LL-37

A chain of 37 amino acids that the human body makes itself. The gene CAMP produces a longer protein called hCAP-18, which is stored in white blood cells and released by skin, lung and gut lining; an enzyme cuts the last 37 building blocks off it, and those 37 are LL-37 — named after its first two letters, both leucine. It is the only cathelicidin humans have. The synthetic version has been developed as a prescription wound medicine under the international name ropocamptide, and is separately sold online as a freeze-dried powder labelled for research use only, promoted for infections, wound healing and gut problems. The sequence can be printed here because it is verifiable letter by letter: it is residues 134 to 170 of UniProt P49913, and PubChem's chemical name for the same molecule shows every residue in the ordinary L form with plain, uncapped ends.

Sold as

In published human studies: a gel applied to a leg ulcer, and an injection into a skin tumour. In the COVID-19 trial, a swallowed engineered bacterium rather than the peptide. Sold online as a freeze-dried powder for injection under the skin — a route for which no human study has ever been published.

Legal status in the EU

Not approved as a medicine in the EU, Sweden, the United States or anywhere else, and no cathelicidin product has been approved in any country. It has been an investigational drug: the Swedish company Promore Pharma ran the phase IIb leg-ulcer trial in Poland and Sweden under EudraCT 2018-000536-10, authorised to proceed in August 2018, and holds the international non-proprietary name ropocamptide for it. In the United States, cathelicidin LL-37 was nominated as a bulk substance for pharmacies to compound; the FDA placed it in the category of substances that may present significant safety risks, and the nomination was subsequently withdrawn by the nominators. In Sweden, selling an unapproved medicine to consumers is prohibited under the Medicinal Products Act (2015:315); an SVT investigation in May 2026 reported that peptides of this kind are nonetheless sold to private buyers and that the Medical Products Agency warned their contents are often unknown.

It is shaped like a rod with a greasy side and an electrically charged side. Bacterial membranes carry a negative charge that human cell membranes largely do not, so the charged side sticks to bacteria and the greasy side pushes into their membrane until it breaks open. That is only half the job: LL-37 also acts as a signal, pulling white blood cells towards an injury, prompting new blood vessels to grow into damaged tissue, and steering the switch between fighting infection and repairing tissue. The same signalling is why the body having too much of it is a problem in its own right — bound to a person's own DNA it turns that DNA into an alarm the immune system reacts to.

Part of the bodyWhat happensHow it is knownKind of evidence

A long-standing open sore on the lower leg

This is the only use of LL-37 that has ever been properly tested against a dummy treatment. A gel containing it was rubbed onto the ulcer twice a week for 13 weeks, on top of the usual compression bandaging. Roughly a quarter of the ulcers closed completely. Roughly a quarter of the ulcers on the dummy gel closed too.

Measured in 148 adults across 15 sites in Poland and Sweden, in a trial where neither patients nor staff knew who got what. Average age 67.6, average wound size 11.6 cm², median time the wound had already been open 20.3 months. A smaller first-in-man trial in 34 people had reported faster shrinking four weeks in, which is what the larger trial was built to confirm and did not.

Measured in people

Source [03]Source [05]

Bacteria living in a wound

LL-37 is an antibiotic the body makes, so the obvious expectation is that painting it on a wound reduces infection. In the 148-person trial it did not do so to any degree the trial could see. Infections of the treated ulcer were rare everywhere: 1 patient on the lower dose, 2 on the higher dose, and 3 on the dummy gel.

Measured in the same 148 adults over 13 weeks of treatment and 16 weeks of follow-up. These are single-digit counts in groups of about fifty, so the trial could not have detected anything but a large effect — but it is the only measurement of LL-37 and infection in people that exists.

Measured in people

Source [03]

Skin, when your own body makes too much of it

This is the direction people rarely hear about. In rosacea — the facial redness and bumps that affect roughly 3 in 100 adults over 30 — the skin carries abnormally high amounts of cathelicidin, and chopped-up forms of it that are not found in normal skin. Injecting those forms into mouse skin produced inflammation. In psoriasis, LL-37 sticks to the person's own DNA and turns it into an alarm signal the immune system reacts to, and two-thirds of people with moderate to severe plaque psoriasis have immune cells that specifically attack LL-37 itself.

