Myostatin inhibition has produced more disappointment than approved medicine. Antibodies, receptor traps, and ligand scavengers have all moved muscle mass in animals; almost none have converted that into durable functional benefit in humans. Against that backdrop, MID-35 is interesting for a narrower reason than the marketing usually suggests. It is not a better drug candidate than the antibodies. It is a demonstration that a 16-residue synthetic peptide, built entirely from D-amino acids, can hold the myostatin-blocking activity of a much larger endogenous protein domain and survive proteolysis long enough to act.
This page covers what the primary literature actually establishes about the MID-35 peptide: where the sequence came from, what its potency numbers mean, where its selectivity breaks down, what is genuinely unknown, and which analytical questions matter most when characterizing a lot. It also documents the errors that circulate about this compound, including two in the primary literature itself.
What MID-35 Is
MID-35 (myostatin inhibitory D-peptide-35) is a 16-mer retro-inverso peptide first reported by Takayama and colleagues in 2022 in ACS Medicinal Chemistry Letters. Its published sequence is lrxkrwirxkiwriyw-amide, where lowercase letters denote D-configuration residues and x denotes D-cyclohexylglycine (D-Chg), an unnatural residue occupying positions 3 and 9. The chain carries a free N-terminus and a C-terminal amide.
| Property | Value |
|---|---|
| Full name | Myostatin inhibitory D-peptide-35 |
| Class | Retro-inverso D-peptide, myostatin (GDF-8) inhibitor |
| Length | 16-mer |
| Sequence | H-D-Leu-D-Arg-D-Chg-D-Lys-D-Arg-D-Trp-D-Ile-D-Arg-D-Chg-D-Lys-D-Ile-D-Trp-D-Arg-D-Ile-D-Tyr-D-Trp-NH₂ |
| Molecular formula | C₁₁₈H₁₈₄N₃₄O₁₇ |
| Molecular weight | 2350.99 g/mol (free base, average mass) |
| C-terminus | Amide |
| Non-natural residues | 2 (D-cyclohexylglycine at positions 3 and 9) |
| Basic sites | 7 (4 × D-Arg, 2 × D-Lys, free N-terminus) |
| Net charge at pH 7 | Approximately +6; no acidic residues |
| CAS number | None assigned |
| Origin | Mouse myostatin prodomain-derived SAR program |
| Parent compound | MIPE-1686 (peptide 8a), via retro-inversion plus two arginine substitutions |
Two details in that table are worth pausing on, because vendor listings across this category get both wrong routinely. First, several sources circulate a molecular weight near 2349.5; recomputing residue-by-residue from the published sequence with a C-terminal amide gives 2350.99. Second, the compound has no assigned CAS number. Numbers do circulate, attached to vendor catalog designations rather than to this molecule, and any listing that supplies one is supplying something it cannot substantiate.
The sequence itself is worth verifying rather than copying, because at least one published figure gets it wrong. See the errors section below.
The Biology: Myostatin as a Brake
Myostatin, also called growth differentiation factor 8 (GDF-8), is a TGF-β superfamily member and the best-characterized negative regulator of skeletal muscle mass. McPherron, Lawler, and Lee established the phenotype in 1997: mice lacking the gene develop dramatically enlarged muscles.
Mature myostatin signals by binding activin type II receptors (ACVR2 and ACVR2B), which recruit type I receptors (ALK4/ALK5) and phosphorylate Smad2/3. The activated complex translocates to the nucleus and suppresses the transcriptional program supporting muscle growth, while promoting the atrophy-associated E3 ubiquitin ligases atrogin-1 (Fbxo32) and MuRF-1 (Trim63). Blocking that axis is the shared logic behind every agent in the class, including the ACE-031 decoy receptor, which intercepts ligands at the receptor rather than at the ligand surface.
Myostatin is synthesized as a latent precursor. The prodomain remains associated with the mature dimer and keeps it inactive until proteolytic release. That prodomain is the natural inhibitor, and it is the sequence source MID-35 was carved from.
One species difference governs how far mouse results travel. Latres and colleagues measured circulating ligand levels across species and found myostatin 4- to 18-fold lower in monkeys, rats, and humans than in mice, while activin A runs 3- to 4-fold higher in monkeys and humans. The consequence was measured rather than theorized: in mice, blocking myostatin alone grew the tibialis anterior 20.1% while blocking activin A alone produced a non-significant 5.0%. In cynomolgus monkeys, neither antibody alone did much, and only the combination worked. Mice run on myostatin. Primates run on both. Any rodent result from a myostatin-directed agent is therefore an upper bound on what the same mechanism would produce in a primate, not a floor.
Binding Mode: Where MID-35 Docks
MID-35 does not bind the receptor. It binds the ligand, at the surface mature myostatin uses to engage its type I receptor. That distinction carries weight in this class, because most of the agents that ran into trouble clinically acted at the type II receptor instead.
Asari and colleagues characterized the binding mode of the 23-mer parent by docking and reported six myostatin residues as critical to the interaction: Leu20, Asn41, Leu52, Leu60, Met101, and Ser109. The 23-mer parent bound with a dissociation constant near 30 nM. Downstream, MID-35 blocks Smad2/3 nuclear translocation, though direct measurement of Smad2 phosphorylation in C2C12 cells showed only a marginal decrease. That gap between reporter-level inhibition and phosphorylation-level inhibition is discussed in the selectivity section.
