LL-37 Research Peptide

Humans carry a single cathelicidin gene, CAMP, and its product is hCAP18: an 18 kDa protein with a cathelin-like domain at the front and the 37-residue LL-37 sequence at the back, stored inactive in the peroxidase-negative granules of neutrophils. Sørensen and colleagues (Blood, 2001) found that the precursor stays intact inside the phagolysosome and is only cut once it has been released from the cell, and that proteinase 3, one of the three serine proteases of the azurophil granules, is the enzyme solely responsible.

Skin handles the precursor differently. Yamasaki and colleagues (FASEB Journal, 2006) showed that kallikrein 5 and kallikrein 7, proteases of the outer skin layer, activate hCAP18 and then trim LL-37 into shorter peptides with activities of their own.

LL-37 carries eleven lysines and arginines against five aspartates and glutamates, a net charge of about +6 at neutral pH, and its hydrophobic and charged residues fall on opposite sides once it folds. Fold is the key word. In pure water at low concentration the peptide is largely disordered; anions such as bicarbonate and sulphate push it into an alpha-helix, and the degree of helicity tracked its antibacterial activity against Gram-positive and Gram-negative bacteria (Johansson and colleagues, Journal of Biological Chemistry, 1998). Shortened versions missing the first four or eight residues needed more anion to fold and were weaker against E. coli.

Membrane studies by Oren and colleagues (Biochemical Journal, 1999) added more detail. LL-37 sits in balance between single molecules and oligomers even at very low concentrations, and it resisted protease breakdown both in solution and when bound to membranes. On bacteria-like membranes its behaviour fitted a detergent-like “carpet” mechanism, and infrared work placed the helix almost flat along mammalian-type membrane surfaces rather than spanning them as a channel. It is cytotoxic to normal eukaryotic cells as well as to bacteria in culture, which is why viability controls feature so heavily in this literature.

Much of the current interest in LL-37 concerns host cells rather than microbes. De Yang and colleagues (Journal of Experimental Medicine, 2000) showed that it attracts human neutrophils, monocytes and T cells by acting on formyl peptide receptor-like 1 (FPRL1, now FPR2), giving it a recruiting role in innate and adaptive immunity.

Whether that host-directed side really matters in a living animal was settled with genetics. Nizet and colleagues (Nature, 2001) deleted the mouse cathelicidin gene and found the animals markedly more vulnerable to necrotic skin infection by group A Streptococcus, confirming cathelicidins as a genuine component of innate defence rather than an in-vitro curiosity. Human genetics points the same way: patients with Kostmann syndrome, a severe congenital neutropenia, were found to lack the peptide in plasma and saliva. Two further strands, the complexes LL-37 forms with the body’s own DNA and the over-processing seen in rosacea skin, are covered in the questions below.

Heilborn and colleagues (Journal of Investigative Dermatology, 2003) found that hCAP18 levels in human skin climb after wounding, peaking around 48 hours and falling back once the wound closes, yet the peptide was missing from the epithelium at the edge of chronic ulcers. In organ-cultured human skin, antibodies raised against LL-37 held back re-epithelialisation in a concentration-dependent way, pointing to a part for the peptide in the migrating wound edge.

That led to topical studies in hard-to-heal venous leg ulcers. A first-in-human trial in 34 participants (Grönberg and colleagues, Wound Repair and Regeneration, 2014) compared three LL-37 concentrations with placebo over four weeks: healing-rate constants were roughly six- and threefold higher in the two lower-concentration groups, significant only for the lowest, while the highest showed no difference from placebo. The follow-up phase IIb HEAL LL-37 trial in 148 patients (Mahlapuu and colleagues, 2021) found no significant improvement in the full study population; a post hoc analysis suggested better healing in the subgroup with wounds of at least 10 cm², a finding that would need confirming in a dedicated study.