Management of bacterial keratitis: beyond exorcism towards consideration of organism and host factors

INTRODUCTION Microbial keratitis is a common, potentially sight threatening ocular infection that may be caused by bacteria, fungi, viruses, or parasites.1 The challenge for the

ophthalmologist is to distinguish clinically microbial keratitis from other noninfectious inflammatory conditions of the cornea resulting from trauma, hypersensitivity, and other

immune-mediated reactions. A further challenge is to develop a clinical suspicion for a bacterial keratitis as opposed to the other potential infectious causes. There are no absolutely

specific clinical signs that confirm infection or suggest definite bacterial cause, yet the clinician should assess and define the distinctive corneal signs based on the status of the

epithelium (intact or ulcerated), type of stromal inflammation (suppurative or nonsuppurative), and the site of the stromal inflammation (focal, diffuse, multifocal, or marginal). When there

is sufficient evidence based on clinical examination to raise the suspicion for a possible infectious etiology, laboratory studies are required to establish the specific causative organism.

Based on the clinical setting and risk factors, features of clinical examination, results from laboratory investigation, and knowledge of the potential corneal pathogens, a therapeutic plan

is initiated.2 Modification of the therapeutic plan is subsequently based on clinical response, results of _in vitro_ susceptibility determination, and tolerance of the antimicrobial

agents.3 With careful clinical monitoring, antimicrobial therapy is then terminated and any residual structural alterations corrected. Recent trends in therapy of suspected bacterial

keratitis have evolved towards an ‘exorcist approach’. Potent, preferably bactericidal, antibacterial agents are poured onto the ocular surface at frequent intervals, sometimes without the

benefit of antecedent diagnostic studies, in an effort to exorcise the invading pathogen(s). Research into ever more potent compounds has steadily progressed, in part stimulated by emerging

patterns of resistance among pathogens, resulting in development of powerful newer commercial antibacterial agents that reinforce the ‘exorcist approach’. Relatively little attention,

however, has been directed towards strategies for preventing destruction of the corneal tissue as a result of release of organism- and host-derived factors. Newer approaches to gain better

understanding of the molecular microbiologic events involved in the infectious process will allow development of a comprehensive treatment strategy focused on not only the organism but

important host factors as well. Tools for the new frontier in therapy of infectious keratitis will likely emerge from microbial genomics and proteomics. Such tools will hopefully identify

newer targets for intervention to prevent corneal destruction as a result of infectious keratitis. Contemporary strategies for management of suspected acute bacterial keratitis must include

consideration of organism- and host-derived factors in order to maximize eradication of the invading pathogen(s) while simultaneously protecting the orderly corneal anatomy in order to

preserve function. PATHOGENESIS The pathogenesis of ocular infectious disease is determined by the intrinsic virulence of the microorganism, the nature of the host response, and the

anatomical features of the site of the infection.4 The avascular clear anatomical structure of the cornea with its specialized microenvironment predisposes to potential alteration and

destruction by invading microorganisms, virulence factors, and host response factors. The intrinsic virulence of organism relates to its ability to invade tissue, resist host defense

mechanisms, and produce tissue damage.5 Penetration of exogenous bacteria into the corneal epithelium typically requires a defect in the surface of the squamous epithelial layer. By virtue

of specialized enzymes and virulence factors, a few bacteria, such as _Neisseria gonorrhea_, _N. meningitidis_, _Corynebacterium diphtheriae_, _Shigella_ and _Listeria_ may directly

penetrate corneal epithelium to initiate stromal suppuration. Bacteria colonize host cells by engaging adhesins at their surface with receptors on the host cell surface. Specific receptors

are often required by many adhesins to achieve binding. Besides adherence, microbial adhesins also contribute to subsequent interactions. Virulence factors may initiate microbial invasion or

secondary effector molecules may assist the infective process. Upregulation or downregulation of host defense mechanisms may be involved. Adhesins may also be toxins.6 Therefore, receptor

recognition is only the first step in the pathogenesis of infection directed by microbial adhesin molecules.7 Many bacteria display several adhesins on fimbriae (pili) and nonfimbriae

structures. Such adhesive proteins may recognize carbohydrates on host cells, or alternatively protein–protein interactions can also occur. Certain bacteria exhibit differential adherence to

corneal epithelium. The adherence of _Staphylococcus aureus_, _S. pneumoniae_, and _Pseudomonas aeruginosa_ to ulcerated corneal epithelium is significantly higher than other bacteria and

may account in part for their frequent isolation.8 _P. aeruginosa_ has many virulence factors that contribute to pathogenesis. Cell-associated structures such as pili9 and flagella10 and

extracellular products, such as alkaline protease,11 elastase,11 exoenzyme S,6 exotoxin A,12 endotoxin,13 slime polysaccharide,14 phospholipase C,11 and leukocidin,11 are associated with

virulence, invasiveness, and colonization. Whereas Gram-positive bacteria, including _S. aureus_, adhere to host tissues via fibronectin and collagen,15 _P. aeruginosa_ attaches to cell

surfaces that lack fibronectin.16 Bacteria adhere to injured cornea,17 to exposed corneal stroma,18 or to immature nonwounded cornea.19 The corneal epithelial receptors for _Pseudomonas_ spp

are glycoproteins.20,21 Pili (fimbriae) are thin (4–10 nm in diameter) protein filaments, which are located on the surfaces of many bacteria. _In vitro_ studies indicate that purified pili

successfully compete for binding with cold bacteria by saturating available binding sites on the ocular surface.22 Monoclonal antibodies specific for _Pseudomonas_ pili and a

peptide-conjugated alkaline phosphatase allowed for identification of host corneal receptor molecules.23 Characterization of these receptor proteins indicates that carbohydrates are

necessary for receptor activity.21 In competitive inhibition experiments, sialic acid was the only aminosugar able to completely inhibit pilus binding to mouse corneal epithelial proteins.

