LL-37 5mg Peptide

LL-37 is the only cathelicidin that humans produce. The CAMP gene on chromosome 3p21.3 encodes a 170-residue precursor called hCAP18, which carries a conserved cathelin-like domain across residues 31 to 131 and the active peptide at its C-terminus, residues 134 to 170 (UniProt P49913). Those 37 residues open with two leucines, and that is the whole of the name: LL, then 37. The free peptide has a molecular weight of 4,493.3 g/mol, a formula of C205H340N60O53, and a net charge near +6 at neutral pH. The charge, paired with a hydrophobic face on the opposite side of the helix, does most of the work.

The precursor is not active. Sorensen and colleagues (2001) showed that hCAP18 survives phagocytosis inside the neutrophil intact and is only cleaved once it has been exocytosed, and that of the three serine proteases in azurophil granules, proteinase 3 alone performs the cut. A second, unrelated activation route exists: seminal plasma carries hCAP18 in its inactive form, and the prostate protease gastricsin cleaves it at vaginal pH to ALL-38, a 38-residue peptide with antimicrobial activity equal to LL-37 (Sorensen et al., 2003). UniProt lists further processed forms including FALL-39, LL-23, FF-33, RK-31, KS-30 and KR-20, which matters to anyone characterising a sample: the tissue does not make one peptide, it makes a family, and only the 134 to 170 species is LL-37.

The literature on this peptide is unusually deep and it cuts in both directions. The same molecule that Nizet and colleagues (2001) showed protects mouse skin from necrotic Group A Streptococcus infection is the one Lande and colleagues (2007) identified as the factor that lets plasmacytoid dendritic cells mistake host DNA for a pathogen in psoriasis, and that Yamasaki and colleagues (2007) found abnormally processed and abnormally abundant in rosacea skin. LL-37 holds no marketing authorisation in any jurisdiction. The furthest it has been taken clinically is a topical Phase 2b programme in venous leg ulcers that missed its primary endpoint, and it is supplied here as a research reagent only.

The bactericidal action of LL-37 has no receptor. It is a physical chemistry problem: a cationic peptide that folds into an amphipathic helix only when the ionic environment permits, and then partitions into any membrane whose surface is negative enough to hold it. Johansson and colleagues (1998) mapped that dependence by circular dichroism. At micromolar concentration in plain water LL-37 is disordered. Adding 15 mM bicarbonate, sulfate or trifluoroacetate drives helix formation with roughly equal efficiency, while 160 mM chloride does so much less efficiently. The transition from disorder to helix is cooperative with rising peptide concentration, which is the signature of oligomer formation, and the extent of helicity tracks antibacterial activity against both Gram-positive and Gram-negative organisms. Below pH 5 the helical content falls away and at pH 2 the peptide is entirely disordered, while at pH above 13 the helix is retained.

Wang (2008) determined the high-resolution NMR structure of uniformly 13C and 15N labelled LL-37 in SDS micelles and found a curved amphipathic helix-bend-helix motif spanning residues 2 to 31, with the bend located between Gly-14 and Glu-16 and a disordered C-terminal tail. In dioctanoylphosphatidylglycerol micelles the aromatic rings of Phe-5, Phe-6, Phe-17 and Phe-27, along with the arginines, gave intermolecular nuclear Overhauser effects with the lipid, direct evidence that the entire helix lies against the anionic surface rather than inserting end-on. The same work identified KR-12, residues 18 to 29, as the shortest fragment retaining antibacterial activity, and showed that unlike the parent peptide it is selectively toxic to bacteria and not to human cells.

How the bound peptide breaks the membrane is best described by the carpet model rather than by a discrete pore. Oren and colleagues (1999) used polarised attenuated total reflectance Fourier-transform infrared spectroscopy and found LL-37 predominantly helical and oriented nearly parallel to the surface of zwitterionic-lipid membranes, an orientation incompatible with a transmembrane channel and consistent with detergent-like disruption once surface coverage passes a threshold. The same study found LL-37 in equilibrium between monomers and oligomers in solution at very low concentrations, significantly resistant to proteolysis both free and membrane-bound, and showed that the N-terminal residues govern proteolytic resistance and haemolytic activity but not antimicrobial potency. That dissociation is the reason N-truncated analogues behave so differently on host cells than on bacteria.

