Introduction

Otitis media (OM) is one of the most common diseases of childhood. In developing countries, it is a significant cause of deafness and chronic suppurative otitis. In developed countries, OM also accounts for more office visits, surgeries, and antibiotic purchases than any other disease, with estimated annual costs of 400 million dollars in Australia, 600 million dollars in Canada, and up to 4 billion dollars in the United States [1–3]. As many as 20% of young children with viral upper respiratory infections develop OM [4]. Acute OM often resolves spontaneously, but chronic OM, recurrent acute OM, and OM with effusion are longer lasting and thereby constitute significant causes of hearing loss. Pathogenesis is multifactorial and includes eustachian tube dysfunction (resulting from small size, functional immaturity, viral respiratory infections, allergy, environmental smoke exposure, and genetic syndromes), invasion of the middle ear space by bacteria and/or viruses, inflammation, and in some cases the formation of biofilms [5, 6•].

In healthy children, the middle ear is generally believed to be sterile. OM is a polymicrobial infection caused by synergism between upper respiratory viruses (most commonly influenza A, respiratory syncytial virus [RSV], adenovirus, and rhinovirus) and several predominant commensal bacteria that commonly colonize the nasopharynx. Streptococcus pneumoniae (SP), nontypeable Haemophilus influenzae (NTHI), and Moraxella catarrhalis (MC) are the most common bacteria isolated by standard culture techniques in acute OM [7•]. The introduction of vaccination against common pediatric strains of SP has resulted in a higher prevalence of NTHI and strains of SP not included in the vaccine [8]. Nonculture polymerase chain reaction techniques confirm the prevalence of these bacteria in acute OM, with the possible addition of a fourth species: Alloiococcus otitidis [9, 10]. In cases of chronic OM and recurrent acute OM, biofilms containing SP, NTHI, coagulase-negative staphylococci, or Staphylococcus aureus appear to be common [11]. Biofilm formation may protect the bacteria from the adaptive and innate immune systems of the host and increase resistance to some antibiotics up to 1,000-fold [12]. Pseudomonas aeruginosa is not a common pathogen in acute OM but is the most common organism isolated in chronic suppurative otitis and in cholesteatomatous otitis [13, 14•, 15].

In 2004, the American Academy of Pediatrics and the American Academy of Family Physicians recommended the option of “watchful waiting,” or observation without antibiotic treatment, for select populations of children with acute OM [16]. A second clinical practice guideline recommended avoidance of antibiotics for OM with effusion [17]. In spite of these recommendations, rates of prescription of antibiotics for acute OM have not decreased [18•]. Unnecessary prescribing of antibiotics for OM is a major contributor to increasing and multiple antibiotic resistance among all three genera of bacteria typically responsible for OM [19]. The need for development of non-antibiotic approaches to prevent and treat OM underscores the importance of understanding the mechanisms by which the healthy middle ear and eustachian tube are protected from bacterial invasion.

The present review begins with a brief overview of the innate immune system in OM, then focuses on defensins and other promising antimicrobial peptides and their roles in protecting the middle ear (Fig. 1).

Fig. 1
figure 1

Innate immunity and the middle ear. Innate immunity includes barrier functions, sensing of pathogenic bacteria (by receptors on and in epithelial cells, dendritic cells, and mast cells), release of antimicrobial peptides and proteins, and recruitment and activity of various effector cells (e.g., neutrophils, macrophages, fibroblasts, natural killer cells, and eosinophils)

Full size image

Innate Immunity of the Middle Ear

Signaling

In the healthy naïve host, the sensors and effectors of the innate immune system are responsible for quickly identifying and clearing pathogens from the middle ear. Epithelial cells, dendritic cells, and mast cells in the uppermost airway, including the eustachian tube and middle ear, express a variety of pattern-recognition receptors that recognize bacterial cell surface molecules and activate a wide range of effector molecules. The importance of several Toll-like receptors (TLRs) and the closely associated MyD88/nuclear factor-κB pathway in OM was reviewed recently [20]. In mice, deficiency or absence of genes that encode for receptors (TLR2, TLR4, TLR9), signaling molecules (MyD88, TRIF), or downstream targets (tumor necrosis factor) leads to the inability to clear OM. Mechanisms for this deficit include persistent thickening of the middle ear mucosa, delayed recruitment of neutrophils and macrophages (though once finally recruited, these cells tend to persist longer), and deficient phagocytosis and intracellular killing of pathogens. The result is persistence of viable bacteria in the middle ear [21, 22••, 23••].