Pili have been used to protect against _P. aeruginosa_ keratitis.22 Pili, however, have considerable antigenic variation between strains.24 Bacteria also possess an array of other virulence

factors, including nonpilus adhesins.25 Some clinical isolates of _P. aeruginosa_ from keratitis are reportedly nonpiliated.26 Flagella are subcellular filamentous organelles (16–18 nm in

diameter) originating in the cell membrane and extending to 15–20 _μ_m from its surface. These organelles are responsible for bacterial motility, which is important for dissemination of

infection.27 The flagellum has a helicofilament composed of cell-assembling subunits of the protein flagellin. More than 95% of _P. aeruginosa_ clinical isolates are flagellated.28 As

flagella contain aminospecific proteins and have large surface exposure available for antibody binding, they have been considered as the basis for vaccine development. Much research

examining flagellum as a virulence factor in _P. aeruginosa_ infection and/or as a vaccine candidate has focused on the burned-mouse model.29 A loss of virulence was observed in

flagella-deficient mutants. Active or passive systemic, as well as topical immunization with flagella, or topical antiflagellar antibody homologous to the infecting strain of bacteria,

protects mice against _P. aeruginosa_ corneal infection.30 In addition to adhesins, the adherence of _P. aeruginosa_ and _N. gonorrheae_ is guided by glycocalyx, a biological slime that

enables them to adhere to susceptible cells producing slime aggregates that are resistant to phagocytosis.31 Similar coatings may form on contact lenses to facilitate the adherence of

bacteria to the lens material.32 Following adherence of microorganism to the cornea and initiation of infection, the complex tissue reactions of the host response occur, including

inflammation, neovascularization, cellular and humoral immune responses, and stromal degradative processes. A variety of cytokines may be released in response to corneal infection. During

inflammation, leucocyte adhesion to the vascular endothelium is enhanced by interleukin-I (IL-I) and tumour necrosis factor (TNF), products of macrophages and T lympocytes.33 IL-I is a known

potent intracellular mediator of inflammation and chemotaxis for neutrophils. TNF stimulates immunocompetent cells and induces release of IL-I and interleukin-VI (IL-VI) from macrophages

and other cells.34 Bacterial exotoxin-A downregulates TNF, IL-I, and lymphotoxin by inhibiting the cells' ability to produce these cytokines.35 Neutrophil infiltration following corneal

infection is a principal host defense mechanism. The host inflammatory response to _Pseudomonas_ spp has been studied in mice designated as susceptible or resistant. Susceptible mice are

unable to clear the cornea, whereas resistant mice may restore corneal clarity within 1 month after ocular challenge.36 Resistant mice appear to have a larger number of corneal leucocytes

present initially. The resistant mice also have a shorter duration of inflammatory cells and bacteria in the cornea compared with susceptible mice. If neutrophils are experimentally

depleted, many resistant animals succumb to lethal _Pseudomonas_ sepsis within 48 h.37 In aged mice, deficiencies of neutrophilic phagocytosis have been observed. These observations may

partially explain the clinically apparent age-related susceptibility of individuals to corneal infection. Preimmunization of rats with phenol-killed _Pseudomonas_ organisms still results in

massive corneal stromal degradation caused by neutrophil migration in the absence of viable bacteria. Naive, unimmunized mice show little stromal destruction during early infection, despite

the presence of numerous bacteria in the cornea. Thus, immune recognition is involved in the host response to _Pseudomonas_ corneal infection and is apparently required for phagocytosis, but

not for neutrophil recruitment.38 Models of experimental _P. aeruginosa_ ocular disease also indicate the importance of the third component of complement (C3) in host resistance to corneal

infection. Resistant mice, experimentally depleted of C3 by Cobra venom and then inoculated with _Pseudomonas_ organisms to the cornea, respond with delayed leucocyte mobility, bacterial

persistence in the cornea, and subsequent scarring and opacification.39 The cellular/cytokine network in response to experimental infection _in vivo_ has begun to be unravelled, and the data

provide substantive evidence for a regulatory role of CD4(+) T cells (Th1 type) contributing directly to persistence of PMN in the cornea of susceptible C57BL/6 (cornea perforates) _vs_

resistant BALB/c (cornea heals) mice. Additionally, in the susceptible mouse model, CD4(+) T cells interact with Langerhans cells (LC) and B7/CD28 ligation appears critical for antigen

presentation and the susceptibility response. Various cytokines and chemokines (eg, MIP-1alpha, IL-1beta, MIP-2, IL-12, and IFN-gamma) and their pattern of sustained upregulation after