The mechanistic consequence is worth stating for anyone comparing modalities. ACE-031 and bimagrumab act at ActRIIB, which also binds BMP9 and BMP10 and regulates vascular function through Smad1/5/8. That cross-inhibition is the mechanism to which the ACE-031 bleeding events in Duchenne trials have been attributed. MID-35 does not bind ActRIIB, so the BMP9 mechanism should not apply. Two independent sources reach that conclusion, and a cachexia study reported no vascular abnormalities such as bleeding after intramuscular administration to mouse hind legs. This is the one clean point in MID-35’s favor on safety mechanism, and it is a mechanistic argument rather than a safety dataset.
Design Lineage: From a Protein Domain to 16 Residues
MID-35 did not appear fully formed. It is the endpoint of roughly a decade of chain-shortening work, and each step is documented.
2015, the minimum prodomain fragment. Takayama et al. identified minimum peptides of 24 and 23 residues from the mouse myostatin prodomain capable of inhibiting human myostatin. The 23-mer was a genuine but weak inhibitor, with a reported IC₅₀ near 3.56 µM. Because murine and human myostatin are identical in mature sequence, the mouse-derived fragment was a legitimate starting point for a human target.
2019, chain shortening plus two design decisions that still show in the final molecule. Structure-activity work produced peptide 8a, later designated MIPE-1686: a 16-mer with unnatural residues and a β-sheet propensity, roughly twice as potent as the 22-mer it descended from, with a reported IC₅₀ near 0.13 µM. Intramuscular injection into mdx mice, a Duchenne muscular dystrophy model, increased tibialis anterior mass by roughly 14% over saline controls at day 42 and improved hindlimb grip strength.
Two supporting experiments from that year explain the final structure. A D-form scan substituted single L-residues with their D-counterparts across the 23-mer parent and tested each at 3 µM. Most single substitutions reduced inhibition, several sharply; only two positions improved on the parent. Piecemeal D-substitution therefore failed, which is why the program later went to full retro-inversion, a transformation that flips every stereocenter at once while preserving the spatial arrangement of the side chains. Separately, an aliphatic residue scan replaced isoleucine and leucine positions with cyclohexylglycine or phenylglycine. Several substitutions matched or beat the parent, and cyclohexylglycine is the residue that survived into MID-35 at positions 3 and 9.
2020, the stability problem. MIPE-1686 held up well against aminopeptidase N, chymotrypsin C, and trypsin 3 in isolated-enzyme work, but a linear L-peptide with an unprotected N-terminus remained a liability for in vivo development.
2022, the retro-inverso solution. The group synthesized a panel of retro-inverso versions of the 16-mer. MID-35 emerged as the arginine-containing member of that panel, formed from the retro-inverso parent by two substitutions, D-Ser5 to D-Arg and D-Gln8 to D-Arg. Those two positions are what distinguish MID-35 from its immediate precursor, and they are the reason the finished molecule carries four D-arginines. Related analogs (MID-36, MID-39) mapped the size floor: below roughly 15 residues, potency in the 0.2 to 0.3 µM band was not reachable.
In the 2022 head-to-head, MID-35 was reported at an IC₅₀ near 0.19 µM against myostatin, against roughly 0.26 µM for MIPE-1686 measured alongside it.
A caution worth stating plainly, because it gets mangled constantly in secondary write-ups: the 0.13 µM figure for MIPE-1686 from 2019 and the 0.26 µM figure from 2022 are the same peptide in different assay runs. Cross-paper IC₅₀ comparisons in this series are not meaningful. Only the within-paper comparison is.
Selectivity: Where the “Selective Myostatin Inhibitor” Label Fails
MID-35 is frequently described as a selective myostatin inhibitor. The primary literature does not support that description. Takayama and colleagues reported dose-response inhibition against four TGF-β superfamily ligands in the same 2022 assay run:
| Ligand | MID-35 IC₅₀ | MIPE-1686 IC₅₀ |
|---|---|---|
| Myostatin (GDF-8) | 0.19 ± 0.05 µM | 0.26 ± 0.04 µM |
| GDF-11 | 0.63 ± 0.05 µM | 1.4 ± 0.2 µM |
| Activin A | 0.89 ± 0.09 µM | 1.4 ± 0.5 µM |
| TGF-β1 | 1.6 ± 0.2 µM | 6.7 ± 0.6 µM |
Read that as a selectivity statement rather than a potency statement. The four values span roughly eightfold. A compound whose weakest measured target sits within an order of magnitude of its strongest is a broad TGF-β family inhibitor with a preference, not a selective one.
Structurally this is unsurprising. GDF-11 shares roughly 90% mature-domain identity with myostatin and uses the same receptor machinery, so a peptide binding the myostatin ligand surface has an obvious route to cross-reactivity. Almost nothing in this class separates the two.
The activin A result is contested, and researchers should know that before designing around it. Hanada and colleagues, reporting in Cancer Science the same year, used HepG2 cells carrying an (SBE)₄-luc Smad-responsive reporter and a single 3 µM peptide concentration. In that assay, MID-35 suppressed myostatin-, GDF-11- and TGF-β-induced reporter activity but did not inhibit activin A. The two results are not reconcilable from the published record. A concentration of 3 µM sits more than threefold above the 0.89 µM IC₅₀ reported by Takayama, so an assay at that concentration should have detected clear inhibition. The author groups overlap. No third measurement exists anywhere in the indexed literature.
