B. Agerberth, Department of Medical Biochemistry and Biophysics (MBB), Karolinska Institutet, SE 171 77 Stockholm, Sweden Fax: +46 8 337 462 Tel: +46 8 524 87781 E-mail: email@example.com
Antimicrobial peptides are present in all walks of life, from plants to animals, and they are considered to be endogenous antibiotics. In general, antimicrobial peptides are determinants of the composition of the microbiota and they function to fend off microbes and prevent infections. Antimicrobial peptides eliminate micro-organisms through disruption of their cell membranes. Their importance in human immunity, and in health as well as disease, has only recently been appreciated. The present review provides an introduction to the field of antimicrobial peptides in general and discusses two of the major classes of mammalian antimicrobial peptides: the defensins and the cathelicidins. The review focuses on their structures, their main modes of action and their regulation.
The first indications of the presence of antimicrobial peptides (AMPs) were in bacteria and fungi , where they were regarded as unique defense molecules in unicellular organisms. Some of these peptides were shown to be synthesized independently of ribosomal translation, instead being derived from enzymatic synthesis, often forming cyclic peptides, and also containing unusual amino acids. However, in 1962, the field broadened to metazoans with the isolation and characterization of the hemolytic bombinin peptide from the toad Bombina variegata . The focus of the initial study in Germany was on hemolytic properties. Despite a follow-up study in 1969 describing its antimicrobial properties , this discovery did not reach the research community in a broader perspective. Hence, the field lay virtually dormant for almost 20 years, during which research focused on adaptive immunity and oxygen-dependent killing. The first isolated and characterized antimicrobial peptides were the cecropins, from hemolymph of the moth Hyalophora cecropia, in 1980 . This discovery also answered the longstanding question of how insects, dispossessed of adaptive immune responses, defend themselves in a world abundant of microbes.
When, in the early 1980s, the α-defensins were isolated and characterized from mammalian leukocytes, including human cells, it became clear that this line of defense could no longer exclusively be associated with organisms that lack an adaptive immune system [5,6]. The third important discovery was the identification of the magainins in the skin of Xenopuslaevis , indicating the widespread presence of AMPs among metazoans. These discoveries initiated research on AMPs as being important defense molecules in many other species. Examples of AMPs that have been discovered are the human cathelicidin LL-37 [8–10], drosomycin from the fruit fly Drosophila melanogaster , the thionins from the plant Arabidopsis thaliana  and the fungal peptide plectasin . These examples demonstrate that AMPs are not restricted to few species, and span almost all eukaryotic life-forms, illustrating their importance in immunity.
General characteristics of AMPs
Although the number of AMPs in nature is large (the antimicrobial peptide database lists ∼ 1700 unique peptides; http://aps.unmc.edu/AP/) and great sequence diversity exists, there are general structural features that are shared by a majority of AMPs . They are commonly around 30 residues in length and are generally cationic in nature (+2 to +9). They carry an average of 40–50% hydrophobic residues arranged so that the folded peptide adopts an amphipathic structure. These properties are important for their microbial killing mechanism: the cationic character of AMPs results in electrostatic attraction to the negatively-charged phospholipids of microbial membranes and their hydrophobicity aids integration into the microbial cell membrane, leading to membrane disruption. Furthermore, the amphipathic structure also allows the peptides to be soluble both in aqueous environments and lipid membranes . There are several different secondary structures of the AMPs, although two forms are predominant: β-sheet structures, often in peptides rich in disulfide bonds (i.e. the mammalian defensins and porcine cathelicidin protegrins) or α-helical structures adopted by linear peptides such as human cathelicidin LL-37 and the frog magainins .
AMPs in mammals
In mammals, there is a plethora of AMPs. However, the two families that have been most thoroughly characterized are the defensins (Fig. 1A,B) and the cathelicidins (Fig. 1C).
The defensins are an evolutionarily ancient class of AMPs present in animals, plants and fungi [16,17]. They are able to kill or eliminate both Gram-positive and Gram-negative bacteria, in addition to fungi, protozoans and viruses. Similar to most AMPs, the defensins have mainly been claimed to interact and disrupt the lipid membranes of microbes by multimerizing and forming pores, with subsequent lysis of the microbes [18,19]. Interestingly, β-defensin peptides were recently shown to interact with Lipid II and interrupt cell wall synthesis in bacteria [20,21]. In addition to acting as antimicrobials, defensins have been shown to act as modulators in the immune system by inducing the production of pro-inflammatory cytokines, act as chemokines for neutrophils and enhance phagocytosis of macrophages [22,23].
The defensin peptides carry six cysteines, forming three intramolecular cystine bonds. The bonds stabilize a predominant β-sheet structure and are essential for the correct proteolytic processing of their pro-sequences [6,24]. However, a recent study indicates that reduction of the cystine bonds of human β-defensin (HBD)-1 results in improved antimicrobial activity of the peptide against a number of microbes, suggesting a further role of the disulfide bonds as regulators of antimicrobial activity .