Measured in patients' own skin and blood, not in anyone given LL-37 as a treatment. The rosacea work compared facial skin from people with rosacea against normal skin and then reproduced the inflammation in mice. The psoriasis work found LL-37-reactive T cells in the blood of about two-thirds of a group of patients with moderate to severe plaque psoriasis, and the number of those cells tracked how active the disease was.

Measured in people

Source [12]Source [13]Source [16]

A tumour it is injected into

Injected straight into a skin tumour once a week, on the theory that it would wake the immune system up inside the tumour. The trial ran for more than five years at one American cancer centre and managed to enrol four people. One of the three who were treated stopped because it was not working.

4 people enrolled between July 2015 and November 2020 at MD Anderson Cancer Center; 3 received injections, 2 at 250 micrograms per tumour and 1 at 500. There was no comparison group and no dummy injection. The results the sponsor posted for tumour response record a number identical to the number of people analysed and carry no label, so they cannot be read as a count of anyone who responded.

One small study

Source [07]

A virus in the body, when the peptide is grown in the gut

The one trial of LL-37 taken by mouth did not give people LL-37. It gave them a swallowed preparation of a Lactococcus lactis bacterium engineered to produce LL-37 inside the gut. The published abstract calls this "Oral LL-37" and does not describe the formulation any further; this register could not read the full paper, which is behind a paywall. Among people started on it within six days of a confirmed COVID-19 diagnosis, the time to a negative swab was shorter than on placebo.

238 adult hospital inpatients at a single Chinese military hospital, randomly assigned, but open-label — everyone knew who was getting what, which matters for a result measured partly by when swabs were taken. The published abstract reports the comparison only within the group treated early, not for everyone who was randomised, and this register could not read the full paper, which is behind a paywall.

One small study

Source [09]

Sperm

LL-37 kills sperm. It breaks open the outer membrane of human sperm in a dish, and mouse sperm exposed to it then failed to fertilise eggs — both in a dish and inside a live female mouse. This is not a side finding somebody stumbled on: one research group is deliberately developing LL-37 as a vaginal contraceptive because of it.

Human sperm from several donors, treated in a dish at concentrations within the range the body itself reaches. Mouse sperm tested both in a dish and by injecting treated sperm into female mice, 26 mice per group. No study has looked at what LL-37 does to fertility in a man who takes it.

Only seen in animals

Source [11]Source [18]

Human cells, at only slightly higher amounts than it takes to kill bacteria

The gap between the amount that kills bacteria and the amount that kills your own cells is narrow. In the classic laboratory study, LL-37 stopped E. coli growing at 5 micromolar and became toxic to several kinds of human and animal cells at 13 to 25 micromolar — a factor of about three to five. The same study found that ordinary blood serum suppresses both effects, the killing of bacteria and the killing of cells alike.

Cells and bacteria in dishes only. Nobody has measured what concentration LL-37 reaches anywhere in a person after any route of administration, so nobody can say where a real dose would sit on that scale.

Only seen in a dish, not in a body

Source [10]

Anywhere, after an injection under the skin

Nothing is known. This is the way LL-37 is most often sold and used outside a hospital, and there is no published human study of it — not for infection, not for gut problems, not for healing, not even a study measuring how much of it ends up in the blood.

A ClinicalTrials.gov API v2 search on 3 August 2026 for LL-37 as an intervention returned 20 studies. Reading them, only two ever gave a person LL-37 itself, both by putting it on or into tissue rather than injecting it under the skin. The FDA's own published statement is that it lacks sufficient safety information on cathelicidin LL-37 to know whether it would cause harm when given to people.

