LL-37: what the trials show, status and safety

Summary

LL-37 is a peptide the human body makes itself, part of the first line of defence against microbes. Under the drug name ropocamptide it was tested as a liquid put on chronic leg wounds. The largest trial, in 149 adults, missed its main goal: 26.5% and 24.7% of wounds closed on the peptide against 25.3% on a dummy treatment. No country has approved it, and the American medicines authority lists it as an ingredient that may present significant safety risks.

Key findings at a glance

• The main trial missed its goal. In 149 adults with hard-to-heal venous leg ulcers, confirmed complete wound closure after 13 weeks was 26.5% on the lower strength, 24.7% on the higher one and 25.3% on the dummy treatment ^[1].
• The FDA names two concrete concerns, not just missing data. Its list of compounding ingredients that may present significant safety risks records that non-clinical findings suggest harm to male reproduction and that LL-37 can be protumorigenic in some tissues ^[2].
• It kills human cells at two to five times the amount that kills bacteria. The concentration that stops E. coli growing is 5 µM, a measure of how much peptide sits in a given volume. Damage to several human cell types begins at 13 to 25 µM ^[3].
• No person has ever received it into a vein or under the skin in a trial. In people it has only been put on wounds or injected straight into skin tumours ^[4], ^[1], ^[5].
• There is no pharmacokinetic data at all. No half-life, no absorption figure, for any route in any species. The FDA states it lacks the safety information needed to know whether the substance would harm people ^[2].
• In psoriasis, the immune system attacks LL-37 itself. T-cells directed against the peptide were found in the blood and skin of roughly two thirds of people with moderate to severe psoriasis ^[6].
• Banned in sport at all times. Class S0 of the 2026 prohibited list covers every drug-like substance that no health authority has approved for human use ^[7].

What it is

LL-37 is a chain of 37 amino acid building blocks that the human body produces on its own. It is the only cathelicidin humans have, a small family of defence peptides found in the white blood cells and surface tissues of many animals. It is made in neutrophils, the most common type of white blood cell. It is also made in the linings of skin, airways, gut and testis, in sweat glands and in several other immune cells ^[8], ^[9].

The name confusion around it is worth untangling once. CAMP is the gene, and hCAP-18 is the 170-building-block protein that gene makes. That protein does not kill bacteria on its own.

LL-37 is the short tail piece of it, positions 134 to 170. An enzyme called proteinase 3 cuts the tail loose after the cell has already pushed the protein out ^[8], ^[10]. Ropocamptide is the drug name given to laboratory-made LL-37 by the Swedish company Promore Pharma ^[1].

Two research groups described the sequence independently in 1995, one at the Karolinska Institute and one in Copenhagen ^[11], ^[12]. No pharmaceutical company invented it, and outside the Promore programme there has been no attempt to turn it into a medicine.

Quick facts

There is no pharmacopoeia entry for LL-37 anywhere, so no official purity or storage standard exists ^[14]. Its stability in the fluid of a non-healing wound was a development problem in its own right ^[15].

How it works

LL-37 has no single receptor and no single job. What it does depends on how much of it is present, on the salt content and acidity around it, and on whether blood serum is there.

• It breaks membranes open, and not selectively. The peptide lies flat along the surface of a membrane and dissolves it the way a detergent does, rather than punching a defined hole. The same property that bursts a bacterium damages a human cell once there is enough of it ^[16].
• Its shape decides its strength. In plain water LL-37 is a floppy, disordered chain. Negatively charged ions and higher concentrations force it into a helix, and the more helix there is, the better it kills bacteria. Below pH 5 the helix unwinds; at pH 2 the structure is gone entirely ^[3].
• It mops up bacterial endotoxin. LL-37 binds lipopolysaccharide, the molecule on the outside of many bacteria that drives the inflammatory reaction to an infection ^[8], ^[17].
• It may act through several immune receptors. Receptors are the docking points a cell uses to pick up a signal. Three candidates have been proposed: formyl peptide receptor 2, the purinergic receptor P2X7 and an indirect switch on the receptor for epidermal growth factor. None of them has been established as the LL-37 receptor ^[18], ^[19].
• It may turn the body's own DNA into an alarm signal. LL-37 clumps otherwise inert human DNA into packages that stay inside immune cells long enough to trigger the sensor TLR9. This is the proposed route by which tolerance to one's own DNA breaks down in psoriasis ^[20].
• It may push tumours along. In lung cancer it has been reported to switch on Wnt/beta-catenin, a growth pathway, through the sensor TLR4. In pancreatic cancer it is reported to keep cancer stem cells alive, and in melanoma to release inflammatory messengers ^[21], ^[19], ^[22].