The defensins are further subdivided into α-, β- or θ-defensins depending on their molecular weight and how their cysteines and disulfide bonds are distributed and arranged in the peptide (α-defensin: C-X1-C-X3–4-C-X9-C-X6–10-C-C; β-defensin: C-X4–8-C-X3–5-C-X9–13-C-X4–7-C-C) and θ-defensins share the common feature that their N- and C-termini are post-translationally joined, yielding a cyclic peptide [6,26,27]. The defensins are widespread in nature, although the α-defensins are exclusively found in vertebrates, including primates, rodents, basal mammals and marsupials, and θ-defensins are only detected in primates [28–33]. α- and β-defensins are synthesized as precursors that are proteolytically cleaved into their antimicrobially active forms [29,34]. However, the θ-defensins precurors undergo a more complex processing, resulting in cyclic peptides through ligation of two truncated α-defensin-like peptides .
In humans, the α-defensins are encoded from one locus on chromosome 8p23.1 . The six human α-defensins are transcribed from three exons with the sequence encoding the active, mature AMP on the third exon. The α-defensins, human neutrophil peptide (HNP)1 to HNP3, are found in high concentrations in the primary granules of neutrophils, as is HNP4 (Table 1), although HNP4 is present in substantially lower concentrations than HNP1–3 . The HNPs are released by degranulation as a response to pro-inflammatory or bacterial stimuli such as the formylated tripeptide formyl-methionyl-leucyl-phenylalanine . Human defensin (HD) 5 and HD6 (Table 1) are present in Paneth cells in the crypts of the small intestine [38,39]. These peptides have been shown to have diminished expression in Crohn’s patients with mutations in the intracellular receptor NOD2 or reduced expression of Wnt-signaling pathway transcription factor Tcf-4, whereas, in coeliac sprue, the expression is enhanced [40–42]. In addition to their antimicrobial function, human α-defensins have also been shown to: degranulate rat peritoneal mast cells, act as chemokines for monocytes (HNP1 and HNP2), inhibit glucocorticoid production (HNP4) and be mitogenic for epithelial cells [22,43–46].
Table 1. Human antimicrobial peptides of the defensin and cathelicidin families.
In several epithelial cells: respiratory tract, gastrointestinal tract, cornea, skin, kidney, etc.
Neutrophils, epithelia of the skin, gut and lungs, monocytes, natural killer cells and mast cells
The HBDs are encoded from several gene clusters where up to 30 different genes for HBDs are present . In addition, one β-defensin cluster encompasses the α-defensin cluster . The β-defensin genes carry two exons, of which the 3′ exon encodes the antimicrobially-active β-defensin peptide. The human β-defensin peptides so far characterized are expressed in a number of epithelial cells (Table 1) and are generally constitutively expressed, such as HBD-1. However, HBD-2 and HBD-3 are induced by wounding, bacterial products (lipopolysaccharide and lipoteichoic acid) or pro-inflammatory cytokines (tumor necrosis factor-α or interleukin-1α) [48–52].
The cathelicidin family of AMPs is not as widespread in nature as the defensin peptides, indicating a more recent ancestry . The cathelicidin genes are transcribed as preproproteins from four exons, of which the first three encode a signal peptide and the cathelin-like N-terminal domain and the fourth exon encodes the C-terminal antimicrobial domain, including the processing site. Genes encoding cathelicidins have been discovered in lizards, birds, fish and a number of mammals. For many species, there are several homologues encoding cathelicidins, such as in porcine, equine, bovine and ovine species, whereas there is only one copy of the cathelicidin gene present in murine species (in rat: rCRAMP and in mouse: mCRAMP), lapines (in rabbit: CAP18) and in primates (in human: CAMP) (Fig. 1C). The sequence of the N-terminal domain (i.e. the cathelin propart) is highly conserved and shares sequence homology with the porcine cysteine protease inhibitor cathelin and, similar to cathelin, has been show to inhibit cathepsin L . Hence, the name cathelicidin originates from cathelin connected with a microbicidal peptide. In addition to its protease inhibitor activity, cathelin is also an anionic antimicrobial polypeptide active against Escherichia coli and methicillin-resistant Staphylococcus aureus . The sequences of the C-terminal antimicrobial domains are highly divergent, encoding AMPs of varying sizes (12–79 amino acid residues) and folds, with the most common conformation being a linear peptide that adopts an amphipatic α-helical structure in contact with lipid membranes [55,56].
The human cathelicidin LL-37
In humans, the sole cathelicidin is encoded from the 2 kb gene CAMP (chromosome: 3p21.3) that encodes the 100 residue pro-protein hCAP18 . It is found in high concentrations in the specific granules of polymorphonuclear neutrophils, although it is also expressed in epithelial cells of the skin and mucosa of the intestinal, respiratory, urinary and genital tracts. Furthermore, its expression has also been detected in natural killer cells, monocytes, B-cells and mast cells (Table 1) [48,57,58]. The C-terminal antimicrobial domain is named LL-37 after its double N-terminal leucine and 37 amino acid length. The peptide is cationic and adopts an amphipathic α-helical structure in contact with lipid micelles or in the presence of mono- or divalent anions [59,60]. A recent study on the mechanism of action indicates that LL-37 reaches the bacterial periplasmic space of Gram-negative bacteria, halting bacterial growth .
