Article tools
  • View PDF
  • Cite this article
  • E-mail this article
  • Send a letter
  • Information on reuse
  • Standard abbreviations
  • Article usage
Author information
Need help?


Defending the cornea with antibacterial fragments of keratin

Michael Zasloff

Georgetown Transplant Institute, Georgetown University School of Medicine, Washington, DC, USA.

Address correspondence to: Michael Zasloff, Professor of Surgery and Pediatrics, Transplant Institute, Georgetown University Hospital, 2 Main, 3800 Reservoir Road NW, Washington, DC 20007, USA. Phone: 484.433.7807; Fax: 202.687.2792; E-mail:

First published September 24, 2012

In addition to its role in refraction, the cornea forms a barrier between the eye and environmental and infectious insults. Corneal infections are surprisingly rare, suggesting that multiple aspects of the immune system are at play in mediating protection. In this issue of the JCI, Tam et al. describe the unexpected role of a structural protein, cytokeratin 6A, in this process.

See the related article beginning on page 3665.

The usual healthy appearance of the cornea and conjunctivae of the human eye should puzzle you. Why is it that this surface looks so healthy, most of the time? How is it that despite the almost certain diversity of microbes that come in contact with it, we so rarely see infection, or its associated sign, inflammation, evidence of the body’s mechanisms that are called forth to fight off microorganisms? Lysozyme in tears can defend the eye, but organisms that inhabit the upper airway, such as Staphylococcus aureus are resistant to this enzyme (1), and other antimicrobial systems must be at work, because in the various dry-eye syndromes, the reduced tear production is not associated with frequent bacterial infections. More perplexing is the fact that a corneal transplant will not necessarily develop infections at the incision or around the suture tracks, and antibiotics are not necessarily required postoperatively (2). Thus, even the wounded cornea seems to handle microbes in some mysterious — and remarkably effective — fashion.

Layers of protection

The cornea is a wonderful, close-up example of a site protected almost completely by the chemical and physical defenses of our innate immune system. Tears contain high concentrations (about 1 to 2 mg/ml) of each of three antimicrobial proteins: lysozyme, lactoferrin, and lipocalin (3). The presence of lactoferrin, which chelates iron, and lipocalin, which captures the iron-transporting siderophores used by many bacteria, tells us that many microbes would discover the microenvironment of the corneal epithelium to be unfavorably iron-depleted. The corneal epithelium secretes several types of mucin that adhere to the corneal surface (4) and form a barrier that both provides a physical shield from invaders and creates an “unstirred” micron-thick fluid layer between itself and the corneal epithelial cell. Antimicrobial peptides are secreted into this barrier and can accumulate without diffusion (or dilution) into to the tear fluids. Furthermore, we have known for some time that the epithelium expresses several well-characterized antimicrobial peptides, both constitutively and induced following injury (5). These include several of the β-defensins and LL-37 (cathelicidin), which are believed to be secreted from the epithelial cell onto the corneal surface beneath the mucin layer (6). Antimicrobial peptides act by binding to microbial membranes and fatally disturbing their permeability (7). An organism that somehow made it through the mucin layer would then encounter the high concentrations of antimicrobial peptides bathing the surface of corneal epithelial cells, be killed rapidly, and fail to adhere or to penetrate.

Finding a new player

In this issue of JCI, Tam et al. describe another antimicrobial defense of the corneal epithelium that has gone unnoticed to date (8). These investigators hypothesized that there might be additional, as yet undiscovered, antimicrobial peptides being expressed by the human corneal epithelium. To search for these putative peptides, they derived an immortal cell line from human corneal epithelium and differentiated it into mature corneocytes by addition of calcium ions. In vitro, a crude extract of these cultured cells exhibited antibacterial activity against Pseudomonas aeruginosa. Fractionation of the extract on the basis of molecular weight demonstrated that the activity resided in two fractions (less than 3,000 Da and between 3000 and 10,000 Da). Analysis of these fractions by LC/MS identified the principal active molecules to be peptide fragments derived from the carboxyl-terminal region of cytokeratin 6A.

To confirm that cytokeratin 6A was the protein from which these peptides originated, Tam et al. demonstrated that reducing cytokeratin 6A expression using siRNA significantly attenuated antibacterial activity (8).

The keratin peptides identified in the extract were synthesized, and their individual activities were studied. Perhaps the most surprising result (for those interested in the details of antimicrobial peptide structure-function) was that several of these short peptides did not exhibit the anticipated sequence and structural characteristics of classical linear antimicrobial peptides, because they were glycine-rich and lacking both a net cationic charge and the tendency to form an α-helix. Several of these peptides exhibited bactericidal activity in vitro against potential ocular pathogens in addition to P. aeruginosa, including S. aureus and Streptococcus pyogenes.

One peptide, a 19-mer with a “classical” cationic amphipathic sequence, was effective in rapidly killing a cytotoxic strain of P. aeruginosa in either water or at physiological ionic conditions (150 mM NaCl), suggesting it would retain activity within the liquid film in contact with the apical surface of the corneal epithelium. This peptide, when introduced into culture medium, effectively protected corneal epithelial cells from invasion and cytotoxicity by P. aeruginosa. Further studies by Tam et al. suggested that this 19-mer bound specifically to the bacterial cytoplasmic membrane and caused it to become leaky, subsequently killing the bacterial cell within minutes of exposure.