Claimed, never tested

Source [06]Source [08]

The one properly controlled trial failed. In 148 adults with hard-to-heal venous leg ulcers — open sores on the lower leg caused by poor blood return through the veins — 13 weeks of LL-37 gel twice a week closed an estimated 26.5 per cent of ulcers at 0.5 mg/mL and 24.7 per cent at 1.6 mg/mL, against 25.3 per cent on a dummy gel (p = 0.45 and p = 0.53). The company's press release led instead with a post-hoc subgroup of 66 patients whose wounds were at least 10 cm² (28.1 per cent versus 8.1 per cent, p = 0.046); the paper itself labels that analysis post-hoc, and no confirming trial has been run. The earlier 34-person dose-finding study (Grönberg et al. 2014) had reported healing about six times faster than placebo at 0.5 mg/mL and no difference at all at the highest strength tested. Beyond leg ulcers the human record is two studies: an intratumoral injection study in melanoma that enrolled 4 people over five years and treated 3, with one stopping for lack of effect (NCT02225366), and a 238-person open-label randomised trial in COVID-19 that did not give LL-37 at all but a swallowed Lactococcus lactis bacterium engineered to make it in the gut. No published human study has ever given LL-37 by injection under the skin, into a vein, or by inhaler. A ClinicalTrials.gov API v2 search on 3 August 2026 for "LL-37" as an intervention returned 20 studies, of which only two administered the peptide itself; neither is recruiting.

For a leg ulcer, the honest answer is that 13 weeks of twice-weekly treatment produced no more closed wounds than a dummy gel. The earlier 34-person study saw faster shrinking within four weeks at the two lower strengths, and that signal did not survive the larger trial. Beyond that, nothing is known about timing, because there is no human pharmacokinetic study of LL-37 by any route — nobody has published how much of it reaches the blood, how long it stays there, or how fast it is broken down in a person. One thing the laboratory does suggest is that it may not travel: in solutions containing ordinary blood serum, LL-37's ability both to kill bacteria and to damage cells was suppressed.

What was measuredWhat happenedIn whomHow it is known

Ulcers that closed completely within 13 weeks

26.5 per cent closed on the lower dose of LL-37 and 24.7 per cent on the higher dose, against 25.3 per cent on the dummy gel. Neither difference came close to being real (p = 0.45 and p = 0.53). This was the question the trial was designed to answer, and the answer was no.

148 adults treated, out of 149 randomly assigned and 190 screened, all with a leg ulcer that had resisted healing; average age 67.6, average wound 11.6 cm², median duration 20.3 months. The percentages above are estimated from the full analysis set of 144 (46, 48 and 50 per group). Everyone, including the dummy group, also got compression bandaging.

Measured in people

Source [03]

Infection of the treated ulcer

1 patient (2.1 per cent) on the lower dose and 2 patients (4.1 per cent) on the higher dose, against 3 patients (5.9 per cent) on the dummy gel. None were serious and none were judged to be caused by the treatment. With counts this small in groups of about fifty, this neither shows nor rules out an antibacterial benefit.

The same 148 adults, over 13 weeks of treatment.

Measured in people

Source [03]

Ulcers that closed, counting only the largest wounds

Among people whose wound was at least 10 cm² to begin with, 28.1 per cent closed on the lower dose and 19.6 per cent on the higher dose, against 8.1 per cent on the dummy gel; the lower dose reached statistical significance (p = 0.046). This is the number the company's press release led with. The paper itself calls the analysis post-hoc — meaning the subgroup was picked out after the results were in, not written into the plan beforehand, which is how chance findings are manufactured.

21, 21 and 24 people in the three groups — a total of 66 out of the 144 in the full analysis set. No trial has since been run to test whether the finding holds.

One small study

Source [03]

How fast ulcers shrank over four weeks, in the first trial in people

The speed of healing was about six times the dummy gel's at 0.5 mg/mL (p = 0.003) and about three times at 1.6 mg/mL (p = 0.088, which is not significant). At the highest dose, 3.2 mg/mL, there was no difference from the dummy gel at all. A treatment that works better at a low dose than a high one is a warning sign, not a bonus.

34 adults with hard-to-heal venous leg ulcers: three weeks on the dummy gel first, then four weeks randomly assigned to one of three LL-37 strengths or the dummy, then four weeks of follow-up. Note that this measured how fast wounds shrank, not whether they closed.

One small study

Source [04]

Days until a COVID-19 swab turned negative

9.80 days on average, plus or minus 2.67, against 14.04 days plus or minus 5.89 on placebo. That comparison covers only the people started on treatment within six days of diagnosis. The published abstract does not report the same comparison for everyone who was randomised, and the trial was open-label, so both patients and staff knew who was getting what.

