Follistatin 344

FS-344

Follistatin 344 (FS-344) is a gh & growth research compound. Binds and neutralizes myostatin (and activin), releasing the brake on muscle growth.

How Follistatin 344 works

Binds and neutralizes myostatin (and activin), releasing the brake on muscle growth.

Proposed benefits

Myostatin inhibition for potential muscle growth (early-stage).

Where to get Follistatin 344

The evidence for Follistatin 344

Graded by what exists behind each claim.

✅ Clinically validated

• Human gene-therapy trials of follistatin have run — in Becker muscular dystrophy and inclusion body myositis — using AAV delivery, with reported improvements in walking distance in small open-label cohorts. That is a completely different intervention from injecting the protein.
• No trial of injected follistatin protein.

📊 Correlative data

• Sold and used as an injectable peptide on the strength of the gene-therapy headlines. There is no human series, and the gap between 'a virus makes your muscle produce follistatin continuously' and 'you inject some' is the whole question.
• Beyond that the record is self-reported: community dosing logs are real information about tolerability and almost none about efficacy.

🧪 Theoretical / extrapolated

• Follistatin binds and neutralizes myostatin, the brake on muscle growth. Myostatin-null animals and the rare human cases are dramatically muscular, so the mechanism is real and vividly demonstrated.
• The problem is pharmacokinetics. Circulating follistatin-344 is cleared in minutes, and it also binds activins involved in reproduction and inflammation — so sustained systemic myostatin blockade is neither easy to achieve nor obviously consequence-free.

What community dosing logs are worth →

How to read these tiers: they say how much human evidence exists, not how well something works — and ✗ flags harm, never a disappointing trial. How the evidence tiers work →

What Follistatin 344 actually does

Start with the number in the name, because almost every page selling this compound gets it wrong. The human follistatin entry in UniProt, accession P19883, is 344 amino acids long. Residues 1 to 29 are the signal peptide, the address label that gets cut off inside the cell before the protein is ever secreted. The mature chain is residues 30 to 344, which is 315 residues, and UniProt names that same isoform FS315 UniProt Consortium 2026.

So FST-344 and FST-315 are not two molecules. They are one isoform counted two ways: 344 if you count the signal peptide, 315 if you count what actually leaves the cell. A vendor selling both as distinct products with distinct properties is selling the same protein twice, and a buyer reasoning that 344 must be ‘more complete’ than 315 is reasoning about 29 residues that no circulating follistatin molecule has ever carried.

The isoform that genuinely is different is FS288, and nobody sells it. Alternative splicing produces a shorter form lacking the acidic 26-residue C-terminal tail that FS315 carries; a third form, FS303, arises by proteolytic trimming of that tail Schneyer 2004. Their biochemistry differs in one specific property — binding to cell-surface proteoglycans — and it ranks FS288 above FS303 above FS315. An FS315-specific immunoassay then confirmed the consequence directly: FS315 is the major circulating isoform in human serum and is undetectable in follicular fluid, while the surface-binding forms sit in tissue compartments Schneyer 2004. The isoform in the vial is the one built to travel, not the one built to stay.

Where the stickiness lives, at the residue level. Follistatin carries a basic heparin-binding sequence at residues 75 to 86, in the first of its three follistatin domains, and it associates with cell-surface heparan-sulfate proteoglycans through it Sidis 2005. Mutational work localized it to two lysine pairs: 75 and 76, and 81 and 82. Mutating either pair partially reduced surface binding; mutating both eliminated it. And here is the part that matters for every engineered version of this protein — replacing that sequence, or the whole of domain 1, abolished cell-surface binding but enhanced activin binding 4- to 8-fold, while the 75/76 mutation reduced activin binding and pituitary bioactivity 5-fold Sidis 2005. The heparin site and the activin site are not separate parts. They overlap.

And the single most useful mechanistic fact on this page, which is about what happens after a capture. The crystal structure of the FS315-activin A complex shows both isoforms inhibit activin similarly, but that FS315 moves within its third domain when the C-terminal extension is bound Lerch 2007. Free FS315 loses heparin affinity sharply as ionic strength rises, because the acidic tail folds back over the heparin site. Once FS315 has captured a ligand, that tail can no longer reach the site, and the complex binds heparin like FS288 does — the authors state that this activin-induced heparin binding is unique to FS315 and may stimulate clearance of FS315 complexes Lerch 2007.

Read that as a dosing statement and it says something no vendor page does. This protein appears to be a stoichiometric trap that marks itself for disposal at the moment it does its job. It is not an enzyme and it is not a catalyst. Each molecule captures, changes shape, becomes sticky, and is removed. How much you need is therefore set by how much ligand there is, not by how potent the protein is.