Alongside the membrane chemistry, LL-37 is a signalling ligand, and here it does use a receptor. Yang and colleagues (2000) showed that it is chemotactic for human monocytes, neutrophils and T lymphocytes, mobilises calcium in monocytes and in FPRL1-transfected HEK293 cells, and is cross-desensitised by an FPRL1-specific agonist, identifying formyl peptide receptor-like 1, now FPR2, as the receptor. Koczulla and colleagues (2003) then found the same receptor on endothelial cells, where LL-37 drove proliferation and vessel-like structure formation. A third mode is neither membranolytic nor receptor-mediated: LL-37 binds nucleic acids and condenses them into aggregates, which changes how the immune system reads them. That single property underlies both its endotoxin-neutralising activity and, on the wrong substrate, its role in autoimmunity.

The immunological output is not simply pro-inflammatory. Scott and colleagues (2002) profiled macrophage gene expression and found LL-37 directly up-regulating 29 genes and down-regulating 20, inducing monocyte chemoattractant protein 1 in macrophages and mouse lung and interleukin-8 in A549 epithelial cells and in whole human blood, while leaving TNF-alpha unchanged. It recruits without igniting. In the same study the peptide blocked macrophage stimulation by lipopolysaccharide, lipoteichoic acid and non-capped lipoarabinomannan, and protected mice against lethal endotoxaemia.

The peptide arrived from two directions at once in the mid-1990s. Larrick and colleagues (1995) cloned human CAP18 from a bone marrow cDNA library using probes designed against the rabbit sequence, described a protein made of a 30-residue signal peptide, a 103-residue N-terminal domain of then-unknown function, and a 37-residue C-terminal antimicrobial domain, and showed that synthetic CAP18(104-140) bound LPS-coated erythrocytes, blocked LPS-induced nitric oxide and tissue factor generation, and protected mice from LPS lethality. In the same period the Swedish group of Agerberth and Gudmundsson identified the same C-terminal region as FALL-39, named for its first four residues and its length as then assigned. The two names converged once the processing site was pinned down.

The cathelin-like domain that occupies most of the precursor is the family signature: every cathelicidin across mammals carries a conserved pro-region of this kind attached to a wildly variable C-terminal effector. In cattle and pigs the effector is released by neutrophil elastase. Sorensen and colleagues (2001) showed that human biology does it differently, using proteinase 3 and doing it outside the cell after exocytosis rather than inside the phagolysosome. Immunoelectron microscopy put hCAP18 and the azurophil granule proteins together in the phagolysosome, yet immunoblotting found no cleavage there; the cut appeared only in exocytosed material. That finding reframed LL-37 from an intracellular killing agent into an extracellular one, which is consistent with everything later established about its signalling roles in the wound bed and on the airway surface.

Bals and colleagues (1998) localised LL-37/hCAP18 transcripts by in situ hybridisation to the surface epithelium of the conducting airway and to serous and mucous cells of the submucosal glands, and recovered antimicrobially active peptide from human airway surface fluid and from a bronchial xenograft model. Nizet and colleagues (2001) then supplied the genetic proof of relevance in vivo, combining mice lacking Cnlp, the gene for the murine homologue CRAMP, with Group A Streptococcus mutants, and showing that cathelicidin is a real component of cutaneous host defence rather than an in vitro curiosity. Those two papers are why the field takes the peptide seriously despite how easily it is inactivated in a test tube.

The activity of LL-37 is conditional in a way that few research peptides are, and understanding the conditions is the difference between an interpretable assay and an uninterpretable one. Ionic strength is the first variable. Turner and colleagues (1998) measured the MIC against E. coli ML-35p rising from 0.6 to 7.6 micrograms per millilitre when 100 mM NaCl was added, a more than tenfold loss, although in radial diffusion underlay gels containing 100 mM NaCl the peptide retained activity against most organisms tested, with MRSA the exception. Bals and colleagues (1998) independently reported NaCl sensitivity of killing at the airway surface, alongside synergy with lactoferrin and lysozyme.

Serum is the second and larger problem. Johansson and colleagues (1998) noted that in solutions matching the ion composition of plasma, interstitial fluid or intracellular fluid, LL-37 is helical and therefore in principle active, but that human serum inhibits both its antibacterial and its cytotoxic activity. Wang and colleagues (1998) identified why. Affinity chromatography on an LL-37 column pulled out a single 26 kDa band, sequenced as apolipoprotein A-I. Surface plasmon resonance put the interaction in the low micromolar range, 50 micromolar apoA-I accounted for about half of the inhibition plasma exerts on 50 micromolar LL-37, and anti-apoA-I IgG completely blocked plasma inhibition up to 25 micromolar peptide. Sorensen and colleagues (1999) extended the picture to the precursor: plasma hCAP18 is carried on lipoproteins, delipidated plasma retains under 1 percent of it, and the binding is mediated by the same C-terminal region.