One point separates LL-37 from most other peptides sold for research. Vitamin D switches the CAMP gene on directly, through a control element in the gene's promoter. That element is specific to primates and is absent in mice, rats and dogs. Rodent work therefore cannot reproduce this part of human biology ^[23].

Raising the body's own output is also not the same as putting the finished peptide somewhere. When the body makes it, the amount, the place and the timing are set by the cutting step ^[10], ^[23].

What the trials found

Wound closure in venous leg ulcers

Phase 2b RCT The trial that matters most is the one that failed. HEAL LL-37 randomised 149 adults with hard-to-heal venous leg ulcers in Poland and Sweden, in three equal groups, after a three-week run-in on the dummy treatment. Average age was 67.6 years, the median wound had been open for 20.3 months, and average wound area was 11.6 cm² ^[1].

The main measurement was confirmed complete closure of the target wound by week 13. Closure was checked again two weeks later and photographed. It came out at 26.5% on 0.5 mg/mL, 24.7% on 1.6 mg/mL and 25.3% on the dummy treatment.

Average healing rate per day was 0.0261, 0.0112 and 0.0204. Average time to closure was 83.1, 90.3 and 87.9 days ^[1].

A subgroup of larger wounds, from 10 cm² upwards, did show differences. That analysis was done after the results were known, which makes it a question for a future trial rather than an answer from this one ^[1].

The earlier and much smaller trial

Phase 1/2 RCT The first study in people enrolled 34 adults with the same kind of ulcer in Sweden, treating them twice a week for four weeks. The wounds closed faster. The measure used was the healing rate constant, which describes how quickly a wound shrinks. It came out about six times higher than on the dummy treatment at 0.5 mg/mL (p = 0.003), and about three times higher at 1.6 mg/mL (p = 0.088). Average wound area fell by 68% and 50% ^[4].

Two things must be read alongside that. The highest strength tested, 3.2 mg/mL, showed no difference from the dummy treatment at all. And the measurement used was a healing rate rather than a closed wound, which is exactly the point on which the larger trial then found nothing ^[4], ^[1].

Diabetic foot ulcers

Phase 2 RCT A single-centre trial in Jakarta compared an LL-37 cream against a dummy cream, on top of standard wound care, for four weeks. The participants were adults with a diabetic foot ulcer and a mild infection. Granulation tissue is the fresh tissue that fills a wound from below. It built up faster in the LL-37 group at all four measurement points (p = 0.031, 0.009, 0.006 and 0.037) ^[24].

The two inflammatory messengers measured, IL-1α and TNF-α, did not fall (p > 0.05), and the count of aerobic bacteria fell further in the dummy group by day 28. The trial planned 40 participants and its registry entry is still marked unknown ^[24].

Injection into melanoma tumours

Phase 1/2, single arm A study at the MD Anderson Cancer Center injected LL-37 straight into skin and under-skin melanoma deposits once a week for up to eight weeks. Four people were recorded as enrolled and three as treated: two at 250 µg per tumour and one at 500 µg. Two finished; one stopped because the treatment was not working ^[5].

Three treated people cannot show whether something works. A published case report from the same study describes a 63-year-old who developed multiple wart-like papules and a blistering lesion about 45 days after starting. Eleven of twelve biopsies showed an abnormal growth of skin cells. Everything cleared within two months of stopping ^[25].

Laboratory and animal work

Animal data In three rat models of severe bloodstream infection with Gram-negative bacteria, 1 mg/kg of LL-37 into a vein reduced deaths compared with salt solution. Endotoxin and TNF-α, an inflammatory messenger, both stayed lower than under conventional antibiotics ^[17]. This is the only substantial evidence for giving LL-37 into the body, and it is in rats.

