MOTS-c: An Evidence Ledger

What the molecule is, and where it is made

MOTS-c is sixteen amino acids long. The verified sequence is MRWQEMGYIFYPRKLR, with a molecular formula of C101H152N28O22S2 and a molecular weight of 2174.6 g/mol for the free base, confirmed against PubChem CID 146675088 together with CAS 1627580-64-6 and UNII A5CV6JFB78. The reading of the SMILES string residue by residue matches the sixteen-residue sequence. The open reading frame encoding it sits inside the mitochondrial 12S ribosomal RNA gene, MT-RNR1, which is the same structural arrangement that produced humanin from the neighbouring 16S region.

Lee and colleagues reported the peptide in Cell Metabolism in 2015, having gone looking for additional short open reading frames in mitochondrial DNA after humanin established that they existed. One detail from that paper is easy to skip and worth keeping. The authors stated that translation of MOTS-c obligatorily occurs in the cytoplasm using the standard genetic code, because reading the same sequence with the mitochondria-specific code yields tandem start and stop codons. The transcript has to leave the organelle to become a peptide.

Naming is a live problem rather than a pedantic one. In its May 2026 evaluation the FDA recorded that MOTS-c is a common name, not a United States Adopted Name, that there is no USP or National Formulary monograph for either the free base or the acetate salt, and that the withdrawn nomination it reviewed named one substance in the title of its certificate of analysis and a different one by CAS number. Free base and acetate salt are distinct substances with distinct properties, and material sold under the single label MOTS-c may be either.

The founding mouse experiments

Lee and colleagues gave intraperitoneal MOTS-c at 0.5 mg/kg/day for eight weeks to adult male CD-1 mice on a diet supplying 60 percent of calories as fat, ten animals per group. Treated animals ended the eight weeks weighing roughly 20 percent less than vehicle-treated animals on the same diet, with increased energy expenditure and less hepatic fat, and the diet-induced rise in insulin was suppressed. Caloric intake did not differ between the groups, which is the observation that makes the result something other than an appetite effect.

A second arm used the hyperinsulinaemic-euglycaemic clamp. Mice treated at 5 mg/kg/day intraperitoneally required an exogenous glucose infusion rate approximately 30 percent higher than vehicle-treated animals to hold euglycaemia under insulin stimulation, and the difference sat in insulin-stimulated glucose disposal rather than hepatic glucose production. Ex vivo, soleus muscle from 12-month-old male C57BL/6 mice given 5 mg/kg/day for seven days took up 2-deoxyglucose at rates comparable to muscle from 3-month-old animals.

Mechanistically the paper traced the effect to the folate cycle and the de novo purine synthesis tethered to it, with accumulation of the intermediate AICAR and downstream AMPK activation. Kim and colleagues later showed that under glucose restriction the peptide translocates to the nucleus in an AMPK-dependent manner and engages antioxidant-response-element genes and NFE2L2/NRF2. The FDA's pharmacology reviewers noted two limits on all of this: the in vivo work has not established dose-response relationships, and the molecular target through which MOTS-c acts remains unknown, which makes it difficult to predict which organs are affected.

Exercise, physical capacity, and what the lifespan curve showed

Reynolds and colleagues sampled ten sedentary healthy young men, mean age 24.5 plus or minus 3.7 years, around a stationary cycle ergometer bout. Skeletal muscle MOTS-c rose 11.9-fold after exercise (P = 0.0098) and stayed elevated after four hours of rest. Plasma rose 1.6-fold during exercise (P = 0.0011) and 1.5-fold immediately after (P = 0.0021), returning to baseline by four hours. Nothing was administered in that experiment; endogenous peptide was measured before and after a physiological stimulus.

Mouse arms in the same paper used intraperitoneal injection at 5 or 15 mg/kg. In 22-month-old C57BL/6N animals, treadmill running time roughly doubled and distance rose 2.16-fold, both at P = 0.000002, with 17 percent of treated animals reaching the final sprint stage against none of the controls. Late-life intermittent treatment, begun at 23.5 months at 15 mg/kg three times weekly, improved grip strength, stride length and a 60-second walking test at beyond 30 months of age.

The lifespan result is the one that travels furthest and holds up least. With 19 control and 18 treated old mice, the authors reported a trend towards increased median lifespan of 6.4 percent and maximum of 7.0 percent, a hazard ratio of 0.654, and a log-rank P of 0.05 up to 31.8 months. Across the whole curve the comparison returned P = 0.23. Their own sentence is that larger cohorts will be needed to confirm the broader significance for overall longevity.

