F tier · safety concern
The only phase 2 trial (NCT01099761, 24 boys with Duchenne muscular dystrophy) was terminated by the sponsor in 2011 after epistaxis and telangiectasias emerged, and Acceleron and Shire permanently abandoned the programme on 2 May 2013 because subsequent nonclinical and toxicology work did not support further development. A documented off-target vascular safety signal plus a defining trial that failed places this at the bottom of the scale.
ace-031
ACE-031 (ramatercept) is a dimeric fusion protein linking the extracellular domain of activin receptor type IIB (ACVR2B/ActRIIB) to the Fc region of human IgG1, designed by Acceleron Pharma to act as a circulating decoy that sequesters myostatin (GDF-8) and related TGF-beta superfamily ligands. It reached phase 2 in Duchenne muscular dystrophy, where it produced favourable body-composition changes but was halted after non-muscular bleeding events. Development was permanently discontinued in 2013 and the molecule has never been approved in any jurisdiction.
the explanation
ACE-031 is a lab-made protein that soaks up myostatin, the natural brake on muscle growth, so muscles can grow larger. In trials it did increase lean mass, but children in the Duchenne study started getting nosebleeds, bleeding gums and small spidery blood vessels in the skin, so the study was stopped and the drug was abandoned in 2013. It has never been approved as a medicine anywhere, and it is banned in competitive sport.
regulatory status
Never approved; development permanently discontinued 2013
ACE-031 received FDA orphan drug designation for Duchenne muscular dystrophy on 19 August 2010, but never progressed beyond phase 2 and holds no marketing authorisation from the FDA, EMA or MHRA. Both DMD trials (NCT01099761 and its extension NCT01239758) were terminated by the sponsor with the registry reason 'Based on preliminary safety data'; dosing was suspended in February 2011 and on 2 May 2013 Acceleron and Shire concluded their collaboration and confirmed the programme would not restart. ACE-031 is explicitly named on the WADA Prohibited List as a 'decoy activin receptor' under the activin receptor IIB section of S4 (Hormone and Metabolic Modulators), prohibited at all times in and out of competition. It is not a dietary supplement or a licensed medicine in the UK; material sold under this name is unlicensed and lawful supply is as a research reagent only, not for human use.
how it works · proposed mechanism
Myostatin (GDF-8) is a secreted TGF-beta superfamily ligand that signals through activin type II receptors to restrain skeletal muscle growth. ACE-031 was engineered as a soluble decoy: the ligand-binding ectodomain of ActRIIB fused to an IgG1 Fc, giving a long-circulating trap that intercepts myostatin before it reaches cell-surface receptors. Acceleron described it as a 'freely circulating, decoy version of ActRIIB which removes proteins, such as GDF-8 (myostatin), that limit the growth and regeneration of muscle'.
Ligand sequestration, not receptor blockade
ACE-031 does not sit on the muscle cell receptor; it binds circulating ligand. Fc fusion confers dimerisation and extended residence time, so a single administration lowers effective free myostatin for a prolonged period. Downstream, reduced ActRIIB engagement means less SMAD2/3 phosphorylation and less transcriptional suppression of the hypertrophy programme.
Promiscuity is intrinsic to the design
The ActRIIB ectodomain binds many TGF-beta superfamily members, not only myostatin: activin A, GDF-11, and BMP9/BMP10 among them. A decoy built from that domain therefore traps a whole ligand family. This breadth is what made it potent in animals and also what made it unsafe in humans.
The BMP9 vascular link
The clinical adverse events were epistaxis, gum bleeding and cutaneous telangiectasias, and have been attributed to inhibition of ligands beyond myostatin and activin A, particularly BMP9. That phenotype closely mirrors hereditary haemorrhagic telangiectasia, a disorder of BMP9/ALK1-endoglin endothelial signalling, which makes the mechanistic attribution biologically coherent rather than coincidental.
Body composition effects were real
In healthy postmenopausal women a single administration increased total body lean mass by 3.3% and thigh muscle volume by 5.1% by day 29 in the highest-exposure group, with changes in bone and fat markers. In the DMD trial there were trends toward increased lean mass and bone density and reduced fat mass. In marmosets, 14 weeks of treatment raised lean mass, increased biceps brachii fibre cross-sectional area (20% type II, 34% type I) and increased force production in extensor digitorum longus.
what’s reported
evidence shape
8 sourcesThe composition of the cited literature by study type. This describes the shape of the evidence, not its quality — and it does not by itself set the grade.
⚠ the catch
ACE-031 was stopped for harm, not for lack of money or interest. Dosing in the 24-patient phase 2 DMD trial was halted after the second administration because of epistaxis and telangiectasias, the registry reason for terminating both that trial and its 11-patient extension reads 'Based on preliminary safety data', and after two years of additional nonclinical and toxicology work the developers stated in May 2013 that the findings 'do not support further development'. The vascular events are attributed to trapping ligands beyond myostatin, particularly BMP9, which is an unavoidable consequence of using the promiscuous ActRIIB ectodomain as the binding moiety, so it is a property of the molecule rather than a fixable artefact. Total human exposure across the entire published programme is roughly 100 subjects, none of them healthy young men, and no functional endpoint was ever met. Separately, analytical work found that grey-market vials labelled ACE-031 generally are not ACE-031: they contained full-length ACVR2B lacking the Fc region, carried a His-tag absent from the clinical product, and were contaminated with numerous other proteins.
key published findings
- In the phase 2 DMD trial (NCT01099761, n=24; Campbell et al., Muscle & Nerve 2017), no serious adverse events occurred, but dosing was stopped after the second administration because of epistaxis and telangiectasias; there were trends toward maintenance of 6-minute walk distance versus decline on placebo, with increased lean mass and bone density and reduced fat mass.