In neutrophils, the mature antimicrobial peptide LL-37 is extracellularly released from the N-terminal domain by proteinase-3, a serine protease stored in the azurophilic granules of neutrophils . However, there is evidence of hCAP-18 being cleaved differentially by serine proteases such as kallikrein-7, generating multiple forms in the skin. In seminal fluid, the protease gastricsin releases a 38-residue antimicrobial peptide (i.e. ALL-38) from the cathelin domain [63–65].
The antimicrobial activity of LL-37 is broad, similar to that of the defensins. It is antimicrobially active not only against a wide variety of Gram-negative and Gram-positive bacteria, but also against fungi and viruses . Functional roles for cathelicidins have been confirmed in gene-deficient mice. The cathelicidin knockout mice are much more susceptible to wound infections by group A Streptococcus than wild-type mice . Interestingly, vaccinia virus infections of the cathelicidin-deficient mice resulted in more severe lesions compared to wild-type mice . This confirms a function of cathelicidin in bacterial infections and indicates a role for cathelicidin in virus immunity.
Immunomodulatory properties and additional activities
In addition to its antimicrobial activities, LL-37 has been found to be influential in the re-epithelialization after wounding, to function as an angiogenic factor and to be a mitogen for endothelial cells [69,70]. LL-37 also acts a chemoattractant for neutrophils, monocytes and T-cells by interacting with the receptor formyl peptide receptor-like 1 [57,71]. In neutrophils, LL-37 and the pro-inflammatory lipid mediator leukotriene B4 can be engaged in a positive feedback loop that leads to an amplification of the inflammatory process . In addition, LL-37 has been proposed to play a key role in activating inflammation in psoriasis because it can complex with extracellular self-DNA and trigger Toll-like receptor 9-dependent interferon responses in plasmacytoid dendritic cells . Thus, LL-37 is involved in autoimmunity . A similar mechanism has recently been reported to take place in systemic lupus erythematosis, indicating a general contribution of mediators in innate immunity to autoimmunity . It has also been established that LL-37 is a potent lipopolysaccharide binding factor and hence can neutralize endotoxin .
Regulation of the CAMP gene
In the normal state, the CAMP gene is expressed constitutively, although it can also be induced. There are several studies showing that the expression can be increased in certain disease states such as psoriasis, cystic fibrosis and lupus erythematosis [77–79]. By contrast, down-regulation of CAMP gene expression has been described in atopic dermatitis, chronic skin ulcers and enteric infections of Shigella spp., enterotoxigenic Escherichia coli and Vibrio cholerae [69,80–82]. This down-regulation has been postulated to be a way for the pathogen to evade host defenses and has led to efforts to find compounds that restore this attenuated expression. Indeed, several compounds that can induce CAMP gene expression in vitro have been identified, such as the bacterial products butyrate and lithocolic acid [83–85]. Recently, in a rabbit model of shigellosis, we showed that the down-regulation of the CAMP orthologue CAP18 could be counteracted by supplementing the rabbits with sodium butyrate . Furthermore, phenylbutyrate, a derivative of butyrate, has been demonstrated to exhibit similar or even stronger inducing capacity than butyrate both in vitro and in vivo [87,88]. Factors such as insulin-like growth factor I and vitamin D have also been shown to induce the CAMP gene [89–92]. A connection between vitamin D and the induction of antimicrobial peptides has been highlighted as part of the defense against tuberculosis . This connection could, however, be a more general mechanism (i.e. strengthening the immune defenses against various infectious microbes).
Recently, a clinical trial validated vitamin D utilization as an adjunctive therapy against tuberculosis. A significant effect was observed in a subgroup of patients with a specific allele encoding the vitamin D receptor . Several additional trials are ongoing using vitamin D, and the outcome will be of general interest. The treatment of other infections based on a similar principle (i.e. the induction of AMPs using phenylbutyrate) has been successfully utilized by our group in a rabbit model of shigellosis  and the effect of phenylbutyrate on other enteric infections is under evaluation. Clearly, the field is open for additional models for the use of AMPs. For example, the induction of endogenous mediators of innate immunity in the treatment of infections could be used in combination with classical antibiotics. Because bacterial resistance is an increasing burden for healthcare and society in general, alternative therapy against infections is urgently needed. Understanding the complex regulation of AMP expression together with relevant internal and external signals will gradually clarify how we can control this arm of innate immunity. Hence, novel treatment regimens based on this concept can emerge for treating infectious diseases.
The authors are supported by the Swedish Research Council (11217 and 20854 a Linnaeus grant CERIC), the Swedish Foundation for Strategic Research, the Torsten and Ragnar Söderbergs Foundation, the Swedish Cancer Society, the Swedish International Development Cooperation Agency (SIDA), Karolinska Institutet, the Icelandic Centre for Research (RANNIS) and the University of Iceland Research Fund.