Those findings together explain the shape of the whole field. LL-37 is an excellent surface-acting peptide in the environments where the body actually deploys it, which are the sweat film, the airway lining fluid, the wound bed and the zone immediately outside a degranulating neutrophil, and a poor one in bulk plasma, where lipoproteins scavenge it. It is also why nearly every attempt to develop the molecule has been topical, and why analogue work concentrates on shortened, stapled or D-substituted derivatives that resist both proteolysis and lipoprotein capture. Ridyard and Overhage (2021) list the four barriers plainly: synthesis cost, reduced activity in physiological environments, proteolytic susceptibility, and host cell toxicity, and add a fifth that is easy to miss, namely bacterial adaptive responses to sub-inhibitory peptide.

Of everything in the LL-37 literature, the regulatory finding is the most striking. Gombart, Borregaard and Koeffler (2005) showed that 1,25-dihydroxyvitamin D3 and three of its analogues induce CAMP transcription in acute myeloid leukaemia lines, immortalised keratinocytes, colon cancer lines, normal human bone marrow-derived macrophages and fresh bone marrow cells, acting through a consensus vitamin D response element in the CAMP promoter that is bound by the vitamin D receptor. The unusual part is comparative. That response element sits inside a short interspersed nuclear element and is conserved across primates but absent from the mouse, rat and canine genomes. Murine cells were not induced, and CAMP mRNA in vitamin D receptor-deficient mouse bone marrow was indistinguishable from wild type.

The functional consequence was demonstrated by Liu and colleagues (2006). Toll-like receptor activation of human macrophages up-regulated both the vitamin D receptor and the vitamin D 1-hydroxylase, which converts circulating 25-hydroxyvitamin D into the active hormone locally, leading to cathelicidin induction and killing of intracellular Mycobacterium tuberculosis. Sera from African-American donors, a group with recognised higher tuberculosis susceptibility, had low 25-hydroxyvitamin D and were inefficient at supporting cathelicidin mRNA induction in that assay.

Two practical implications follow for anyone designing work with this compound. First, a rodent model cannot be used to study vitamin D-driven cathelicidin induction, because rodents lack the response element; humanised CAMP transgenic mice were built specifically to work around that gap. Second, endogenous cathelicidin is a moving background in any human sample, varying with vitamin D status, infection, smoking and periodontal disease. That is exactly what most of the registered clinical studies naming LL-37 are measuring.

The clinical case for LL-37 in wounds was built on an absence rather than a presence. Heilborn and colleagues (2003) showed hCAP18 rising in human skin after injury, peaking at 48 hours and falling to pre-injury levels once the wound closed, present in both the inflammatory infiltrate and the epithelium migrating over the wound bed, and then showed that in chronic ulcers hCAP18 levels are low and immunoreactivity is absent from the epithelium at the ulcer edge. Blocking the peptide with an affinity-purified antibody in a non-inflammatory ex vivo model of organ-cultured human skin inhibited re-epithelialisation in a concentration-dependent manner, and Ki67 staining disappeared from the epithelium of the blocked wounds. The peptide is not merely present during repair; the repair depends on it.

The first-in-man trial followed that logic. Gronberg and colleagues (2014) enrolled 34 participants with hard-to-heal venous leg ulcers, ran a three-week open-label placebo lead-in, then randomised to four weeks of twice-weekly topical LL-37 at 0.5, 1.6 or 3.2 mg/mL or placebo, with four weeks of follow-up. Healing rate constants at 0.5 and 1.6 mg/mL were approximately six-fold and three-fold that of placebo (p = 0.003 and p = 0.088), mean ulcer area fell 68 percent and 50 percent respectively, and the highest concentration, 3.2 mg/mL, was indistinguishable from placebo. That non-monotonic pattern is what the concentration-dependent host cytotoxicity of the peptide predicts, and it is one of the more informative results in the whole field.

