LL-37 is the only cathelicidin antimicrobial peptide humans make, and the honest bottom line is that its biology splits by concentration: a broad microbicide at the high local levels inside an abscess or at a degranulating neutrophil, and a signaling alarmin at the far lower levels of normal tissue. It holds no regulatory approval anywhere for any indication, and the controlled human evidence amounts to two topical venous leg ulcer trials and a four-patient early-phase intratumoral melanoma study. The same reactivity that drives wound repair also makes the peptide the established initiator of psoriasis, so the molecule is useful in a narrow window and harmful outside it.
LL-37 is the sole human cathelicidin, released from the inactive hCAP18 precursor by proteinase 3 in neutrophils and by kallikreins in skin, and it carries no marketing approval from any regulator for any indication.
The name is literal, and the more consequential fact sits behind it: no cell transcribes LL-37 directly. The CAMP gene encodes hCAP18, an inactive precursor built from a conserved cathelin prodomain and a variable C-terminal antimicrobial domain, and an active peptide exists only once a protease releases that domain. Which protease does the cutting depends on the tissue, and that detail decides whether the released product defends or damages.
| Criteria | Neutrophil route | Skin route |
|---|---|---|
| Precursor storage | Specific (secondary) granules | Keratinocytes, produced on demand |
| Cleaving protease | Proteinase 3 | Kallikrein-5 and kallikrein-7 |
| Released product | LL-37 itself | Shorter derivatives such as RK-31 and KS-30 |
| Reported trigger | Degranulation | Injury or infection |
| Consequence when disordered | Loss of the microbicidal burst | Disordered kallikrein activity produces disease, not defense |
The CAMP gene on chromosome 3p21.3 encodes the hCAP18 precursor, and humans carry exactly one cathelicidin gene where cattle and pigs carry a dozen or more, which leaves LL-37 without a functional backup.
Killing starts with charge rather than with molecular recognition. Bacterial surfaces are strongly anionic, carrying lipopolysaccharide or lipoteichoic acid, while the outer leaflet of a human plasma membrane is zwitterionic and stiffened with cholesterol, so a peptide carrying roughly six positive charges is drawn to the first and largely ignores the second. That electrostatic difference is the entire basis of the selectivity, and it is also the reason the potency measured in dilute buffer does not survive the move into blood.
Physiological salt at 150 millimolar and serum components, particularly apolipoprotein A-I, cut LL-37 potency by an order of magnitude or more, so the in vitro minimum inhibitory concentrations of a few micrograms per milliliter measured in dilute buffer do not carry into blood.
At the one to five micrograms per milliliter typical of normal tissue, LL-37 kills almost nothing yet stays highly active as a host defense signal, which is why much of the field now prefers host defense peptide over antimicrobial peptide. Whether it is pro- or anti-inflammatory is a malformed question: against endotoxin it is suppressive, against double-stranded RNA and self-nucleic acids it is powerfully amplifying, and what else is present sets the direction.
At the one to five micrograms per milliliter found in normal tissue, LL-37 is essentially non-microbicidal yet still neutralizes endotoxin and recruits neutrophils and monocytes through FPR2, so its signaling functions, not its killing, are what human tissue reliably achieves.
The link is structural, written into the DNA rather than inferred from a loose nutritional association. A functional vitamin D response element sits in the CAMP promoter, and it arrived on an Alu short interspersed nuclear element inserted in a primate ancestor roughly 55 to 60 million years ago, so it exists in humans and other primates and does not exist in rodents. Any mouse experiment on vitamin D and cathelicidin is therefore studying an animal that lacks the pathway, which is a standing caution against reading rodent data into human physiology here.
Serum 25-hydroxyvitamin D is rate-limiting for cathelicidin induction, so sera from vitamin D deficient donors support the induction poorly and are rescued when the substrate is added back, and human supplementation raises cathelicidin most convincingly in people who were deficient to begin with.
Injury switches cathelicidin on within hours: expression climbs at the wound edge, peaks through the proliferative phase, and falls as the epithelium closes, which is the first clue that the peptide is a repair signal and not only a disinfectant. The clinical interest follows from the wounds that stall, since chronic venous leg ulcers degrade the peptide in a protease-rich bed and diabetic wounds blunt its induction in the first place. What the trial record then shows is that more of it is not better.
LL-37 transactivates EGFR on keratinocytes through metalloproteinase-mediated release of heparin-binding EGF and promotes angiogenesis through FPR2 on endothelial cells by a route independent of VEGF, and blocking cathelicidin in experimental wounds delays closure by a margin that infection alone does not explain.
The mechanism that makes LL-37 pathogenic is the one that makes it useful, pointed at the wrong ligand. Extracellular self-DNA is normally ignored because nucleases degrade it before it can reach an endosomal sensor, but a cationic peptide and a polyanion condense into ordered aggregates that resist nuclease attack and are efficiently taken into plasmacytoid dendritic cells. What follows is not a peripheral observation about psoriasis; it is the current model of how the disease starts.
