Antimicrobial Peptides: A Guide to Cathelicidins, Defensins, Magainin, and Clinical AMPs

Antimicrobial resistance has been described by the WHO as one of the top ten global public health threats. The pipeline of conventional small-molecule antibiotics has been thin for two decades — only a handful of truly novel antibiotic classes have been approved since the 1980s, and most recent approvals have been modifications of existing scaffolds. Meanwhile, carbapenem-resistant Enterobacterales, multi-drug-resistant Pseudomonas aeruginosa, and extensively drug-resistant Acinetobacter baumannii continue to spread in hospital settings, and mcr-1-mediated plasmid-borne polymyxin resistance has been documented on six continents.

Against this backdrop, AMPs offer three distinct advantages. First, their primary mechanism — membrane disruption — is harder to evolve resistance to than target-specific mechanisms like ribosomal or cell-wall inhibition, though resistance is not impossible and has been documented (including through lipid A modifications that reduce cationic peptide binding). Second, many AMPs have immunomodulatory activity that can complement direct antimicrobial killing. Third, AMP-design space is enormous — synthetic modifications (D-amino acid substitution, cyclization, lipidation) can tune selectivity and stability in ways that are well-suited to modern peptide chemistry.

The counterweights are also real: oral bioavailability is poor, serum stability is limited, manufacturing costs are higher than small molecules, and the therapeutic window (antimicrobial efficacy vs host cytotoxicity) can be narrow. The clinical AMP field is littered with failed late-stage trials — pexiganan for diabetic foot ulcers, iseganan for ventilator-associated pneumonia, omiganan for catheter-site infections — often because topical efficacy was insufficient rather than because the peptide was unsafe. The 2020s have seen renewed investment alongside more realistic expectations.

Most AMPs share a common starting point: they are cationic (net positive charge) and amphipathic (containing distinct hydrophobic and hydrophilic faces). Bacterial membranes are rich in negatively-charged phospholipids such as phosphatidylglycerol and cardiolipin on their outer leaflet, whereas mammalian membranes are dominated by zwitterionic phosphatidylcholine on the outer leaflet. This asymmetric charge distribution is the basis for AMP selectivity — electrostatic attraction draws cationic AMPs toward bacterial surfaces preferentially over host surfaces.

After initial binding, AMPs insert into the membrane via their hydrophobic face and disrupt bilayer integrity through one of several proposed mechanisms: the barrel-stave model (peptides aggregate into transmembrane pores with hydrophilic interiors); the toroidal-pore model (peptides induce curvature and co-mingle with lipid headgroups to form mixed-lipid pores); and the carpet model (peptides accumulate on the membrane surface until they reach a critical density that disrupts bilayer structure micelle-like). These models are not mutually exclusive — many AMPs appear to act through combinations or context-dependent mechanisms.

AMPs can also have intracellular targets once they traverse the membrane: inhibiting protein synthesis, DNA replication, or cell-wall biogenesis. And many AMPs — particularly mammalian cathelicidins and defensins — have immunomodulatory effects beyond direct killing, including chemokine induction, macrophage activation, and modulation of adaptive immunity. This multifunctionality is part of why "host defense peptide" is sometimes preferred over the narrower term "antimicrobial peptide."

Cathelicidins are an AMP family defined by a conserved N-terminal cathelin-like domain that is proteolytically cleaved to release the active antimicrobial peptide from a larger precursor protein. Humans have a single cathelicidin gene, CAMP, which encodes the precursor hCAP-18. Proteolytic cleavage by proteinase-3 in neutrophils and other proteases at epithelial surfaces releases the mature active peptide: LL-37, a 37-amino-acid linear peptide beginning with two leucine residues.

LL-37 is expressed by neutrophils, monocytes, macrophages, NK cells, mast cells, epithelial cells of the skin, gut, airway, and reproductive tract, and keratinocytes during wound healing. Its concentrations rise sharply during infection and inflammation. Beyond direct bactericidal activity against a broad range of Gram-positive and Gram-negative organisms, LL-37 has well-documented immunomodulatory activity — chemoattracting neutrophils, monocytes, and T cells; promoting wound healing through keratinocyte migration and angiogenesis; modulating TLR signaling; and inducing autophagy.

LL-37 is studied in contexts well beyond classical infection. Downregulation is linked to atopic dermatitis, and dysregulation is implicated in psoriasis and rosacea. Low LL-37 levels have been associated with increased susceptibility to respiratory infection and, in some studies, worse outcomes in tuberculosis and COVID-19. Research-grade LL-37 is widely available, and clinical candidates have included topical formulations for chronic wounds. As of 2026, no LL-37-based therapeutic has received FDA approval, but several analogs and fragment derivatives are in development. LL-37 is not an appropriate self-administered peptide — its immunomodulatory effects are dose- and context-dependent, and systemic administration has not been adequately characterized in humans.