Observational A study of 31 adults with severe sepsis measured several antimicrobial peptides in blood. Every one of them was raised except LL-37. It was also the only one that did not track the cardiovascular part of the organ failure score ^[26].

What is still unknown

• Whether LL-37 does anything measurable in a person when given anywhere other than on a wound ^[1], ^[5].
• How the body handles it: no half-life, no absorption, no distribution data, for any route ^[2].
• Whether it triggers antibody formation against itself. No trial has reported anti-drug antibodies, and the FDA names immunogenicity as an open risk for certain routes ^[2].
• What happens beyond 13 weeks, which is the longest treatment period ever studied ^[1].
• Whether the tumour-promoting findings in animals and cells apply to people ^[2], ^[21], ^[19].
• Anything about pregnancy, breastfeeding or children.

Side effects and safety

The honest summary is that there is no body of human safety data for LL-37 given into the body. What exists comes from wound treatment and from three people who had it injected into tumours.

• In the phase 2b trial, tolerability looked acceptable. At least one adverse event occurred in 20, 24 and 20 people across the two LL-37 groups and the dummy group. Events rated as possibly or probably related to the treatment numbered 4, 1 and 1. Serious events numbered 4, 7 and 1 ^[1].
• One signal in that trial points one way. Erysipelas, a spreading skin infection, appeared as a serious event only in the two LL-37 groups, once and twice. In no case was it at the treated wound ^[1].
• In the tumour study, no deaths and no serious events were reported. Other events recorded across the three people included a squamous cell carcinoma and an actinic keratosis, which is a sun-damaged patch of skin. Also recorded were skin depigmentation, an underactive thyroid, anaemia, falling lymphocyte and white cell counts, and rising bilirubin ^[5].
• Toxicity to human cells sits close to the useful range. Cells are damaged from 13 to 25 µM, against 5 µM to stop bacteria growing, and the mechanism is explicitly not selective for bacteria ^[3], ^[16].
• Cells need an active defence against it. A protein called p33 catches LL-37 inside the cell. Reduce p33 and the peptide becomes more toxic, so cells short of it are more vulnerable ^[27].
• It affects sperm. LL-37 punctures the membrane of human and mouse sperm in an amount-dependent way, peaking at 18 µM for human and 3.6 µM for mouse sperm. Treated mouse sperm fertilised nothing, in a dish or in an animal ^[28]. It is being developed as a vaginal contraceptive precisely because of this ^[29].
• It is involved in two skin diseases rather than helping them. People with rosacea carry abnormally high cathelicidin levels in facial skin, in badly processed forms, and injecting those forms into mice caused inflammation ^[30]. In psoriasis the immune system treats LL-37 itself as an enemy: T-cells directed against it are a confirmed finding ^[6]. In lupus, antibodies form against the peptide-DNA complexes ^[31].

The FDA's own wording is the clearest safety statement available. Its list of bulk drug substances that may present significant safety risks says three things about LL-37.

Compounded drugs containing it "may pose risk for immunogenicity for certain routes of administration". The agency "lacks sufficient safety-related information" to judge it. And non-clinical findings "suggest detrimental effects on male reproduction and that this drug can be protumorigenic in some tissues" ^[2].

Data gaps are wide. Nothing has been published on how the body handles LL-37, and no carcinogenicity, reproductive or genotoxicity testing to regulatory standard exists. No trial measured immunogenicity. There is no pharmacopoeia entry, and nothing at all on pregnancy, breastfeeding or children ^[2], ^[14].

Doses used in studies

The doses below are simply what the cited studies gave, listed so that the results can be understood. They are not advice on how to use anything.

Four points frame the table. In people, LL-37 has only ever been put on a wound or injected into a tumour. A concentration in mg/mL applied to a wound surface cannot be converted into an injected amount.

In the only dose-finding study, the highest strength was the one with no effect. And no human pharmacokinetic data exists to bridge any of it ^[4], ^[2].

No authority anywhere has set a dose, because no product is licensed anywhere ^[14].