infection in susceptible _vs_ resistant mice also are being understood in the light of an _in vivo_ cytokine network. T-cell-mediated pathogenic mechanisms are of importance in development

of the susceptible response to _P. aeruginosa_ ocular infection. In the absence of T-cell infiltration into the cornea, PMNs do not persist in the stroma, and cytokines and chemokines are

better balanced, resulting in decreased stromal destruction and the resistance response.40 Previous experimental studies have shown that extended-wear contact lens usage results in a

centripetal migration of LC from the conjunctiva into the central cornea. To test the consequences of this, LC were induced into the cornea before _P. aeruginosa_ infection in BALB/c mice

that are normally resistant (the cornea heals) and in C57BL/6 mice that are susceptible (the cornea perforates) to bacterial challenge. Mean clinical scores, slit-lamp examination, adenosine

diphosphatase (ADPase), and acid phosphatase staining as well as immunostaining with DEC-205, B7-1, CD4, and interleukin-2 receptor (IL-2R) antibodies and histopathologic, reverse

transcription-polymerase chain reaction (RT-PCR), and delayed-type hypersensitivity (DTH) analyses were used to examine the effects on bacterial disease after polystyrene bead induction of

LC into the cornea before bacterial challenge. No difference in disease response was observed in bead- _vs_ sham-treated C57BL/6 mice after bacterial infection; however, significant

differences leading to corneal perforation were seen in BALB/c mice that included an increased number of LC in the central cornea at 1 and 6 days after infection, an increased number of

B7-1+(mature) LC at 6 days after infection, CD4+and IL-2R+T cells at 5 days after infection, enhanced DTH, and increased mRNA levels for IFN-gamma in cornea and cervical lymph nodes.

Alternately, levels of IL-4 were significantly higher in the cornea and cervical lymph nodes of sham- _vs_ bead-treated animals. These data provide evidence that LC are critical in the

innate immune response to _P. aeruginosa_ and provide new information regarding the mechanisms governing resistance _vs_ susceptibility to bacterial infection with this opportunistic

pathogen.41 The role of macrophage inflammatory protein-1alpha (MIP-1alpha) in cell infiltration into _P. aeruginosa_-infected cornea and subsequent disease was examined. Greater amounts of

the chemokine (protein and mRNA) were found in the infected cornea of susceptible B6 (‘cornea perforates’) _vs_ resistant BALB/c (‘cornea heals’) mice from 1 to 5 days postinfection.

Treatment of BALB/c mice with recombinant (r) MIP-1alpha exacerbated disease and was associated with an increased number of neutrophils (PMNs) in the cornea. Treatment of BALB/c mice with

rMIP-1alpha also induced recruitment of activated CD4+T cells into the affected cornea, converting resistant to susceptible mice. Depleting CD4+T cells in r-treated BALB/c mice significantly

decreased PMNs in cornea tissue, suggesting that T cells regulate persistence of PMNs at this site. In B6 mice, administration of neutralizing MIP-1alpha polyclonal antibody also

significantly reduced PMN numbers and pathology. Collectively, evidence is provided that MIP-1alpha directly contributed to CD4+T-cell recruitment and indirectly to PMN persistence in the

infected cornea.42 Macrophage migration inhibitory factor (MIF) is a recently rediscovered proinflammatory cytokine, and has been shown to play a role in the regulation of neutrophil

chemokines and angiogenesis. Corneal epithelial and endothelial cells have been shown to express MIF. A study evaluated the expression of MIF during _Pseudomonas_ keratitis in mice and _in

vitro_ using a corneal epithelial cell line.43 Three strains of _P. aeruginosa_, 6294 (invasive strain), 6206 (cytotoxic strain), and Paer1 (noninfectious strain), were used. Both cytotoxic

and invasive strains were isolated from human corneal ulcers and the Paer1 strain was isolated from a noninfectious condition. Following challenge in mouse corneas or a corneal epithelial

cell line, corneal homogenates or lysed corneal epithelial cells were used to isolate RNA. MIF mRNA expression in the mouse cornea or human corneal epithelial cells was examined by RT-PCR

analysis, and was found to be expressed as early as 4 h after the injury (scratch controls) or infection in the mouse corneas. MIF mRNA in scratch controls and Paer1-inoculated corneas

showed peak levels at 4 h postchallenge and this dropped by 24 h postchallenge. Corneas challenged with invasive and cytotoxic strains showed peak 24 h postchallenge. MIF mRNA levels were

significantly higher in invasive and cytotoxic strain inoculated corneas compared to Paer1-inoculated corneas. Challenging the corneal epithelial cell line with _Pseudomonas_ 6294 and 6206

strains induced peak expression at 8 h and levels were decreased by 12 h. Epithelial cells inoculated with recombinant human interleukin-1beta protein induced very high levels of MIF mRNA at

all time points compared to infected and control corneal epithelial cells. High expression of MIF in the infected corneas suggests that it may have a role in the pathogenesis of corneal

disease induced by invasive and cytotoxic strains of _P. aeruginosa_.43 Evidence suggests that _P. aeruginosa_ stromal keratitis and corneal perforation (susceptibility) is a CD4(+)