238 adult inpatients at one hospital in China during an Omicron BA.5.1.3 wave. What they swallowed was a Lactococcus lactis bacterium engineered to make LL-37, not LL-37 itself.

One small study

Source [09]

Sperm's ability to fertilise an egg (mice)

Mouse sperm treated with LL-37 did not fertilise eggs, either in a dish or after being placed in a live female mouse; untreated sperm did. Human sperm treated in a dish had their outer membrane broken open in a dose-dependent way compared with untreated sperm.

26 female mice in each of the two conditions — sperm plus LL-37, and sperm alone — using sperm retrieved from 26 males; human sperm donated by fertile men, tested in a dish only.

Only seen in animals

Source [11]

There is no safety data for any route except a gel on a leg ulcer, where it was well tolerated over 13 weeks: 49 and 45 adverse events on the two LL-37 strengths against 28 on the dummy gel, mostly mild, and 4 and 7 serious events against 1 on the dummy gel, with no death and nothing serious attributed to treatment. The FDA's published risk statement for cathelicidin LL-37 says it lacks sufficient safety information to know whether the drug would cause harm in people, names a risk of the body mounting an immune reaction against the peptide by some routes, flags unresolved questions about peptide impurities and how the ingredient is characterised, and states that non-clinical findings suggest detrimental effects on male reproduction and that it can be pro-tumour in some tissues. Independent work supports both flags: LL-37 kills human sperm in a dish and blocked fertilisation in mice, and blocking LL-37 slowed the growth of ovarian tumours in mice. It is also a molecule the body can have too much of — it is abnormally raised in rosacea skin, it converts a person's own DNA into an immune alarm signal in psoriasis, and roughly two-thirds of people with moderate-to-severe plaque psoriasis carry T cells that attack LL-37 itself. In the laboratory the gap between the concentration that kills bacteria (5 µM against E. coli) and the concentration that kills human cells (13–25 µM) is only three- to fivefold, and both effects are suppressed by ordinary blood serum. Nothing is known about pregnancy or breastfeeding in any species.

EffectWhat it feels likeHow often

An immune reaction against LL-37 itselfSerious

The FDA names immunogenicity — the body learning to attack an injected peptide — as a risk for compounded LL-37 by some routes, and says it has no information on what impurities these powders contain. On top of that general risk, LL-37 is one of the few peptides already known to be a target of the immune system in a common disease: in psoriasis it is an autoantigen, and the number of LL-37-attacking cells in the blood tracks how active the disease is. Nobody has tested what happens if such a person is given more of it.

Never measured in anyone given LL-37 as a treatment. But two-thirds of people with moderate to severe plaque psoriasis already carry immune cells that attack it.

Source [06]Source [13]

Serious events during the leg ulcer trial, none blamed on the treatmentSerious

They included three episodes of erysipelas — a spreading skin infection — in two patients, plus single cases of cellulitis, heart failure, a heart attack, two fractures, two kidney problems, a stroke and asthma. None were assessed as related to the treatment, which is what you would expect in a group whose average age was 67 and who were being treated for a circulation problem. The count is nonetheless lopsided: five people stopped treatment or left the study because of an adverse event, and all five were in the higher-dose LL-37 group.

12 serious events in 11 of the 148 patients (7.4 per cent), split 4 on the lower dose and 7 on the higher dose against 1 on the dummy gel. No deaths.

Source [03]

More reported side effects overall than on the dummy gel

Across the whole trial, 122 events were reported by 64 of the 148 patients. Most were mild; 22 were moderate and 8 were severe. The same number of people reported something on the lower dose as on the dummy gel — they just reported more things each. The authors' own conclusion was that the drug was well tolerated.

49 events in 20 patients on the lower dose and 45 events in 24 patients on the higher dose, against 28 events in 20 patients on the dummy gel.

Source [03]

Redness, swelling and scaling around the treated wound

Mostly mild or moderate. These figures are for everyone in the trial together, including the group getting the dummy gel and compression bandaging, so they describe what happens around a treated leg ulcer generally rather than an effect of LL-37 by itself.