Cellular Response

The cells of the innate immune system can be grouped according to their major functions: phagocytosis (neutrophils, macrophages), barriers (keratinocytes, epithelial cells), signaling (dendritic cells, epithelial cells, fibroblasts, macrophages), and secretion (epithelial cells, eosinophils, basophils, mast cells, and natural killer cells). The influx of neutrophils early in acute OM follows opsonization by complement and leads to rapid phagocytosis of bacteria. Within the neutrophil, the phagosome fuses with the lysosome, which has an acidic environment and contains free radicals, acid hydrolases, lactoferrin, lysozyme, and several defensins. The neutrophil is short-lived and is subsequently phagocytosed by macrophages.

Several of the pathogens of OM have developed protective mechanisms against destruction by neutrophils. SP produce at least three exoglycosidases, which deglycosylate human complement, thus preventing opsonization [24]. NTHI are able to avoid phagocytosis by potently stimulating the host to form neutrophil extracellular traps, a niche in which this pathogen is relatively protected [25].

Interactions Between Innate and Adaptive Immunity

The innate immune system is primarily responsible for immediate and nonspecific responses to infection, whereas adaptive immunity provides long-term protection based on memory of a specific invader. Interactions and overlap between the two components of mammalian immunity are complex. The protective benefit of pneumococcal vaccination against OM caused by the strains contained in the vaccine is strong evidence of the importance of adaptive immunity in this disease. The essential nature of interactions between innate and adaptive immunity is demonstrated by the attenuated responses in mucosal IgA, systemic IgG, and T-helper type 1 (Th1) immune responses in the TLR4 mutant mouse compared with the wild type following vaccination with an outer membrane protein of NTHI [26••]. The increased prevalence of OM in patients with allergy or asthma is further evidence of the important interactions between innate and adaptive immunity in the middle ear.

Cytokines and chemokines interconnect components of the innate and adaptive immune systems. A classic example is viral activation of epithelial cells, smooth muscle cells, and fibroblasts (through recognition of viral double-stranded RNA [dsRNA] by TLR3 and other mechanisms) to produce cytokines and chemokines. For instance, dsRNA stimulates human nasal fibroblasts to produce CCL5 and CXCL8 [27], and rhinovirus infection causes smooth muscle cells to produce interleukin (IL)-6 and CXCL8 [28]. In the presence of Th2 cytokines, IL-4 and IL-13, dsRNA stimulates fibroblasts to produce CCL17, further increasing the Th2 response [29].

Defensins

Defensins are broad-spectrum antimicrobial peptides expressed by all studied species of mammals, birds, invertebrates, and plants. They are an important weapon in the arsenal of the innate immune system, providing rapid response protection from pathogens and shaping the microbiota of several important anatomic niches [30, 31]. They share the following common structural characteristics: small peptides (30–45 amino acids), rich in cationic amino acids so that they have a net positive charge, and stabilized by 3 disulfide bonds that protect them from degradation by proteases [32–34]. In humans, α-defensins are produced in large quantities in the granules of neutrophils and in the Paneth cells of the small intestine, whereas β-defensins are produced by epithelial cells in the upper and lower airway (including the middle ear), as well as by colonic enterocytes, keratinocytes, corneal and conjunctival cells, and the genitourinary tract.

Most defensins are directly antimicrobial at micromolar concentrations and have activity against a variety of bacteria, viruses, fungi, and protozoa. Killing mechanisms may be multifaceted; the predominant mechanism is likely formation of a pore within the microbial membrane. In addition, some defensins can inhibit bacterial toxins, including the cytolysin PLY produced by SP [35]. Some of the defensins have proinflammatory activities, stimulating cytokine and chemokine expression or acting as chemoattractants for neutrophils, mast cells, memory T cells, and dendritic cells [36]. The recent description of self-assembling nanonets of one of the human Paneth cell α-defensins to immobilize Salmonella spp in the lumen of the small intestine is a compelling example of the diversity of biological mechanisms of these peptides (Chu, unpublished data). The defensins and other antimicrobial peptides described in this review are summarized in Table 1.

Table 1 Antimicrobial molecules in otitis media
Full size table

Defensins and Animal Models of Otitis Media

The chinchilla is an exotic but compelling model of OM due to its anatomic similarities to the human and the possibility of creating polymicrobial infections [37•]. The chinchilla expresses a defensin in the middle ear (chinchilla β-defensin 1 [CBD1]), which is orthologous to human β-defensin (HBD) 3 and has potent antimicrobial activity against NTHI, SP, and MC [38]. Some middle ear pathogens are able to influence expression of CBD1. When the chinchilla is infected with RSV, expression of CBD1 is suppressed at the mRNA and the protein levels, resulting in increased colonization with NTHI in the upper respiratory tract. Chinchilla middle ear epithelial cells infected with RSV, MC, or SP express lower levels of CBD1, but this effect is not seen with influenza A, adenovirus, or NTHI [39]. Furthermore, intranasal treatment of chinchillas with recombinant CBD1 significantly decreased NTHI colonization in nasopharyngeal lavage fluid [40••].