T-cell-regulated inflammatory response following experimental _P. aeruginosa_ infection. A study examined the role of LC and the B7/CD28 costimulatory pathway in _P. aeruginosa_-infected

cornea and the contribution of costimulatory signaling by this pathway to disease pathology.44 After bacterial challenge, the number of LC infiltrating the central cornea was compared in

susceptible C57BL/6 (B6) _vs_ resistant (cornea heals) BALB/c mice. LC were more numerous at 1 and 6 days postinfection, but were similar at 4 days postinfection, in susceptible _vs_

resistant mice. Mature, B7-positive-stained LC in the cornea and _P. aeruginosa_-associated LC in draining cervical lymph nodes were also increased significantly following infection in

susceptible mice. To test the relevance of these data, B6 mice were treated systemically and subconjunctivally with neutralizing B7 (B7-1/B7-2) mAbs. Treatment decreased corneal disease

severity and reduced significantly the number of B7-positive cells as well as the recruitment and activation of CD4(+) T cells in the cornea. IFN-gamma mRNA levels were also decreased

significantly in the cornea and in draining cervical lymph nodes of mAb-treated mice. When CD28(−/−) animals were tested, they exhibited a less severe disease response (no corneal

perforation) than wild-type B6 mice and had a significantly lower delayed-type hypersensitivity response to heat-killed _P. aeruginosa_. These results support a critical role for B7/CD28

costimulation in susceptibility to _P. aeruginosa_ ocular infection.44 In _P. aeruginosa_ ocular infection, T-helper cell 1-responsive mouse strains are susceptible (the cornea perforates),

and neutralization of IFN-gamma before infection has been shown to delay the onset of perforation. IFN-gamma is the predominant cytokine induced by IL-12, and positive regulation of IL-12 by

IFN-gamma, if unchecked, leads to excessive cytokine production and toxicity. Despite its potential importance, the role of IL-12 in ocular infection with _P. aeruginosa_ remains unexplored

and was the purpose of this study. IL-12 knockout mice, histopathology, RT/PCR and ELISA analyses, immunocytochemistry, and quantitation of viable bacteria in cornea were used to examine

the role of IL-12 in IFN-gamma production and the susceptibility phenotype. To test directly the effect of IL-12 on IFN-gamma production, IL-12 knockout and wild-type C57BL/6 mice were used.

Both groups of mice were susceptible to infection, with corneal perforation seen at 5–7 days after infection. RT-PCR and ELISA analyses confirmed that IL-12 message and protein levels were

elevated after infection only in the wild-type mouse cornea. Other differences between the two groups were detected. Knockout _vs_ wild-type mice showed a significant decrease in IFN-gamma

mRNA levels in the cornea and cervical lymph nodes and decreased TNF-alpha protein levels in cornea. Corneas of knockout mice also had a significant increase in bacterial load at 5 days

after infection when compared with wild-type mice. These data provide evidence that IL-12 is important in IFN-gamma production, and in the absence of the cytokine, both IFN-gamma and

TNF-alpha levels in cornea are significantly decreased, resulting in unchecked bacterial growth and perforation.45 In addition to organism factors, host lysosomal enzymes and oxidative

substances produced by neutrophils, keratocytes, and epithelial cells may significantly contribute to the destruction caused by _Pseudomonas_ keratitis.46 Two eukaryotic gelatinase species

have been characterized, including a type IV collagenase (72 kDa) and a type V collagenase (90–100 kDa).47 Corneal epithelial cells produce predominantly the 92 kDa form of progelatinase,

whereas stromal fibroblasts synthesize predominantly the 72 kDa progelatinase.48 The progelatinase is cleaved by _Pseudomonas_ elastase to produce the active form.49 A variety of bacterial

toxins and enzymes may be produced during corneal infection to contribute to destruction of the corneal substance. _P. aeruginosa_ produce many toxic substances capable of causing necrotic

central corneal ulceration. Toxin-A inhibits protein synthesis much as diphtheria toxin by catalysing the transfer of the adenosine 5′-diphosphate ribose (ADPR) portion of nicotinamide

adenine dinucleotide to mammalian elongation factor II. With Staphylococcal keratitis, alpha-toxin, but not protein A, is a major virulence factor mediating corneal destruction.50 In

summary, the pathogenesis of bacterial keratitis initially requires the adherence of bacteria to disrupted or normal corneal epithelium. Adhesins are microbial proteins that direct the

high-affinity binding to specific cell-surface components. These adhesins are able to promote bacterial entry into the host cell, derange leucocyte migration, activate plasmin, and induce

cytokine production. In addition, they may act as toxins directly. Adhesins recognize carbohydrate and protein moieties on the host cell surface. Most bacteria can display a number of

adhesins. Although the cognate oligosaccharides for bacterial adhesins are known, the molecules bearing these determinants are not well characterized. Integrins are a family of glycoproteins

mediating cell–cell and cell–extracellular matrix recognition. Many bacterial pathogens have co-opted the existing integrin-based system masking ancillary ligand recognition in a form of

mimicry. Once the bacterial pathogen has adhered to the corneal epithelial surface, the next step in establishing infection is invasion into the corneal stroma. Bacterial invasion is

facilitated by proteinases that degrade basement membrane and extracellular matrix and cause cell lysis. Proteinases may be derived from bacteria, corneal cells, and migrating leucocytes.