Development and approval status

• 1995 — the peptide is first described, under two different names, by groups in Stockholm and Copenhagen ^[11], ^[12].
• 2014 — the first trial in people reports a positive signal on a healing rate in 34 adults ^[4].
• 2021 — the larger phase 2b trial reports that wound closure did not differ from the dummy treatment ^[1].
• 2026 — the FDA's list of compounding ingredients that may present significant safety risks carries cathelicidin LL-37, under the heading of substances nominated but withdrawn ^[2].

Status was checked on 7 September 2026. For the United States and Australia the finding rests on full-text checks of the relevant documents ^[2], ^[33]. For the other markets it rests on the investigational status stated in the publications and on the absence of any approval notice. That is weaker evidence ^[14].

Two further points shape what can be said. Neither Promore Pharma trial is registered on ClinicalTrials.gov, so searching that register alone misses the two most important studies ^[34].

LL-37 was also not among the substances the FDA's compounding advisory committee examined in July 2026. None of the detailed agency assessments that exist for several other research peptides exist for this one ^[35].

Anti-doping

LL-37 is banned in sport at all times, in and out of competition. It is not printed on the 2026 prohibited list by name. It falls under class S0, which covers "any pharmacological substance which is not addressed by any of the subsequent sections of the List and with no current approval by any governmental regulatory health authority for human therapeutic use" ^[7]. A substance in clinical development with no approval anywhere is exactly that case. The examples printed under S0 are explicitly not exhaustive.

That the body makes LL-37 itself changes nothing. What is banned is taking an unapproved pharmacological preparation, not the presence of the peptide in the body. The American college sports association does not name it either, and states that its list is neither complete nor exhaustive ^[36]. Wider context sits under peptides banned in sport.

Compared with related peptides

Comparing LL-37 with the peptides it is usually grouped with is only honest along the axes that the same documents cover. The table below uses the FDA compounding list, the July 2026 advisory committee agenda and the prohibited list, because all four substances appear in those records.

Sources for the table: ^[2], ^[35], ^[7].

One difference is real and worth stating plainly. Of these four, LL-37 is the only one with a randomised, placebo-controlled trial in people, published in a peer-reviewed journal, with a hard measurement ^[1]. That is a higher standard of evidence than the others have reached, and the result was that the treatment made no difference.

The second difference runs the other way. All four sit in the same withdrawn group. For the other three the agency records missing data and the general difficulties of peptide preparations, but no specific harm. For LL-37 it names two ^[2]. Better evidence and a worse risk profile are not a contradiction here. They are two consequences of the same thing: LL-37 has actually been studied.

Common misconceptions

• "It is made by the body, so it must be safe." The body keeps it under tight control. It is released by one specific cut, only after the cell has already pushed the precursor out, and a protein called p33 mops up what stays inside ^[10], ^[27]. Supplied from outside, it damages human cells from 13 to 25 µM. In psoriasis the immune system attacks it, and the FDA names two concrete laboratory harms ^[3], ^[6], ^[2].
• "LL-37 could replace antibiotics." It kills a wide range of bacteria in a dish. That activity is suppressed by human serum, and it depends on the salt and acidity around the peptide ^[3]. It has never been tested as a whole-body anti-infective in people. In severe sepsis, LL-37 was the one antimicrobial peptide whose blood level was not raised ^[26].
• "LL-37, ropocamptide, hCAP-18 and CAMP are four different things." They are four levels of the same thing. CAMP is the gene, hCAP-18 the inactive protein it makes, LL-37 the active tail piece cut from it, and ropocamptide the drug name for laboratory-made LL-37 ^[8], ^[10], ^[1].
• "Studies on the trial registers prove there is clinical evidence." Of ten interventional entries linked to LL-37, seven give no LL-37 at all. They measure it as a marker or try to raise the body's own output using vitamin D, sodium butyrate or topical ivermectin. Three of those were withdrawn and never ran ^[34].
• "More cathelicidin is good for the skin." In rosacea the opposite is documented. Affected skin carries abnormally high levels in badly processed forms, and injecting those forms caused inflammation in mice ^[30]. In psoriasis, LL-37 is what turns the body's own DNA into an alarm signal ^[20].
• "It was trialled against melanoma, so it fights cancer." One single-arm study treated three people and reached no conclusion ^[5]. In parallel, independent work reports LL-37 promoting pancreatic, lung and melanoma tumours, and the FDA has taken that finding into its own risk statement ^[19], ^[21], ^[22], ^[2].
• "Injectable LL-37 offered online is the trial material." It differs in four ways. It has never been given under the skin in a trial. The FDA names a risk of immune reactions for certain routes, and difficulty in characterising the substance at all. No pharmacopoeia sets a purity standard for it. And the nomination for the American compounding list was withdrawn, so there is no legal compounding route either ^[2], ^[14].