Different cell backgrounds, HepG2 against HEK293, remain the only proposed explanation, and nobody has tested it. Cite whichever source you use, state which assay it came from, and treat the activin A question as open.
The consequence is not academic. Given the species data above, if MID-35 does not touch activin A then every in vivo result on file comes from the one species where myostatin-only blockade looks best. If it does inhibit activin A at 0.89 µM, it covers both of the ligands that matter in primates, but it does so without selectivity, and the safety argument in this class has always rested on selective dual blockade outperforming broad blockade.
A second nuance from the Hanada work deserves attention. When Smad2 phosphorylation was measured directly in C2C12 cells, MID-35 produced only a marginal decrease, while the ALK4/5/7 kinase inhibitor SB-431542 abolished it. A ligand-binding peptide and a receptor-kinase inhibitor are not interchangeable tools. Reporter-level inhibition does not translate proportionally to suppression of the immediate phosphorylation event.
Preclinical In Vivo Record
All published in vivo work is rodent, and essentially all of it is local intramuscular administration rather than systemic dosing. The complete published record is small enough to tabulate.
| Report | Route | Model | Reported outcome |
|---|---|---|---|
| Takayama 2022 | Intramuscular, left tibialis anterior, single administration | 8-week male C57BL/6J | Treated tibialis anterior 133 ± 10% of the contralateral saline muscle at day 28 |
| Morito 2026 | Intramuscular, left tibialis anterior, single administration | Young 8-week, adult 20-week, aged 72-week | Separation from control from day 14, sustained to day 84 |
| Morito 2026, lower level | Intramuscular, left tibialis anterior, single administration | Young and adult | Significant in young mice at day 84; not significant in adult mice at day 28 |
| Hanada 2022 | Intramuscular, gastrocnemius, three administrations | Lewis lung carcinoma cachexia, C57BL/6J | Gastrocnemius weight, grip strength and fiber area all increased |
| Michiue 2023 | Transdermal iontophoresis, three applications | C57BL/6J | Tibialis anterior 1.25-fold at day 42; gastrocnemius in the same limb unchanged |
Every rodent study uses the same internal control. Compound goes into the left tibialis anterior and saline into the right tibialis anterior of the same animal, with paired statistics. Same animal, same circulation. Across young, adult and aged mice, checked on weight, myofiber diameter, centralized nuclei, Pax7 staining, transcript panels and lipid panels out to 84 days, the saline limb held as a flat baseline. Michiue adds that gastrocnemius in the same treated limb did not grow significantly while the muscle under the electrode did.
That design has a known limit worth stating: it cannot detect an effect that hits both limbs equally, and no MID-35 study has compared a saline-injected limb against a fully untreated animal. Work on other molecules shows the design class is not blind by construction. In one experiment on a different compound family, a locally injected receptor trap moved the uninjected limb while a locally injected ligand trap did not, in the same animals and the same readouts. The control has been run successfully elsewhere. It has not been run on MID-35.
Kinetics and durability. The most informative pharmacology arrived in 2026, when Morito and colleagues published a detailed time-course in ACS Pharmacology & Translational Science across young, adult, and aged mice. Differentiation markers (Pax7, Myod1, Myog) rose and atrophy markers (Trim63, Fbxo32) fell within 3 days. Measurable weight gain did not appear until day 14. The hypertrophy was sustained for 12 weeks after a single treatment.
The growth signature is regenerative, and the supporting figures quantify it. In healthy adult muscle, myofiber nuclei sit at the fiber periphery. Nuclei in the center mark fibers that have been broken down and rebuilt, so counting centralized nuclei counts recently remodeled fibers. In the supporting data, treated muscle reached roughly 35% centralized nuclei in young mice at day 84 and roughly 53% in adult mice at day 28, against roughly 5 to 6% in the saline limb. Both limbs received a needle, so injection trauma is controlled. Values are read from the published figures rather than printed in text.
The transcript timeline matches. At day 7, Pax7 message rose roughly 65% and myogenin roughly eightfold, while Myod1 was unchanged. By day 28, Pax7 immunostaining had risen about 1.5-fold. Satellite cells activate in the first week and structural remodeling continues for months.
Duration of effect is not duration of exposure. A single local injection still separated from control at 12 weeks. That is easy to misread as the peptide persisting for three months, and it is not what the data show. The compound triggers a satellite-cell-associated remodeling process, and the remodeled tissue persists after the compound is gone. No published work measures how long MID-35 itself remains in tissue or in circulation.
Age dependence. The 2026 work also showed that the lower of the two administered levels separated from control in young animals but not in adult animals at the earlier timepoint, while the higher level worked in both. Sensitivity to this compound declines with age in the same assay.
The sphingolipid finding and the aging signal. The same study tracked sphingosine 1-phosphate (S1P), a bioactive sphingolipid implicated in muscle mass regulation, using quantitative LC-MS/MS with internal standards across eleven analytes. S1P rose significantly at day 3 in young and adult mice, consistent with a role in satellite cell activation. In aged mice, that increase did not occur. Since sarcopenia is the rationale most often attached to this compound, a mechanism that partially fails in aged tissue is a consequential open question.