The Phase 2b trial did not confirm it. Mahlapuu and colleagues (2021) randomised 149 patients across 15 sites in Poland and Sweden, mean age 67.6 years, median ulcer duration 20.3 months, mean wound size 11.6 square centimetres, to twice-weekly 0.5 mg/mL, 1.6 mg/mL or placebo for 13 weeks. Confirmed complete closure sustained at the two-week post-closure visit occurred in 26.5, 24.7 and 25.3 percent of the three arms. Nothing separated them, and the mean healing rates, 0.0261, 0.0112 and 0.0204 per day, did not either. An exploratory analysis of ulcers of 10 square centimetres or larger did separate: 28.1 percent closure against 8.1 percent for placebo (p = 0.0458), healing rate 0.0367 against 0.0093 per day (p = 0.0439), time to closure 87.4 against 97.5 days (p = 0.0066), and 47.2 percent against 16.2 percent achieving at least a 70 percent area reduction (p = 0.0149). Local reactions, mostly mild redness and oedema, occurred in 39 to 43 percent of patients with no consistent difference between the active arms and placebo. A subgroup finding from a trial that missed its primary endpoint is a hypothesis rather than a result, and the authors say so.

A second registered wound study, NCT04098562, is a Phase 2 trial of LL-37 cream at 0.5 mg/mL applied twice weekly for four weeks in diabetic foot ulcers, run by Universitas Indonesia, with granulation index, aerobic bacterial counts and IL-1 alpha and TNF-alpha as endpoints in 40 planned participants. No results have been posted to the registry.

Lande and colleagues (2007) asked why plasmacytoid dendritic cells, which normally ignore host DNA, respond to it in psoriatic skin, and found LL-37 to be the answer. The peptide binds host DNA into aggregated, condensed structures that are delivered to and retained within early endocytic compartments of the pDC, where they engage TLR9 and drive type I interferon production. The DNA has not changed; its packaging has. This is the same nucleic-acid-condensing property that underlies endotoxin neutralisation, operating on the wrong substrate, and it sits at the centre of a whole branch of psoriasis immunology. Later work extended the same logic to RNA associated with neutrophil extracellular traps.

Rosacea has an adjacent story with a different lesson. Yamasaki and colleagues (2007) found that people with rosacea express abnormally high cathelicidin in facial skin, and, more importantly, that the processed forms present are different from those in normal skin, because stratum corneum tryptic enzyme activity is elevated in the epidermis and cuts the precursor differently. Injecting the rosacea-associated peptide forms into mouse skin, adding the protease, or increasing protease activity by deleting the serine protease inhibitor gene Spink5 each increased inflammation, and mice with Camp deleted did not develop it. Processing, not abundance alone, is the determinant.

There is also a straightforward toxicity ceiling that no amount of formulation removes. Johansson and colleagues (1998) reported LL-37 cytotoxic to several eukaryotic cell types at 13 to 25 micromolar, against a MIC of 5 micromolar for E. coli, giving a window under a factor of five in the absence of serum. Oren and colleagues (1999) showed the peptide is haemolytic and that the N-terminal residues, which are dispensable for antimicrobial activity, are what confer that haemolysis. This is why fragment engineering has been so productive: KR-12, residues 18 to 29, retains antibacterial activity while losing toxicity toward human cells (Wang, 2008), and stapled KR-12 derivatives such as KR-12(Q5, D9) increase helical content and net charge, resist proteolysis, and raise the therapeutic index relative to the parent in an infected-wound mouse model (Zhang et al., 2024).

Zheng and colleagues (2007) complete the picture from the other side: LL-37 makes human neutrophils generate reactive oxygen species through NADPH oxidase, produce IL-8 under p38 and ERK control, and release the alpha-defensins HNP1 to 3. In an infected wound that is host defence. In psoriatic skin, where neutrophil infiltration is a defining feature, the same activity is amplification. Nothing about the molecule changes between those two settings, which is the honest summary of why a peptide this potent has taken thirty years to fail its way toward a topical indication.

LL-37 sits in the same catalogue category as BPC-157, TB-500 and GHK-Cu, but it works on a different axis and the comparison is worth making precisely. BPC-157 and TB-500 are studied as regenerative signalling peptides with no direct antimicrobial component, and neither is a human gene product with a defined receptor pharmacology. GHK-Cu carries a metal and acts largely on matrix gene expression. LL-37 is the only one of the group whose primary described activity is direct physical destruction of a microbial membrane, and the only one whose endogenous counterpart is under transcriptional control by a vitamin.

The nearest functional neighbour is not in the Healing and Recovery category at all but among the immune peptides: thymosin alpha-1, another human-derived peptide studied as an immunomodulator. The two differ in evidence weight in a way worth stating plainly. Thymosin alpha-1 has been through dozens of randomised trials and holds marketing authorisation in a number of countries. LL-37 has two published randomised topical wound trials, one positive at 34 participants and one null at 148, plus a Phase 1/2 intratumoral melanoma study at MD Anderson (NCT02225366) that enrolled 4 of a planned 36 participants before completing in November 2020, reporting no dose-limiting toxicities in the 3 evaluable participants across cohorts at 250 and 500 micrograms per tumour.

