LL-37 is a short human defense peptide studied for its ability to help skin cells cover a wound as well as for its antimicrobial and immune effects. The most informative treatment studies applied synthetic LL-37 directly to chronic ulcers. A small venous-ulcer trial found faster early wound shrinkage at its lowest tested concentration, but a larger trial found similar complete-closure estimates with LL-37 and placebo across the full study population. A separate diabetic-foot study found more new tissue filling the wound without establishing greater wound-area reduction. Together, these findings make the type of wound, formulation and definition of healing central to understanding LL-37.
Overview
The body produces LL-37 by cutting a larger protein, human cathelicidin antimicrobial protein 18 (hCAP18). The released peptide contains 37 amino-acid residues. In neutrophils, a type of white blood cell, Sørensen and colleagues showed that this processing occurred in material released outside the cells and identified proteinase 3 as the responsible enzyme. LL-37 and its precursor therefore refer to related but distinct molecules. Sørensen and colleagues, 2001
Molecular profile
- Form
- Synthetic human LL-37; the larger venous-ulcer trial used the acetate salt in a polyvinyl alcohol preparation Source
- Length
- 37 amino-acid residues Source
- Natural precursor
- hCAP18, a larger human cathelicidin protein from which LL-37 is released Source
- Clinical research route
- Direct application to persistent venous leg ulcers in a defined topical formulation Source
- Diabetic-foot preparation
- LL-37 cream applied directly to ulcers, compared with matching cream without peptide Source
Forms and derivatives
hCAP18 and LL-37
hCAP18 is the larger precursor protein. LL-37 is the released 37-residue peptide; a tissue measurement of the precursor is not automatically a measurement of free active LL-37.
Topical LL-37 acetate in PVA
HEAL LL-37 tested synthetic LL-37 acetate in a viscous polyvinyl alcohol preparation against PVA without LL-37. All groups received dressings and compression for venous ulcers.
LL-37 cream
The diabetic-foot trial used LL-37 in a base cream, compared with matching cream without the peptide. Its formulation methods state 0.5 mg/g; other passages and the registry use mg/mL, so those units are not treated as interchangeable.
Oral LL-37 bacterial product
The COVID-19 report used recombinant LL-37-producing Lactococcus lactis. Its delivery system and virologic outcomes are separate from those of purified peptide applied to an ulcer.
Gallium-67-labeled LL-37 variant
The mouse imaging study attached a radioactive label through a chelator to follow distribution. The tracer measurements do not establish human disposition of unmodified LL-37.
The venous-ulcer clinical program used a synthetic peptide with the same amino-acid sequence as endogenous human LL-37. In the larger trial, the drug substance was LL-37 acetate in a preparation containing polyvinyl alcohol (PVA), a polymer that makes the liquid viscous. The control was PVA without LL-37, and both groups received wound dressings and compression bandaging. This formulation matters because the trial tested an addition to structured wound care in people with persistent ulcers. Mahlapuu and colleagues, 2021, methods
Uses and research areas
Venous leg ulcers are persistent open sores associated with impaired return of blood through the leg veins. Compression helps address that circulation problem. The LL-37 trials enrolled people whose ulcers had remained open despite previous care and asked whether adding the peptide improved healing. The smaller study measured how quickly ulcer area declined during four weeks of treatment. The larger study emphasized complete closure: new skin covering the wound, with no drainage or need for a dressing, sustained at a confirmation visit two weeks later. A wound can become smaller without reaching that more demanding outcome. Grönberg and colleagues, 2014; Mahlapuu and colleagues, 2021
The diabetic-foot trial addressed another setting: uninfected or mildly infected ulcers in adults with diabetes. It measured wound area, the fraction occupied by granulation tissue, inflammatory proteins in wound fluid and bacteria grown from wound swabs. Granulation tissue is new connective tissue and small blood vessels filling the wound bed; it supports repair before new skin completely covers the surface. Its appearance, wound shrinkage and bacterial clearance answer different questions. Miranda and colleagues, 2023
Other programs use substantially different preparations. A randomized COVID-19 study tested an oral product containing recombinant LL-37-producing Lactococcus lactis, a bacterial delivery system. A melanoma registry describes injections directly into tumors. Those preparations and disease settings need separate interpretation from purified peptide applied to an ulcer. Zhao and colleagues, 2023; melanoma study registry
Mechanism of action
Skin repair requires cells at the wound edge to move across the exposed surface and rebuild its covering. Heilborn and colleagues found that hCAP18 increased after acute human skin injury and was present in the migrating epithelium, the sheet of cells covering the wound. Chronic ulcer-edge epithelium lacked the same staining for hCAP18/LL-37. In separately cultured human skin wounds, antibodies directed against LL-37 inhibited resurfacing. This linked the endogenous peptide system to repair and supplied a reason to test replacement at chronic wounds. It did not establish that adding LL-37 would close an ulcer in a patient. Heilborn and colleagues, 2003
