ACE-031: A Decoy Receptor Is Only as Selective as the Receptor It Copies

Most compounds in a research catalogue are molecules you can draw. ACE-031 is a recombinant fusion protein: the extracellular domain of the activin type IIB receptor joined to an immunoglobulin G1 Fc region [1].

The design is a decoy. ActRIIB normally sits on the cell surface and receives myostatin, so a soluble copy of that domain mops up the ligand before it reaches the real receptor. The logic is clean. The problem sits in the premise: ActRIIB does not bind only myostatin.

That promiscuity is why the clinical programme stopped, why the newer molecules in this class are engineered to be pickier, and why the compound remains a useful reference material for anyone studying ligand traps. Kimera supplies ACE-031 as 1 mg of lyophilised powder.

Everything below reports findings from animal studies and clinical trials. This material is for research use only, not for human or veterinary use.

What ACE-031 is

The molecule is a protein, not a peptide, and that distinction runs through everything in this article.

Identity data, and what is missing

Three fields sit empty here: no CID, no InChIKey, no molecular formula that means anything. A glycosylated dimeric protein has a mass distribution rather than a mass, because glycan occupancy and composition vary between molecules in one batch.

Why the Fc is there

The Fc region does two jobs. It drives dimerisation, which presents two ligand-binding domains and raises avidity. It also engages the neonatal Fc receptor, which recycles the protein and extends its circulating half-life from minutes to days.

That second property is the whole reason a fusion protein exists rather than a soluble domain on its own.

Ramatercept, and the naming problem

Three strings reach this compound: ACE-031, ramatercept as the nonproprietary name, and ACVR2B-Fc or ActRIIB-Fc as the descriptive form. PubMed indexes them separately, and the descriptive forms return the widest set because academic groups rarely use the development code (PubMed).

Search all four strings. A search on the development code alone finds the trial and misses the biology.

Why a decoy receptor is not selective

ActRIIB is a type II receptor in the TGF-beta superfamily, and it receives several ligands: myostatin, activin A and activin B, GDF11, and some bone morphogenetic proteins.

The trap inherits the receptor’s promiscuity

A soluble copy of the binding domain binds what the domain binds. Selectivity for myostatin was never a property of this molecule, and every downstream result has to be read with the other ligands in view.

The 2026 literature makes the point directly. One group built a modified trap, RKER-012, specifically for BMP-sparing properties [5]. That design goal only matters if the unmodified version fails to spare them.

The engineered successors

A 2021 study described ActRIIB:ALK4-Fc, a heteromeric trap built to narrow the ligand profile. It alleviated muscle dysfunction in murine models of neuromuscular disorders [4]. The design intent is explicit: keep the muscle effect, drop the rest.

Reading that engineering effort backwards tells you what the original molecule’s limitation was.

The ligand list decides the safety profile

Ligand traps fail in ways that follow their binding profile rather than their intended target. BMP signalling maintains vascular endothelium, so a trap that catches BMPs has a plausible route to vascular effects.

That reasoning is retrospective and it is not proof. It does explain why the successor molecules were engineered around BMP sparing rather than around potency [4][5], which is where a design team goes when the problem is selectivity rather than strength.

The myostatin rationale

Myostatin is a TGF-beta family member that negatively regulates skeletal muscle mass. Remove it genetically and animals gain muscle, which is the observation the whole field is built on.

Several ways to block it

A 2011 review surveys the approaches: blocking antibodies, decoy receptors, propeptides [15]. It also records that a clinical trial of the anti-myostatin antibody MYO-029 failed to reach a significant outcome in muscular dystrophies. That context belongs in any reading of this class.

Why ACE-031 came out of that programme

Antibodies bind one ligand. A decoy receptor binds whatever its parent receptor binds, which looked like an advantage when the goal was maximum pathway blockade rather than precision.

ACE-031 therefore represents a deliberate design choice rather than an oversight. The field chose breadth, measured the consequences in people, and then spent a decade engineering the breadth back out [4][5].

Combination approaches exist

A 2017 study combined myostatin-targeting siRNA with an ActRIIB-Fc fusion protein and reported increased masseter muscle mass and fibre size [16]. Two mechanisms against one pathway, in one experiment.

What the animal work shows

Every result below belongs to its model.

Primates

A 2026 study in common marmosets reported increased muscle mass and strength [3]. Non-human primate data is rare in this catalogue, and it is the strongest species evidence available for the compound.

Cachexia and wasting models

A 2019 study characterised the systemic activin response to pancreatic cancer and drew implications for cachexia therapy [7]. Two years earlier, a study reported that ACVR2B/Fc counteracted chemotherapy-induced loss of muscle and bone mass [8]. A 2020 study reported attenuated muscle wasting in ischaemic heart failure without compromising cardiac function [6].