Cancer cachexia. In tumor-bearing mice, intramuscular MID-35 alleviated skeletal muscle atrophy and outperformed the earlier prodomain-derived peptide-2. Combined with anamorelin, a ghrelin receptor agonist mechanistically adjacent to growth hormone secretagogues such as ipamorelin, the pairing increased food intake, maximized grip strength, and extended survival relative to either agent alone. Appetite and growth-hormone-axis stimulation and direct ligand blockade appear to address non-overlapping arms of the wasting phenotype.
A useful negative from the same work: in a cell viability assay, MID-35 at 3 µM left Lewis lung carcinoma cell growth at roughly 85% of control, which was not significant. In a cachexia context the obvious concern is that a growth-promoting agent also feeds the tumor. In that assay it did not. One cell line, one concentration, in vitro only.
Delivery research. Because intramuscular injection is invasive, Michiue and colleagues tested iontophoresis, transdermal delivery driven by weak electric current, in Pharmaceuticals (2023). Fluorescently labeled peptide reached at least 1000 µm into muscle from the skin surface. Three applications produced a statistically significant 1.25-fold tibialis anterior mass increase at day 42. The comparison is honest in the source: the iontophoresis quantity was substantially higher than that used for direct injection to reach a similar effect size, and gastrocnemius weight did not change significantly.
Proteolytic Stability, and What It Does Not Mean
The measured stability is real. MID-35 was reported 99% and 97% intact after 400 minutes in bovine trypsin and α-chymotrypsin respectively. Unlike MIPE-1686, no non-specific binding was observed in that work. Related stability work on the parent 16-mer series measured 14 days intact in human serum, the harshest common matrix, which makes the figure conservative rather than fragile: peptides are generally least stable in serum and most stable in fresh blood.
Protease resistance and circulating half-life are different measurements, and the class literature is explicit that one does not deliver the other. Substituting arginine with D-arginine took one antimicrobial peptide from a 25-minute serum half-life to over 8 hours, which is the encouraging case. The discouraging cases are equally documented. Dermorphin analogues carrying extra D-substitutions were cleaved faster than the parent. D-substituted GRF(1-29)-amide analogues showed no significant difference in plasma half-life after intravenous dosing in rats. Cetrorelix carries five D-amino acids and resists chymotrypsin, pronase and nargase for up to 50 hours at 37 °C in vitro, which is an enzyme-stability figure and not a half-life.
The clearest documented payoff of this design strategy in an approved drug is octreotide, where shortening somatostatin from 14 to 8 residues and swapping L for D residues raised plasma half-life from a few minutes to roughly 1.5 hours. Hours, not days. Treat any claim that D-configuration confers long circulation as unsupported unless a half-life was actually measured.
Disposition: What Is Known at Class Level
Nothing in this section was measured on MID-35. Every item is class-level literature on molecules of similar size, charge and stereochemistry, and it is included because the MID-35 record itself is empty here. Read it as context for designing an experiment, not as a description of this compound.
Serum protein binding. Anti-doping method development on MIPE-1686, MID-35’s immediate precursor, recovered only 7.7% and 8.4% of spiked peptide from human serum at two concentrations, against 20 to 48% for five comparator peptides. Degradation was ruled out because the peptide survived 14 days intact, and tube binding accounted for no more than 15%. The authors attributed the loss to combined binding to serum proteins and sample tubes. Several related peptides were undetectable at neutral and basic pH for the same reason. MID-35 is more cationic than MIPE-1686. It has never been run through this assay.
Renal filtration. Molecules under roughly 5 kDa that are not bound to plasma proteins pass the glomerular filter completely and are excreted renally, while molecules above roughly 50 kDa are barely filtered. At 2.35 kDa, MID-35 sits far below that threshold. The unbound condition matters and pulls against the paragraph above: either the peptide is free, in which case it is filtered quickly, or it is protein-bound, in which case it is not free drug. Both routes point the same way on persistence while disagreeing on mechanism.
Cationic peptides behave distinctly in the proximal tubule. Below roughly 1.8 nm effective radius, molecules are freely filtered, and cationic species are filtered more efficiently than anionic ones. In the proximal tubule, cationic peptides are then reabsorbed more efficiently, attributed to positive groups binding anionic sites on the brush border membrane. The effect is established enough that co-infusing lysine and arginine is standard kidney protection in clinical peptide radionuclide therapy. Radiopharmaceutical groups building residualizing labels around three D-arginines reported very high kidney uptake and engineered away from that chemotype deliberately. Peptides differing only in D versus L configuration showed strikingly different renal accumulation and retention. MID-35 carries four D-arginines and two D-lysines.
Lysosomal handling. The normal fate of a peptide taken up by the proximal tubule is lysosomal breakdown to amino acids. D-configured peptides substantially resist that breakdown. A single D-residue at the C-terminus conferred complete resistance to rat liver lysosomal extract. Cathepsin B, the principal lysosomal endopeptidase, cleaved an all-L construct within an hour and did not degrade the all-D version at all. A retro-inverso D-peptide, the same chemotype as MID-35, showed exceptional stability in lysosomal homogenate. In live animals, a single D-phenylalanine substitution slowed metabolism inside renal cells enough that intermediate species remained detectable. Counterweights exist: peptidase-resistant peptides delivered to the cytosol are eventually degraded in lysosomes, radiolabel retention is partly a property of the chelate rather than the peptide, and D-residues are not intrinsically toxic, with more than a dozen D-amino-acid-containing drugs approved.