Cell, rodent, human — and where it stops

Step one, in mice, with a gene rather than a syringe. Transgenic mice expressing follistatin from a skeletal-muscle-specific promoter showed dramatic increases in muscle mass comparable to myostatin knockout mice, as did mice expressing the myostatin propeptide or a dominant-negative activin type IIB receptor Lee 2001. Independent transgenic lines, three different molecular strategies, one result. The target is not in doubt. What produced the result was a gene switched on in every muscle cell from conception.

Step two, in pigs, and this is the only large-animal experiment run on this exact isoform. Transgenic Duroc pigs expressing human follistatin-344 specifically in muscle showed an increased proportion of skeletal muscle and reduced body fat Chang 2017. The number is worth quoting because it is modest: lean meat percentage was 72.95% plus or minus 1.0% in the transgenic F1 generation against 69.18% plus or minus 0.97% in wild type, with N equal to 16 and P less than 0.05. Myofiber hypertrophy appeared in longissimus dorsi. The signaling readout was clean: reduced Smad2 phosphorylation and increased Akt Ser473 phosphorylation, which is exactly the pattern myostatin blockade predicts. No cardiac muscle hypertrophy and no reproductive abnormality were observed. A lifelong transgene in every myocyte moved lean meat by 3.8 percentage points.

Step three, in humans, and this is where the compound's whole reputation comes from. Six patients with Becker muscular dystrophy received AAV1.CMV.FS344 by direct bilateral intramuscular injection into the quadriceps Mendell 2015. Three received 3 times 10 to the 11th vector genomes per kilogram per leg and walked 58 meters further, 125 meters further, and no further, respectively, on the 6-minute walk test. Two more at double the dose improved by 108 meters and 29 meters; one did not improve. No adverse effects were encountered. Muscle histology corroborated the effect with reduced endomysial fibrosis, reduced central nucleation and a more normal fiber size distribution, especially at the higher dose.

Step four, in sporadic inclusion body myositis. Six subjects received rAAV1.CMV.huFS344 at 6 times 10 to the 11th vector genomes per kilogram into the quadriceps of both legs Mendell 2017. Annualized 6-minute walk performance improved by 56.0 meters per year in treated subjects against a decline of 25.8 meters per year in 8 untreated subjects matched for age, sex and baseline, with p equal to 0.01. Four of the six improved by 58 to 153 meters; two improved by 5 to 23 meters.

The obstacles, and the first one is the whole page. (1) Every human result above is gene therapy. A virus was placed in the muscle and the muscle then manufactured follistatin continuously, for as long as the transduced fibers survived. Injecting a protein produces a bolus that is captured, changed and cleared Lerch 2007. Those are not two doses of the same treatment; they are two different pharmacologies, and only one of them has ever been tested in a person. (2) The isoform was chosen to be the less sticky one: the trial team states that it used the alternatively spliced FS344 to avoid potential binding to off target sites Mendell 2015, which is a deliberate choice of the circulating form and works against the idea of a locally acting depot. (3) The sIBM comparison was against a matched untreated group rather than a randomized placebo arm, and the protocol included an exercise regimen Mendell 2017, so the contrast is treated plus supervised exercise against untreated. (4) The total human experience with this isoform is 12 people across 2 open-label trials, with one or two non-responders in each. (5) The route was direct intramuscular injection into quadriceps under a surgical protocol, not a subcutaneous shot.

And the comparison that is already on this site. The closest thing to a test of ‘inject follistatin protein into a muscle and see what happens’ is ACE-083, a follistatin-based protein injected intramuscularly in two randomized placebo-controlled trials. It increased muscle volume by 13.5% to 16.4% and produced no consistent functional improvement. The class review that covers 20 years of this target is subtitled the story of great expectations and limited success Nielsen 2021, and follistatin protein injection is the part of the story where the expectations were never tested at all.

Follistatin 344 pharmacokinetics — how much of it actually gets in

This site's own data card says the half-life is roughly 35 to 38 hours. That number has no published source attached to it here, and it is worth saying so plainly rather than repeating it.

What clears it, and this is a real mechanism rather than a guess. Free FS315 in plasma has its heparin-binding site masked by its own acidic C-terminal tail. When it captures a ligand the tail can no longer reach that site, the complex acquires FS288-like heparin affinity, and the authors of the structure propose that this activin-induced heparin binding stimulates clearance of FS315 complexes Lerch 2007. If that is right, then the pharmacokinetics of this protein are ligand-dependent: an unloaded molecule circulates, a loaded one is removed. The practical consequence is that the half-life measured in a fasted, unstimulated person is an upper bound, and the half-life during actual ligand capture is shorter by an unknown amount. Nobody has measured either in a human given follistatin protein.

