Follistatin (FST)
FST, FS, activin-binding protein, FS-288, FS-315, FST317, FST344, follistatin-344, AAV1-FS344 (gene therapy construct)
Follistatin is a naturally occurring protein that traps and neutralises members of the TGF-beta superfamily, most notably activins and myostatin. Because myostatin restrains muscle growth, blocking it with follistatin produces dramatic muscle hypertrophy in animals. In humans, follistatin has only ever been delivered as a gene therapy in very small open-label studies, not as an injectable protein, and the injectable protein sold online has no human evidence behind it whatsoever.
Mechanism
Follistatin is a secreted glycoprotein that works by ligand sequestration rather than receptor antagonism. Two follistatin molecules wrap around a single dimeric TGF-beta superfamily ligand, and the crystal structures make the mechanism unambiguous: the follistatin N-terminal domain occludes the type I receptor binding site while the follistatin domains block the type II receptor site. The ligand is completely encircled and cannot engage either receptor. Thompson and colleagues showed this for the follistatin-activin complex in 2005, and Cash and colleagues resolved the myostatin:FS-288 complex in 2009.
The ligands so neutralised include activin A and activin B, myostatin (GDF-8), GDF-11 and several bone morphogenetic proteins, with affinities differing markedly between them. Sidis and colleagues showed that the isoforms differ not only in cell-surface retention but in relative specificity across activin, myostatin and BMPs. This promiscuity is central to understanding follistatin, and it is what distinguishes it from a selective anti-myostatin antibody. Blocking myostatin adds muscle. Blocking activins simultaneously affects the reproductive axis (activins regulate follicle-stimulating hormone secretion), inflammation, wound healing, erythropoiesis and glucose handling. Follistatin is not a myostatin inhibitor that happens to have off-target effects; it is a broad TGF-beta superfamily trap of which myostatin neutralisation is one consequence.
Downstream, removing myostatin and activin from the system silences ActRIIB and ActRIIA signalling through ALK4/ALK5, which stops phosphorylation of SMAD2 and SMAD3. De-repressed muscle then shows increased AKT-mTORC1 activity, reduced FoxO-driven atrogene transcription, and satellite cell activation. The result in animals is hypertrophy of a magnitude no conventional pharmacological agent matches.
The two isoforms are not interchangeable, and the distinction drove the design of the clinical gene-therapy programme. FS-288 carries a heparin-binding sequence that anchors it to cell surfaces, giving intense local action; FS-315 lacks it and circulates. The Nationwide Children's Hospital group selected the FS344 transgene, which yields the circulating FS-315 form, on the reasoning that the cell-surface-retained isoform carries greater potential for off-target effects on reproductive tissue.
What the research shows
The preclinical case is genuinely striking. Amthor and colleagues established in 2004 that follistatin complexes myostatin and antagonises myostatin-mediated inhibition of myogenesis. Follistatin-overexpressing mice develop muscles larger than myostatin-null mice, which is consistent with follistatin's ability to block additional ActRII ligands beyond myostatin alone. Kota and colleagues then took the crucial step of testing AAV1-delivered FS344 in non-human primates in 2009, reporting sustained increases in muscle size and strength with no adverse pathology, the study that justified moving to humans.
The human data are much thinner than the enthusiasm around them suggests. Two open-label studies, both from Mendell's group at Nationwide Children's Hospital, delivered rAAV1.CMV.huFollistatin344 by bilateral intramuscular injection into the quadriceps under the same registration, NCT01519349 (total enrolment 15, listed as completed). In Becker muscular dystrophy, six patients received either 3 x 10^11 or 6 x 10^11 vector genomes per kilogram per leg; six-minute walk distance improved by 58 m and 125 m in two low-dose patients with no change in the third, and by 108 m and 29 m in two high-dose patients with one non-responder. Muscle biopsies showed reduced endomysial fibrosis, reduced central nucleation and more normal fibre size distribution. In sporadic inclusion body myositis, six subjects received 6 x 10^11 vg/kg and showed a median one-year change of +56.0 m/year against -25.8 m/year in eight untreated comparators (p = 0.01); four of six treated subjects improved by 58 to 153 m and two were minimally improved (5 to 23 m). No adverse effects were reported in either study.