Very common across the whole trial: redness of the wound in 42.6 per cent, swelling in 39.2 per cent, warmth in 20.3 per cent; on the skin next to the wound, redness in 54.1 per cent and scaling in 52.7 per cent.

Source [03]

Blood count, thyroid and skin changes in the melanoma study

The registry lists anaemia in both cohorts, and single reports of a drop in white cells and lymphocytes, raised bilirubin, an underactive thyroid, loss of skin colour, a new skin cancer and a sun-damage skin lesion. No serious events were posted. All three had advanced melanoma and prior treatments, so nothing here can be pinned on LL-37.

Reported in the three people who ever received LL-37 injected into a tumour. There was no comparison group.

Source [07]

Damage to your own cells at high enough concentrations

LL-37 works by breaking open membranes, and human cell membranes are not immune to it. In the laboratory it was toxic to several kinds of eukaryotic cells at 13 to 25 micromolar, against 5 micromolar to stop E. coli. Whether an injected or applied dose ever gets near that range in a person has never been measured, because there is no human pharmacokinetic study of LL-37 by any route.

Not measured in people. Seen in dishes at about three to five times the antibacterial amount.

Source [10]

These are the amounts the studies below gave their participants, and they are here as facts about those studies. They are not a recommendation, not a starting point and not a range to pick from. What a person should take, if anything, is a question for a doctor who knows them.

The phase IIb trial in hard-to-heal venous leg ulcers (HEAL LL-37, EudraCT 2018-000536-10), the trial that failed its main test. Two strengths of gel against a dummy gel, all on top of compression bandaging.

0.5 mg/mL or 1.6 mg/mL applied to the ulcer, twice a week

13 weeks of treatment, then 16 weeks of follow-up; 148 people treated

Source [03]

The first trial in people (Grönberg et al. 2014), a dose-finding study in venous leg ulcers with a three-week run-in on the dummy gel first.

0.5, 1.6 or 3.2 mg/mL applied to the ulcer, twice a week

4 weeks of treatment after a 3-week run-in, then 4 weeks of follow-up; 34 people

Source [04]

The melanoma study at MD Anderson Cancer Center (NCT02225366), injected directly into a skin or under-skin tumour at least 1 cm across.

250 micrograms per tumour for the first two people, 500 micrograms for the third

Once a week for up to 8 weeks; 4 people enrolled over five years, 3 treated

Source [07]

What people report

• Whether the one positive result — closure of wounds of 10 cm² or more — is real: it came from a subgroup picked out after the results were in, and the trial designed to test it has never been run.
• What an injection under the skin does, which is how it is most often sold and used: no published human study of any kind by that route.
• How much of it reaches the blood after any route, how long it lasts there and how fast it is broken down: there is no human pharmacokinetic study of LL-37 at all.
• Whether injecting it does anything useful for an infection: the only randomised infection trial used a swallowed engineered bacterium, not the peptide.
• Whether it helps any gut condition: no published human study, and the peptide would be digested like any other protein if swallowed on its own.
• Whether the body learns to make antibodies against it: never studied in anyone given it, despite the FDA naming immunogenicity as a risk.
• What happens if someone with psoriasis, rosacea or lupus takes it — conditions in which their own LL-37 is already part of the disease.
• Whether it affects a man's fertility: it kills human sperm in a dish and blocked fertilisation in mice, and nobody has looked at people.
• Whether it is safe in pregnancy or breastfeeding: no data in any species.
• Which way it would push an existing cancer: the FDA says animal findings show it can be pro-tumour in some tissues, and blocking it shrank ovarian tumours in mice — but it has also been reported to suppress tumours elsewhere, and it was tried as a melanoma treatment in three people with nothing readable coming out of it.
• What it does to the lungs when inhaled: work on airway inflammation is in rats only.
• How often a vial sold as LL-37 contains what the label says, in what purity, and free of bacterial toxins: never measured for this substance, and the FDA states it lacks the information to judge peptide-related impurities in compounded LL-37.
• Whether anything at all happens beyond 13 weeks: no human study has run longer.

Does LL-37 actually work?