Tokumaru and colleagues then examined cultured human keratinocytes, the principal cells of the outer skin layer. Cells exposed to LL-37 moved more in a migration assay. Blocking activation of the epidermal growth factor receptor, which helps cells respond to growth signals, prevented that migration response; the experiments also implicated the downstream signaling protein STAT3. These cell experiments explain one possible route from peptide exposure to resurfacing. The clinical ulcer trials measured healing outcomes and did not show that this pathway accounted for an individual patient's response. Tokumaru and colleagues, 2005
Immune effects depend on the surrounding tissue and the material available to bind LL-37. In research involving psoriasis and human immune cells, Lande and colleagues found that LL-37 bound the body's own DNA and enabled it to trigger interferon production by plasmacytoid dendritic cells. These cells ordinarily respond to microbial nucleic acids while ignoring much of the body's own DNA. The result helps explain how an endogenous defense peptide can also participate in inflammatory disease; increasing its activity does not have a single predictable consequence throughout the body. Lande and colleagues, 2007
Pharmacokinetics
The wound trials did not establish a human blood concentration curve, absorption fraction or elimination half-life for administered LL-37. Applying peptide to an ulcer exposes the wound surface, but the application schedule does not reveal how much intact peptide reaches the circulation or how long it remains active in the tissue. A graph based on a presumed human half-life would give more precision than these reports support. Venous-ulcer trial; diabetic-foot trial
Two experiments help explain why local persistence and blood clearance require different measurements. Grönberg and colleagues added synthetic LL-37 to diluted fluid collected from nine non-healing venous ulcers. They found no apparent peptide cleavage during incubation for up to 24 hours, although a separate enzyme experiment showed susceptibility to trypsin. This was a laboratory stability test of collected fluid; it did not measure retention on a living wound or provide a human plasma half-life. Grönberg and colleagues, 2011
Esposito and colleagues studied a gallium-67-labeled LL-37 variant in healthy mice using nuclear imaging. After intravenous administration, the tracer left blood quickly and accumulated mainly in the liver, with temporary lung distribution. After subcutaneous administration, clearance was predominantly through the kidneys. Free gallium-67 scans served as an additional control. These observations concern a chemically modified tracer in mice; they do not supply a measured human half-life for the unmodified therapeutic peptide. Esposito and colleagues, 2024
Research findings
A small trial found faster early shrinkage at the lowest concentration
The first randomized venous-ulcer study enrolled 34 people. After a three-week period using placebo, participants received topical LL-37 at one of three concentrations or placebo during a four-week blinded treatment period, followed by four weeks of observation. The investigators fitted a mathematical model to the shrinking wound area. The estimated healing-rate constant was approximately six times the placebo value with 0.5 mg/mL LL-37 (P = 0.003), and approximately three times the placebo value with 1.6 mg/mL (P = 0.088). The latter comparison remained too uncertain to establish faster healing. Grönberg and colleagues, 2014
Mean ulcer area fell by 68% and 50% in those two groups, respectively. The highest tested concentration, 3.2 mg/mL, did not establish faster shrinkage than placebo. The favorable model estimate at the lowest concentration justified a larger trial, while the short treatment period and small sample left complete closure and durability unresolved. A sixfold rate constant describes the fitted change in area; it does not mean six times as many people had a closed wound.
The larger trial found similar closure estimates in the full population
HEAL LL-37 randomized 149 people at 15 sites in Poland and Sweden. One never received treatment; 148 contributed safety observations, and 144 met the criteria for the main efficacy analysis. Participants received one of the two lower LL-37 concentrations or PVA placebo for 13 weeks alongside compression and dressings. The primary outcome required complete closure during treatment that remained closed at confirmation two weeks later. Mahlapuu and colleagues, 2021, methods and results
The estimated proportions reaching confirmed closure were 26.5% with 0.5 mg/mL LL-37, 24.7% with 1.6 mg/mL and 25.3% with placebo. The paper reports 90% confidence intervals of 17.1–38.7%, 15.8–36.4% and 16.4–36.9%, respectively. The active-placebo comparisons did not establish improved closure (one-sided P = 0.4453 and 0.5274). The proportions were estimated with a statistical model from groups of 46, 48 and 50 in the full analysis, rather than calculated as simple fractions of every randomized participant. The closely spaced estimates and their uncertainty provide no evidence of an overall closure advantage under this trial's conditions; they do not prove that the true effect is exactly zero.