Corneal matrix metalloproteinases are excreted in an inactive form, but are activated during infection by bacterial proteinase. Corneal proteinase production may also be induced during the

course of infection. The invasion of bacteria into the cornea is facilitated by a number of exotoxins, including _P. aeruginosa_ phospholipase, heat-stable haemolysin, and exotoxin-A, which

leads to stromal necrosis. Once the bacterial invasion into the cornea has ensued, infection is further facilitated by a complex sequence leading to interruption of the host immune response.

Exopolysaccharide formation by both Gram-positive and Gram-negative bacteria results in local immunosuppressive effects. Certain bacteria with capsular polysaccharide also have

immunosuppressive properties, including interference with phagocytosis. Proteases degrade complement components, immunoglobulins, and cytokines and may inhibit leucocyte chemotaxis and

lymphocyte function. Toxin-A inhibits protein synthesis much as diphtheria toxin by catalysing the transfer of ADPR portion of nicotinamide dinucleotide to mammalian elongation factor-II.

Exoenzyme-S is another ADP-ribosyl transferase that may act as an adhesin and also contribute to dissemination of the organism. Two specific bacterial proteases, elastase and alkaline

protease, cause marked destruction of the cornea when injected intrastromally.51,52 Intrastromal injection of purified elastase alone also results in severe corneal damage. Inhibition of

elastase activity with 2-mercaptoacetyl-L-phenylalanine-L-leucine53 prevents keratolysis. The proteases contribute to the pathogenesis of keratitis by degrading basement membrane,54 laminin,

proteoglycans, extracellular matrix,55 and collagen.56 In addition, the bacterial proteases inhibit host defense systems by degrading immunoglobulins, interferon, complement, IL-I, IL-II,

and TNF.51 Such interference results in decreased neutrophil chemotaxis, T-lymphocyte function, and NK cell function. Mutants deficient for alkaline protease do not establish corneal

infection, suggesting that this protease is an important initiating factor.57 A bacterial heat-labile phospholipase C has been shown in antibody/substrate specificity studies to be produced

in mouse ocular infections suggesting its role as a potential virulence factor.58 Bacterial lipopolysaccharide (LPS) stimulates neutrophil migration and infiltration into the cornea with

subsequent corneal scarring and opacification.59 Bacterial exotoxins are released by actively replicating organisms and some endotoxins are released only after the death of the organism.

These enzymes and toxins have been shown to persist in the cornea for a protracted period and continue to cause stromal destruction after the death of the pathogen. Most of the bacterial

exotoxins are thermal labile and have antigenic properties. Gram-positive bacteria elaborate a variety of biologically active and immunologically distinct toxins. Coagulase-positive strains

of Staphylococci are the most pathogenic and elaborate other extracellular enzymes, such as staphylokinase, lipase, hyaluronidase, DNase, coagulase, and lysozyme. Coagulase-negative

Staphylococci, including _S. epidermidis_ also produce potentially destructive toxin.60 Streptococcal toxins include Streptolysin O and S, erythrogenic toxin and the enzymes hyaluronidase,

streptodornase, and streptokinase. The invasiveness of _S. pneumoniae_ is aided by collagenase activity,61 although the organism may be inherently invasive without toxin production. The

lipopolysaccharides composing the endotoxins within the cell wall of Gram-negative bacteria may be released upon the death of the organism. These lipopolysaccharides may result in the

production of stromal rings. These rings have been shown to consist of polymorphonuclear leucocytes within the corneal stroma, which have been chemoattracted by the alternative complement

pathway.62,63 In addition to occurring with Gram-negative bacterial keratitis, such ring infiltrates have been described in fungal, viral, and acanthamoebal keratitis. In nonbacterial

corneal infection, these stromal rings are thought to result from antigen–antibody precipitants (immune rings). TREATMENT In general, because of the potential rapid destruction of corneal

tissue that may accompany bacterial keratitis, if there is a clinical suspicion suggestive of a bacterial pathogen, the patient should be treated appropriately for bacterial keratitis until

a definitive diagnosis is established. The objective of therapy in bacterial keratitis is to eliminate the infective organism in a rapid fashion, reduce the inflammatory response, prevent

structural damage to the cornea, and promote healing of the epithelial surface.3 With suspected infectious keratitis, the clinician has the option based on the clinical impression and

severity of the keratitis of initiating specific directed or broad-spectrum antimicrobial therapy, or deferring treatment pending the results of laboratory investigation, or monitoring

clinical signs.64,65 The initial therapy selection is based on the clinical features, antecedent risk factors, and familiarity with the most likely responsible corneal pathogen(s) and their

respective antimicrobial susceptibility patterns.66 A basic plan for therapy of severe suppurative keratitis depends on the results of Gram stains on smears from diagnostic corneal

scrapings.2 Inherent in such a therapeutic plan is confidence in the quality of specimen obtained and the technical proficiency of the microbiology laboratory in the processing and

interpretation of the clinical material. Evaluation of diagnostic smears with Gram stain is a relatively insensitive method for diagnosis of bacterial keratitis. With optimal conditions,

pathogen(s) can be identified with microscopic analysis of Gram-stained smears in about 75% of monobacterial keratitis and 37% of polybacterial keratitis cases.3 Screening of smears from

