Every other category on this catalog is organized around a target. The metabolic shelf sorts by receptor count. The growth-hormone shelf sorts by tier on an axis. The immune shelf sorts by direction of effect. In each case the interesting question is what does this molecule bind, and what happens next.
The cosmetic shelf does not work that way — and the reason is worth understanding before reading a single paper in it. For a compound intended to act in skin, the binding question is often the second question. The first is whether the molecule ever reaches the cell it is supposed to signal. The stratum corneum is the class-defining variable, and it is a variable that simply does not exist for a compound studied in a well or delivered systemically.
That single fact reorganizes how the literature should be read.
One product on the shelf, four mechanisms in the field
Exactly one catalog entry is filed under cosmetic: SNAP-8. As with the immune class primer, the useful discussion requires pulling in compounds filed elsewhere by mechanism — GHK-Cu from healing-recovery, Melanotan 1 and Melanotan 2 from sexual-health, and LL-37 and KPV from healing-recovery again.
Dermatology has its own taxonomy for this space, and it predates the research-peptide market. The standard four-way scheme comes from Lupo and Cole's 2007 review in Dermatologic Therapy (20(5):343–349), and it sorts topical peptides by mechanism:
- Signal peptides (matrikines) — short fragments of extracellular matrix proteins that act as messengers instructing fibroblasts. Palmitoyl pentapeptide-4 (KTTKS), a fragment of the collagen I propeptide, is the canonical example.
- Carrier peptides — sequences whose job is to deliver a trace metal to an enzyme that requires it. This category exists essentially because of GHK-Cu.
- Neurotransmitter-inhibiting peptides — sequences designed to interfere with the exocytosis machinery at the neuromuscular junction. SNAP-8 and acetyl hexapeptide-8 (Argireline) sit here.
- Enzyme-inhibiting peptides — sequences that inhibit matrix-degrading proteases rather than stimulate synthesis.
That last category is worth naming precisely because this shelf has no member of it. A taxonomy is more useful when it shows you what is absent.
Branch one: the exocytosis mimics
SNAP-8 — INCI name acetyl octapeptide-3 — is a mimic of the N-terminal region of SNAP-25, one of the three proteins forming the SNARE complex that drives synaptic vesicle fusion. The design logic is competitive interference with SNARE assembly, reducing acetylcholine release at the neuromuscular junction.
That mechanism, its evidence state, and the important caveats around it were covered in full in the SNARE complex and vesicle fusion explainer, and are not re-argued here. Two points carry forward into the class discussion:
First, the mechanism is best described as proposed rather than established — the in vitro work demonstrating catecholamine-release inhibition in chromaffin cells used concentrations in the tens of micromolar, and no in vivo study has confirmed inhibition of muscle contraction by this route.
Second, and central to this primer: the 2026 safety framework for cosmetic peptides published by Bjerke and colleagues in Current Research in Toxicology (10:100291) reports that for acetyl hexapeptide-8, 99.7% of applied material remained at the application site. That figure is a safety finding — it argues against meaningful systemic exposure. Read as a delivery finding, it says something less comfortable about the intended mechanism, because a neuromuscular junction is not at the application site.
Branch two: signal and carrier, both of them GHK
GHK-Cu is the tidiest illustration of why the four-way taxonomy is a description rather than a partition: it belongs to two categories at once.
As a signal peptide, GHK (glycyl-L-histidyl-L-lysine) is a matrikine — a sequence liberated from larger matrix proteins during tissue injury — and its reported activity profile includes upregulation of collagen, elastin, glycosaminoglycans and decorin in fibroblast culture, plus modulation of matrix metalloproteinases. Its angiogenic strand, including VEGF and FGF-2 upregulation, was covered in the angiogenesis and VEGF pathway explainer.
As a carrier peptide, its function is the copper. GHK has high affinity for Cu(II), and the complex is studied as a delivery vehicle for a cofactor that several matrix enzymes — lysyl oxidase among them — genuinely require.
Two honesty notes belong on this compound. The frequently quoted claim that GHK-Cu "resets the expression of roughly a third of human genes" traces to a narrow originating research lineage (largely Pickart and colleagues) built substantially on gene-expression profiling, and the evidence hierarchy piece covers why a transcriptome-wide percentage is a weaker claim than it sounds. And GHK-Cu's short circulating half-life — under thirty minutes in the reported literature — means the systemic version of this compound has a delivery problem too.
Why molecular weight is necessary but not sufficient
The conventional gatekeeper in transdermal pharmacology is the 500 Dalton rule, set out by Bos and Meinardi in Experimental Dermatology in 2000 (9:165–169): compounds much above 500 Da do not passively cross intact stratum corneum in useful amounts.
Here is where vendor copy routinely gets it backwards. GHK is 340.4 Da as the free tripeptide, and the copper complex sits around 400 Da. Both are comfortably under the 500 Da ceiling — and GHK-Cu still penetrates poorly. Molecular weight was never the whole rule. Bos and Meinardi also specified a lipophilicity window, and GHK-Cu is a hydrophilic, positively charged complex facing a barrier built of lipid lamellae and negatively charged corneocytes. Charge and polarity, not size, are the obstacle.
