Each entry describes what the published animal and cell-culture work reports, names the species and the model it came from, and states plainly where the evidence stops.
Last reviewed 12 August 2026 · Maintained by the editorial desk
Standing note: the entries below summarise laboratory research in rodents and cultured cells. They are not medical advice, describe no human use, and contain no dosing, preparation, or sourcing information. Rodent results are hypothesis-generating. They are not evidence of an effect in people.
Subject areas
Tissue repair
Immunomodulation
Dermal
Metabolic
Host defence
Neuro
BPC-157 · Gastric pentadecapeptide · Primary model: rat (Sprague-Dawley, Wistar)
BPC-157
A fifteen-residue sequence with an unusually large rodent literature and an unusually small independent one.
What it is
BPC-157 is a synthetic peptide of fifteen amino acids. The published descriptions present it as corresponding to a partial sequence of a protein identified in human gastric juice, and note that the synthetic form is stable in gastric acid — a property that shaped how it came to be studied. No discrete endogenous peptide of this exact sequence has been isolated as a circulating molecule; what exists in the literature is a laboratory construct derived from a larger parent protein and studied on its own terms.
What the animal literature reports
The substantial majority of this work comes from a research group based in Zagreb, across the 1990s, 2000s, and 2010s, together with laboratories affiliated with it. The models are almost entirely rat: Achilles tendon transection, medial collateral ligament transection, muscle crush injury, gastric mucosal lesion models, injury induced by non-steroidal anti-inflammatory compounds, chemically induced colitis, and a range of vascular occlusion preparations. Reported outcomes include faster closure on histological scoring, differences in biomechanical measures such as load-to-failure in transected rat tendon, and altered vascular patterning described as collateral recruitment. A smaller body of cell-culture work reports effects on tendon fibroblast outgrowth and migration. The proposed mechanism has shifted across the literature — nitric oxide signalling, focal adhesion kinase and paxillin activity, and vascular endothelial growth factor receptor pathways have each been advanced at different points, which is itself worth noticing when assessing how settled the biology is.
Where the evidence thins
The dominant weakness is structural rather than statistical: the same laboratory tradition designed the models, ran the animals, scored the outcomes, and published the results. Reporting of randomisation and of blinded outcome assessment is inconsistent across the corpus, and cohorts are typically six to twelve animals per arm. Independent groups with no connection to the original programme have rarely attempted the flagship tendon and ligament models with matched methods. Endpoints vary enough across papers that pooling them into a single quantitative claim is not defensible. There is no published controlled human efficacy trial for any indication, and citation volume in this area should not be mistaken for confirmation — a great many of those citations point back to the same small set of primary papers.
Evidence snapshot
Model organisms
Rat (Sprague-Dawley and Wistar) across tendon, ligament, muscle and gastrointestinal models; limited mouse work; tendon fibroblast culture
Study scale
Small single-centre cohorts, repeated across many separate papers
Human trial status
No published controlled efficacy trial
Replication status
Overwhelmingly concentrated in one research tradition; independent replication sparse
Thymosin β4 and the TB-500 fragment · Actin-sequestering peptide · Primary model: mouse (wound, cardiac)
Thymosin β4 and the TB-500 fragment
A well-characterised intracellular protein, and a short fragment that the secondary literature keeps treating as the same molecule.
What it is
Thymosin β4 is a 43-residue peptide found in most mammalian cells, where its principal established role is binding monomeric G-actin and regulating the pool available for filament assembly. That intracellular function is not in dispute. A separate and less settled literature describes extracellular signalling activity, which is where the repair and cardiac work sits. “TB-500” is a label applied to a short synthetic fragment associated with the actin-binding region rather than to the full-length peptide, and the two are routinely conflated outside the primary literature. They are not interchangeable, and the volume of evidence behind them is very different.
What the animal literature reports
Full-length thymosin β4 has been examined in murine full-thickness excisional wound models, in corneal epithelial wound models in mice and rats, and in rodent myocardial infarction preparations, where reported effects include cardiomyocyte survival and activation of epicardial progenitor populations. Additional work covers angiogenesis assays and cultured endothelial cells and keratinocytes. Unlike some entries in this Library, this material is spread across multiple unaffiliated laboratories, which raises its standing considerably.