Elimination route, from the one tracking study in this chemotype. An all-D retro-inverso peptide of 2860 Da was radiolabeled and followed in rats after intravenous administration. Whole blood fell from 2.262 to 0.189 SUV between 5 and 20 minutes, roughly a 92% drop, while liver held flat. The circulating material was almost entirely intact parent, 96% at 5 minutes and 76% at 20 minutes, with no other labeled species detected. Uptake ranked liver first, then bladder, kidneys, spleen, bone marrow, lungs, bone and brain, with bladder signal continuing to climb past liver by the end of the scan. The authors describe hepatobiliary clearance as the main route, but both hepatic and renal elimination are plainly active. That molecule is not MID-35: it is larger, it carries a cell-penetrating motif far more cationic than anything in MID-35, and it was studied in rats. It is the closest published behavior of the chemotype and nothing more.
Analytical Considerations: Purity Is Not Content
This is where most peptide characterization in this category quietly falls apart, and where MID-35 is a particularly good illustration.
Purity and content are different measurements. HPLC purity reports the fraction of peptide-related material represented by the main peak. It says nothing about how much peptide is actually in the vial. A lot can be 98% pure by HPLC and still be substantially less than 98% peptide by mass, because the remainder is counterion, residual water, and residual solvent.
MID-35 makes this gap unusually wide. The molecule carries seven basic sites: four D-Arg, two D-Lys, and a free N-terminus. Peptides purified by reverse-phase HPLC in trifluoroacetic acid systems isolate as TFA salts, with counterion loading tracking basic site count. At a typical 3 to 5 TFA equivalents, net peptide content lands somewhere near 80 to 87% of vial mass. A 10 mg nominal vial characterized only by HPLC purity may contain closer to 8 to 8.7 mg of actual peptide. For a quantitative inhibition assay, that is the difference between a clean IC₅₀ and an unexplained rightward shift.
This is not a hypothetical concern with this compound family. The primary literature reports every peptide in the series as purified and weighed as a TFA salt, with yields calculated on that basis. Published quantities are therefore salt-form quantities.
Orthogonal methods resolve it. HPLC establishes chromatographic purity. Mass spectrometry confirms identity against the calculated free base of 2350.99 Da, and note that observed salt-form mass will read higher, which is expected rather than a red flag. Elemental analysis provides an independent quantification anchor for net peptide content that chromatography cannot supply. Sequence-level confirmation matters especially here, since D-amino acid content and the two unnatural D-Chg positions are not verifiable by mass alone. L to D substitution is mass-silent, and a peptide synthesized with the wrong stereochemistry weighs exactly the same as the right one while behaving completely differently in a protease-rich environment.
No validated bioanalytical assay exists. As of this writing there is no validated method for quantifying MID-35 in serum, plasma or urine. The anti-doping panel that covers this peptide family names MID-35 explicitly and excludes it, with a stated intention to include it in future studies. That statement is now several years old and remains unfulfilled. Anyone building a quantitative assay for this compound is building it from scratch, and should expect the nonspecific binding losses described above.
Every Kimera lot ships with third-party COA verification, and lot-specific documentation is archived in the public COA archive. Purity figures should always be read against the specific lot rather than a catalog-level claim, and read off the certificate itself rather than from a summary.
Handling. Treat MID-35 as a standard lyophilized research peptide: store the dry solid sealed at controlled room temperature, protected from moisture and light, and limit freeze-thaw cycles. Reconstitution and storage protocols should be validated in-house against the intended assay. Storage guidance is a house recommendation. Analytical documentation is per-lot release testing.
Anti-Doping and Regulatory Context
Agents that block activin receptor signaling are prohibited in sport. The 2026 WADA Prohibited List addresses them under S4.3, “Agents preventing activin receptor IIB activation.” That section is prefaced “including, but not limited to,” and its myostatin-inhibitor examples are agents reducing or ablating myostatin expression, myostatin-binding proteins such as follistatin and myostatin propeptide, and myostatin- or precursor-neutralizing antibodies. MID-35 is not named on the List, and as a 16-mer synthetic peptide it matches none of those three examples literally. It falls under the non-exhaustive language of S4.3, and unambiguously under S0, Non-Approved Substances, having no therapeutic approval in any jurisdiction. Researchers working with athlete populations should treat it as prohibited and confirm current status against the List directly.
Doping-control laboratories have responded accordingly. Walpurgis and colleagues published detection methodology for myostatin inhibitory peptides in Drug Testing and Analysis (2023), with the Takayama peptide series explicitly in scope. For method-development researchers, MID-35 functions as a reference analyte. The same work documented significant analytical losses from nonspecific binding across this peptide class and recorded the working assumption behind its design, that these peptides are administered intramuscularly and therefore probably produce comparably low blood levels. That is an assumption held by the people whose job is finding them in blood, not a measurement.