For the one thing it has been properly tested on — long-standing open sores on the lower leg — no. In a trial of 148 adults, about a quarter of the ulcers closed on LL-37 gel and about a quarter closed on a dummy gel. A result that survived only after the researchers went back and looked at the biggest wounds on their own is a lead worth chasing, not a finding. For everything else it is sold for, including infections and gut problems, no published human study exists.

Source [03]Source [04]

Is LL-37 legal, and can you buy it?

It is not an approved medicine anywhere, so it cannot legally be sold to you as one. In the United States it was put forward as an ingredient that pharmacies could mix up for patients; the FDA placed it in the category of substances that may present significant safety risks, and the people who nominated it later withdrew the nomination. In Sweden and the rest of the EU, selling an unapproved medicine to consumers is not allowed, which is why online sellers label these vials "for research use only". That label describes the paperwork, not the contents.

Source [06]Source [15]

What are the side effects of LL-37?

The only real safety information comes from putting it on a leg ulcer, where it was well tolerated: more events were reported on LL-37 than on the dummy gel, but they were mostly mild, and nothing serious was blamed on the treatment. For every other way of taking it, side effects have never been measured. The FDA's published risk statement for LL-37 flags a possible immune reaction against the peptide, and adds that animal findings suggest harm to male reproduction and that it can help tumours grow in some tissues.

Source [03]Source [06]

LL-37 vs BPC-157 — which one has better evidence?

They fail in opposite directions. BPC-157 has never had a proper trial: a 2025 systematic review of its use in sports medicine found 36 studies, 35 of them in animals or cells, and the single human one was a look back at 12 patients given a knee injection, with no comparison group. LL-37 did get a proper trial — randomised, blinded, placebo-controlled, 148 people — and lost to the dummy gel on the question it was designed to answer. A compound that was tested and did not beat placebo tells you more than one that was never tested; what it tells you is discouraging.

Source [03]Source [19]

Can you inject LL-37?

People do, and there is no published study of it. Every trial that gave LL-37 itself to a person either put it on a wound or injected it into a tumour. Nobody has published an injection under the skin, into a vein or into a muscle, so there is no dose, no timing, no measurement of how much reaches the blood, and no safety data for that route. The FDA states plainly that it lacks enough safety information about LL-37 to know whether it would cause harm when given to people.

Source [06]Source [08]

Does LL-37 help with Lyme disease, biofilms or chronic infections?

No published human study has tested LL-37 for Lyme disease or any long-term infection. The one randomised infection trial gave people a swallowed bacterium engineered to make LL-37 in the gut, for COVID-19, and reported a faster negative swab only within the group treated early, in a trial where everyone knew who was getting what. LL-37 does kill bacteria in a dish, but ordinary blood serum suppresses that — which is exactly the gap between a laboratory result and a treatment.

Source [09]Source [10]

Is LL-37 banned in sport?

It is not named anywhere on the World Anti-Doping Agency's 2026 prohibited list. That is not the same as permitted. Section S0 of that list prohibits, at all times, any drug with no current approval from any health authority anywhere — which is what LL-37 is — and section S2.3 catches substances affecting the growth of blood vessels or the body's capacity to repair itself. An athlete should assume a positive test would be argued under one of those, not that the absence of the name is a clearance.

Source [14]

My body already makes LL-37, so isn't more of it safe?

That reasoning does not hold for this one. The diseases where LL-37 has been measured most carefully are diseases where there is too much of it, not too little: it is abnormally high in the facial skin of people with rosacea, and in psoriasis it binds a person's own DNA and turns it into an alarm signal that keeps the inflammation going. Two-thirds of people with moderate to severe psoriasis have immune cells that attack LL-37 directly. A molecule the body raises to fight infection is not automatically something to add more of.

Source [12]Source [13]Source [16]

Each letter represents one amino acid.

