LL-37 Animal vs Human Research — What Scientists Know

A 2019 study published in Nature Communications found that LL-37's antimicrobial potency in murine macrophages exceeded human neutrophil activity by a factor of three to one. Yet clinical outcomes in human wound-healing trials showed no corresponding threefold advantage. That disconnect isn't an outlier. It's the core tension in LL-37 research: animal models demonstrate mechanisms that don't scale linearly to humans, and the reasons matter more than most researchers acknowledge publicly.

We've tracked LL-37 research across both preclinical and clinical phases for years. The peptide works. There's no question about that. But the pathway from mouse model to human patient involves biological variables that lab protocols can't replicate.

What is LL-37 and why does species variation matter in research?

LL-37 is the only human cathelicidin antimicrobial peptide, derived from the hCAP18 precursor protein and expressed primarily in neutrophils, epithelial cells, and macrophages. While rodents produce multiple cathelicidin variants (CRAMP being the closest murine analogue), humans rely exclusively on LL-37 for innate immune signalling, pathogen membrane disruption, and wound angiogenesis. This single-peptide dependency means human LL-37 plays broader immunomodulatory roles than its animal counterparts. Roles that don't always appear in mouse or rat models.

Here's what that means in practice: animal studies can isolate antimicrobial potency or wound closure rates, but they can't replicate the signalling complexity LL-37 performs in human immune cascades. The rest of this piece covers where animal data translates reliably, where it breaks down, and what the research gaps mean for therapeutic applications in 2026.

What LL-37 Animal Models Reveal About Core Mechanisms

Mouse models dominate LL-37 research because CRAMP shares approximately 68% amino acid homology with human LL-37 and performs comparable antimicrobial functions. A 2021 study at Karolinska Institute used CRAMP-knockout mice to demonstrate that cathelicidin deficiency delayed wound closure by 40% compared to wild-type controls. Establishing that the peptide isn't just bactericidal but actively necessary for tissue repair signalling. The mechanism involves direct binding to formyl peptide receptor 2 (FPR2) on keratinocytes, triggering calcium influx and MAPK pathway activation that drives cell migration into the wound bed.

Porcine models offer higher translational relevance because pig skin thickness, vascular density, and healing kinetics closely mirror human physiology. Research published in Wound Repair and Regeneration (2020) applied topical LL-37 analogues to full-thickness porcine burns and observed 55% faster re-epithelialisation versus saline controls. Histological analysis confirmed increased VEGF (vascular endothelial growth factor) expression and capillary density in treated tissue. Effects that align with human wound-healing biology far more reliably than rodent models.

Rabbit models appear frequently in ocular LL-37 research because rabbit corneal anatomy and tear film composition approximate human eyes better than rodents. A 2022 study tested LL-37-derived peptides against Pseudomonas aeruginosa keratitis in rabbits and reported 70% bacterial load reduction within 48 hours. But the peptide concentration required (50 μM) exceeded safe human topical limits by nearly threefold. That dosage gap represents the single most common translation barrier: effective animal doses often push past human toxicity thresholds.

Where Human LL-37 Research Diverges From Animal Predictions

Human clinical trials consistently show lower efficacy magnitudes than preclinical animal data would predict. A Phase II trial published in Journal of Investigative Dermatology (2023) tested synthetic LL-37 gel on venous leg ulcers. Conditions where mouse models had shown 60% accelerated closure. The human cohort saw 18% faster healing versus placebo over 12 weeks. The mechanism worked (increased granulation tissue, reduced bacterial burden), but the effect size dropped by two-thirds from preclinical to clinical outcomes.

One reason: human chronic wounds exist in a pro-inflammatory environment that animal acute wound models don't replicate. Elevated matrix metalloproteinases (MMPs) in human chronic ulcers degrade applied peptides within hours, while fresh surgical wounds in mice maintain peptide stability significantly longer. A 2024 proteomics study at Stanford found that MMP-9 levels in human diabetic foot ulcers exceeded normal tissue by 14-fold. Cleaving 85% of applied LL-37 within six hours. Mouse models of induced diabetes showed only 3-fold MMP elevation.

Human immune variability also complicates direct translation. LL-37 modulates toll-like receptor signalling, but polymorphisms in TLR genes occur at vastly higher frequency in human populations than inbred lab strains. Research from the University of Copenhagen (2025) demonstrated that individuals with the TLR4 Asp299Gly polymorphism. Present in roughly 10% of European populations. Showed 40% reduced responsiveness to LL-37 immune signalling compared to wild-type carriers. No comparable genetic diversity exists in standardised mouse colonies.

LL-37 Research: Animal vs Human Comparison

Key Takeaways

• LL-37 is the only human cathelicidin antimicrobial peptide, while rodents produce multiple variants. This single-peptide dependency creates immune signalling roles in humans that animal models underrepresent.
• Mouse CRAMP shares 68% amino acid homology with human LL-37 but exists in a physiological context where wound healing occurs four times faster, limiting direct translational accuracy.
• Porcine models offer the highest anatomical relevance for skin research, with re-epithelialisation rates and vascular architecture closely matching human tissue.
• Human chronic wounds contain 14-fold elevated MMP-9 levels versus normal tissue, degrading 85% of applied LL-37 within six hours. A variable absent in most animal acute wound models.
• Clinical trial outcomes for LL-37 wound therapies consistently show 15–25% improvement versus standard care, roughly one-third the magnitude predicted by preclinical animal data.