The Class Translation Record
MID-35 has no clinical data. Its class does, and the pattern is consistent enough to be worth stating precisely rather than gesturing at.
| Agent | Reported outcome |
|---|---|
| MYO-029 (stamulumab), anti-myostatin antibody | Discontinued, lack of efficacy |
| Domagrozumab, anti-myostatin antibody | Two phase 2 Duchenne studies terminated for lack of efficacy |
| Landogrozumab, anti-myostatin antibody | Missed primary endpoint in hip replacement; no survival benefit in pancreatic cancer cachexia |
| PINTA-745, peptibody | Discontinued, lack of efficacy |
| RG6206, adnectin | Discontinued, unlikely to reach primary endpoint |
| ACE-031, ActRIIB decoy receptor | Phase 2 Duchenne terminated after bleeding events attributed to BMP9/BMP10 cross-inhibition |
| Bimagrumab, anti-ActRIIB antibody | Inclusion body myositis program failed. In a sarcopenia dose-ranging study, lean mass rose significantly while every functional endpoint failed, and the indication was discontinued |
| ACE-083, locally acting follistatin trap | Two phase 2 programs terminated. Details below |
| Luspatercept | Approved, for anaemia associated with β-thalassaemia |
The single most relevant comparator is ACE-083, because it is the only agent in this class designed for local intramuscular action rather than systemic exposure, which is the same delivery logic behind every MID-35 rodent study. Its record is now complete and it is unflattering in an instructive way.
In a phase 1 study in healthy volunteers, unilateral injection produced detectable serum drug in only 14 of 252 samples, all on day 1, and the contralateral non-injected muscle behaved like placebo. Muscle volume rose. Strength did not.
Two randomized placebo-controlled phase 2 trials followed, one in facioscapulohumeral muscular dystrophy and one in Charcot-Marie-Tooth disease. Both increased injected-muscle volume significantly. In the muscular dystrophy trial, treated muscle volume rose 14.20% against 5.46% for placebo in the lower limb and 20.41% against 1.07% in the upper arm, and isolated biceps strength by dynamometry rose 32.58% against a 3.54% decline. In the neuropathy trial, total muscle volume rose 13.5% over placebo and contractile volume 23.3%. Then every functional endpoint failed in both: walking distance, walk and run time, stair climb, upper-limb performance, quality of life, and fall risk. Both programs were terminated.
The containment finding from those trials is the part most relevant to a local-delivery compound. Serum concentrations after repeated high-level local injection stayed in the tens of nanograms per millilitre and did not accumulate across six months of dosing; values at day 86 were lower than at day 2 in both muscle groups. In the extension study, the sponsor reported that pharmacokinetic parameters could not be determined because too little quantifiable material was present. The neuropathy trial separately reported no notable effects on the systemic biomarkers this class uses. Local delivery does produce low, measurable, non-accumulating systemic exposure. It does not produce none.
The line that follows from the whole table is short. Muscle mass and muscle function are not the same endpoint, and the only agent in the myostatin pathway approved anywhere is approved for anaemia rather than for muscle.
What the Evidence Does Not Show
Stating the limits precisely is more useful than repeating the positive findings.
- No human data of any kind. There are no clinical trials of MID-35 and no record of human administration. The entire record is cell-based assays and rodent models.
- No pharmacokinetics. No half-life, clearance, Cmax or bioavailability by any route in any species, and no published measurement of MID-35 in blood, plasma, serum or urine. An independent anti-doping group stated in 2023 that pharmacokinetic data for myostatin inhibitory peptides are not available. That remains the position.
- No toxicology of any kind. No LD50, no acute or repeat-dose study, no genotoxicity, no safety pharmacology, in any species. Only one non-muscle tissue has ever been measured in any MID-35 study: heart weight relative to body weight, which was unchanged.
- No immunogenicity data. None, in any system. Class-level work on retro-inverso peptides is mixed and does not support a claim in either direction.
- No serum stability study on MID-35 itself. The longest observation on this specific molecule is 400 minutes in buffered enzyme solution. The 14-day serum figure belongs to the parent series.
- No validated bioanalytical method in any matrix.
- Local, not systemic. Efficacy has been demonstrated by direct intramuscular injection and by transdermal iontophoresis into the tissue underneath. No published work establishes systemic exposure producing whole-body muscle effects. No intravenous, subcutaneous, intraperitoneal or oral study has ever been published.
- Selectivity is incomplete and the activin A result is contested, as described above.
- The aged-tissue signal is unresolved. The S1P response that accompanies hypertrophy in young and adult animals was absent in aged mice, and the lower administered level lost significance in adult animals.
- Class-level translation has been poor, as set out in the table above.
Only five papers in the indexed literature name MID-35 at all. This is a small evidence base, and its size is itself a finding.
Errors That Circulate About MID-35
Two of these are in the primary literature, not in secondary summaries. Anyone building a reference file for this compound should be aware of all four.
- A published supporting figure gives the wrong sequence. One 2022 paper includes a supporting figure captioned as the sequence of MID-35 that shows serine at position 5 and glutamine at position 8. Those are the two positions that define MID-35: it is formed from its precursor by substituting both with arginine. The figure therefore depicts the precursor rather than MID-35. The correct sequence is lrxkrwirxkiwriyw-amide, and it is confirmed independently by the compound’s own patent family and by the detection literature describing the two arginine substitutions explicitly.