Defensins are a second major family of mammalian AMPs, distinguished by a conserved disulfide-bridged beta-sheet structure. Three defensin subfamilies exist: alpha-defensins, beta-defensins, and theta-defensins. Humans express alpha- and beta-defensins; theta-defensins are found in non-human primates only (rhesus macaques, for example) because the human theta-defensin gene has been pseudogenized by a premature stop codon.

Human alpha-defensins include human neutrophil peptides 1-4 (HNP1-4), stored in neutrophil primary granules and released into the phagosome upon pathogen engulfment, and HD5 and HD6, produced by Paneth cells in the small intestinal crypts where they shape the gut microbiome and provide a barrier to enteric pathogens. Human beta-defensins (HBD1, HBD2, HBD3, HBD4) are expressed by epithelial cells of skin, airway, gut, and urogenital tract. HBD1 is constitutively expressed; HBD2 and HBD3 are induced by pro-inflammatory signals and infection.

Defensins have broad antimicrobial activity against bacteria, fungi, and some enveloped viruses, and they also modulate immune responses — chemoattracting dendritic cells, activating macrophages, and bridging innate and adaptive immunity. Several defensin-related therapeutic candidates have been explored, and defensin expression has been pursued as a biomarker in inflammatory bowel disease (Paneth-cell defensin deficiency is a feature of Crohn's disease). Like cathelicidins, defensins are predominantly research peptides with no FDA-approved direct therapeutic analog as of 2026.

Magainin is one of the archetypal AMPs of research history. It was identified in 1987 by Michael Zasloff while he was working at the NIH, from the skin of the African clawed frog Xenopus laevis. Zasloff had noticed that frogs used in laboratory experiments recovered from non-sterile surgical incisions without developing infection, and tracked the activity to a previously uncharacterized peptide in frog skin mucus. Magainin-1 and magainin-2 are 23-amino-acid linear cationic amphipathic peptides with broad-spectrum antimicrobial activity.

The discovery catalyzed decades of AMP research and inspired one of the most ambitious attempts to translate an AMP into a human drug: pexiganan (MSI-78), a synthetic 22-amino-acid magainin analog with improved potency. Pexiganan reached Phase 3 trials as a topical cream for infected diabetic foot ulcers in the 1990s and 2000s, with MacroChem and Access Pharmaceuticals developing the program. The Phase 3 trials ultimately failed to demonstrate superiority over standard of care (oral ofloxacin), and FDA approval was not granted. The clinical lesson was subtle — pexiganan appeared non-inferior to oral antibiotic, but regulatory requirements at the time demanded superiority, and the economic case for a new topical in an established indication was weak.

Pexiganan's story is often cited as the cautionary tale of AMP translation: a molecule that worked but could not overcome competition from cheap, effective, orally bioavailable small molecules. More recent magainin-derived candidates have been explored but the program has not been revived commercially. Magainin itself remains a reference research AMP widely used in membrane-biophysics studies and as a positive control in AMP screens.

Cecropin is to insect immunity what cathelicidins are to mammalian innate defense. First isolated from the pupae of the giant silk moth (Hyalophora cecropia) by Hans Boman's group in Sweden in 1980, cecropins are linear amphipathic alpha-helical peptides of approximately 35–37 amino acids with broad antimicrobial activity against Gram-negative and Gram-positive bacteria. They are found across insects, including Drosophila melanogaster, and also in limited form in some non-insect invertebrates.

Cecropins are notable in research for two reasons. First, they helped establish the general importance of peptide-mediated innate immunity at a time when the adaptive immune system dominated immunological thinking. Boman's work on Drosophila immunity contributed to the broader recognition — eventually underlying the 2011 Nobel Prize in Physiology or Medicine to Jules Hoffmann, Bruce Beutler, and Ralph Steinman — that innate immune pathways are phylogenetically ancient and conserved from insects to humans. Second, cecropins have been used extensively as a design template for synthetic hybrid peptides: cecropin-melittin hybrids, in particular, have been engineered for improved selectivity and potency.

As a therapeutic, cecropin itself has not achieved clinical translation — its stability and selectivity profile is not ideal for systemic human use. Cecropin-based and cecropin-inspired peptides continue to appear in preclinical AMP development pipelines and in antimicrobial material engineering (surface coatings, wound dressings). The peptide remains a staple of academic research.