- In 48 healthy postmenopausal women (Attie et al., Muscle & Nerve 2013), a single administration was generally well tolerated and increased total body lean mass by 3.3% and thigh muscle volume by 5.1% at day 29 in the highest-exposure group.
- Both DMD studies are recorded on ClinicalTrials.gov as TERMINATED with the reason 'Based on preliminary safety data'; dosing was suspended in February 2011 and the programme was permanently ended on 2 May 2013 when Shire concluded its collaboration with Acceleron.
- The vascular adverse events have been attributed to inhibition of ligands beyond myostatin and activin A, particularly BMP9, consistent with the hereditary-haemorrhagic-telangiectasia-like presentation.
- Preclinical efficacy is reproducible: in common marmosets, 14 weeks of ACE-031 increased lean body mass, raised biceps brachii fibre cross-sectional area by 20% (type II) and 34% (type I), and increased absolute and specific force in extensor digitorum longus, with no adverse effects observed in that study.
limitations of the evidence
- Only two ACE-031 human studies have been published in peer-reviewed journals; the second phase 1 study (NCT00952887, 70 subjects) has no full publication, so a substantial share of the human safety dataset has never been independently scrutinised.
- The phase 2 trial was designed as a safety and pharmacodynamic study of 24 patients and was truncated after the second dose, so it was never capable of demonstrating a functional or clinical benefit; the 6-minute walk result was a trend, not a met endpoint.
- All human exposure was in boys with Duchenne muscular dystrophy and in postmenopausal women. There is no controlled human data in healthy adults of the population that uses grey-market myostatin inhibitors, and no long-term follow-up beyond the truncated dosing period.
- No dose- or exposure-response safety threshold was ever established for the vascular events, so there is no evidence base for any 'safer' level of exposure.
- Basic chemistry attributes cannot be pinned down from public authoritative sources: no CAS registry number, no defined molecular formula and no published amino acid sequence for the clinical construct, and as a glycosylated Fc-fusion dimer it has no single exact molar mass. No verified human half-life figure is published in an accessible primary source.
documented safety signals
- Epistaxis (nosebleeds), gum bleeding and cutaneous telangiectasias (small dilated skin blood vessels) in the phase 2 Duchenne trial, sufficient for the sponsor to halt dosing after the second administration; the Muscular Dystrophy Association reported these resolved fully on discontinuation, but they nonetheless ended the programme.
- Unspecified adverse findings in subsequent nonclinical and toxicology studies conducted during the 2011-2013 review, which the developers said 'do not support further development' — the specific toxicology results have never been published, so the full risk profile is not publicly characterised.
- Mechanism-level risk that is not avoidable by formulation: ActRIIB ectodomain traps BMP9/BMP10 alongside myostatin, and disruption of BMP9-ALK1 endothelial signalling is the established basis of hereditary haemorrhagic telangiectasia.
- Product-identity risk in grey-market supply: analytical characterisation of 14 black-market products found 12 contained ACVR2B-related protein but consisted of full-length ACVR2B without the Fc region, all carried a His-tag not present in the clinical molecule, one was mislabelled follistatin, and samples contained many additional proteins.
identity
| full name | Ramatercept (ACE-031), a soluble activin receptor type IIB extracellular domain fused to human IgG1 Fc |
| category | Anabolic & Receptor |
| modality | other |
laboratory handling
| storage | A glycosylated Fc-fusion protein: laboratory handling is refrigeration at 2-8 °C for working solutions and -80 °C for long-term storage, with freeze-thaw cycles avoided because Fc-fusion dimers aggregate readily. Handling information only. |
| solubility | Supplied and handled as a protein rather than a synthetic peptide; reconstituted in sterile buffer, with carrier protein often added to reduce adsorption losses. |
| co-studied with | No human combination data exist for ACE-031 with any other agent; the entire clinical programme was monotherapy in Duchenne muscular dystrophy or single-agent phase 1. Any discussion of combining it with other compounds is unsupported by evidence and outside the scope of a discontinued investigational biologic. |
Handling information describes laboratory practice for a research reagent. It is not a preparation guide for use in humans.
the receipts
8 citedClinicalTrials.gov, US National Library of Medicine · 2011 · official
ClinicalTrials.gov, US National Library of Medicine · 2011 · official
World Anti-Doping Agency, Scientific Research · 2021 · official
Drug Testing and Analysis 17(10) (Reichel C, Filip T, Gmeiner G, Thevis M) · 2025 · peer-reviewed
Frontiers in Pharmacology 12:735912 (Markati T, De Waele L, Schara-Schmidt U, Servais L) · 2021 · review
Muscle & Nerve 55(4):458-464 (Campbell C, McMillan HJ, Mah JK, et al.) · 2017 · randomized
Muscle & Nerve 47(3):416-423 (Attie KM, Borgstein NG, Yang Y, et al.) · 2013 · randomized
PLOS ONE (Cadena SM, Bogdanovich S, Khurana TS, et al.) · 2026 · preclinical
others in Anabolic & Receptor