Within the LL-37 family itself the meaningful comparison is to its own fragments. KR-12 is the shortest sequence retaining antibacterial activity and is bacterially selective in a way the parent is not. KR-20, KS-30, RK-31 and FF-33 arise from further skin protease processing and each carries a different balance of antimicrobial to inflammatory activity, which is precisely what Yamasaki and colleagues found dysregulated in rosacea. ALL-38, the gastricsin product from seminal plasma, is 38 residues with equal antimicrobial potency. A vial labelled LL-37 that has been mis-synthesised or partially degraded is likely to contain one of these, which is why the analytical question below is not a formality.

LL-37 is supplied as a white lyophilised powder and is hygroscopic in that state. The lyophilised material is the well-characterised form: held desiccated at -20 degrees Celsius, protected from light and moisture, and brought to room temperature before the vial is opened so that condensation does not reach the powder. Aqueous stocks are held at 2 to 8 degrees Celsius, aliquoted for single use, and not cycled repeatedly through freezing and thawing. Two properties specific to this peptide deserve attention beyond the generic advice. It is strongly cationic, so it adsorbs to glass and to untreated polypropylene, and dilute working solutions lose measurable material to the container wall; low-binding tubes and a carrier protein or non-ionic detergent are the standard countermeasures in the membrane literature. And its conformation is anion-dependent, so the buffer is not neutral with respect to the result: a stock prepared in a bicarbonate or sulfate-containing buffer is more helical, and therefore more active, than the same nominal concentration in plain water.

The pH sensitivity documented by Johansson and colleagues (1998) is the other handling consequence. Helical content declines below pH 5 and is entirely lost at pH 2, which matters because trifluoroacetate is the standard counterion left by reverse-phase purification, and residual TFA both acidifies a stock and, at the concentrations it can reach, independently promotes helix formation. A peptide characterised as an acetate salt and one characterised as a trifluoroacetate salt are not interchangeable in a circular dichroism experiment, and net peptide content differs between them for a molecule carrying six positive charges.

Analytically, LL-37 poses a harder identity problem than most catalogue peptides because so many of its truncations are themselves biologically active and chromatographically close. Reverse-phase HPLC gives purity as a percentage of peak area; on its own it does not distinguish LL-37 from a des-Ser C-terminal deletion or from the shorter processed forms. Mass spectrometry is what separates them, and the average mass expected from C205H340N60O53 is 4,493.3, with a peptide carrying six positive charges giving a straightforward electrospray charge envelope. Because the sequence contains four phenylalanines and no tryptophan or tyrosine, ultraviolet quantitation at 280 nanometres is unavailable, and concentration must be established by amino acid analysis or by far-ultraviolet absorbance instead. That missing 280 nanometre chromophore is a real and frequently overlooked difference between LL-37 and most other peptides in this catalogue.

Thirteen studies on ClinicalTrials.gov name LL-37, and reading the list is instructive because only a minority administer it. NCT02225366 (Phase 1/2, MD Anderson Cancer Center, intratumoral injection in stage IIIB, IIIC and IV M1a melanoma, completed November 2020) and NCT04098562 (Phase 2, diabetic foot ulcer cream, Universitas Indonesia) are the interventional entries. The venous leg ulcer programme that produced the 2014 and 2021 publications is the substantive clinical dataset, and it was run outside those two registrations.

The remaining registrations measure endogenous LL-37 as a biomarker. Several concern periodontal and peri-implant disease, where cathelicidin in gingival crevicular fluid and saliva is tracked against disease stage, smoking status and vitamin D level: NCT04404335, NCT04861493, NCT06867250, NCT07280754 and NCT04946617 sit in that group. NCT04292548 and NCT03639376 measure salivary LL-37 in children exposed to passive smoking. NCT03270709 was a Phase 1 high-dose vitamin D3 study in smokers with and without HIV that was terminated, and NCT02138591, a Phase 3 preoperative vitamin D supplementation trial, was also terminated.

The shape of that record is the honest position of the compound. LL-37 is far better established as a measurable component of human innate immunity, and as a biomarker of it, than as an administered agent. Its two randomised administration trials were both topical, both in chronic wounds, and produced a positive small trial followed by a null larger one. Anything claimed about systemic use is extrapolation from cell and rodent work.