Despite the difficulty of translating AMPs into human medicine, two AMP-class drugs are in routine clinical use and have been for decades.

Polymyxin-B (and its close relative colistin, also called polymyxin-E) is a cyclic lipopeptide produced by Bacillus polymyxa. It is one of the last-line antibiotics for multi-drug-resistant Gram-negative infections, including carbapenem-resistant Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter baumannii. Polymyxin-B binds the lipid A moiety of lipopolysaccharide on the outer membrane of Gram-negative bacteria, displacing divalent cations and disrupting membrane integrity. Clinical use is constrained by significant nephrotoxicity and neurotoxicity, which led to its near-abandonment in the 1970s when safer alternatives were available; the rise of carbapenem resistance in the 2000s has forced a return to polymyxin use despite the toxicity profile. Emergence of mcr-1-mediated plasmid-borne polymyxin resistance is a serious concern.

Gramicidin is a linear pentadecapeptide (gramicidin D is a mixture of gramicidin A, B, and C) produced by Bacillus brevis. It was isolated by René Dubos in 1939 and was one of the first antibiotics used clinically. It forms ion-channel pores in bacterial membranes but is too toxic for systemic use and is today used exclusively in topical ophthalmic and otic preparations, often in combination with other antibiotics (Polysporin, Neosporin Ophthalmic). Gramicidin-S, a cyclic peptide produced by Bacillus brevis var. G.-B., is used in some countries as a topical antiseptic but not in the United States.

These two drugs demonstrate that AMPs can be clinical antibiotics — but only within narrow niches (topical application or last-line systemic use for otherwise untreatable infections) where their toxicity profile is acceptable.

Antiviral peptides are a related but mechanistically distinct class from bacterial AMPs. Rather than disrupting microbial membranes broadly, the most successful clinical antiviral peptides target specific viral entry or fusion events.

Enfuvirtide (Fuzeon, T-20) is the paradigm case. It is a 36-amino-acid synthetic peptide derived from the heptad repeat 2 (HR2) region of the HIV-1 gp41 envelope glycoprotein. It binds the HR1 region of gp41 during the six-helix-bundle formation that drives viral-to-host membrane fusion, preventing the conformational change required for HIV to enter the target CD4+ T cell. FDA-approved in 2003, enfuvirtide was the first entry inhibitor approved for HIV and is one of the largest peptide drugs ever manufactured — its 36-amino-acid length required novel large-scale peptide synthesis.

Enfuvirtide's clinical role has narrowed substantially since approval. It is administered twice-daily by subcutaneous injection, which causes near-universal injection-site reactions and is inconvenient compared with oral antiretrovirals. The newer integrase strand-transfer inhibitors (dolutegravir, bictegravir) and long-acting injectable antiretrovirals have displaced enfuvirtide from most treatment algorithms. It retains a role in heavily treatment-experienced patients with multi-class-resistant HIV, but is no longer a front-line agent. Enfuvirtide remains important as a proof of concept: a rationally designed fusion-inhibitor peptide derived from the viral sequence itself can be a durable antiviral.

The persistent gap between AMP research and AMP therapeutics reflects a cluster of well-characterized development hurdles. Oral bioavailability is essentially zero for most AMPs — they are rapidly degraded by gastric proteases and the acidic pH of the stomach, and the amphipathic architecture that makes them antimicrobial also makes them poor candidates for enteric absorption. This pushes AMP development toward parenteral, topical, inhaled, or locally delivered formulations.

Serum stability is also limited. Plasma proteases degrade linear peptides on timescales of minutes to hours, requiring structural modifications — cyclization, D-amino acid substitution, backbone N-methylation, PEGylation, or lipidation — to achieve pharmacokinetically useful half-lives. Selectivity is the other core challenge: the membrane-disruptive activity that kills bacteria can also lyse red blood cells and damage mammalian membranes at therapeutic doses. Optimizing the therapeutic index requires careful tuning of cationicity, hydrophobicity, and secondary structure.

Manufacturing is a further hurdle. AMPs are typically 20–50 amino acids long, which is in the cost-intensive range for solid-phase peptide synthesis at scale. Biomanufacturing (recombinant expression) is possible but introduces its own complexity with cationic toxic-to-host peptides. These manufacturing costs limit the price competitiveness of AMPs against small-molecule antibiotics in most indications — which is why current AMP clinical development is concentrated in niches where those constraints matter less (topical, inhaled, last-line infection) rather than in broad-spectrum oral antibiotic markets.