LL-37 is the established initiator of psoriasis through self-DNA complexes that trip TLR9 and the established driver of rosacea through aberrant kallikrein-5 processing, so exogenous administration introduces a molecule whose documented mode of action in disease is not a theoretical adverse event drawn from a safety table.
The selectivity that spares human cells is relative rather than absolute, and it fails as concentration rises, because a carpet mechanism that dissolves lipid bilayers has no way to spare a zwitterionic membrane once enough peptide accumulates on it. No dataset establishes how much of the resulting risk materializes in people, since the controlled human evidence consists of small topical wound studies and a four-patient early-phase study of intratumoral injection in melanoma.
No trial has established the safety of subcutaneous or systemic LL-37 in humans for any indication, so claims about such use rest on extrapolation rather than data.
The most accurate statement about LL-37 dosing is that no validated human protocol exists, and the numbers in circulation come from settings that do not translate into one another. Controlled human dosing covers topical wound application and one injected route only, that route being a four-patient early-phase melanoma study giving 250 micrograms per tumor by intratumoral injection every seven days for up to eight weeks. Handling shifts the numbers as well, since a strongly cationic peptide adsorbs avidly to glass and untreated plastic, so a solution prepared without low-binding tubes or a carrier can lose much of its nominal concentration to the container.
| Criteria | Laboratory antimicrobial work | Immunomodulatory work | Controlled human topical trial |
|---|---|---|---|
| Concentration | 1 to roughly 32 micrograms per milliliter | Roughly 1 to 5 micrograms per milliliter | 0.5, 1.6, and 3.2 milligrams per milliliter |
| Medium or route | Dilute low-salt buffer | Cell systems at tissue-relevant levels | Hydrogel on hard-to-heal venous leg ulcers |
| Schedule | Single-exposure assay | Single-exposure assay | Twice weekly over four weeks |
| Reported outcome | Minimum inhibitory concentrations that collapse in saline or serum | Signaling activity without killing | Significant healing at the lowest concentration only |
The most fully characterized controlled human dosing on record is a topical hydrogel at 0.5, 1.6, and 3.2 milligrams per milliliter twice weekly over four weeks in venous leg ulcers, and animal dosing translates poorly because rodents express CRAMP rather than LL-37 and lack the primate vitamin D response element in the cathelicidin promoter.
LL-37 holds no marketing approval from the FDA, the EMA, or any comparable regulator, for any indication, in any formulation. Clinical development has not progressed beyond Phase II, so there is no approved product to point to and no approved label from which a dose or an indication could be read. What exists commercially is a research chemical trade, and the protection that trade rests on is narrower than it looks.
LL-37 holds no marketing approval from the FDA, the EMA, or any comparable regulator for any indication in any formulation, its clinical development has not progressed beyond Phase II, and availability through a compounding pharmacy is not evidence of legal or regulatory standing.
Humans deploy two families of antimicrobial peptide, and LL-37 is the entire cathelicidin side of that pair, with the defensins forming the other. Against conventional antibiotics the comparison is favorable in principle and disappointing in practice: targeting membrane architecture rather than a single enzyme means killing is fast, works on dormant cells, and offers no single mutable target, yet pexiganan, iseganan, and omiganan all reached late-stage trials and none reached approval.
| Criteria | LL-37 (cathelicidin) | Defensins | Conventional antibiotic |
|---|---|---|---|
| Structure | Linear amphipathic helix, no cysteines | Compact beta-sheet locked by three disulfide bonds | Small molecule |
| Target | Membrane lipid architecture | Membrane lipid architecture | A specific enzyme or pathway |
| Trade-off | Conformational flexibility at the cost of protease lability and salt sensitivity | Disulfide stabilization partly escapes both | No lipid selectivity problem, one mutable target |
| Resistance route | Surface charge remodeling via mprF and teichoic acid D-alanylation, plus proteolysis | Partly shielded by disulfide bonds | Point mutation at the target |
| Approved products | None | None | Many |
Pexiganan, iseganan, and omiganan all reached late-stage trials without approval, defeated above all by a failure to show clear efficacy or superiority over conventional treatments, and LL-37 itself has gone no further, which is why engineering has concentrated on derived fragments such as KR-12, FK-13, and GF-17.
Educational use only. This article describes what the published scientific and clinical literature reports about LL-37. It is not medical advice, and it does not recommend, prescribe, or tell anyone to use anything described here. The regulatory status shown at the top of this page reflects what the record showed on the date given there and can change. mdpep.com does not sell any substance described here, does not endorse human use of it, and does not direct anyone to obtain it.
This is not guidance for your situation. Nothing here accounts for your medical history, your current medications, or anything else specific to you, and none of it should be used to make a decision about your own health.
Affiliate disclosure. Some links on this site are affiliate links, and mdpep.com may earn a commission when they are used. That never affects what is written here, it is not an endorsement of any vendor, and it is not a statement that anything described on this page is available to buy.
Every claim here ties to a named primary source and a date, written as plain documentation with nothing for sale. That is how MD PEP covers the whole peptide market.