Frequently asked questions

Is LL-37 approved as a medicine anywhere?

No. As of 7 September 2026 no country has approved LL-37, under that name or as ropocamptide. The furthest it reached was a completed phase 2b trial that missed its main measurement. In the United States it appears on the FDA's list of compounding ingredients that may present significant safety risks, in the group of substances nominated but withdrawn.

What is LL-37?

It is a chain of 37 amino acid building blocks that the human body makes itself, and the only cathelicidin humans have. It is cut from a larger protein called hCAP-18 after white blood cells release it. Its normal job is part of the innate immune response, damaging bacterial membranes and shaping inflammation.

What did the LL-37 wound healing trials find?

The largest one found no difference. In 149 adults with hard-to-heal venous leg ulcers, confirmed complete wound closure after 13 weeks was 26.5% and 24.7% on the two LL-37 strengths against 25.3% on the dummy treatment. An earlier trial in 34 people had reported a faster healing rate, which the larger trial did not confirm.

Is LL-37 safe just because the body produces it?

No, and the two questions are separate. Toxicity to human cells starts at 13 to 25 µM, only two to five times the 5 µM that stops bacteria, and the mechanism is not selective. In psoriasis the immune system treats LL-37 as a target, and the FDA names harm to male reproduction and tumour promotion as non-clinical concerns.

What is the difference between LL-37, ropocamptide and hCAP-18?

They describe the same molecule at different stages and under different naming conventions. CAMP is the gene, hCAP-18 the 170-building-block protein it produces, which is not antibacterial on its own, and LL-37 the 37-building-block tail cut from it. Ropocamptide is the international drug name for laboratory-made LL-37.

Can LL-37 be used instead of antibiotics?

There is no evidence for that in people. It kills bacteria in a dish, but human serum suppresses that activity, and the effect depends heavily on acidity and salt content. The only evidence for giving it into the body comes from rat sepsis models at 1 mg/kg into a vein.

How long does LL-37 stay in the body?

Nobody knows. No half-life, absorption or distribution data has been published for LL-37 by any route in any species. That gap is one of the reasons the FDA states it lacks the safety information needed to judge whether the substance would harm people.

Has LL-37 ever been injected under the skin in a trial?

No. In people it has only been applied to wound surfaces as a liquid or cream, and injected directly into skin tumours in one small study of three treated people. There is no registered trial of LL-37 given under the skin or into a vein in humans.

Is LL-37 banned in sport?

Yes, at all times, in and out of competition. It is not named on the 2026 prohibited list. Class S0 covers every pharmacological substance that no health authority has approved for human use, drugs in clinical development included. The fact that the body produces LL-37 naturally does not change that.

What is still unknown about LL-37?

Almost everything about giving it to people beyond the wound surface. There is no pharmacokinetic data, no systemic dosing experience, no immunogenicity measurement from any trial and nothing beyond 13 weeks of treatment. Whether the tumour-promoting findings from cells and animals apply to people is also unresolved.

Cite this page

Facts on this page were verified on 7 September 2026, and the regulatory and anti-doping documents behind the status tables were read in full on that date. LL-37 is an unapproved substance under research, so the version and date matter as much as the content.

myPeptides Research & Editing. (2026). LL-37: what the trials show, status and safety. Version 1.0, 7 September 2026. myPeptides Peptide Register. Retrieved from https://mypep.app/peptides/ll-37

How pages in this register are compiled and graded is described under methodology; the full register is at peptides.

Last verified: 7 September 2026. Next review: on any new clinical trial of LL-37, on a change to the FDA compounding list, or on any change in the anti-doping classification.