The mouse expresses multiple α-defensins in the small intestine and at least four β-defensins in epithelial cells, but its neutrophils do not express defensins. In a murine model involving induction of OM by inoculation of the middle ear with Escherichia coli endotoxin, none of the α-defensins are detected in the healthy middle ear or in OM. Mouse β-defensin 1 is expressed constitutively in healthy middle ear mucosa and in eustachian tube mucosa and does not change in this model of acute OM. Mouse β-defensins 2, 3, and 4 are not expressed in the healthy middle ear or eustachian tube mucosae but are highly induced in these tissues in endotoxin-mediated OM [41]. Recombinant mouse β-defensin 3 (the ortholog of HBD2) is highly effective at killing filamentous fungi and yeasts and moderately effective at killing Gram-positive bacteria, with marginal activity against Gram-negative bacteria [42]. Recombinant mouse β-defensin 2 has potent antiviral activity against influenza A [43•].

Human Defensins and Otitis Media

HBD2 is a potent antimicrobial peptide. Killing of NTHI, MC, and two strains of SP by HBD2 at low concentrations has been demonstrated by radial diffusion assay, liquid broth assay, and with electron microscopy [44]. HBD2 expression is induced in human middle ear epithelial cells by at least two distinct mechanistic pathways. IL-1α induces HBD2 expression through the Raf-MEK1/2 extracellular signal-related kinase mitogen-activated protein (MAP) kinase pathway [45]. In addition, NTHI is recognized by TLR2 and induces HBD2 expression through the MyD88-IRAKI-TRAF6-MKK3/6-p38 MAP kinase signal transduction pathway [46]. Direct measurement of HBDs in the middle ear fluid from patients with acute OM has not yet been reported.

Human α-defensins 1 to 4 are expressed in neutrophil granules and are important in destroying phagocytosed bacteria. Of the four, human neutrophil peptide-1 is the most potent antimicrobial, with activity comparable to that of HBD2 against several strains of NTHI [47]. In a rat model of OM caused by direct inoculation with NTHI, recruitment of neutrophils to the middle ear peaked on day 4, with macrophage recruitment peaking on day 7, and maximal lymphocyte recruitment on day 14 [48]. Infants with severe congenital neutropenia have a high incidence of OM. The precise role of the neutrophil defensins in OM has not been explored.

Other Antimicrobial Molecules

Cathelicidins are cationic peptides with widely diverse activities. In addition to being potent antimicrobials, they are chemotactic for neutrophils, monocytes, and dendritic cells, and are able to inactivate endotoxin and stimulate release of histamine from mast cells. In the chinchilla, a cathelicidin orthologous to the mouse CRAMP has been identified that is expressed throughout the upper respiratory tract, including the middle ear and eustachian tube, and is highly effective at killing NTHI and MC but less potent against SP [39]. Some middle ear pathogens are able to alter expression of cathelicidins, as has been described with β-defensins. In chinchilla middle ear epithelial cells, infection with influenza A suppressed CRAMP expression, as did infection with MC, while infection with RSV and adenovirus had no effect, and infection with NTHI and SP caused an increase [39]. A synthetic peptide, P60.4-Ac, based on the structure of the human cathelicidin LL-37, neutralized the effects of endotoxin and lipoteichoic acid on sinus epithelial cells in culture [49]. Whether direct instillation of such a peptide into the nasopharynx would be effective in treatment of antibiotic-resistant OM is unclear.

Lysozyme

Lysozyme is an antimicrobial enzyme that degrades the peptidoglycan of the bacterial cell wall. It is expressed by many mucosal surfaces, including the gut, eye, and upper respiratory tract. Lysozyme has synergistic killing effects against SP when combined with HBD2 [44]. In a mouse depleted of lysozyme, susceptibility to OM caused by injection of SP into the middle ear increased, suggesting that middle ear lysozyme plays a role in protection against OM [50].

Bacteriocins

Bacteriocins are peptides synthesized by bacteria with antimicrobial activity predominantly against closely related species. Several strains of α-hemolytic streptococci produce bacteriocins. Among them, Streptococcus salivarius has been proposed as an ideal probiotic agent to alter the oral flora [51•]. A decade ago, a clinical trial involved spraying α-hemolytic streptococci directly into the noses of 130 children with recurrent acute OM. The treated children had fewer episodes of acute OM [52]. Subsequent studies have demonstrated fewer bacteriocin-producing bacteria (α-hemolytic streptococci, nonhemolytic streptococci, peptostreptococci, and Prevotella spp) in the adenoidal tissue of otitis-prone children [53]. Peptide ST4SA is an amphiphilic cationic peptide produced by Enterococcus mundtii; it is unusual in that it interacts with the negatively charged cell membranes of both Gram-positive and Gram-negative organisms, creating pores and causing leakage of small intracellular molecules such as DNA and enzymes [54]. This peptide is stable over a wide pH and temperature range, is bacteriostatic against many strains of SP and Pseudomonas spp, and suppressed growth of bacteria recovered from middle ear aspirates [54].