corneal scrapings with acridine orange and fluorescent microscopy is more sensitive than Gram stain,67,68 raising the sensitivity to approximately 80%. Given the 20–30% possibility of a

false-negative interpretation of smears from corneal scrapings, the ophthalmologist must base initial therapy on the clinical assessment of severity. If the Gram stain is equivocal or there

is uncertainty in interpretation of diagnostic smears, broad-spectrum antibiotic coverage should be initiated in the initial treatment of all cases of severe suppurative microbial keratitis

since the consequences of inappropriate or inadequate therapy can be devastating.65 The design for drug administration in severe suppurative keratitis includes antibiotics administered

frequently. Owing to the rapid evolution to perforation in keratitis because of virulent pathogens and visual loss secondary to central corneal scarring, many patients with bacterial

keratitis having significant ulceration may require hospitalization in the absence of adequate support and assistance. The high frequency and intense dosage scheduling of antimicrobial

therapy often requires assistance of trained nursing personnel. Frequent antibiotic administration required for therapy of severe bacterial keratitis can often fatigue even highly motivated

patients and family members, resulting in decreased compliance. In the absence of severe fulminant keratitis with impending perforation, patients having mild to moderate keratitis may be

managed as outpatients with close monitoring.69 Indeed, the availability of highly potent topical antibacterial solutions has permitted outpatient management of bacterial keratitis in all

but the most severe cases. With the goal of rapid cessation of replication and elimination of the bacterial pathogen, selection of appropriate therapy requires an effective drug having

minimal toxicity. The results of laboratory investigations assist the clinician in selecting the most specific and efficacious treatment based on results from microbial culture and

antimicrobial susceptibility data. INITIAL TOPICAL ANTIBIOTIC THERAPY The large number of active antimicrobial drugs available to the treating clinician offers the patient with bacterial

keratitis a greater chance for cure with less drug-related toxicity while providing alternative choices despite the continuing emergence of drug-resistant pathogenic organisms. In addition

to considering the possible organism(s) involved, most effective classes of antibacterial agents, and the possibility of organism resistance, variation in patient response, that is

predicting the response of an individual patient to both the organism and the proposed therapy, aids the ophthalmologist in determining the most effective, least toxic agent(s) for

treatment. Additional criteria required to identify the actual drug(s) of choice for a particular patient include outpatient _vs_ in-patient management and cost. Although some organisms

continue to be predictably susceptible to selected antimicrobial agents, the development of clinically important resistant isolates is common because of the high selective pressures applied

by intense antibiotic usage. Mechanisms of organism resistance include plasmid transfer, either by conjugation or transduction and chromosomal mutations. The rate of the emergence of

mutation and the extent of resistance may vary between organisms and from drug to drug. Depending upon the mechanism and degree of resistance, the administration of larger dosages of drug

that do not cause serious adverse effects may be adequate for clinical cure. With corneal infection, an advantage is direct accessibility via topical route of administration to achieve high

tissue concentrations without significant systemic or local side effects. In contrast, effective treatment of certain infections, such as nontuberculous Mycobacterial keratitis, often

requires combinations of antimicrobial drugs to avoid the rapid development of highly resistant strains that appear when single agent therapy is used. Antimicrobial susceptibility testing is

recommended whenever there is doubt about the potential susceptibility of a given pathogen. Patterns of antimicrobial susceptibility may vary from one geographic area to another, among

different hospitals in a given area, between clinics within a single building, and between community-acquired and hospital-acquired infections. With community-acquired bacterial keratitis,

antibiotic resistance is infrequently encountered and usually surmounted with intensive topical dosing. Nosocomial corneal infections as with those sustained in patients maintained on

chronic mechanically assisted ventilation with exposure keratopathy in intensive care or burn units should be monitored closely for high-level organism resistance. Apart from the virulence

of the organism, severity of corneal infection is also determined in part by the age, genetic makeup, and general health of the patient. Deficiencies of humoral and cellular defense

mechanisms negatively impact patient response to therapy. Factors that may limit the use of a particular agent in an individual include a reliable history of an allergic reaction to the

antibiotic (rash, urticaria, angioedema, wheezing, or anaphylaxis), potential adverse interactions, and predictable adverse effects in certain clinical situations (eg use of tetracyclines in

young children or pregnant women). Outpatient _vs_ in-patient drug therapy modes are distinguished primarily by the severity of the keratitis and the anticipated compliance with treatment.

Among possible efficacious alternatives, low toxicity and ease of administration are the most important factors in selection of drug therapy for outpatients. Potency of the anti-infective

agent with requirement for less frequent administration may improve compliance. With severe keratitis and the potential for permanent structural alteration, the rapid attainment of

inhibitory, preferably bactericidal concentration within corneal tissues is the primary consideration often over-riding factors such as convenience and localized toxicity. The impact of a

drug cost is an increasingly important variable in therapeutic choices. Owing to the potential devastating effects of severe keratitis, the ophthalmologist's primary responsibility is

selection of the antimicrobial agent(s) that is most likely to effect a complete cure in the shortest time. In some instances, selection of an appropriate, equally effective, but less

expensive, preparation may ensure compliance in patients having mild to moderate keratitis. The ophthalmologist should make therapeutic comparisons based on unbiased sources of information