Those results have to be read with the design in mind. Both studies were open-label with no placebo arm and no randomisation, in conditions where the six-minute walk test is highly effort- and expectation-dependent, with six participants each and non-responders in both cohorts. The inclusion body myositis comparison used untreated subjects rather than randomised controls. A separate Phase 1/2 study delivered the same construct to three boys with Duchenne muscular dystrophy (NCT02354781, completed, enrolment 3). These are hypothesis-generating findings from a serious academic group, not evidence of efficacy, and no randomised controlled trial of follistatin gene transfer has been completed.
Separately, an injectable follistatin plasmid was administered to 43 participants in a Phase 1 study conducted in Roatan, Honduras, by the company Minicircle (NCT06411366, listed as completed, conditions frailty and ageing), outside the jurisdiction of FDA or EMA. It should be treated with the scepticism appropriate to an unblinded longevity-marketed intervention delivered in a jurisdiction chosen partly for its regulatory environment, and no peer-reviewed publication of its results was located during this audit.
None of this bears on the product sold as 'follistatin' or 'follistatin-344' by online peptide vendors. That is a recombinant protein for injection, an entirely different modality from AAV or plasmid gene transfer, and it has never been administered to a human in a published study.
Evidence assessment
Limited evidence
Human data exist but consist only of small open-label, unblinded, uncontrolled gene-therapy studies with historical or untreated comparators; follistatin as an injectable protein, the form sold commercially, has never been given to a human in a published study.
Tiers are applied consistently across the library and re-checked when new trials read out. Read the grading method.
Key studies
Follistatin complexes Myostatin and antagonises Myostatin-mediated inhibition of myogenesis Preclinical only
Follistatin forms a complex with myostatin and reverses its inhibition of myogenic differentiation.
The structure of myostatin:follistatin 288: insights into receptor utilization and heparin binding Preclinical only
Two follistatin molecules encircle the myostatin dimer, occluding both type I and type II receptor binding surfaces, and the structure defines the heparin-binding element that distinguishes FS-288 from FS-315.
Biological activity of follistatin isoforms and follistatin-like-3 is dependent on differential cell surface binding and specificity for activin, myostatin, and bone morphogenetic proteins Preclinical only
FS-288 and FS-315 differ substantially in cell-surface retention and in relative specificity for activin, myostatin and BMPs, so the two isoforms are not biologically equivalent.
Follistatin gene delivery enhances muscle growth and strength in nonhuman primates Preclinical only
Sustained increases in muscle size and strength without adverse pathology; the study that enabled human trials.
A phase 1/2a follistatin gene therapy trial for becker muscular dystrophy Limited evidence
Six-minute walk distance improved by 58 m and 125 m in two low-dose patients (no change in the third) and by 108 m and 29 m in two high-dose patients (no improvement in the third); biopsies showed reduced endomysial fibrosis, reduced central nucleation and more normal fibre size distribution; no adverse effects were reported.
Follistatin Gene Therapy for Sporadic Inclusion Body Myositis Improves Functional Outcomes Limited evidence
Median one-year change in six-minute walk distance was +56.0 m/year in treated subjects versus -25.8 m/year in untreated subjects (p = 0.01); four of six treated subjects improved by 58-153 m and two were minimally improved (5-23 m).
Safety
The published human safety record consists of twelve people in two open-label gene-therapy studies plus three in a Duchenne study, all reporting no adverse effects, followed for periods measured in months to a few years. That is far too small a denominator to characterise the safety of a broad TGF-beta superfamily trap.
The mechanistic concerns follow from follistatin's promiscuity. Activins regulate follicle-stimulating hormone secretion and gonadal function, so systemic activin blockade has predictable reproductive endocrine consequences; this is precisely why the clinical programme chose the circulating FS-315 isoform over the cell-surface-binding FS-288. Activin signalling also has roles in wound healing, inflammatory resolution, erythropoiesis, glucose homeostasis and, importantly, in tumour suppression in several tissues; activin A is growth-inhibitory in a number of epithelial cell types, and neutralising it systemically is not an obviously benign proposition. GDF-11 blockade has separate implications, since GDF-11 is involved in axial patterning and has been implicated, contentiously, in age-related tissue maintenance.
