What cosmetic peptides are
Cosmetic peptides are short chains of amino acids made for skin creams, serums and eye products. Most are synthetic, many carry a fatty acid such as palmitic acid to help them into the outer layer of the skin, and on packaging they appear under their INCI names, the standard names for cosmetic ingredients, such as palmitoyl tripeptide-1 or pentapeptide-18. Suppliers sell them to cosmetics makers under trade names, often as dilute ready-made solutions.
They differ from most other compounds in this wiki in three ways. They are made only for use on intact skin; in the EU they are cosmetic ingredients, not medicines, investigational drugs or research material; and the amounts in a finished product are tiny. For the group it assessed in 2018, which includes palmitoyl tripeptide-1, palmitoyl hexapeptide-12 and palmitoyl tetrapeptide-7, the US Cosmetic Ingredient Review panel found typical use concentrations below 10 parts per million. Whatever a peptide does in a dish of skin cells, the open question is how much of it gets through the skin barrier to the cells it is meant to act on.
Reviews sort cosmetic peptides by their intended action: signal peptides, meant to prompt skin cells to make more collagen and other matrix proteins; neurotransmitter-inhibiting peptides, meant to relax the small muscles behind expression lines; carrier peptides, which deliver a trace metal such as copper; and enzyme-inhibiting peptides. Reviews do not always agree on which peptide belongs where. This guide covers twelve ingredients the wiki has no separate article for. Matrixyl, GHK-Cu, AHK-Cu, Argireline and Snap-8 have their own articles.
Signal peptides
Signal peptides are fragments of the skin's own structural proteins, such as collagen and elastin, or are modelled on them. In cultured skin cells such fragments can increase the production of collagen and other matrix proteins, and the idea is that on the skin they act as the same repair signal. The best-studied member, palmitoyl pentapeptide-4, has its own article under Matrixyl. For the four below, the human evidence is thin:
- Palmitoyl tripeptide-1 (pal-GHK). Palmitic acid attached to glycine-histidine-lysine, the same three amino acids as in GHK-Cu, here without copper; it used to be sold under the INCI name palmitoyl oligopeptide, a name retired because it covered more than one peptide. Together with palmitoyl tetrapeptide-7 it makes up Matrixyl 3000. The most informative published study comes from the supplier Sederma itself (Mondon et al., 2015): with a blend of palmitoyl tetrapeptide-7 and 'palmitoyl oligopeptide', the band of low-density tissue just under the epidermis became thinner and denser on ultrasound, and the dermal matrix looked better organised on in vivo confocal microscopy than with placebo. The abstract gives neither the number of volunteers nor effect sizes. We found no independent trial of either peptide on its own.
- Palmitoyl tripeptide-38 (Matrixyl synthe'6; pal-Lys-Met(O2)-Lys). The one human study is open-label, with no comparison group (Lintner et al., 2020): 35 women aged 49 to 74 applied a serum with 15% vitamin C, vitamin E and 5 ppm palmitoyl tripeptide-38 once a day for 56 days. Skin roughness fell by 8 to 9%, redness by 9% and evenness of tone rose by 8%, with no side effects attributed to the serum. The authors were consultants for Alumier Labs, which makes the serum, and with that much vitamin C in the same product the peptide's share cannot be separated out.
- Palmitoyl hexapeptide-12 (Biopeptide EL; pal-Val-Gly-Val-Ala-Pro-Gly). Palmitic acid attached to a six-amino-acid sequence from elastin, the protein that lets skin spring back. The best data are not a trial: on donated full-thickness human skin, researchers at the University of Nottingham working with Estée Lauder showed by mass-spectrometry imaging that it entered the stratum corneum and epidermis, formed a gel there, and helped strengthen skin whose lipids had been stripped (Ligorio et al., 2025). We found no published efficacy trial in people.