The large-ulcer finding came from an exploratory analysis
The investigators subsequently examined participants whose wounds measured at least 10 cm² at randomization. In this subgroup, the estimated closure proportions were 28.1% with the lower LL-37 concentration, 19.6% with the higher concentration and 8.1% with placebo. The lower-concentration comparison favored LL-37 (P = 0.0458); the higher-concentration closure comparison remained uncertain (P = 0.1393). Each group contained only 21–24 people. Mahlapuu and colleagues, 2021, Table 4 and discussion
Wound size had been used to balance allocation during randomization, but this analysis of treatment benefit in large wounds was post hoc, meaning it was undertaken after the main analysis rather than serving as the trial's planned primary test. Several related healing measurements were examined, and the closure comparisons used one-sided tests that assessed improvement in the expected direction. The observation identifies a population worth testing in a trial designed for that question; it does not replace the full-population result. Attendance at the later follow-up visits was also low, limiting conclusions about whether healed wounds stayed closed.
HEAL LL-37 compared two topical LL-37 preparations with polyvinyl alcohol (PVA), the viscous preparation without peptide, alongside compression and dressings. Points show model-estimated confirmed closure by 13 weeks and bars show 90% confidence intervals. The full efficacy analysis included 46, 48 and 50 participants; the exploratory subgroup with ulcers ≥10 cm² included 21, 21 and 24. Closure required new skin without drainage or dressing need and confirmation two weeks later. The full-population result did not establish a closure benefit; the post hoc subgroup needs a confirmatory trial.
See the figure data
| Analysis | Group | Estimate | Lower CI | Upper CI | Analysis n |
|---|---|---|---|---|---|
| Full efficacy population | LL-37 0.5 mg/mL | 26.5 | 17.1 | 38.7 | 46 |
| Full efficacy population | LL-37 1.6 mg/mL | 24.7 | 15.8 | 36.4 | 48 |
| Full efficacy population | PVA without LL-37 | 25.3 | 16.4 | 36.9 | 50 |
| Ulcers ≥10 cm² at randomization | LL-37 0.5 mg/mL | 28.1 | 14.9 | 46.6 | 21 |
| Ulcers ≥10 cm² at randomization | LL-37 1.6 mg/mL | 19.6 | 8.8 | 38.0 | 21 |
| Ulcers ≥10 cm² at randomization | PVA without LL-37 | 8.1 | 2.5 | 23.1 | 24 |
The diabetic-foot trial separated tissue filling from wound shrinkage
Miranda and colleagues randomized 25 adults with uninfected or mildly infected diabetic foot ulcers in Jakarta: 13 received LL-37 cream and 12 received matching base cream without the peptide. Most ulcers were uninfected. Both groups received saline cleansing, removal of dead tissue and dressings. Over four weeks, the fraction of the wound occupied by granulation tissue increased more with LL-37. At day 21, the mean increase in this fraction from baseline was 0.64 ± 0.30 with LL-37 and 0.30 ± 0.26 with placebo (mean ± standard deviation; P = 0.006). Those values describe changes in a tissue-coverage index, not the proportion of ulcers that closed. Miranda and colleagues, 2023, methods, results and Table 2
Wounds also became smaller in both groups. By day 28, the mean reduction in area was 2.35 ± 3.86 cm² with LL-37 and 1.54 ± 0.94 cm² with placebo. The comparisons at each visit were too uncertain to establish greater area reduction with LL-37 (P > 0.05). The study likewise did not establish a reduction in the measured inflammatory proteins, IL-1α and TNF-α, or in aerobic bacteria grown from wound swabs. The new-tissue finding therefore supports a particular repair response in a small, short study; it does not demonstrate complete healing or clinical eradication of infection.