Bone and fat

In ovariectomised mice, a 2022 study reported reduced bone loss and reduced fat gain [9]. Bone effects recur across this literature, which is consistent with the trap catching BMP-family ligands as well as myostatin.

Heterotopic ossification

Two studies delivered ACVR2B-Fc from engineered stem cells in a mouse model of fibrodysplasia ossificans progressiva, reporting reduced heterotopic ossification [13]. A 2025 follow-up added rapamycin to the same approach [14]. That application uses the BMP binding as the point rather than treating it as an off-target problem.

The human studies

Healthy volunteers

A 2013 single ascending-dose study examined the compound in healthy volunteers [2]. It established the pharmacokinetic and pharmacodynamic profile that the later trial built on.

The Duchenne trial, and why it stopped

A randomised, double-blind, placebo-controlled ascending-dose trial gave the compound subcutaneously to ambulatory boys with Duchenne muscular dystrophy every two to four weeks [1]. Safety was the primary objective.

No serious or severe adverse events occurred in that trial. The study stopped after the second dosing regimen because of potential safety concerns: epistaxis and telangiectasias [1]. A trend toward maintenance of six-minute walk distance appeared in the treated group. The trial ended before anyone could test it properly.

Reading the stop decision

Nosebleeds and small dilated vessels are vascular effects, not muscle effects. They point at ligands other than myostatin, which brings the discussion back to the trap’s binding profile. The clinical result and the mechanistic criticism are the same observation seen from two directions.

How laboratories detect it

This is where a fusion protein differs most from every other compound in the catalogue, and the methods come from anti-doping science.

Immunoaffinity plus mass spectrometry

A 2018 study detected two related ActRII-Fc fusion proteins in serum by immunoaffinity purification, tryptic digestion and LC-MS/MS [11]. The method captures the protein, cuts it into peptides, and identifies it from those peptides rather than from an intact mass.

Immunoassay and electrophoresis

A 2022 study reported an automated capillary immunoassay for activin receptor type IIA and IIB-Fc fusion proteins [10]. A 2018 study validated a sensitive electrophoretic method for the same class in human blood [12].

Three orthogonal methods exist here, because no single one suffices for a protein whose mass varies with glycosylation.

Designing an experiment with ACE-031

Decide which ligand you are blocking

The trap catches several ligands, so an experiment attributing an effect to myostatin needs an independent line of evidence (PubMed). A myostatin-specific antibody arm, or a comparison against a narrowed trap, does that work.

Account for the Fc

The Fc region is not inert. It binds Fc receptors on immune cells and the neonatal Fc receptor in endothelium, so an Fc-only control separates trap effects from Fc effects. That control is standard in fusion protein work and often missing from catalogue-material experiments.

Confirm activity, not just presence

A protein can be present and inactive. Aggregation, denaturation and freeze-thaw damage all leave the protein detectable by immunoassay while removing its function. Run a binding or reporter assay before concluding that a null result means the biology is absent.

Include a positive control the assay can see

A ligand trap produces no signal of its own. Its effect appears only as the absence of ligand signalling, which makes a positive control mandatory: add the ligand, show the response, then show ACE-031 removing it.

Without that three-arm design, a flat result cannot distinguish a working trap from a dead one.

Report the material fully

State the source, the expression system if known, the lot, and the reconstitution history. For a protein those details carry the reproducibility burden that a structure carries for a small molecule.

Characterising a protein, not a molecule

What identity means here

A small molecule’s identity is its structure. For a recombinant protein, identity is a sequence plus a set of post-translational modifications plus a higher-order structure.

A certificate that reports only purity by SDS-PAGE describes how much of the sample is the right size. It says nothing about sequence, glycosylation or folding.

Why the certificate question is sharper here

A small-molecule certificate can be checked against a public record. This compound has no such record, so the certificate is the only identity statement available, and its completeness is the whole basis for trusting the material (PubMed).

That asymmetry is worth stating plainly. Every other article in this library tells readers to check a supplier claim against PubChem. This one cannot.

Size-exclusion chromatography earns its place

One measurement does more work than any other for a protein: size-exclusion chromatography reports monomer, dimer and higher aggregate in a single run.

For an Fc fusion the expected species is the disulfide-linked dimer. A trace showing substantial higher-order material means the preparation aggregated, and aggregation is both a potency problem and an immunogenicity problem in animal work.