- Hexapeptide-9 (Collaxyl; Gly-Pro-Gln-Gly-Pro-Gln). A collagen-like sequence without a fatty acid. The only controlled trial tested a cyclised version the authors had developed, not hexapeptide-9 itself (Chang et al., 2025): 96 adults aged 30 to 55 in China were randomised, double-blind, to a serum with 0.002% of the cyclised peptide, a serum with 0.002% retinol or the plain serum for 56 days, and 91 finished. The peptide serum reduced the number, area and roughness of crow's feet and forehead wrinkles significantly more than the plain serum, and more than retinol on all those wrinkle measures except crow's-feet roughness, with no treatment-related side effects. The authors work for the Chinese cosmetics companies Shanghai Chicmax and Shanghai Kans Biotech.
Neurotransmitter-inhibiting peptides
These peptides are meant to work like a very weak, surface version of botulinum toxin: by damping the nerve signal that makes the small facial muscles contract, expression lines should soften. The best known are Argireline (acetyl hexapeptide-8) and its longer relative Snap-8 (acetyl octapeptide-3), which have their own articles.
The weak point of the whole group is delivery, because the nerve endings that move facial muscles lie in the muscle, beneath the epidermis and the dermis. When FDA scientists left an emulsion with 10% acetyl hexapeptide-8 on human donor skin for 24 hours, most of it washed off: 0.22% of the dose stayed in the stratum corneum, 0.01% reached the epidermis and none was found in the dermis (Kraeling et al., 2015). Two further peptides in this group:
- Pentapeptide-18 (Leuphasyl; Tyr-D-Ala-Gly-Phe-Leu). Chemically this is [D-Ala2]-leucine-enkephalin, a stabilised version of enkephalin, one of the body's own opioid peptides; it is described as mimicking enkephalin at the nerve ending to reduce the release of acetylcholine. The supplier is Lipotec. The best human data come from a Polish study run partly by employees of the cream maker (Pawłowska et al., 2024): 20 women aged 30 to 65 tested a night cream with retinol and pentapeptide-18 in lipid nanoparticles against the plain base cream for 8 weeks. Wrinkle counts fell by 29% on the forehead and 18% around the eyes with the active cream, against 6% and 9% with the base. Because the cream also contained retinol, the result says nothing about the peptide alone.
- Dipeptide diaminobutyroyl benzylamide diacetate (SYN-AKE; β-Ala-Pro-Dab-NH-benzyl as the diacetate salt). A small synthetic molecule modelled on waglerin-1, a peptide from the venom of Wagler's pit viper (Tropidolaemus wagleri) that blocks the acetylcholine receptor on adult muscle. It contains no venom. Its only human data are two L'Oréal studies of a serum that also contained acetyl hexapeptide-8 (Argireline), gluconolactone, niacinamide and kelp extract (Zhu et al., 2026). In the first, with 50 people over 12 weeks, clinical scores for static wrinkles improved by 35 to 69%; in the second, with 42 people, scores for expression lines improved by 10 to 13%. The abstract reports change from baseline, not a comparison with another serum, and the effect cannot be put down to any one ingredient.
Carrier, enzyme-inhibiting and other peptides
The remaining peptides fit neither of the first two groups, or fit several. Carrier peptides deliver a trace metal to the skin; the copper peptides GHK-Cu and AHK-Cu are the main examples and have their own articles. The others here are meant to organise collagen, drain the under-eye area, block pigment signals, act as a growth factor or neutralise oxidation:
- Tripeptide-10 citrulline (Lys-Asp-Ile-citrulline amide). Developed by the Spanish supplier Lipotec to mimic decorin, a protein that binds collagen fibres and controls how they assemble. It has one academic trial, from the Technological Educational Institute of Athens (Raikou et al., 2017): 24 volunteers were randomised to a cream with tripeptide-10 citrulline, with acetyl hexapeptide-3 (Argireline), with both, or with neither, for 60 days. The authors report significant differences in skin-roughness measures between some of the groups at 20 and 60 days and conclude that both peptides smoothed wrinkles. With 24 people spread over four groups, the evidence is thin.