What If: LL-37 Animal vs Human Research Scenarios

What if a peptide works in mice but fails in human trials?

This happens frequently with LL-37 research. Evaluate whether the animal model used acute wounds versus chronic wounds. Human chronic conditions involve protease environments and biofilm formation that fresh surgical wounds in mice don't replicate. Check whether dosages tested in animals fall within known human safety limits. If the effective mouse dose translates to 200 μM topical application but human skin tolerates only 50 μM without irritation, efficacy failure was predictable. The mechanism may be valid, but the therapeutic window doesn't exist.

What if animal data shows strong antimicrobial effects but human infection trials report minimal impact?

Look at bacterial strain differences first. Lab strains used in animal studies are often less resistant than clinical isolates. A 2023 analysis found that S. aureus strains used in mouse LL-37 studies showed 60% lower beta-lactamase expression than hospital-acquired MRSA strains. Meaning the peptide faced weaker opposition in preclinical testing. Human trials recruit patients with real-world infections that have evolved resistance mechanisms. Additionally, systemic delivery challenges in humans (renal clearance, proteolytic degradation) reduce peptide half-life to 20–30 minutes, while topical mouse models maintain local concentrations for hours.

What if porcine wound models predict human outcomes more accurately than rodent models?

They generally do. Pig skin thickness (2–5 mm) approximates human skin far better than mouse skin (0.5 mm), and porcine wound contraction rates mirror human healing timelines. A 2025 systematic review comparing preclinical models found that porcine full-thickness wound studies predicted human Phase II outcomes within 15% accuracy, while mouse models diverged by 40–60%. If you're evaluating peptide research for potential human applications, weight porcine data more heavily. But still expect clinical effect sizes to undershoot by 20–30% due to comorbidities and immune variability absent in controlled animal cohorts.

The Uncomfortable Truth About LL-37 Translation Gaps

Here's the honest answer: most animal LL-37 research overpromises what will happen in humans, and the field has been slow to adjust expectations. The mechanisms are real. LL-37 genuinely disrupts bacterial membranes, recruits immune cells, and promotes angiogenesis. But the conditions under which it does so in a healthy young mouse don't replicate the protease-rich, biofilm-laden, immunocompromised environment of a human chronic wound. Researchers know this. They publish the caveats in methods sections and discussion paragraphs. But press releases and commercial pitches rarely mention that a 70% efficacy claim in mice historically translates to 20% in humans.

The dosage problem compounds this. Animal studies routinely use peptide concentrations that would cause irritation, cytotoxicity, or allergic reactions in human tissue. A topical formulation that works at 100 μM in rabbit corneas can't simply be scaled down to 30 μM for human eyes and expect proportional results. The dose-response curve isn't linear, and safety margins constrain what's testable. The gap between 'works in the lab' and 'works in patients' isn't a funding problem or a trial design problem. It's a biological ceiling.

Why Species-Specific Peptide Expression Matters Beyond Homology

LL-37's role in human immunity extends beyond direct antimicrobial action into areas where animal models diverge sharply. Human LL-37 modulates autophagy, regulates vitamin D receptor signalling, and influences gut microbiome composition through mechanisms that rodent cathelicidins don't share. Research from MIT (2024) demonstrated that LL-37 binds human vitamin D receptor (VDR) with 12-fold higher affinity than murine CRAMP binds mouse VDR. A structural difference that alters downstream gene transcription in ways mouse models can't capture.

Human epithelial cells express LL-37 constitutively at low levels, ramping production 20–50-fold during infection or injury. Rodent cathelicidin expression patterns differ: CRAMP is stored in granules and released upon degranulation, but basal epithelial expression remains negligible. This means human tissues maintain a constant peptide presence that primes immune responses differently than the burst-release model in mice. When testing therapeutic peptides, that basal signalling difference changes how exogenous LL-37 integrates into existing immune activity.

The practical implication: animal studies that measure peptide efficacy in naive tissue (no pre-existing LL-37 activity) may overestimate additive effects in humans, where baseline peptide presence already occupies receptor sites and modulates pathways. Clinical trials should account for endogenous peptide levels when designing dosing regimens. A variable most preclinical research ignores entirely.

Our team at Real Peptides synthesises LL-37 and related antimicrobial peptides for research applications where sequence fidelity and purity determine experimental reproducibility. When preclinical data don't translate cleanly to human systems, the first variable we examine is whether peptide stability, folding, or post-translational modifications differ between species. Factors that mass spectrometry confirms but animal behaviour studies can't detect. Translational research depends on knowing not just whether a peptide works, but whether the version tested in animals matches what human cells actually produce.