What degrades it. This is a 315-residue secreted glycoprotein, so it is handled by cellular uptake and lysosomal proteolysis rather than by a single plasma peptidase, and it is assembled and exported by the secretory pathway — pulse-chase labeling shows FST315 is the fastest secreted of the isoforms while FST288 leaves more slowly with some retained inside the cell Saito 2005.

The oral barrier, which is absolute for this one. A 315-residue glycoprotein swallowed intact meets gastric acid, then pancreatic proteases, then brush-border peptidases, and is digested to amino acids. There is no transporter for a protein of this size across the gut wall, and hepatic first-pass extraction would take anything that crossed. Oral bioavailability here is not low, it is zero, and no oral follistatin product has ever been tested in anything.

The route, and how far it is from the trials. The research market sells this for subcutaneous or intramuscular injection at 100 to 300 micrograms. Every human study of FS344 delivered 6 times 10 to the 11th vector genomes per kilogram of an adeno-associated virus directly into quadriceps muscle Mendell 2017. There is no conversion between those two units, and that is not a rhetorical point: one produces months of continuous local synthesis and the other produces a single clearable bolus.

What would have to be true, and how you would know it was not

Three predictions. The first is the one this compound's own discovery history hands you and nobody uses it.

1. Follistatin was named for suppressing FSH, so check FSH. The bioassay used to grade engineered follistatin mutants is suppression of pituitary follicle-stimulating hormone secretion, and mutations that reduced activin binding reduced that bioactivity 5-fold Sidis 2005. So if an injected dose reaches the circulation in a pharmacologically meaningful amount, FSH should fall. Draw a luteinizing hormone and FSH panel at baseline and at 4 to 6 weeks. This is the cleanest available test of whether anything systemic is happening at all: no change means either the protein is not getting in, is not intact, or is not there. A fall means it is working and that you have suppressed a reproductive axis.

2. IGF-1 should not move, and if it does the explanation is elsewhere. Follistatin acts on the TGF-beta arm, through Smad2 phosphorylation and Akt Chang 2017, not through the growth hormone axis. So IGF-1 (Insulin-like Growth Factor 1) at baseline and 8 weeks should be flat. A rise on a stack containing a growth hormone secretagogue is the secretagogue. A rise on follistatin alone would be a genuinely new finding and worth reporting rather than assuming.

3. The prediction that cuts against the compound: a 4-week course will not change whole-body lean mass. Every positive result on this page came from continuous expression — a lifelong transgene in mice Lee 2001, a lifelong transgene in pigs that moved lean meat by 3.8 percentage points Chang 2017, or a virus manufacturing protein inside the muscle for months Mendell 2015. A bolus that is consumed by binding and cleared on capture Lerch 2007 is a different exposure entirely. Take a DEXA scan before and after a full course. The prediction is no change in lean mass beyond measurement error, and that is the single most useful number a person running this could contribute, because it does not exist anywhere.

What nobody has tested yet

Five experiments. Two of them are blood draws.

1. Nobody has published a pharmacokinetic curve for injected follistatin protein in a human. Not a maximum concentration, not a half-life, not a time to peak, by any route. Every number this market quotes for the half-life of this product is unsourced. One subject, one dose, six timed draws and a follistatin immunoassay would produce the first such curve in existence.

2. Nobody has measured FSH across a course. It is the original bioassay for this protein Sidis 2005, it is on a standard hormone panel, and it has never been reported for anyone using the research-market product. A dozen paired draws would settle whether systemic exposure is happening.

3. Nobody has asked which isoform is in the vial. An FS315-specific immunoassay already exists and was validated on human serum and follicular fluid Schneyer 2004. Run it against a vendor lot alongside a total-follistatin assay and you learn whether the product is FS315, FS288, a mixture, or a fragment. That comparison has never been published for any research-market preparation, and it is the difference between the circulating isoform and the tissue-binding one.

4. Nobody has looked for anti-drug antibodies. Repeated injection of an exogenous protein is the classic setting for an immune response that neutralizes the drug and, in the worst case, cross-reacts with the endogenous version. No study of repeated follistatin protein dosing in humans exists, so no antibody data exists. For a protein with an endogenous counterpart that regulates the reproductive axis, that is not a small blank.

5. Nobody has measured the ligand. The mechanism says the trap is consumed by capture Lerch 2007. Serum myostatin and activin A are measurable. A before-and-after pair would show whether free ligand actually falls, which is the difference between a compound that engages its target in a person and a compound that is degraded before it reaches one.

Follistatin 344 — its own safety story, not its class's

The shared block on this page warns about removing a brake the body installed deliberately. That is the right instinct and it is not specific enough, because this molecule's own risk is that it does not only block myostatin — and unlike its engineered relatives, that is not a design flaw, it is what follistatin is.