- Acetyl tetrapeptide-5 (Eyeseryl; Ac-β-Ala-His-Ser-His). Marketed for puffiness and bags under the eyes, on the strength of the supplier's own tests. In a 2021 Chinese laboratory study it was tested on human skin fibroblasts for its effect on angiotensin-converting enzyme, alone and in peptide mixtures, but we found no peer-reviewed study of it in people.
- Nonapeptide-1 (Melanostatine 5; Met-Pro-D-Phe-Arg-D-Trp-Phe-Lys-Pro-Val-NH2). A modified fragment of α-MSH, the hormone that tells pigment cells to make melanin, built to block its receptor (MC1R) rather than switch it on, and sold for dark spots. Researchers have used it as a reference α-MSH blocker in cell studies of pigmentation, but we found no published trial of it in people.
- sh-Oligopeptide-1 (recombinant human EGF). Despite the INCI name, this is a protein of up to 53 amino acids: human epidermal growth factor, made by E. coli bacteria from a synthetic copy of the human gene. It is not the same as oligopeptide-1, which CosIng describes as a short peptide of glycine, histidine and lysine, although the two names are often mixed up. The best trial is small (Abud et al., 2026): 20 women aged 50 to 65 were randomised, double-blind, to a serum with recombinant EGF or a serum with human-derived growth factors for 12 weeks. Investigators rated texture and firmness better with the EGF serum, wrinkle severity barely changed in either group, and there was no placebo. The study was funded by DRMTLGY, which makes the EGF serum.
- Palmitoyl tetrapeptide-7 (pal-Gly-Gln-Pro-Arg), the partner of palmitoyl tripeptide-1 in Matrixyl 3000. GQPR is a fragment of immunoglobulin G, and in laboratory tests the peptide lowered the release of interleukin-6, a messenger of inflammation; the idea is that calmer skin breaks down less collagen. Its human data are the Matrixyl 3000 studies described under signal peptides, in which it was always combined with palmitoyl tripeptide-1.
- Carnosine (β-alanyl-L-histidine). A dipeptide the body makes itself, found in large amounts in muscle. In skincare it is used as an antioxidant and against glycation, the binding of sugars to collagen that stiffens it. On donated human skin kept alive in the laboratory, topical carnosine reduced glycation products and damage from simulated sunlight in studies by the skincare company ISDIN and the ingredient maker Symrise. In people it has only been tested in mixtures: in four open-label ISDIN studies with 117 people in total, a night cream with melatonin, carnosine and Helichrysum italicum extract reduced wrinkle counts by up to 18.9% over up to 12 weeks (Granger et al., 2020).
How to read the evidence
Across these twelve ingredients the pattern repeats: plausible cell biology, little proof that the peptide reaches its target in living skin, and human studies that are small, short and usually run or paid for by the company selling the ingredient or the product. A few questions help to weigh any claim:
- What was tested. Most human studies used a finished product with several actives, such as vitamin C, retinol or niacinamide, so the result belongs to the product, not to the peptide.
- What it was compared with. A study without a comparison group cannot separate the peptide from the base cream, the effect of daily application or the passage of time. A placebo or vehicle cream, on the other half of the face or in another group, is the minimum.
- Size and length. The studies above enrolled about 20 to 100 people each, for 8 to 12 weeks. Studies that small and short can miss uncommon side effects and say nothing about years of use.
- Who ran it. Supplier data in brochures and on websites are not peer reviewed and usually leave out the number of volunteers and the comparison.
- What was measured. Wrinkle depth on a skin replica, a clinician's score or a questionnaire is not the same as more collagen, which almost none of these studies measured in people.
- Which name. A trade name can stand for a dilute solution rather than pure peptide, and similar names can hide different substances: oligopeptide-1 and sh-oligopeptide-1 are not the same, and 'palmitoyl oligopeptide' once covered more than one peptide.