- A published molecular weight of 2.100 kDa is wrong. It appears in one delivery paper. Recomputation from the published sequence gives 2350.99 g/mol.
- Cross-paper IC₅₀ comparison. Secondary write-ups routinely compare the 2019 MIPE-1686 figure with the 2022 MID-35 figure. Those are different assay runs. Only within-paper comparisons carry information.
- Invented CAS numbers. Numbers do circulate for this compound. The ones we have traced attach in supplier documents to vendor catalog designations rather than to MID-35, sitting adjacent to unrelated compounds in the same list. No CAS has been assigned.
Researchers comparing modalities in this space may find the broader peptides research library useful for adjacent compound profiles.
Frequently Asked Questions
What is the MID-35 peptide?
MID-35 is a 16-residue retro-inverso peptide built entirely from D-amino acids, derived from the mouse myostatin prodomain and designed to bind mature myostatin at its type I receptor interface. It carries two non-natural D-cyclohexylglycine residues and a C-terminal amide.
Is MID-35 selective for myostatin?
No. Published dose-response data show inhibition of GDF-11, activin A and TGF-β1 within roughly eightfold of the myostatin value. It is a broad TGF-β family inhibitor with a preference for myostatin, and the activin A result is contested between two papers.
What is a retro-inverso peptide?
A retro-inverso peptide reverses the sequence order and inverts every stereocenter from L to D. The result approximately reproduces the parent’s side-chain arrangement in space while presenting a backbone that mammalian proteases cleave poorly. The strategy works best on short, unstructured peptides and often fails on structured ones.
Does MID-35 have a known half-life?
No. There is no published pharmacokinetic study of MID-35 in any species by any route. Protease resistance measured in isolated enzyme solutions is a stability figure, not a half-life, and the class literature contains documented cases where D-substitution did not extend in vivo half-life at all.
Why did the effect last 12 weeks in mice after one injection?
Because the readout is remodeled tissue rather than circulating compound. The supporting data show satellite cell activation within a week and a large rise in centrally nucleated fibers, which is the signature of fibers that have been rebuilt. Persistence of the effect says nothing about persistence of the peptide.
What is MID-35’s CAS number?
It does not have one. Any supplier listing that provides a CAS number for MID-35 is providing something it cannot substantiate.
Research Use Only
MID-35 is supplied strictly for laboratory research and analytical use. It is not a drug, food, cosmetic, or dietary supplement. It is not for human or veterinary use and must not be administered to humans or animals. No statement here has been evaluated by the FDA, and nothing here describes or implies a therapeutic use. Sales are limited to qualified research professionals and institutions.
References
Takayama K, Noguchi Y, Aoki S, et al. Identification of the minimum peptide from mouse myostatin prodomain for human myostatin inhibition. J Med Chem. 2015;58(3):1544–1549. doi:10.1021/jm501170d
Asari T, Takayama K, Nakamura A, Shimada T, Taguchi A, Hayashi Y. Structural basis for the effective myostatin inhibition of the mouse myostatin prodomain-derived minimum peptide. ACS Med Chem Lett. 2017;8(1):113–117. doi:10.1021/acsmedchemlett.6b00420
Takayama K, Asari T, Saitoh M, et al. Chain-shortened myostatin inhibitory peptides improve grip strength in mice. ACS Med Chem Lett. 2019;10(6):985–990. doi:10.1021/acsmedchemlett.9b00174
Takayama K, et al. Enzymatic stability of myostatin inhibitory 16-mer peptides. Chem Pharm Bull. 2020;68(6). doi:10.1248/cpb.c20-00158
Asari T, et al. Structure-activity relationship study on the binding mode of myostatin inhibitory peptides. Bioorg Med Chem. 2021;40:116181. doi:10.1016/j.bmc.2021.116181
Takayama K, Hitachi K, Okamoto H, et al. Development of myostatin inhibitory D-peptides to enhance the potency, increasing skeletal muscle mass in mice. ACS Med Chem Lett. 2022;13(3):492–498. doi:10.1021/acsmedchemlett.1c00705
Hanada K, Fukasawa K, Hinata H, et al. Combination therapy with anamorelin and a myostatin inhibitor is advantageous for cancer cachexia in a mouse model. Cancer Sci. 2022;113(10):3547–3557. doi:10.1111/cas.15491
Michiue K, Takayama K, Taniguchi A, Hayashi Y, Kogure K. Increasing skeletal muscle mass in mice by non-invasive intramuscular delivery of myostatin inhibitory peptide by iontophoresis. Pharmaceuticals. 2023;16(3):397. doi:10.3390/ph16030397