An oral bacterial product was studied separately in COVID-19
Zhao and colleagues reported an open-label, single-center randomized trial in 238 adult inpatients using an oral Lactococcus lactis product that produced recombinant LL-37. The paper reported that viral RNA tests became negative earlier, compared with oral placebo, in participants treated within six days after confirmation of infection. This outcome concerns viral detection during Omicron BA.5.1.3 infection, and its timing-dependent analysis requires care. It does not establish concentrations of intact LL-37 in human blood, the effects of a purified peptide injection, or a general benefit across respiratory infections. Zhao and colleagues, 2023
Safety and tolerability
The topical trials supply limited, formulation-specific tolerability information. In HEAL LL-37, at least one adverse event was reported by 20 of 48 lower-concentration participants, 24 of 49 higher-concentration participants and 20 of 51 placebo participants. There were 12 nonfatal serious events in 11 people and no deaths; investigators judged none of the serious events related to the study drug. Six less serious events in four people were considered possibly related to the study preparation, including wound infection and dermatitis in the lower-concentration group, peripheral swelling in the higher-concentration group and deep phlebitis in the placebo group. The appearance of an event during treatment and an investigator's judgment about its cause are separate pieces of information. Mahlapuu and colleagues, 2021, safety results and Table 5
In the diabetic-foot trial, one LL-37 participant developed mild irritant dermatitis attributed to excess wound fluid; the report did not attribute a side effect directly to LL-37. With only 25 participants and four weeks of treatment, that observation offers little information about uncommon or delayed harms. Miranda and colleagues, 2023
LL-37 remains an investigational treatment. FDA's current compounding safety page lists cathelicidin LL-37 among substances whose nominations were withdrawn and retains concerns about immune reactions, peptide-related impurities and verification of the manufactured ingredient’s identity. It also cites laboratory or animal findings suggesting harm to male reproduction and promotion of tumor growth in some tissues. Those nonclinical findings do not establish the frequency of such harms in people, while short topical trials cannot resolve them. The regulatory listing is separate from approval of a medicine and does not validate an injectable LL-37 product. FDA compounding safety assessment, accessed October 15, 2026
What the research means
The wound studies show why healing needs a concrete definition. Faster shrinkage in an early model, more granulation tissue in the wound bed and durable coverage by new skin represent different stages of repair. LL-37 has biological reasons to influence these stages, and some small studies detected favorable changes. The larger venous-ulcer study did not establish better complete closure across its full population, despite a favorable exploratory result in larger wounds.
The next clinical test would need to define the wound population and closure outcome in advance, retain a comparable wound-care control and follow healed wounds long enough to assess recurrence. The existing results are most useful for designing that test. They do not establish a broadly effective systemic treatment for wound healing, infection or immune enhancement.
FAQs
Did LL-37 close more venous leg ulcers than placebo?
The larger HEAL LL-37 trial did not establish better complete closure across its full study population. At 13 weeks, estimated confirmed closure was 26.5% with the lower LL-37 concentration, 24.7% with the higher concentration and 25.3% with polyvinyl alcohol without peptide. All groups also received compression and dressings. A favorable analysis in larger ulcers was exploratory and needs a trial designed to confirm it. Trial report
What did the diabetic-foot trial improve?
Among 25 adults with uninfected or mildly infected diabetic foot ulcers, LL-37 cream increased the fraction of the wound covered by granulation tissue more than matching cream without peptide. Granulation tissue is new tissue filling the wound bed. Wounds became smaller in both groups, but the differences in area reduction remained too uncertain to establish a benefit from LL-37. The study also did not establish reduced bacterial colonization. Trial report
Does the sixfold result mean wounds healed six times faster?
The sixfold estimate came from a mathematical model of declining wound area in the lowest-concentration group of a 34-person venous-ulcer trial. It describes a healing-rate constant over a short treatment period. It is not a sixfold increase in the number of closed ulcers or a demonstrated sixfold reduction in time to complete healing. Earlier trial
Is there a measured human half-life for LL-37?
The clinical wound reports reviewed here do not establish a human concentration-time curve or elimination half-life for administered LL-37. Laboratory stability in collected wound fluid and imaging of a labeled variant in mice answer other questions. Neither can be substituted for blood measurements of intact LL-37 in people. Wound-fluid experiment; mouse imaging study
Does being a natural human peptide establish treatment safety?
The body releases LL-37 in specific tissues and circumstances. Administering a manufactured preparation changes that exposure. The short topical trials provide some local tolerability observations, while FDA retains concerns about immune reactions, impurities and nonclinical reproductive and tumor-related effects for compounded LL-37. These sources do not establish a general safety profile for injected products. FDA assessment
Comparisons and related reading
References
The wound-healing papers distinguish changes in wound area, granulation tissue and confirmed closure. The mouse imaging study used a labeled variant to track distribution, while the oral bacterial-product study measured viral detection. These experiments answer different questions about repair, movement through the body and infection testing.