What EU law says
In the EU, every peptide in this guide is a cosmetic ingredient under Regulation (EC) No 1223/2009. A cosmetic product is one intended for the external parts of the body, such as the epidermis, to clean, perfume, protect or keep them in good condition or to change their appearance, and the regulation states that a substance intended to be swallowed, inhaled, injected or implanted is not a cosmetic. These peptides are sold as cosmetic ingredients, not as medicines.
The ingredients need no approval before use unless they appear in the regulation's annexes of banned or restricted substances, and none of the twelve does. Instead, each finished product needs a safety assessment and must be notified in the EU cosmetics portal (CPNP) before it goes on sale; in the Netherlands the NVWA supervises this. Ingredient lists use the common ingredient names, which are based on INCI. The Commission's CosIng database lists all twelve, mostly with the function skin conditioning; the regulation says its glossary of ingredient names is not a list of authorised substances, so a listing is not an approval.
Claims are regulated as well. Article 20 forbids presenting a cosmetic as having characteristics or functions it does not have, and Regulation (EU) No 655/2013 requires claims to be supported by adequate and verifiable evidence. It also forbids implying that a finished product has the properties of one of its ingredients when it does not, so a study of an ingredient does not prove a claim about a cream that contains it.
In the United States, cosmetics other than colour additives need no FDA approval before sale, and a product intended to change the structure or function of the body is legally a drug. That is why labels there promise a smoother look rather than new collagen.
Frequently asked questions
- Do peptides in creams actually work?
- Some show modest effects, but the evidence is weak for most. Only two randomised trials in this guide compared a peptide cream with the same cream without peptide: one of a cyclised form of hexapeptide-9 in 96 people and one of tripeptide-10 citrulline in 24. Most other human data come from products with several active ingredients, often studied by the company that sells them, so the result cannot be credited to the peptide. Matrixyl (palmitoyl pentapeptide-4) and Argireline have somewhat more placebo-controlled data, described in their own articles.
- Are cosmetic peptides safe?
- On intact skin, at the small amounts used in cosmetics, the record so far is good. The trials in this guide that reported on tolerability found no serious or treatment-related side effects, and in 2018 the US Cosmetic Ingredient Review panel concluded that palmitoyl tripeptide-1, palmitoyl hexapeptide-12 and palmitoyl tetrapeptide-7 are safe as used, at typical concentrations below 10 ppm. The studies were short, so they say little about years of daily use, and none of the safety data covers any route other than the skin.
- Can cosmetic peptides be injected?
- None of the twelve ingredients in this guide was developed for injection, and as cosmetic ingredients they have no safety data for it. Under the EU Cosmetics Regulation a product intended to be injected is not a cosmetic by definition, and the safety assessments behind these ingredients concern small amounts on intact skin. Some are biologically active molecules in their own right: pentapeptide-18 is an analogue of the body's opioid peptide enkephalin, and sh-oligopeptide-1 is human epidermal growth factor.
- Does SYN-AKE contain snake venom?
- No. Its ingredient name is dipeptide diaminobutyroyl benzylamide diacetate, a small synthetic molecule built on two amino acids. It was modelled on waglerin-1, a peptide from the venom of Wagler's pit viper that blocks the acetylcholine receptor on adult muscle, but it contains no venom. There is no published study of it on its own in people.
- How do I recognise peptides on an ingredient list?
- Look for INCI names with 'peptide' in them. The prefix gives the number of amino acids (tri is three, penta five, nona nine); the number after the hyphen only tells apart different sequences of the same length, so tripeptide-1 and tripeptide-10 are unrelated. 'Palmitoyl' or 'acetyl' means a fatty acid or acetyl group is attached, and 'sh-' marks a protein made from a synthetic copy of a human gene. Trade names such as Matrixyl 3000 or Leuphasyl do not appear on EU ingredient lists.