Experience from the wider ActRII pathway is directly relevant. ACE-031, a soluble ActRIIB decoy that traps a similar ligand set, was halted in Duchenne muscular dystrophy over epistaxis and telangiectasia, a vascular signal that emerged only once the drug reached a paediatric population. Sotatercept, an ActRIIA-Fc fusion licensed for pulmonary arterial hypertension, carries warnings for erythrocytosis, thrombocytopenia and serious bleeding. Any agent that broadly blocks this pathway should be assumed to carry vascular and haematological risk until shown otherwise.
Gene therapy adds its own hazards not shared by protein administration: immune responses to AAV capsid and to the transgene product, pre-existing neutralising anti-AAV antibodies limiting eligibility, hepatotoxicity, and irreversibility, since expression cannot be stopped once a vector is delivered.
For recombinant follistatin protein sold online, there is no safety data at all. Follistatin is a large, glycosylated, disulfide-rich protein; a version expressed in bacteria will not be correctly glycosylated and may not fold correctly. Whether such material is active, inactive or immunogenic is unknown, and antibodies raised against injected follistatin could in principle cross-react with the endogenous protein.
Regulatory status
| Jurisdiction | Status |
|---|---|
| United Kingdom | No MHRA marketing authorisation for any follistatin product. Follistatin gene therapy would be regulated as an advanced therapy medicinal product requiring a clinical trial authorisation; no such UK trial was identified during this audit. Supplying recombinant follistatin for human use without authorisation breaches the Human Medicines Regulations 2012. Not controlled under the Misuse of Drugs Act 1971. |
| United States | No FDA-approved follistatin product exists in any form. AAV-delivered follistatin gene therapy is investigational, studied under academic investigational new drug applications at Nationwide Children's Hospital; no marketing application has been filed. Recombinant follistatin protein sold by peptide vendors is an unapproved new drug and misbranded when marketed for human use, and is not a dietary supplement. The Minicircle follistatin plasmid study (NCT06411366) was conducted in Honduras, outside FDA jurisdiction, and has no US regulatory status. Not DEA-scheduled. |
| WADA (sport) | Prohibited at all times. The WADA Prohibited List, section S4.3 (Agents preventing activin receptor IIB activation), names follistatin explicitly as an example of a myostatin-binding protein. The same section separately prohibits agents reducing or ablating myostatin expression, which captures gene-therapy approaches, and gene doping is additionally prohibited under M3. Substances in S4.3 are Specified Substances. A combined LC-HRMS/MS method for detecting inhibitors of activin receptor signalling pathways was published in 2025 (Sakellariou et al.). |
Questions
The only human evidence comes from gene therapy, not from injecting follistatin protein. Two open-label studies of six patients each, using an AAV vector carrying the FS344 gene, reported improvements in six-minute walk distance in Becker muscular dystrophy and sporadic inclusion body myositis. Both had non-responders, neither was blinded or placebo-controlled, and the walk test is highly susceptible to expectation effects in unblinded trials. No randomised controlled trial of follistatin has been completed in anyone.
No, and the difference is fundamental. The trials delivered a gene, an adeno-associated viral vector carrying FS344, injected into muscle so that the fibres themselves produce follistatin continuously. Products sold online are recombinant follistatin protein, a different modality that has never been given to a human in a published study. Trial results cannot be transferred from one to the other.
They are alternatively spliced isoforms of the same gene. FS-288 carries a heparin-binding region that anchors it to cell-surface proteoglycans, so it acts locally and is cleared from circulation quickly. FS-315 lacks that region and circulates freely. Sidis and colleagues also showed they differ in relative specificity across activin, myostatin and BMPs. The distinction is not academic: the clinical gene-therapy programme deliberately used the FS344 transgene, which produces the circulating FS-315 form, because the cell-surface-bound isoform was judged to carry greater risk of effects on reproductive tissue.
Yes, at all times. The WADA Prohibited List names follistatin by name under section S4.3, Agents preventing activin receptor IIB activation, as an example of a myostatin-binding protein. Gene-therapy approaches are covered both by that section's ban on agents reducing myostatin expression and by the separate gene doping prohibition, M3. Detection methods for this class have been published.