1. The reproductive axis is the exposed one, and it is exposed by definition. Follistatin was characterized as an activin-binding and neutralizing protein whose bioactivity is measured as suppression of pituitary FSH secretion Sidis 2005. Activin and follistatin regulate FSH in both sexes. A systemic dose large enough to do anything to muscle is, on the same mechanism, large enough to act at the pituitary, and no human study has ever reported gonadotropins on this product. This is the risk that is specific to full-length follistatin and is absent from an engineered fragment that has genuinely lost activin binding.

2. What the large-animal data says, including the reassuring part. Pigs carrying a muscle-specific human FST-344 transgene showed no cardiac muscle hypertrophy and no reproductive abnormality Chang 2017. That is a real negative finding and it deserves stating. It also has a boundary: the transgene was muscle-specific, so the pituitary and gonads were never exposed to elevated follistatin the way an injection exposes them. The experiment that is reassuring about the heart is silent about the axis.

3. The class precedent for what happens when this pathway is blocked systemically. A soluble activin type IIB receptor trap given subcutaneously to boys with Duchenne muscular dystrophy was stopped after its second dosing regimen because of nosebleeds and telangiectasias, vascular events rather than muscle ones Campbell 2017. Follistatin is a different molecule with a different ligand profile, and there is no report of that syndrome with follistatin. But the general lesson holds: the TGF-beta superfamily arm this compound blocks is a vascular and connective-tissue regulator as well as a muscle one.

4. What the human safety record actually consists of. Twelve people, in two open-label gene-therapy trials, who received an adeno-associated virus rather than a protein Mendell 2015 Mendell 2017. No adverse effects were encountered in the Becker trial. That is a genuine observation and it is a statement about a virus injected into a quadriceps under a protocol. It cannot be transferred to repeated subcutaneous dosing of a recombinant protein of unverified isoform, and there are 0 published adverse events for the latter because there are 0 published studies of it.

Sources read for this page

• UniProt Consortium. UniProtKB P19883 (FST_HUMAN) - Follistatin: signal peptide 1-29, chain 30-344, sequence length 344, isoform 1 named FS315. UniProtKB, entry retrieved from rest.uniprot.org on 4 September 2026
• Schneyer AL, et al. Differential distribution of follistatin isoforms: application of a new FS315-specific immunoassay. Journal of Clinical Endocrinology and Metabolism 2004 · PMID 15472207
• Sidis Y, et al. Heparin and activin-binding determinants in follistatin and FSTL3. Endocrinology 2005 · PMID 15471966
• Lerch TF, et al. Structural and biophysical coupling of heparin and activin binding to follistatin isoform functions. Journal of Biological Chemistry 2007 · PMID 17409095
• Saito S, et al. Differential biosynthesis and intracellular transport of follistatin isoforms and follistatin-like-3. Endocrinology 2005 · PMID 16150905
• Mendell JR, Sahenk Z, Malik V, Gomez AM, Flanigan KM, Lowes LP, et al. A phase 1/2a follistatin gene therapy trial for becker muscular dystrophy. Molecular Therapy 2015 · PMID 25322757
• Mendell JR, Sahenk Z, Al-Zaidy S, Rodino-Klapac LR, Lowes LP, et al. Follistatin Gene Therapy for Sporadic Inclusion Body Myositis Improves Functional Outcomes. Molecular Therapy 2017 · PMID 28279643
• Chang F, Fang R, Wang M, Zhao X, Chang W, Zhang Z, Li N, Meng Q. The transgenic expression of human follistatin-344 increases skeletal muscle mass in pigs. Transgenic Research 2017 · PMID 27787698
• Lee SJ, McPherron AC. Regulation of myostatin activity and muscle growth. Proceedings of the National Academy of Sciences 2001 · PMID 11459935
• Nielsen TL, et al. Antimyostatin Treatment in Health and Disease: The Story of Great Expectations and Limited Success. Cells 2021 · PMID 33802348
• Campbell C, et al. Myostatin inhibitor ACE-031 treatment of ambulatory boys with Duchenne muscular dystrophy: Results of a randomized, placebo-controlled clinical trial. Muscle & Nerve 2017 · PMID 27462804

Follistatin 344 — safety, predicted from mechanism

Follistatin 344 — interference & stacking

Bloodwork to run alongside Follistatin 344

Baseline first, then again at 8–12 weeks.

The “Running GH Peptides or MK-677” panel covers these in one order — 9 markers, $132.30 with the discount applied.

Check results you already have → · All 103 markers A–Z

Follistatin 344 — frequently asked questions

Follistatin 344 inside a finished plan

One arm of 1 Protocol Blueprint, free to read in full.

What Follistatin 344 is used for

Follistatin 344 appears under 2 goals in the goal router.

Related GH & Growth compounds

← Explore the full Protocol Vault

↑ Back to on this page