1. [01]UniProt P49913 (CAMP_HUMAN) — FASTA of the hCAP-18 precursor; the peptide feature at residues 134–170 gives LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES, 37 residues, no terminal modification
2. [02]PubChem: LL-37, CID 16198951 — C205H340N60O53, MW 4493, CAS 154947-66-7, INN ropocamptide; systematic name shows all-L residues with free ends
3. [03]Mahlapuu M et al. Evaluation of LL-37 in healing of hard-to-heal venous leg ulcers: a multicentric prospective randomized placebo-controlled clinical trial. Wound Repair Regen 2021;29(6):938-950 (148 treated; primary endpoint not met, 26.5% and 24.7% vs 25.3% placebo; ≥10 cm² subgroup analysis stated as post-hoc) (2021)
4. [04]Grönberg A, Mahlapuu M, Ståhle M, Whately-Smith C, Rollman O. Treatment with LL-37 is safe and effective in enhancing healing of hard-to-heal venous leg ulcers: a randomized, placebo-controlled clinical trial. Wound Repair Regen 2014;22(5):613-21 (n=34; no difference from placebo at 3.2 mg/mL) (2014)
5. [05]EU Clinical Trials Register, EudraCT 2018-000536-10 — Promore Pharma AB, protocol LL-37002, phase IIb double-blind randomised placebo-controlled trial in hard-to-heal venous leg ulcers, start 13 August 2018, results posted (2018)
6. [06]FDA, Certain Bulk Drug Substances for Use in Compounding that May Present Significant Safety Risks — cathelicidin LL-37 among substances nominated but withdrawn; immunogenicity, impurity and characterisation risks, detrimental effects on male reproduction, pro-tumorigenic in some tissues (page current as of 22 April 2026) (2026)
7. [07]ClinicalTrials.gov API v2: NCT02225366, Induction of Antitumor Response in Melanoma Patients Using the Antimicrobial Peptide LL37 — phase 1/2, 4 enrolled 2015–2020, 3 treated, single group, results posted (2020)
8. [08]ClinicalTrials.gov API v2, quoted intervention query for "LL-37": totalCount 20, retrieved 3 August 2026 (2026)
9. [09]Zhao Y et al. Efficacy and safety of Oral LL-37 against the Omicron BA.5.1.3 variant of SARS-CoV-2: a randomized trial. J Med Virol 2023;95(8):e29035 — 238 inpatients, open-label; the product was recombinant LL-37 Lactococcus lactis, not the peptide (2023)
10. [10]Johansson J et al. Conformation-dependent antibacterial activity of the naturally occurring human peptide LL-37. J Biol Chem 1998;273(6):3718-24 — MIC 5 µM against E. coli, cytotoxic to eukaryotic cells at 13–25 µM, both activities inhibited by human serum (1998)
11. [11]Srakaew N et al. Antimicrobial host defence peptide, LL-37, as a potential vaginal contraceptive. Hum Reprod 2014;29(4):683-96 (2014)
12. [12]Yamasaki K et al. Increased serine protease activity and cathelicidin promotes skin inflammation in rosacea. Nat Med 2007;13(8):975-80 (2007)
13. [13]Lande R et al. The antimicrobial peptide LL37 is a T-cell autoantigen in psoriasis. Nat Commun 2014;5:5621 (2014)
14. [14]WADA, World Anti-Doping Code International Standard — Prohibited List 2026 (in force 1 January 2026); neither LL-37 nor cathelicidin appears anywhere in the document, while S0 covers any substance with no current regulatory approval (2026)
15. [15]SVT investigation: peptides contain banned substances — buyers can be prosecuted (in Swedish) (2026)
16. [16]Lande R et al. Plasmacytoid dendritic cells sense self-DNA coupled with antimicrobial peptide, Nature 2007;449(7162):564-9 — "LL37 converts inert self-DNA into a potent trigger of interferon production" (2007)
17. [17]Coffelt SB et al. The pro-inflammatory peptide LL-37 promotes ovarian tumor progression through recruitment of multipotent mesenchymal stromal cells, PNAS 2009;106(10):3806-11 — neutralising LL-37 inhibited tumour growth in mice (2009)
18. [18]Kiattiburut W et al. Antimicrobial peptide LL-37 and its truncated forms exert spermicidal effects and microbicidal activity against Neisseria gonorrhoeae. Hum Reprod 2018;33(12):2175-2183 (2018)
19. [19]Vasireddi N et al. Emerging Use of BPC-157 in Orthopaedic Sports Medicine: A Systematic Review. HSS J 2025 — 36 studies included, 35 in animals or cells (2025)