Where the evidence thins
The fragment problem comes first. The great majority of the animal literature concerns the full-length 43-residue peptide; the short fragment has a thin published record of its own, and describing findings from one as findings about the other is a category error that appears constantly in non-specialist writing. Beyond that: the rodent cardiac results have had mixed success under independent replication, outcome measures differ substantially between groups, and the wound-healing endpoints — closure area, histological grade, tensile measurement — are not directly comparable. Early-phase human trials of the full-length peptide have been conducted in ophthalmic and dermatologic settings; the published results did not establish efficacy, and no approved product followed. We note that as trial history, not as evidence of benefit.
Evidence snapshot
Model organisms
Mouse and rat (dermal, corneal, cardiac); cultured endothelial, epithelial and cardiac cells
Study scale
Small to moderate cohorts; multiple laboratories for the full-length peptide, minimal for the fragment
Human trial status
Early-phase trials of the full-length peptide reported; efficacy not established
Replication status
Partial and contested for the full-length peptide; effectively absent for the short fragment
Figure plate 03 — Culture-dish schematic. A large share of the evidence in this Library was generated in cells rather than animals, and the distance between the two is where most overstatement occurs.
The most substantial cell-culture record in this Library, attached to the least resolved question about intact skin.
What it is
GHK is glycyl-L-histidyl-L-lysine, a tripeptide first isolated from human plasma in the early 1970s. It binds copper(II) with high affinity, and the copper complex — GHK-Cu — is the form nearly all of the research concerns. Reports describe declining plasma concentrations with age, which is frequently cited as motivation for the work rather than as a finding about what supplementation would do.
What the animal literature reports
The strongest material is in vitro. In cultured human dermal fibroblasts, GHK-Cu has been reported to increase collagen and glycosaminoglycan synthesis, and transcriptional studies at microarray scale have described broad shifts in gene expression associated with matrix remodelling. In vivo work is smaller and more scattered: rat and mouse dermal wound models reporting faster closure and altered histology, some rabbit and porcine skin work, and hair-follicle studies in mice. Several groups have contributed, which distinguishes this entry from the more single-source material elsewhere in the Library.
Where the evidence thins
A fibroblast in a dish is bathed directly in whatever the investigator adds; intact skin has a barrier specifically evolved to prevent that, and the literature has not resolved what proportion of the in vitro effect is achievable through it. Concentrations used in culture are not straightforwardly translatable to any in vivo condition. A meaningful fraction of the work has cosmetic-industry funding, which does not invalidate it but does warrant declaration. In vivo endpoints are heterogeneous — wound area, tensile strength, histological grade — and cannot be pooled. The transcriptional findings, which do a great deal of rhetorical work in secondary writing, have limited independent replication. No controlled human trial with hard clinical endpoints appears in the published record.
Evidence snapshot
Model organisms
Cultured human dermal fibroblasts and keratinocytes; rat, mouse, rabbit and porcine dermal models
Study scale
Numerous small in vitro studies; a considerably smaller in vivo literature
Human trial status
No controlled trial with hard clinical endpoints in the published record
Replication status
In vitro effects reported by several groups; in vivo findings much less replicated
Three amino acids, one disease model, and a great deal resting on how faithfully that model behaves.
What it is
KPV is lysine-proline-valine, the C-terminal tripeptide of α-melanocyte-stimulating hormone. The research interest is that it appears to retain part of the anti-inflammatory activity described for the parent hormone while lacking its pigmentary signalling — a small, cheap, chemically simple fragment that might do one job of a larger molecule.
What the animal literature reports
Nearly all of the in vivo work uses chemically induced murine colitis: dextran sulphate sodium in drinking water, or trinitrobenzene sulphonic acid instilled directly. Reported outcomes include lower disease activity index scores, preservation of colon length, reduced histological damage grades, and lower pro-inflammatory cytokine concentrations in colon tissue. A large fraction of recent papers pair the peptide with a delivery system — nanoparticle formulations, hydrogels, targeted carriers — and report the combination. Supporting cell-culture work uses intestinal epithelial and macrophage lines, with effects described on NF-κB signalling.
Where the evidence thins
Dextran sulphate sodium colitis is a chemical injury delivered on a fixed schedule to genetically uniform mice housed in controlled conditions. It reproduces epithelial damage and the acute response to it. Human inflammatory bowel disease is chronic, relapsing, genetically heterogeneous, and shaped by a microbiome no vivarium replicates. Endpoints in these studies are surrogate measures, not clinical outcomes. Cohorts are small. Where the peptide is tested inside a novel delivery vehicle, its independent contribution is difficult to isolate from the vehicle’s. No completed controlled human trial of KPV alone appears in the published record.