Surfactant Proteins

Surfactant proteins opsonize, aggregate, and promote phagocytosis of bacteria. Some also have direct broad-spectrum antimicrobial activity. In a guinea pig model of OM with effusion, nebulized bovine surfactant administered intranasally decreased edema and reduced the passive opening pressure of the eustachian tube [55•]. The demonstrated effect is likely predominantly mechanical rather than antimicrobial. SPLUNC1 is a newly described antimicrobial protein secreted by the mucus cells and submucosal glands of the upper airway. This protein has broad-spectrum antimicrobial activity, reduces biofilms produced by P. aeruginosa, and functions as a surfactant to reduced surface tension at the air–liquid interface in the upper airway, including the eustachian tube. In the chinchilla model, inhibition of SPLUNC1 with selective small inhibitory RNA did not increase the bacterial load of NTHI within the middle ear but led to eustachian tube dysfunction, likely due to decreased mucociliary clearance [56•].

Cationic Peptides

Cationic peptides that bind to bacterial cell surfaces can be identified by studying the affinity of large peptide phage display libraries for a specific organism. This approach identified five small peptides that bind to the cell surface of NTHI and are bactericidal in the low micromolar range against all strains of H. influenzae tested [47]. All five peptides were highly active when expressed as part of the phage coat; however, the linearized forms lacked activity, dampening enthusiasm for their development as synthetic antimicrobials. A similar approach may be useful in identifying antimicrobial peptides capable of binding to multidrug-resistant organisms.

Halocidin

Halocidin is a small (3.4 kDa) peptide identified in the hemocytes of the tunicate Halocynthia aurantium. It consists of two subunits covalently bound by a single cysteine disulfide bond. Synthetic peptides of the monomeric and dimeric forms of one of the subunits are highly effective at killing both methicillin-resistant S. aureus and multidrug-resistant P. aeruginosa [57]. Compared with topical ofloxacin, the dimeric form (Di-K19Hc) was significantly more potent against these drug-resistant organisms and better at clearing infected middle ear fluid. Importantly, unlike gentamicin, there was no evidence of damage to the hair cells of the organ of Corti with Di-K19Hc [58•].

Xylitol

Xylitol is a five-carbon sugar alcohol that inhibits attachment of H. influenzae and SP to nasopharyngeal cells and inhibits growth of α-hemolytic streptococci, including SP. One mechanism of action appears to be inhibition of capsule formation and biofilm formation by SP [59•]. Clinical trials of xylitol-containing chewing gum have demonstrated decreases in OM of 30%–40% in otitis-prone children [60], although from a practical standpoint, the patients at highest risk of OM are too young to chew gum. Delivering xylitol with a “slow-release pacifier” to prevent caries in infants has been explored [61], but such an approach for infants with recurrent OM has not yet been studied.

Resistance to Antimicrobial Peptides

Many pathogenic bacteria have developed methods of resistance to killing by antimicrobial peptides. One example is particularly pertinent to OM: NTHI that contain the sap (sensitivity to anti-microbial peptides) operon are able to resist killing by chinchilla β-defensin, whereas a mutation in one of the sap genes confers susceptibility to defensin-mediated killing [62]. The sap gene products are a family of at least six proteins that detect the presence of β-defensin, upregulate production of the sap genes, and mediate potassium uptake, thereby protecting the pathogen [63].

A second example is the ability of S. aureus and P. aeruginosa to alter their peptidoglycan via the addition of lysine or alanine, respectively, altering the net surface charge and decreasing affinity for binding of cationic antimicrobial peptides [64]. Whether the more common organisms of acute OM are able to avoid killing by antimicrobial peptides in this way is not known.

Conclusions

Otitis media is a common, expensive, and vexing infection. As antibiotic resistance has increased, guidelines for minimizing use of antibiotics for OM have been published. However, adherence to these guidelines has been poor [65•]. Vaccine development is promising [66], but experience suggests that new strains may replace the strains included in new vaccines; this is particularly relevant for SP. Antimicrobial peptides and proteins are evolutionarily ancient, with minimal resistance developing in spite of long use by virtually all plants, vertebrates, and invertebrates. Defensins and other antimicrobial molecules play an important role in protecting the middle ear and responding to infection of this unique niche. The extremely high prevalence of OM in young children suggests that the innate immune system is easily circumvented in this vulnerable population. The development of formulations of antimicrobial molecules that could be directly instilled into the nasopharynx and their potential use in the treatment of resistant and chronic OM may represent a novel therapeutic approach, thereby adding a hopeful chapter to what heretofore has been the troublesome saga of OM.