What a full characterisation would include

Intact mass by mass spectrometry, peptide mapping for sequence coverage, glycan profiling, size-exclusion chromatography for aggregation state, and a binding or cell-based assay for potency. Potency for a protein needs measuring rather than inferring from purity. That single point separates this material from everything else in the catalogue.

Batch documentation for catalogue material sits in the certificate of analysis database.

Endotoxin is a protein-specific question

Recombinant proteins carry endotoxin risk that synthetic peptides do not, because they come from cell culture. Endotoxin activates immune pathways at picogram levels and confounds any inflammatory or immune readout.

Ask whether endotoxin was tested and at what limit. For cell work with immune cells, that number matters more than the last decimal place of purity.

Expression system matters

Glycosylation depends on the cell line that made the protein. Material from CHO cells differs from material from HEK cells in glycan structures, and those differences change half-life and can change immunogenicity. A protein without a stated expression system stays incompletely described.

Physicochemical properties and handling

Reconstitution for a protein differs from a peptide

A peptide tolerates rough handling that a protein does not. Peptides have no tertiary structure to lose, so a vortexed peptide solution is still the same molecule. A vortexed protein solution may contain denatured material that no longer binds anything.

Add diluent slowly against the vial wall and let the cake dissolve without agitation. Give it time rather than energy: ten minutes of standing beats ten seconds of shaking, and the difference shows up in every downstream measurement.

Aggregation is the failure mode

Proteins aggregate at air-liquid interfaces, under shear, and during freeze-thaw. Aggregates are not simply lost material: they can change what an assay reports, and they can raise immunogenicity in animal work.

Swirl rather than vortex. Avoid repeated freezing. If a reconstituted solution turns cloudy or shows visible particles, discard it rather than filtering and hoping.

Adsorption at low concentration

At cell-work concentrations, a protein adsorbs to tube walls the same way a peptide does. A 1 mg vial leaves little margin for that loss. Carrier protein in the diluent blocks it.

Reading the trial literature carefully

Trial results belong to the trial

The Duchenne study enrolled ambulatory boys, used a defined dosing schedule and ended early [1]. Its findings describe that population under those conditions.

The healthy-volunteer study answers a different question again: what the compound does to pharmacodynamic markers in people without the disease [2]. Neither study licenses a general claim about muscle.

The trial that did not happen

Nothing followed the 2017 report for this molecule. The programme stopped, and the compound’s clinical record ends there.

That absence shapes what the research material is good for. It is a well-documented tool for studying ligand traps, with a published human safety signal attached, and it is not a candidate anyone is still advancing.

Where the animal and human results diverge

Animal models produced consistent muscle gains across species, including marmosets [3][6][8]. The human programme stopped for reasons unrelated to muscle. A compound can work on its intended endpoint and still fail on a different axis, and this one is a clean example.

What a replication would need

Any new in vivo work should measure the vascular endpoints alongside the muscle ones. The 2017 trial made those findings visible only because clinicians were watching for adverse events [1], and an animal study that measures muscle alone cannot see them.

Where ACE-031 sits among catalogue peptides

The size range across those four spans fifty-fold, and the characterisation requirements scale with it. Further reading sits in the peptides research library.

What this literature does not establish

That the muscle effect is myostatin-specific

The trap binds several ligands, and no published experiment with this molecule isolates the myostatin contribution. The engineered successors exist to address that [4][5].

That the vascular findings are understood

Epistaxis and telangiectasias stopped a trial [1]. Nobody has established the mechanism behind them for this molecule, and attributing them to a specific ligand remains inference.

That research material behaves like the clinical protein

Sequence identity does not guarantee equivalence for a glycoprotein. Two preparations of the same sequence from different expression systems differ in glycan structures, and those differences change clearance and can change activity.

That animal muscle gains predict function

Mass and strength rose in marmosets [3]. Muscle mass and useful function are different endpoints, and the trial that could have tested the distinction in people ended early [1].

That catalogue material matches trial material

Clinical ACE-031 was a defined manufactured product with a controlled expression system and full characterisation. Research-grade material meets a different standard. A paper’s results belong to the protein that paper used.

Frequently asked questions

Is ACE-031 a peptide?

No. It is a recombinant fusion protein of around 110 kDa, roughly twenty times the size of the largest peptide in this catalogue.

What does it bind?

Myostatin and other ActRIIB ligands including activins, GDF11 and some BMPs. Selectivity for myostatin was never established [4][5].

Why did the Duchenne trial stop?

Potential safety concerns, specifically epistaxis and telangiectasias, after the second dosing regimen [1]. The published report records none that were serious or severe.

Is ACE-031 the same as luspatercept or sotatercept?