Evidence snapshot
Model organisms
Mouse (dextran sulphate sodium- and trinitrobenzene sulphonic acid-induced colitis); intestinal epithelial and macrophage cell lines
Study scale
Small cohorts; frequently reported alongside delivery-vehicle development
Human trial status
No completed controlled trial of the peptide alone in the published record
Replication status
Reported by several groups, but tied closely to a single class of model
Absence of evidence gets written here as absence of evidence. Leaving a polite silence where a human trial should be is exactly how a rodent result turns into a rumour.
Mara Ellison, Editor
Figure plate 04 — Counting chamber. Most of the in vivo work summarised here rests on cohorts small enough to fit in a single rack of cages.
An endogenous defence peptide whose measured behaviour depends heavily on what else is in the dish.
What it is
LL-37 is the 37-residue C-terminal peptide released by cleavage from the human cathelicidin precursor hCAP18. It is expressed by neutrophils and by epithelial cells at barrier surfaces. Structurally it is cationic and amphipathic, and its described activities fall into two groups: direct membrane-active antimicrobial effects, and immunomodulatory signalling including chemotaxis, cytokine modulation, and involvement in wound repair.
What the animal literature reports
Mouse models dominate: skin and soft-tissue infection, pulmonary infection, sepsis preparations, and excisional wound models. Some of the most informative work uses cathelicidin-knockout mice, which show greater susceptibility to particular infectious challenges than wild-type littermates. That design is genuinely strong evidence that the endogenous peptide matters to host defence — it is a different question from what administering the peptide does, and the two are regularly merged in summary writing. Cell-culture work spans antimicrobial assays against a wide range of organisms, plus keratinocyte and monocyte studies.
Where the evidence thins
Antimicrobial activity measured in low-salt buffer is substantially attenuated at physiological salt concentrations and in the presence of serum, which has complicated this entire field for decades. The peptide is cytotoxic to mammalian cells above certain concentrations, narrowing the window in which any effect could be described as selective. Results on inflammation point in both directions depending on tissue and context, and LL-37 has also been implicated in the pathology of psoriasis — a finding that sits awkwardly with any straightforward beneficial reading. Antimicrobial peptides as a class have a poor record in late-stage development, with several candidates failing after promising preclinical work. Knockout evidence establishes physiological importance, not therapeutic utility.
Evidence snapshot
Model organisms
Mouse (infection, wound, cathelicidin-knockout); human primary cells and cell lines
Study scale
A moderate literature spread across many independent laboratories
Human trial status
No established efficacy; antimicrobial peptides as a class have repeatedly failed in late-stage development
Replication status
Core biology confirmed across laboratories; applied findings context-dependent and often contradictory
The one entry here with a real clinical development history, and the one where our scope limit matters most.
What it is
Thymosin α1 is a 28-residue, N-terminally acetylated peptide derived from the precursor protein prothymosin α, originally identified in thymic tissue preparations. The described activity is immunomodulatory rather than directly antimicrobial: effects on innate signalling including Toll-like receptor pathways, on dendritic cell maturation, and on T-cell development markers.
What the animal literature reports
The preclinical record is broad and old. Mouse and rat models of fungal and bacterial infection, sepsis preparations, chemically or surgically induced immunosuppression, and tumour models in mice all appear. Cell-culture work covers T-cell maturation markers, dendritic cell function, and cytokine profiles in human and murine immune cells. The material spans decades and many unaffiliated groups.
Where the evidence thins
This entry carries a factual note the others do not: a synthetic form of this peptide has been through registered human trials and is licensed as a medicine in a number of countries. We record that as regulatory history. Human indications, prescribing, and outcomes sit outside this publication’s scope, we take no position on them, and nothing here should be read as endorsement or as a recommendation to anyone. Confined to the preclinical literature, the weaknesses are real: immune endpoints are famously context-dependent, a result in one infection model routinely fails to transfer to another, and the sheer heterogeneity of designs across decades makes direct comparison between papers difficult. Older work in particular often predates current expectations for reporting randomisation and blinding.
Evidence snapshot
Model organisms
Mouse and rat (infection, immunosuppression, tumour); human and murine immune cell culture
Study scale
Broad and heterogeneous, accumulated across several decades
Human trial status
Registered trials exist and a synthetic form is licensed in some jurisdictions. Recorded as regulatory fact; human use is outside our scope
Replication status
Multi-laboratory, but endpoint variation limits direct comparison
A young field with real mechanistic interest and very little settled ground beneath it.