Drift from that finding is documented, and not only on commercial pages. The FDA's own 2026 briefing document states that late-life intermittent treatment increased not only physical capacity but also lifespan, and prints median and maximum figures of 912 and 1047 days for vehicle against 970 and 1120 days for treated animals without reproducing the P = 0.23 attached to the curve. The underlying numbers are correct. The characterisation is stronger than the statistics in the source support.

The measurement problem

Knoop, Thomas and Thevis built a liquid chromatography-mass spectrometry assay for MOTS-c in human plasma for doping control, validated to the World Anti-Doping Agency's International Standard for Laboratories, with a lower limit of detection of 100 pg/mL and coverage of four in vitro metabolites and two oxidation products. To set endogenous reference limits they compared it against a commercially available ELISA in a reference population of 20 healthy subjects. The ELISA returned 45.9 to 218.5 ng/mL. Those levels could not be confirmed by LC/MS.

The size of that gap deserves stating plainly. An ELISA reading of 45.9 ng/mL is 459 times the assay's stated detection limit. A method capable of seeing 100 pg/mL and finding nothing where an immunoassay reports hundreds of times that concentration is not a marginal disagreement between techniques. Either the immunoassay is detecting something other than intact MOTS-c, or the mass-spectrometry method is failing on a matrix effect the validation did not catch. The paper does not resolve it, and no published resolution was located for this page.

That unresolved discrepancy sits underneath the human observational literature, which is built on immunoassays. D'Souza and colleagues measured 26 healthy men per age group and reported circulating MOTS-c lower in middle-aged and older men than in young men by 11 and 21 percent respectively, using an in-house sandwich ELISA developed by the laboratory that first described the peptide. That is the source of the 21 percent figure now quoted widely without its assay, its sample size or its restriction to men.

Direction across human cohorts is not consistent either. Ramanjaneya and colleagues found serum MOTS-c significantly lower in 225 subjects with type 2 diabetes than controls (p < 0.007) and negatively correlated with age. Sequeira and colleagues, in 125 Chinese and 34 European participants without diabetes, found plasma MOTS-c elevated in those with metabolic syndrome and positively associated with android and liver fat. D'Souza found muscle expression roughly 1.5-fold higher in older men while plasma was lower. These studies are describing different tissues, different populations and different assays, and averaging them into one statement about what MOTS-c does with age would misrepresent all three.

Route, degradation, and the analogue substitution

No in vivo pharmacokinetic or toxicokinetic study of MOTS-c exists in any species. The FDA searched for one in preparing its evaluation and stated that neither the nominator nor the Agency identified any. The single pharmacokinetic-adjacent measurement is in vitro: Knoop and colleagues incubated human whole blood with MOTS-c at 5 micrograms per millilitre at 37 degrees Celsius for 15 or 60 minutes and recovered the truncated fragments MOTS-c(2-16), MOTS-c(3-16), MOTS-c(4-16) and MOTS-c(5-16), describing the proteolytic hydrolysis as rapid and not requiring long incubation.

Jiang and colleagues met the same wall from the therapeutic side. In a dextran sulfate sodium colitis model, intraperitoneal MOTS-c reduced weight loss, colon shortening and histological damage, while oral administration produced no significant improvement. Their response was to build a different molecule: MOTS-c linked through a linker at its C-terminus to the cell-penetrating peptide (PRR)5, designated MP. That conjugate, not MOTS-c, is what the oral arm of that literature actually tests.

An earlier paper from the same group reported that peripherally administered MOTS-c does not cross the blood-brain barrier. Central administration enhanced object and location recognition memory in their model and the effect was blocked by the AMPK inhibitor dorsomorphin, but the intranasal and intravenous results that make the cognitive claim portable were generated with MP. The distinction is the same one that runs through the humanin literature, where analogue data is routinely reported under the parent peptide's name.

The regulatory and trial record

MOTS-c is named in the 2026 World Anti-Doping Code Prohibited List at S4.4.1, among activators of AMP-activated protein kinase, alongside AICAR and BAM15. Substances in class S4.4 are non-Specified Substances, and section S4 is prohibited at all times, in and out of competition. It appears in the list index under both MOTS-c and its full expansion.