Experiments with human neutrophils distinguished the precursor hCAP18 from released LL-37 and identified extracellular processing by proteinase 3.
Human wound staining and separate cultured human skin experiments linked hCAP18/LL-37 to resurfacing. Antibodies directed against LL-37 inhibited repair in the cultured tissue model.
Cultured keratinocytes moved more after LL-37 exposure. Blocking EGFR activation prevented that migration response, supporting a possible repair mechanism.
Work involving psoriasis and human immune cells showed LL-37 enabling self-DNA to stimulate interferon production, explaining a context in which its immune activity can promote inflammation.
Synthetic LL-37 resisted apparent cleavage in diluted fluid from nine non-healing venous ulcers during laboratory incubation, while a separate enzyme comparison showed susceptibility to trypsin. This is local-fluid stability work, not human pharmacokinetics.
The 34-person trial found faster model-estimated early wound shrinkage at the lowest tested concentration. The middle-concentration rate comparison remained uncertain, and the highest concentration did not establish faster shrinkage than placebo.
- Mahlapuu M, Sidorowicz A, Mikosinski J, Krzyżanowski M, Orleanski J, Twardowska-Saucha K, Nykaza A, Dyaczynski M, Belz-Lagoda B, Dziwiszek G, Kujawiak M, Karczewski M, Sjöberg F, Grzela T, Wegrzynowski A, Thunarf F, Björk J, Ekblom J, Jawien A, Apelqvist J. Evaluation of LL-37 in healing of hard-to-heal venous leg ulcers: A multicentric prospective randomized placebo-controlled clinical trial. Wound Repair and Regeneration. 2021;29(6):938–950. doi:10.1111/wrr.12977. · PubMed record · doi.org
HEAL LL-37 found similar complete-closure estimates across its full study population. The favorable large-ulcer finding was exploratory. The report distinguishes 149 randomized people, 148 treated, 144 in the main efficacy analysis and 12 serious events in 11 people.
- Miranda E, Bramono K, Yunir E, Reksodiputro MH, Suwarsa O, Rengganis I, Harahap AR, Subekti D, Suwarto S, Hayun H, Bardosono S, Baskoro JC. Efficacy of LL-37 cream in enhancing healing of diabetic foot ulcer: a randomized double-blind controlled trial. Archives of Dermatological Research. 2023;315(9):2623–2633. doi:10.1007/s00403-023-02657-8. · PubMed record · Full text (PMC)
The 25-person diabetic-foot trial found greater increases in the fraction of wound area occupied by granulation tissue. Wound-area reduction, inflammatory-protein changes and aerobic bacterial counts did not establish a benefit over matching base cream.
A separate oral LL-37-producing Lactococcus lactis product was tested in 238 adult inpatients. The report described earlier negative viral RNA testing with early treatment; this delivery system does not establish exposure to purified injected peptide.
The registration identifies a separate investigation of LL-37 injected directly into melanoma lesions. It is cited for research setting and route, without an efficacy claim.
In healthy mice, a gallium-67-labeled LL-37 variant left blood quickly after intravenous administration and accumulated mainly in the liver. After subcutaneous administration, clearance was predominantly through the kidneys. The experiment does not establish a human half-life.
The current FDA page lists concerns about immune reactions, peptide-related impurities and verifying the manufactured ingredient’s identity. It also cites nonclinical reproductive and tissue-specific tumor-growth concerns. LL-37 appears under nominations withdrawn, not the current Category 2 table.
Glossary
- Cathelicidin
- A family of defense proteins; human hCAP18 is processed to release LL-37.
- Re-epithelialization
- New surface cells moving and growing across a wound to rebuild its skin covering.
- Granulation tissue
- New connective tissue and small blood vessels filling a wound bed before skin covers the surface.
- PVA
- Polyvinyl alcohol, the polymer used to make the venous-ulcer trial preparation viscous. The placebo contained PVA without LL-37.
- Post hoc analysis
- An analysis undertaken after the main analysis rather than serving as a trial's planned primary test.
- Confidence interval
- A range showing uncertainty around an estimate under a stated statistical method. The HEAL LL-37 table reports 90% intervals.
- Keratinocyte
- The principal cell of the outer skin layer, which helps rebuild the covering of a wound.