assessing the cost of equally effective preparations. ROUTES OF ADMINISTRATION One of the fundamental principles of pharmacotherapy is to maximize the amount of drug that reaches the site of

action in sufficient concentration to cause a beneficial therapeutic effect.70 Topical application is the main stay of ocular drug delivery systems and the topical route is the preferred

method of application of antibiotics in therapy for bacterial keratitis.71,72,73 Eyedrops are the most common route of antibiotic delivery to the eye. Other topical preparations, including

ointments, gels, and sustained release vehicles, are used to achieve higher concentrations of antibiotics in the corneal stroma. Drug penetration into the cornea may be increased with higher

concentrations, greater lipophilicity, more frequent applications, and enhanced contact time using certain vehicles.74,75 The corneal epithelium represents a potential barrier to antibiotic

penetration and absence of the epithelium, as with ulcerative keratitis, often enhances drug penetration. Fortified concentrations of antibiotics are more effective and preferred to the

commercial strength of many antibiotics.71 Fluoroquinolone antibiotics may be effective at their commercial concentrations in therapy for bacterial keratitis given their relative high

potency.20 Fortified topical antibiotics applied every 30 min are equally effective to administration every 15 min.76 To achieve rapidly peak corneal concentrations, a ‘loading’ dose of

antibiotic is initiated with frequent, repetitive administration.77 Antibiotic solutions are preferred to ointment formulations because of the ease of altering the concentration of the

solution as opposed to the ointment form. Ointments have the theoretical advantages of prolonged contact time, lubricating properties for the ocular surface, high viscosity to resist

dilution with reflex tearing, and ease of administration. Peak corneal drug concentrations may be limited compared to achievable levels with antibiotic solutions since the ointment must

dissolve in the preocular tear film before it can penetrate into the cornea.78 Antibiotic ointment preparations do not appear to adversely deter wound healing.79 Fortified antibiotics are

prepared by diluting the desired amount of parenteral compound with an appropriate vehicle, such as artificial tear solution or balance salt solution. Caution should be exercised to avoid

mixing two antibiotics together. Different antibiotic chemical structures may result in varying stabilities in aqueous solution. Fortified cephalosporins and aminoglycosides remain stable

and maintain potency in aqueous solution for at least 1 week without significant loss of antibiotic activity.80 The initial loading dose of antibiotic is achieved with topical administration

of one drop every 2 min for five applications. Antibiotics are then administered every 30 min (alternating drugs when multiple antibiotics are employed) for the first 24–36 h, depending on

initial clinical severity judgement. If there is difficulty obtaining fortified antibiotic formulations, therapy should be initiated with very frequent commercial strength antibiotic

application without delay. An advantage of the available fluoroquinolone ophthalmic solutions is their high potency, stability at room temperature, and broad spectrum of antibacterial

activity at commercial strengths. Subconjunctival injection of antibiotics can result in high corneal drug concentrations via diffusion and by leakage through the injection site.81 Topical

fortified tobramycin therapy resulted in a greater reduction of _Pseudomonas_ organisms than did subconjunctival injection of tobramycin in experimental keratitis.31 Topical administrations

of fortified antibiotics appear just as effective as subconjunctival antibiotic injection for therapeutic effect.82,83 Subconjunctival administration of antibiotics may provoke patient

anxiety, pain, subconjunctival haemorrhage, and conjunctival scarring. Depending on the agent administered, pH, and osmolarity, injection may result in significant toxicity to the

conjunctiva. Toxicity may lead to conjunctival scarring, which may make any subsequent conjunctival surgery more difficult. With conjunctival chemosis and inflammation, the potential for

inadvertent penetration and intraocular administration is an additional concern. Based on the clinical evidence, there appears to be no therapeutic advantage of subconjunctival antibiotic

administration _vs_ topical fortified application.84,85,86 Subconjunctival antibiotic injections may be indicated in certain clinical situations, such as impending corneal perforation, where

fortified topical antibiotics cannot be reliably delivered because of patient compliance. The pain from subconjunctival injections may be reduced by adjusting the pH of the antibiotic

solution if possible. The conjunctiva may be anaesthetized prior to administration with topical proparacaine and injection of subconjunctival xylocaine 1% solution (0.25 ml delivered through

a 30-gauge needle) prior to antibiotic administration in the same area. Subconjunctival injections may be repeated every 12–24 h at separate sites during the initial 24–48 h. Continuous

lavage of antibiotic may deliver a high concentration and mechanical irrigation to remove potential virulence factors.87,88 Continuous lavage or infusion via a scleral contact lens

(Mediflow) is costly and impractical because of the requirement of patient immobilization. In addition, such therapy may result in epithelial trauma, cumulative toxicity of the ocular

surface, or systemic toxicity. Hydrophilic soft contact lenses may act as a tear film antibiotic retention device and enhance penetration by prolonging the relative contact time.89,90 If

corneal ulceration is marked, the temporary use of a therapeutic soft contact lens may facilitate stromal repair and promote re-epithelialization by protecting the corneal surface from

mechanical trauma of lid movement. Collagen corneal shields soaked in antibiotic solutions have also been shown to increase antibiotic penetration compared with therapeutic soft contact

lenses.91,92 Collagen corneal shields have been effective adjuncts in the therapy of experimental bacterial keratitis.93 Polymer inserts have also been designed to prolong the presence of

drug in the preocular tear film.94 Temporary intracanalicular collagen implants may also prolong the retention of antibiotics in tears and increase stromal concentration to enhance bacterial

killing in experimental keratitis.95 Liposomal systems have been designed to improve the interaction of drugs with the corneal surface to enhance potentially the safe delivery of

antibiotics.96 Transcorneal iontophoresis of antibiotics has also been employed to increase attainable drug concentrations and increase efficacy of antibacterial therapy.97,98 Antibiotics

administered via the parenteral route are relatively poorly absorbed into the noninflamed eye. Bactericidal corneal tissue levels can only be achieved at the risk of systemic toxicity.