No, though all three are ActRII-Fc traps and the detection literature covers them together [11]. Luspatercept is a modified ActRIIB-Fc and sotatercept is ActRIIA-Fc, and both are approved medicines that this catalogue does not carry.

Why is there no PubChem entry?

Recombinant proteins are not registered as small molecules. Identity comes from sequence and characterisation data rather than from a structure record.

How do laboratories confirm it?

Immunoaffinity capture with tryptic digestion and LC-MS/MS, capillary immunoassay, or validated electrophoresis [10][11][12].

What is the most common handling error?

Vortexing. Shear and foaming denature protein, and the damage does not show up until an assay reads low.

References

1. Campbell C, McMillan HJ, Mah JK, 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;55(4):458-464. PubMed DOI
2. Attie KM, Borgstein NG, Yang Y, et al. A single ascending-dose study of muscle regulator ACE-031 in healthy volunteers. Muscle Nerve. 2013;47(3):416-23. PubMed DOI
3. Cadena SM, Bogdanovich S, Khurana TS, et al. ACE-031, a soluble activin type IIB receptor, increases muscle mass and strength in the common marmoset. PLoS One. 2026;21(2):e0342666. PubMed DOI
4. Li J, Fredericks M, Cannell M, et al. ActRIIB:ALK4-Fc alleviates muscle dysfunction and comorbidities in murine models of neuromuscular disorders. J Clin Invest. 2021;131(4):e138634. PubMed DOI
5. Babbs RK, Ishimwe J, Materna C, et al. RKER-012, a modified ActRIIB-Fc ligand trap with BMP sparing properties, attenuates pathological features of experimental pulmonary arterial hypertension. Front Cardiovasc Med. 2026;13:1827438. PubMed DOI
6. Szabo Z, Vainio L, Lin R, et al. Systemic blockade of ACVR2B ligands attenuates muscle wasting in ischemic heart failure without compromising cardiac function. FASEB J. 2020;34(8):9911-9924. PubMed DOI
7. Zhong X, Pons M, Poirier C, et al. The systemic activin response to pancreatic cancer: implications for effective cancer cachexia therapy. J Cachexia Sarcopenia Muscle. 2019;10(5):1083-1101. PubMed DOI
8. Barreto R, Kitase Y, Matsumoto T, et al. ACVR2B/Fc counteracts chemotherapy-induced loss of muscle and bone mass. Sci Rep. 2017;7(1):14470. PubMed DOI
9. Puolakkainen T, Rummukainen P, Pihala-Nieminen V, et al. Treatment with soluble activin type IIB receptor ameliorates ovariectomy-induced bone loss and fat gain in mice. Calcif Tissue Int. 2022;110(4):504-517. PubMed DOI
10. Desharnais P, Naud JF. Detection of activin receptor type IIA and IIB-Fc fusion proteins by automated capillary immunoassay. Drug Test Anal. 2022;14(11-12):1938-1951. PubMed DOI
11. Walpurgis K, Thomas A, Lange T, et al. Combined detection of the ActRII-Fc fusion proteins sotatercept and luspatercept in serum by means of immunoaffinity purification, tryptic digestion, and LC-MS/MS. Drug Test Anal. 2018;10(11-12):1714-1721. PubMed DOI
12. Martin L, Zouhiri N, Audran M, et al. A validated, sensitive electrophoretic method for the detection of activin receptor type II-Fc fusion proteins in human blood. Drug Test Anal. 2018;10(11-12):1722-1730. PubMed DOI
13. Gao P, Inada Y, Hotta A, et al. iMSC-mediated delivery of ACVR2B-Fc fusion protein reduces heterotopic ossification in a mouse model of fibrodysplasia ossificans progressiva. Stem Cell Res Ther. 2024;15(1):83. PubMed DOI
14. Gao P, Inada Y, Lopez-Iniesta MJ, et al. Combined rapamycin and iPSC-derived mesenchymal stromal cell-mediated delivery of ACVR2B-Fc fusion protein reduces heterotopic ossification in a mouse model of fibrodysplasia ossificans progressiva. JBMR Plus. 2025;9(6):ziaf068. PubMed DOI
15. Sunada Y. [Anti-myostatin antibody therapy for myopathies]. Rinsho Shinkeigaku. 2011;51(11):1157-9. PubMed DOI
16. Bayarsaikhan O, Kawai N, Mori H, et al. Co-administration of myostatin-targeting siRNA and ActRIIB-Fc fusion protein increases masseter muscle mass and fiber size. J Nutr Sci Vitaminol (Tokyo). 2017;63(4):244-248. PubMed DOI