Walpurgis K, et al. Myostatin inhibitory peptides in sports drug testing. Drug Test Anal. 2023;15(11-12):1477–1487. doi:10.1002/dta.3473
Morito K, Nishikawa N, Hitachi K, et al. Myostatin inhibitory D-peptides induce skeletal muscle hypertrophy along with alteration of bioactive sphingolipid metabolism. ACS Pharmacol Transl Sci. 2026;9(6):1544–1553. doi:10.1021/acsptsci.6c00124
Ojima C, et al. Peptide-2 from mouse myostatin precursor protein alleviates muscle wasting in cancer-associated cachexia. Cancer Sci. 2020;111(8):2954–2964. doi:10.1111/cas.14520
McPherron AC, Lawler AM, Lee SJ. Regulation of skeletal muscle mass in mice by a new TGF-β superfamily member. Nature. 1997;387(6628):83–90. doi:10.1038/387083a0
Walker RG, McCoy JC, Czepnik M, et al. Molecular characterization of latent GDF8 reveals mechanisms of activation. Proc Natl Acad Sci USA. 2018;115(5):E866–E875. doi:10.1073/pnas.1714622115
Latres E, Mastaitis J, Fury W, et al. Activin A more prominently regulates muscle mass in primates than does GDF8. Nat Commun. 2017;8:15153. doi:10.1038/ncomms15153
Suh J, Lee YS. Myostatin inhibitors: panacea or predicament for musculoskeletal disorders? J Bone Metab. 2020;27(3):151–165. doi:10.11005/jbm.2020.27.3.151
Glasser CE, Gartner MR, Wilson D, Miller B, Sherman ML, Attie KM. Locally acting ACE-083 increases muscle volume in healthy volunteers. Muscle Nerve. 2018;57(6):921–926. doi:10.1002/mus.26113
Statland J, Campbell C, Desai U, et al. Randomized phase 2 study of ACE-083, a muscle-promoting agent, in facioscapulohumeral muscular dystrophy. Muscle Nerve. 2022;66(1):50–62. doi:10.1002/mus.27558
Thomas FP, Brannagan TH, Butterfield RJ, et al. Randomized phase 2 study of ACE-083 in patients with Charcot-Marie-Tooth disease. Neurology. 2022;98(23):e2356–e2367. doi:10.1212/WNL.0000000000200325
Rooks D, Swan T, Goswami B, et al. Bimagrumab vs optimized standard of care for treatment of sarcopenia in community-dwelling older adults. JAMA Netw Open. 2020;3(10):e2020836. doi:10.1001/jamanetworkopen.2020.20836
Golan T, Geva R, Richards D, et al. LY2495655, an antimyostatin antibody, in pancreatic cancer: a randomized, phase 2 trial. J Cachexia Sarcopenia Muscle. 2018;9(5):871–879. doi:10.1002/jcsm.12331
Wagner KR. The elusive promise of myostatin inhibition for muscular dystrophy. Curr Opin Neurol. 2020;33(5):621–628. doi:10.1097/WCO.0000000000000853
Werle M, Bernkop-Schnürch A. Strategies to improve plasma half life time of peptide and protein drugs. Amino Acids. 2006;30(4):351–367. doi:10.1007/s00726-005-0289-3
Vegt E, de Jong M, Wetzels JFM, et al. Renal toxicity of radiolabeled peptides and antibody fragments. J Nucl Med. 2010;51(7):1049–1058. doi:10.2967/jnumed.110.075101
Böttger R, Hoffmann R, Knappe D. Differential stability of therapeutic peptides with different proteolytic cleavage sites in blood, plasma and serum. PLoS One. 2017;12(6):e0178943. doi:10.1371/journal.pone.0178943
Doti N, Caporale A, Monti A, et al. New mimetic peptides by retro-inverso design. Int J Mol Sci. 2021;22(16):8677. doi:10.3390/ijms22168677
Tugyi R, Uray K, Iván D, Fellinger E, Perkins A, Hudecz F. Partial D-amino acid substitution: improved enzymatic stability and preserved Ab recognition of a MUC2 epitope peptide. Proc Natl Acad Sci USA. 2005;102(2):413–418. doi:10.1073/pnas.0407677102
Chu DSH, Johnson RN, Pun SH. Cathepsin B-sensitive polymers for compartment-specific degradation and nucleic acid release. J Control Release. 2011;157(3):445–454. doi:10.1016/j.jconrel.2011.10.016
Wei X, Zhan C, Shen Q, et al. A D-peptide ligand of nicotine acetylcholine receptors for brain-targeted drug delivery. Angew Chem Int Ed. 2015;54(10):3023–3027. doi:10.1002/anie.201411226
Akizawa H, Arano Y, Uezono T, et al. Renal metabolism of 111In-DTPA-D-Phe1-octreotide in vivo. Bioconjug Chem. 1998;9(6):662–670. doi:10.1021/bc9702258
Pruszynski M, Kang CM, Koumarianou E, Vaidyanathan G, Zalutsky MR. D-amino acid peptide residualizing agents for protein radioiodination. Molecules. 2018;23(5):1223. doi:10.3390/molecules23051223
Morris O, Gregory J, Kadirvel M, et al. Development and automation of a novel [18F]F prosthetic group and its application to an amino(oxy)-functionalised Aβ peptide. Appl Radiat Isot. 2016;116:120–127. doi:10.1016/j.apradiso.2016.07.023
Weinstock MT, Francis JN, Redman JS, Kay MS. Protease-resistant peptide design: empowering nature’s fragile warriors against HIV. Biopolymers. 2012;98(5):431–442. doi:10.1002/bip.22073
World Anti-Doping Agency. The Prohibited List 2026, S4. Hormone and Metabolic Modulators. wada-ama.org/en/prohibited-list