This is the current state of the question rather than a settled one. Ogórek and colleagues, writing in Molecules in 2025 (30(1):136), reviewed transport-assessment methods for cosmetically active compounds and concluded that liposomal encapsulation of GHK-Cu — the obvious workaround — has received little attention, and that the methodology for measuring it is itself underdeveloped. Their title is a question for a reason.
The practical consequence for anyone reading this literature: "does it work" is at minimum two independent questions. Does the molecule do the biology, and does the molecule arrive. A cell-culture paper answers the first and is silent on the second. The Bjerke framework notes the same pattern for palmitoyl pentapeptide-4 — well characterized, favorably tolerated, non-penetrating, and rapidly degraded by skin proteases.
Branch three: the compounds where the barrier argument does not apply
A second group gets grouped under "cosmetic" for its endpoint rather than its route, and mixing the two groups is the most common category error in this space.
Melanotan 1 and Melanotan 2 are melanocortin receptor agonists studied for melanogenesis, and in the published literature they are administered systemically, not topically. The stratum corneum argument is irrelevant to them; a completely different evidence standard applies. Their receptor pharmacology, the MC1R-weighted versus nonselective distinction, and the approved reference point of afamelanotide in erythropoietic protoporphyria are covered in the Melanotan 1 vs Melanotan 2 comparison and the melanocortin system explainer.
Similarly, BPC-157 and TB-500 appear in dermal wound-healing models, but their literature is repair-primary and belongs to the healing-recovery class, not this one.
The skin already makes peptides, and more is not better
The most instructive members of this class may be the two the skin produces itself.
LL-37 is a keratinocyte product, cleaved from hCAP18 by kallikreins in skin. It is not a cosmetic ingredient — it is a host-defense molecule. And the reason it belongs in this primer is what happens when the system runs hot: Yamasaki and colleagues, in Nature Medicine in 2007 (13:975–980), reported that rosacea skin carries both elevated cathelicidin and elevated kallikrein 5, producing abnormally processed peptide forms, and that intradermal LL-37 induced rosacea-like inflammation in mice.
KPV, the C-terminal tripeptide of α-MSH, points the other way — it is studied as an anti-inflammatory brake, with activity that does not appear to require melanocortin receptor signaling.
Together they make the class's least marketable point: in skin, the endogenous peptide systems are bidirectional and tuned, and "more signaling" is not a coherent research goal.
Documentation notes specific to this shelf
Three things differ here from the rest of the catalog, and they are worth flagging rather than re-arguing.
Naming runs on two parallel systems. INCI names describe cosmetic ingredients; research names describe molecules. Acetyl hexapeptide-3 and acetyl hexapeptide-8 are the same compound under an INCI renumbering, while acetyl octapeptide-3 is a different, longer molecule. Nothing about the numeral communicates potency or generation. See the nomenclature primer.
The modification is part of the identity. N-terminal acetylation adds 42 Da and C-terminal amidation subtracts about 1 Da versus the free acid. A certificate of analysis for these compounds must report the mass of the modified peptide, and des-acetyl and free-acid impurities sit close enough in mass that intact-mass confirmation alone is weak — MS/MS sequencing is the stronger check, as discussed in HPLC vs mass spec.
GHK-Cu carries a stoichiometry question no purity percentage answers. Copper content and complexation state are a separate analytical axis from peptide purity. More on documentation generally at /quality/.
The class's defining asymmetry
Here is the closing observation, and it inverts the pattern seen everywhere else on this site.
On the drug-development shelves, safety data is the scarce commodity and mechanism is abundant. On the cosmetic shelf it is the reverse. The safety literature is comparatively mature — cosmetic ingredient review panels, established use-level ceilings, and now the 2026 bioinformatic weight-of-evidence framework screening peptide sequences against toxin, allergen and bioactivity databases before they reach a formulation. Meanwhile the efficacy literature is dominated by small, often uncontrolled appearance studies, and the delivery question underneath them is, by the reviewers' own assessment, not yet reliably measurable.
That is a coherent state of affairs for consumer ingredients used at parts-per-million in finished products. It is a different thing entirely when the same sequences appear as neat lyophilized powder on a research shelf. The safety framework that makes these peptides comfortable in a formulation was built around the assumption that 99.7% of the material stays where it was put.
FAQ
Is the cosmetic class a real pharmacological category? No. It is an outcome category. The members span SNARE-complex interference, matrix signaling, metal delivery and melanocortin receptor agonism — four unrelated mechanisms. What the members share is an intended tissue, not a target.
Why is skin penetration treated as a bigger issue here than bioavailability elsewhere? Because for systemically studied compounds, delivery failure shows up as a pharmacokinetic number that can be measured. For topical compounds, the measurement itself is contested — the 2025 Molecules review's central finding was a methodological gap, not a result.
Does an approved melanocortin drug validate the topical peptides? No, and the two branches should not be pooled. Afamelanotide's regulatory record concerns a systemically delivered melanocortin agonist in a specific rare-disease population. It says nothing about whether a topically applied octapeptide reaches a neuromuscular junction. Browse the full catalog at /library/.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.