Antimicrobial Peptide (AMP)

Antimicrobial peptides are typically twelve to fifty residues with a net positive charge from lysine and arginine, folding on contact with a membrane into a structure that segregates hydrophobic residues onto one face and cationic residues onto the other. Selectivity comes from membrane composition rather than from a receptor: bacterial surfaces present anionic lipopolysaccharide, lipoteichoic acid and phosphatidylglycerol, while the outer leaflet of a mammalian membrane is largely zwitterionic and rich in cholesterol.

Humans make the cathelicidin LL-37 alongside the defensins. Approved drugs from the class are old and narrow: polymyxin B, colistin, gramicidin, bacitracin and the lipopeptide daptomycin, approved in 2003. The modern record is poor. Pexiganan, a magainin analogue for infected diabetic foot ulcers, failed to beat its comparator and was not approved despite two attempts; omiganan failed late-stage trials, and murepavadin was stopped in phase 3 for nephrotoxicity.

A minimum inhibitory concentration is only interpretable with its conditions attached, because activity is frequently lost at physiological ionic strength, with divalent cations present, or once serum proteases arrive, so a value from dilute buffer can overstate real potency by an order of magnitude. The informative number is the ratio of killing concentration to the concentration that lyses red cells, and much published work reports only the first half of it.

Two claims recur and both fail. The first is that these peptides cannot select resistance because they attack the membrane; bacteria resist by remodelling surface charge, by efflux and by proteases, and colistin resistance travels on mobile plasmids. The second is that the label names a therapeutic category. It names a killing mechanism, and that mechanism's non-specificity is why systemic development keeps failing on toxicity.