 ↗

11. Li M, Tang X, You W, et al. HMEJ-mediated site-specific integration of a myostatin inhibitor increases skeletal muscle mass in porcine. Mol Ther Nucleic Acids. 2021 Jun 24;26:49–62.

    PMID 34513293 ↗

12. Mendias CL, Awan TM. Safety and efficacy of approved and unapproved peptide therapies for musculoskeletal injuries and athletic performance. Sports Med. 2026.

    PMID 41966639 ↗