Follistatin

The honest framing of follistatin has four parts, and the peptide page is written around them. First, the rodent preclinical evidence is genuinely strong — the myostatin-knockout and follistatin-overexpression phenotypes produce muscle masses far beyond what any pharmacological intervention has produced in humans, and the mechanism is reproducible across multiple species. The cattle, dog, and human counterparts of the rodent phenotype confirm that the pathway operates across mammals: Schuelke et al., N Engl J Med 2004, 350(26):2682–2688 described a child born homozygous for a myostatin splice-donor-site loss-of-function mutation with extreme muscle hypertrophy and normal development through age 4.5; McPherron and Lee, PNAS 1997, 94(23):12457–12461 characterized the bovine myostatin mutations underlying the Belgian Blue and Piedmontese double-muscled cattle phenotypes; Mosher et al., PLoS Genet 2007, 3(5):e79 showed the "bully whippet" phenotype in dogs homozygous for a two-base-pair MSTN deletion. The biology of myostatin restraint is one of the better-established cross-species genetic stories in muscle physiology.

Second, the peptide form sold in research-channel markets has essentially no published human pharmacology. There is no Phase 1 PK study of subcutaneous recombinant follistatin in humans, no Phase 2 efficacy trial of subcutaneous follistatin for any indication, and no Phase 3 program. The published human follistatin-protein dataset effectively reduces to its use as an assay analyte in physiological-regulation studies and to native concentrations in disease-state characterization — not as an exogenously administered therapeutic. The native protein's roughly 2-hour serum half-life makes systemic peptide dosing pharmacokinetically awkward, and the biotherapeutic-engineering literature has responded by building Fc-fusion variants with half-lives extended ~100-fold rather than pursuing the native protein as a parenteral drug. The peptide commonly sold as "Follistatin 315" or "Follistatin 344" by research-chemical channels is mechanistically the same molecule that the preclinical literature describes, but the assumption that a subcutaneous injection of recombinant follistatin powder produces the muscle-hypertrophy phenotypes of follistatin-transgenic mice is an inference from rodent transgenic biology, not a demonstration of human pharmacology.

Third, related programs in the myostatin pathway have a substantial late-phase failure record across modalities. The anti-myostatin monoclonal antibody MYO-029 (stamulumab, Wyeth) advanced to a 116-subject double-blind placebo-controlled Phase 1/2 trial in adults with muscular dystrophy; the program was acceptably safe but did not produce statistically significant improvements in muscle strength or function, and Wyeth discontinued muscular-dystrophy development in March 2008 (Wagner et al., Ann Neurol 2008, 63(5):561–571). The soluble activin-receptor-IIB decoy ACE-031 (Acceleron) entered Phase 2 in ambulatory Duchenne boys and was halted in 2011 and permanently discontinued in May 2013 after participants developed epistaxis, gingival bleeding, and cutaneous telangiectasias — vascular adverse events attributed to off-target inhibition of BMP9/BMP10 signaling through the ActRIIB receptor (Campbell et al., Muscle Nerve 2017, 55(4):458–464). The anti-ActRIIB antibody bimagrumab (Novartis BYM338) failed its primary endpoint in the 251-patient Phase 2b RESILIENT trial in sporadic inclusion-body myositis at 52 weeks (Hanna et al., Lancet Neurol 2019, 18(9):834–844). Across the three modalities that have actually reached late-phase human testing in the myostatin pathway — anti-myostatin antibody, soluble decoy receptor, anti-receptor antibody — primary endpoints in the muscle-wasting indications they were designed for have not separated from placebo, and one program produced a vascular adverse-event pattern severe enough to end development outright. The translation from "the rodent phenotype is dramatic" to "human muscle function improves on the prespecified primary endpoint" has been substantially harder than the preclinical signal suggested. The failed peptide trials archive catalogs the broader pattern of this kind of late-phase miss across the peptide space.

Fourth, the only recent positive late-phase signal in the broader myostatin-pathway space is bimagrumab in obesity, and it is at the antibody level rather than the peptide level. Heymsfield et al., JAMA Netw Open 2021, 4(1):e2033457 randomized 75 adults with type 2 diabetes, BMI 28–40, and HbA1c 6.5–10.0% to intravenous bimagrumab or placebo every four weeks over 48 weeks. The trial reported a ~21% reduction in body fat mass on bimagrumab versus 0.5% on placebo, a ~3.6% gain in lean mass versus a 0.8% loss on placebo, and a net 6.5% body-weight reduction versus 0.8% — a body-composition profile that flips the usual GLP-1-class trade-off (significant lean-mass loss alongside fat-mass loss). The mechanism is exactly the pathway follistatin acts on: ActRII blockade releases the myostatin/activin brake while a separate caloric-deficit or appetite-suppression mechanism is reducing fat mass. The bimagrumab program in obesity, including a combination Phase 2 with semaglutide, has continued to advance — this is, at the level of published 2026 evidence, the most clinically grounded data point in the entire myostatin-pathway story. But the data is on a monoclonal antibody administered IV in a controlled trial, not on subcutaneous gray-market follistatin peptide. The translation from "bimagrumab works in obesity" to "subcutaneous follistatin peptide works in bodybuilding" is exactly the kind of cross-modality inference the preceding 20 years of muscle-pathway pharmacology should have made the corpus more cautious about, not less.

The closest human evidence on follistatin specifically comes from the gene-therapy side, and is small. Mendell et al., Mol Ther 2015, 23(1):192–201 ran a Phase 1/2a open-label trial of AAV1.CMV.FS344 delivered by direct intramuscular quadriceps injection to six ambulatory adults with Becker muscular dystrophy. The trial reported no adverse effects, histological evidence of reduced endomysial fibrosis and increased fiber size, and 6-minute walk test improvements ranging from 58 to 125 meters in four of six patients. Al-Zaidy et al., J Neuromuscul Dis 2015, 2(3):185–192 reported an additional ambulation analysis of the same cohort, with an 11.5% average 6-minute-walk improvement at six months (p = 0.02) and elevated serum follistatin levels following intramuscular injection. Six open-label patients are insufficient to establish efficacy and the trial is not blinded — but the published follistatin-specific human dataset that does exist in any form is gene-therapy-delivered AAV-FS344 to skeletal muscle in a small dystrophy cohort, not subcutaneous recombinant FS-315 or FS-344 peptide in healthy or trained adults. The non-human-primate gene-therapy work that preceded it (Kota et al., Sci Transl Med 2009, 1(6):6ra15) showed durable 15% quadriceps-circumference increases and 11.8–77.9% increases in twitch and tetanic force in cynomolgus macaques over 15 months without immune or organ-system pathology, which is the strongest non-rodent preclinical readout in the pathway.

The biohacker stacking case with the GH-secretagogues is theoretical. MK-677, Ipamorelin, and Tesamorelin act through IGF-1-axis induction rather than myostatin-pathway sequestration; follistatin peptide has no published human pharmacology of its own, and no human evidence base for the combination exists. The sarcopenia and peptides dossier and muscle preservation decision guide walk why the GH-axis layer alone produces mass without strength on its best published evidence. Adding a second pharmacological layer with zero human pharmacology to a layer that already does not translate to functional endpoints is mechanistic optimism, not an evidence-based protocol.