Palmitoyl pentapeptide-4 (INCI name; formerly palmitoyl pentapeptide-3), Pal-KTTKS
Palmitoyl pentapeptide-4 is KTTKS, a 5-amino-acid piece of type I procollagen, with a palmitic acid tail added to help it get into skin. In lab studies it signals skin cells to make collagen; Matrixyl is the supplier’s trade name.
Sport status follows the WADA Prohibited List in force since January 1, 2026; athletes should confirm with their anti-doping organization. Checked October 5, 2026. Status of every compound.
In cell studies, Matrixyl’s peptide, KTTKS, made fibroblasts (the skin’s collagen-making cells) produce more collagen and fibronectin, another structural protein; it was the smallest piece of type I procollagen, the raw form of skin’s main collagen, to do so. Its palmitic acid tail helps it into skin: in mouse-skin tests a little reached the dermis, the deeper layer, though skin enzymes broke it down quickly.
AnimalMainly from cell and animal studies; not shown in people.
Whether enough reaches the dermis of living human skin to raise collagen, and whether the trial’s wrinkle gains come from that, haven’t been shown.
Robinson 2005
Moisturizer with 3 ppm (0.0003%) pal-KTTKS on one half of the face for 12 weeks.
Lintner patent 2003
Developer test: 0.005% (50 ppm) pal-KTTKS around one eye, twice daily for 28 days.
Continuous, shown over 26 weeks
The ways Matrixyl is most often run, each tagged with where it comes from.
Robinson 2005, 93 women aged 35–55: fine lines and wrinkles improved vs the same moisturizer without it; well tolerated.
Patent data, not peer-reviewed: fold depth 18%, fold thickness 37% and skin rigidity 21% lower.
Applied to clean facial skin, often around the eyes and forehead. Topical only; never inject it.
Nothing special. Skip it and use the next application as usual.
Developer test around the eye: wrinkle fold depth 18% lower (patent data, not peer-reviewed).
TrialFine lines and wrinkles improved more on the treated half of the face than on the placebo half, by image analysis and expert grading (Robinson 2005).
TrialSkin irritation or redness (well tolerated in the 12-week trial)
Patch-test first. Stop if it doesn’t settle.
Eye irritation
Keep it out of the eyes. Rinse with water if it gets in.
Infection or harm from injecting a cosmetic serum
Never inject topical products; they aren’t sterile.
Stop and see a doctor for a spreading rash, swelling, blistering or hives, or eye pain after contact.
No product label covers Matrixyl, so these come from human studies and from precautions based on how it works. Most important first.
Precaution: follows from how it works or from a related drug, and hasn’t been tested. This isn’t a complete list: a pharmacist or doctor can check a specific medicine or condition.
Usually run on its own.
The published human study we found, with how many people took part. Animal studies aren’t listed here.
Fine lines and wrinkles improved significantly more on the peptide side by image analysis and expert grading, and users noticed it too; well tolerated. The abstract gives no effect sizes.
Robinson et al., Int J Cosmet Sci 2005, PMID 18492182
Only one peer-reviewed trial exists: 93 women aged 35–55, 12 weeks, one half of the face, run by scientists at a skin-care maker. Other human results come from the developer’s patent tests and products that mix it with other actives. Its collagen effect rests on cell and mouse-skin studies; stronger mixes and long-term use are untested.
Matrixyl is used on the skin, so the math is a dilution, not a syringe dose.
| Powder | Serum volume | Strength | Percent | ppm |
|---|---|---|---|---|
| 10 mg | 100 mL | 0.1 mg/mL | 0.01% | 100 ppm |
| 10 mg | 200 mL | 0.05 mg/mL | 0.005% | 50 ppm |
| 10 mg | 500 mL | 0.02 mg/mL | 0.002% | 20 ppm |
| 10 mg | 1,000 mL | 0.01 mg/mL | 0.001% | 10 ppm |
The 12-week trial used 3 ppm, about 10 mg in 3.3 liters of product. 10 mg in 200 mL gives 50 ppm, the level in the developer’s 28-day eye test.
Every compound side by side: the half-life chart and the storage chart.
Palmitoyl pentapeptide-4 is KTTKS, a 5-amino-acid piece of type I procollagen, with a palmitic acid tail added to help it get into skin. In lab studies it signals skin cells to make collagen; Matrixyl is the supplier’s trade name.
In cell studies, Matrixyl’s peptide, KTTKS, made fibroblasts (the skin’s collagen-making cells) produce more collagen and fibronectin, another structural protein; it was the smallest piece of type I procollagen, the raw form of skin’s main collagen, to do so. Its palmitic acid tail helps it into skin: in mouse-skin tests a little reached the dermis, the deeper layer, though skin enzymes broke it down quickly. Whether enough reaches the dermis of living human skin to raise collagen, and whether the trial’s wrinkle gains come from that, haven’t been shown.
Not approved as a drug in the US or EU; used as a cosmetic ingredient.
From human studies (Robinson 2005): Moisturizer with 3 ppm (0.0003%) pal-KTTKS on one half of the face for 12 weeks. Developer test: 0.005% (50 ppm) pal-KTTKS around one eye, twice daily for 28 days (Lintner patent 2003). These are the amounts sources reference, not recommendations.
No product label covers Matrixyl, so these come from human studies and from precautions based on how it works. Anyone thinking of injecting it: It is a cosmetic ingredient, not made to injectable standards, and has never been tested by injection. Anyone mixing pure powder at high strength: The only peer-reviewed trial used a very low strength; stronger mixes haven’t been tested for benefit or safety. Pregnancy or breastfeeding: No human safety data exist.
Before mixing: Freezer (about −20 °C) for long-term storage, kept dry. After mixing: Fridge (2–8 °C); use within about 4 weeks. This is common practice; no product label covers it.
No. Not named on the 2026 list. It’s a topical cosmetic ingredient, and USADA lists palmitoyl pentapeptide-4 as not prohibited.
Yes. The human studies section lists one published study. The largest, from 2005, included 93 people. Fine lines and wrinkles improved significantly more on the peptide side by image analysis and expert grading, and users noticed it too; well tolerated. The abstract gives no effect sizes.
For how it works, who should avoid it and the limits of the evidence