What it is
MOTS-c is a 16-residue peptide encoded within the mitochondrial 12S ribosomal RNA region — one of a small set of sequences described as mitochondrial-derived peptides since the early 2000s. That origin is the interesting part: the mitochondrial genome producing short regulatory peptides that act outside the organelle was not part of the standard picture. Reported activity centres on cellular metabolic signalling, including AMP-activated protein kinase activation and the folate–methionine cycle, with descriptions of nuclear translocation under metabolic stress.
What the animal literature reports
Mouse work dominates: high-fat-diet cohorts with reported differences in weight trajectory, insulin sensitivity measures, and glucose tolerance testing; ageing cohorts examining age-related metabolic decline; and exercise-related skeletal muscle studies. Cell-culture work uses myotubes and hepatocytes. Human data exists but is observational — associations between circulating peptide measurements and metabolic or fitness variables — not interventional.
Where the evidence thins
The field is barely a decade old, and that shows. Questions have been raised in the literature about antibody specificity and assay reliability for measuring endogenous circulating peptide, which matters a great deal because several downstream claims depend on those measurements being accurate. Mouse cohorts are small. Independent replication of the flagship metabolic findings is limited, and much of the work involves overlapping groups. Observational human associations cannot establish direction of effect. There are no efficacy trials.
Evidence snapshot
Model organisms
Mouse (high-fat diet, ageing, exercise); myotube and hepatocyte culture
Study scale
Small cohorts within a young and concentrated literature
Human trial status
No efficacy trial in the published record; human work is observational
Replication status
Early. Core findings still await independent confirmation
Humanin · Mitochondrial-derived peptide · Primary models: rodent neuroprotection and metabolic work
Humanin
Most of what is described as humanin’s effect was produced by an engineered analogue that is not humanin.
What it is
Humanin is a 24-residue peptide encoded within the mitochondrial 16S ribosomal RNA region, first described in the early 2000s from work on tissue from brains affected by Alzheimer’s disease. Reports describe interactions with apoptotic regulatory proteins and with a receptor complex at the cell surface, and characterise it broadly as cytoprotective under cellular stress.
What the animal literature reports
Rodent work covers models of neuronal insult, including amyloid-β toxicity preparations and cerebral ischaemia models, alongside a smaller body of metabolic studies in mice. The cell-culture literature is considerably larger and concerns resistance to apoptosis across a range of stressors. A critical detail: a large share of the in vivo work uses engineered analogues rather than the native sequence — most commonly a single-residue substitution variant reported to be far more potent than the parent peptide.
Where the evidence thins
The analogue problem is the headline. Results generated with a modified, substantially more potent peptide are routinely reported downstream as findings about humanin itself, and readers of secondary summaries have no way to detect the substitution. Beyond that: the native peptide has a short half-life, delivery to the central nervous system is unresolved, and endogenous expression levels have been disputed. Rodent neuroprotection models as a category have one of the worst translation records in biomedical research — a long list of compounds that protected neurons in rats and did nothing in people. No efficacy trials exist.
Evidence snapshot
Model organisms
Mouse and rat (neuronal insult, metabolic); extensive cell culture
Study scale
Small in vivo cohorts; a larger in vitro literature
Human trial status
None in the published record; human work is observational
Replication status
Cell-culture findings widely reported; in vivo work limited and largely analogue-dependent
Why the hedging is deliberate, and why the gaps are stated rather than skipped.
Every entry above follows the same order on purpose. Chemistry and origin first, because a surprising amount of confusion in this subject comes from people arguing about different molecules. Then what the animal and cell-culture literature reports, with the species and the model named each time, because “improved healing” means nothing until you know it was a histological damage score in a rat with a surgically transected tendon. Then what is missing, which is usually the most useful section on the page.
The hedged verbs are not timidity. “Reported,” “observed in,” and “has been described” mark the difference between what a paper claims and what the field has established, and that difference is the whole subject. When we write that something remains unreplicated, we mean no unaffiliated laboratory has published a matching result — not that we could not find one in an afternoon.
We state absence of evidence explicitly, in a full sentence, every time. That is a house rule with a reason behind it. A summary that describes ten rodent studies and then goes quiet about human trials leaves the reader to assume the silence is incidental. It usually is not. For most compounds in this Library the honest sentence is that no controlled human efficacy trial has been published, and writing it plainly is the single most informative thing a page like this can do.
Finally, the entries do not converge on a verdict, because the literature does not. A compound can have interesting mechanistic work in cells, a thin and single-source animal record, and no human evidence at all, all at once. Holding those three facts together without collapsing them into a recommendation is the job.