Perhaps most remarkably, Tam et al. demonstrated that subcorneal injection of cytokeratin 6A siRNA into the cornea of a live mouse, reducing expression of cytokeratin 6A by about 25%, resulted in a 5-fold increase in the adherence of P. aeruginosa introduced onto the treated cornea.

The conclusion from these studies is that a cytokeratin, a protein that plays a structural role within cells, appears to protect the surface of the eye. By some mechanism, not as yet understood, fragments of this protein are generated that find their way to the surface of the cornea and kill microbes.

A new role for structural proteins?

Precedents for structural intracellular proteins serving extracellular antimicrobial functions in vertebrates exist. A notable example has been described in the stomach of the Asian toad (9). The 39–amino acid antimicrobial peptide buforin is the N-terminal fragment of histone 2A that appears to play an antimicrobial role on the gastric mucosa of this species. Buforin is produced by the action of gastric pepsin, which acts on unacetylated cytoplasmic histone 2A that has been secreted into the gastric lumen (9).

The antimicrobial defense system described by Tam et al. has the capacity to be rapidly mobilized, since it uses an abundant protein already present within the cell. In contrast, the expression of the induced antimicrobial peptides would likely occur over the course of hours. Thus, this system might be critical in the initial phases of microbial assault (Figure 1). Many questions are provoked by this discovery. What mechanisms are involved in the fragmentation of cytokeratin 6A, and what are the proteases involved in the process? Where within the corneal epithelial cell are the fragments generated? What signals the fragmentation and/or secretion of these fragments? Since the various keratin fragments appear to have different antimicrobial specificities, might assault by different bacteria result in different patterns of fragmentation? Furthermore, cytokeratin 6A is present in many other sites in the body exposed to microbes, such as the skin, hair, teeth, and various mucosal surfaces, and it is unclear how widely used this keratin-based antimicrobial defense might be. Hopefully, some of these questions, and others not posed, will be answered in the future. In addition, as Tam et al. suggest (8), these keratin-derived antimicrobial peptides appear to be exciting new biocompatible candidates for development as human anti-infective therapeutics.

Principal antimicrobial defenses of the human eye.Figure 1

Principal antimicrobial defenses of the human eye. The epithelium of the cornea is depicted with an overlying mucus layer. A hypothetical progressive sequence is presented in which successive antimicrobial defenses are recruited. The tear film proteins, lysozyme, lactoferrin, and lipocalin create an environment that is nonoptimal for microbial growth. The mucin layer, derived from mucins secreted by the epithelium, creates a physical barrier that microbes must overcome in order to invade the cornea. (i) Should a microbe penetrate the mucin layer, it faces constitutively expressed antimicrobial peptides, such as human β-defensin-1 (hBD-1). (ii) If the microbe gains access to the epithelium, we imagine that cytokeratin fragmentation, as described by Tam et al., generated by cleavage of a preexisting abundant protein, is rapidly mobilized. (iii) Expression of a battery of antimicrobial peptides is subsequently induced over the course of hours.


Conflict of interest: The author has declared that no conflict of interest exists.

Citation for this article:J Clin Invest. 2012;122(10):3471–3473. doi:10.1172/JCI65380.

See the related article beginning on page 3665.


  1. Bera A, et al. Influence of wall teichoic acid on lysozyme resistance in Staphylococcus aureus. J Bacteriol. 2007;189(1):280–283.
    View this article via: PubMed CrossRef
  2. Liu Y, et al. Topical application of FTY720 and cyclosporin A prolong corneal graft survival in mice. Mol Vis. 2012;18:624–633.
    View this article via: PubMed
  3. Yamada M, Mochizuki H, Kawai M, Tsubota K, Bryce TJ. Decreased tear lipocalin concentration in patients with meibomian gland dysfunction. Br J Ophthalmol. 2005;89(7):803–805.
    View this article via: PubMed CrossRef
  4. Mantelli F, Argueso P. Functions of ocular surface mucins in health and disease. Curr Opin Allergy Clin Immunol. 2008;8(5):477–483.
    View this article via: PubMed CrossRef
  5. McDermott AM. The role of antimicrobial peptides at the ocular surface. Ophthalmic Res. 2009;41(2):60–75.
    View this article via: PubMed CrossRef
  6. Jäger K, Nielitz A, Garreis F, Sel S, Nave H, Paulsen FP. In vitro evidence of involvement of the epithelial y+ transporter in beta-defensin production on the ocular surface. Ann Anat. 2011;193(6):479–485.
    View this article via: PubMed CrossRef
  7. Zasloff M. Antimicrobial peptides of multicellular organisms. Nature. 2002;415(6870):389–395.
    View this article via: PubMed CrossRef
  8. Tam C, Mun JJ, Evans DJ, Fleiszig SMJ. Cytokeratins mediate epithelial innate defense through their antimicrobial properties. J Clin Invest. 2012;122(10):3665–3677.
    View this article via:
  9. Kim HS, et al. Pepsin-mediated processing of the cytoplasmic histone H2A to strong antimicrobial peptide buforin I. J Immunol. 2000;165(6):3268–3274.
    View this article via: PubMed