In May 2026 the FDA evaluated MOTS-c free base and MOTS-c acetate for the section 503A bulk drug substances list. The original nomination, from a compounding pharmacy network, proposed 5 mg and 10 mg lyophilised vials for subcutaneous injection for insulin resistance, obesity, osteoporosis, vascular calcification, muscle and fat metabolism, and longevity. That nomination was withdrawn; the Agency proceeded at its own discretion. Its findings were that no acute toxicity, repeat-dose toxicity, genotoxicity, developmental and reproductive toxicity, or carcinogenicity study of either substance could be identified; that FAERS searches through 9 March 2025 retrieved no reports; that outsourcing facilities reported no compounding of either substance between January 2017 and December 2025; and that no published study used compounded MOTS-c in humans. FDA proposed that neither form be added to the list.

The advisory committee that heard the review voted against that proposal. Reporting of the July 2026 meeting places the vote in favour of adding MOTS-c at 7 to 5 with two abstentions; that tally was not located in an FDA document and is recorded here as trade-press attribution. The vote is advice. Addition to the list would still require notice-and-comment rulemaking, and the substance remains unapproved for human use.

ClinicalTrials.gov holds one interventional study administering MOTS-c to people. NCT07505745 is a Phase 2a randomised, quadruple-blind, placebo-controlled trial in adults aged 18 to 65 with prediabetes and a BMI of 27 to 40, planned enrolment 120, subcutaneous injection over 12 weeks, primary endpoints the Matsuda index at week 12 and treatment-emergent adverse events through week 16. The sponsor is Hudson Biotech, the single site is Peking University Shenzhen Hospital, it began on 2 February 2026, and primary completion is estimated for February 2027. No results are posted. Other registry records naming MOTS-c measure it as an outcome variable.

One retraction, in the bone literature

Weng and colleagues published a paper in European Review for Medical and Pharmacological Sciences in December 2019 reporting that MOTS-c accelerated bone fracture healing by stimulating osteogenesis of bone marrow mesenchymal stem cells through FOXF1 and the TGF-beta pathway. PubMed types it as a Retracted Publication. The retraction notice appeared in the same journal in March 2021, at the authors' request, citing inaccuracies during the process of organising the images.

A sweep of PubMed publication types and correction links across every paper cited on this page returned no other retraction, expression of concern or erratum. The two surviving bone papers are Hu and Chen, who reported that 1 micromolar MOTS-c increased osteogenic marker expression and calcified nodule formation in cultured rat bone marrow mesenchymal stem cells, and Yan and colleagues, who reported reduced bone erosion after local injection into the calvaria of adult male C57BL/6 mice in a polyethylene-particle osteolysis model. Both are preclinical. Neither addresses fracture healing, which was the retracted paper's subject.

What is not known

The largest gap is the simplest to state: MOTS-c has never been administered to a human being in any completed study. The FDA searched for one in May 2026 and reported that no published literature described compounded or non-compounded MOTS-c used in humans by any route, and that outsourcing facilities had reported no compounding of it between January 2017 and December 2025. Everything human in the file measures the endogenous peptide. Consequently there is no human pharmacokinetic data, no dose-ranging, no exposure-response relationship and no adverse-event profile; FAERS searches through 9 March 2025 retrieved no reports at all, which reflects the absence of a reporting pathway from 503A compounders rather than a demonstration of safety. The nonclinical file has the same shape of hole. No acute toxicity, repeat-dose toxicity, genotoxicity, developmental and reproductive toxicity, or carcinogenicity study of either the free base or the acetate has been identified, and no in vivo pharmacokinetic or toxicokinetic study exists in any species. Dose-response relationships have not been established in vivo, and the molecular target through which the peptide acts is unknown, which the FDA reviewers noted makes it difficult to predict which organs would be affected. Assay validity is unresolved and undercuts the observational literature: the immunoassay concentrations that the human cross-sectional studies rest on could not be confirmed by a validated mass-spectrometry method. Populations skew male — the founding metabolic work, the exercise and physical-capacity work, and the osteolysis model all used male rodents, with Lu's ovariectomy study the main exception — and exposure windows ran from three days to twelve weeks, with the late-life mouse arm the only long one. Immunogenicity has not been assessed for a sixteen-residue peptide proposed for repeated subcutaneous injection, and the aggregation and impurity profiles that would bear on it are absent from the public record for both salt forms.

Questions