Bacterial keratitis with accompanying ocular inflammation may enhance the ocular penetration of systemic antibiotics.99 Severe gonococcal keratoconjunctivitis should be treated with systemic

antibiotics, in addition to topical antibiotics to prevent possible ulcerative keratolysis and perforation. Severe keratitis due to _H. influenzae_ or _P. aeruginosa_ in young children

should be treated with parenteral, as well as local antibiotics to reduce the risk of systemic spread.100 Parenteral antibiotics are indicated in severe keratitis with impending perforation,

perforated infections with potential for intraocular spread, following perforating injuries to the corneosclera, or with contiguous scleral involvement. ANTIBIOTIC THERAPY The objective for

initial antibiotic selection in therapy for bacterial keratitis is rapid elimination of the corneal pathogen(s). No single topical antibiotic is effective against all potential organisms

causing bacterial keratitis. Thus, selection of an antimicrobial agent having a broad spectrum of activity, including the most likely Gram-positive and Gram-negative corneal pathogens, is

desirable. Historically, a combination of two compounds with one directed against the Gram-positive pathogens and the other against the Gram-negative organisms was considered a rational

initial treatment choice.3 Although penicillin-G is superior in activity against _S. pneumoniae_ and other Streptococci, the frequency of penicillinase-producing Staphylococci and other

organisms requires a penicillinase-resistant agent. Cephalosporins evolved as the drug of choice against unidentified Gram-positive cocci. Cefazolin (50 mg/ml) is well tolerated by the

ocular surface with topical and subconjunctival routes of administration. Additionally, it can be employed in therapy of selected patients having a prior history of allergy to penicillin.

Aminoglycosides were the preferred initial antibiotic choice for therapy of suspected Gram-negative keratitis. Gentamicin evolved as the initial aminoglycoside agent of choice owing to its

favourable pharmacokinetics and excellent activity against _Pseudomonas, Klebsiella, Enterobacter_, and other Gram-negative species. With the emergence of gentamicin-resistant strains of _P.

aeruginosa_, tobramycin has become an alternate initial choice. Approximately 10% of corneal isolates of _Pseudomonas_ may be resistant to aminoglycosides, with some strains being resistant

to gentamicin, but sensitive to tobramycin.101 In addition to coverage of Gram-positive cocci, the cephalosporin may provide some activity against Gram-negative rods. The third-generation

cephalosporins provide greater Gram-negative coverage, yet have less Gram-positive activity than first-generation cephalosporins. Combined therapy with topical fortified cefazolin (or other

cephalosporin) and tobramycin (or gentamicin) became the rational initial therapeutic recommendation for polybacterial keratitis or when results of Gram stain were equivocal.3

Aminoglycosides, especially tobramycin, have considerable toxicity when administered topically at frequent intervals for a prolonged period. The fluoroquinolone antibiotics have potent

bactericidal activity against the broad spectrum of Gram-negative aerobic bacteria and many Gram-positive bacteria, including penicillinase-producing and methicillin-resistant Staphylococci.

The fluoroquinolones have now been shown in two separate independent trials to provide a safe and effective therapy for acute bacterial keratitis when compared with combination fortified

antibiotic treatment.102,103 Penicillin-G at a concentration of 100 000 U/ml remains the initial drug of choice for keratitis caused by Streptococcal species, including _S. pneumoniae_. Some

penicillin-resistant pneumococcal strains have recently been identified. Penicillinase-producing strains of _N. gonorrhoeae_ have been introduced and observed in ocular infections. Owing to

a relatively high prevalance of _N. gonorrhoeae_ resistant to penicillin, alternate drug choices, such as ceftriaxone, should be selected. If Gram-positive branching filaments are observed

on microscopic examination of smears from corneal scrapings, penicillin may be an appropriate choice for the possible Actinomycetales organism pending final identification. MODIFICATION OF

THERAPY The results of microbial culture and antimicrobial susceptibility testing data may suggest a modification from the initial therapeutic plan. If the identified organism is likely to

be significantly more susceptible to an antibiotic other than the one originally selected, or if antibacterial susceptibility testing confirms resistance, the more effective drug may be

substituted. If combination broad-spectrum therapy was initially selected and a single organism is isolated, the less effective agent may be discontinued. If microbial culture fails to grow

a causative organism, clinical judgement must be exercised to guide further antimicrobial therapy. Any modification of therapy should be based on the clinical response and patient tolerance

to initial therapy, as well as the laboratory data. The initial judgement of clinical severity also impacts decisions to modify therapy. If there is substantial clinical improvement on the

initial antibiotic course, there may be no advantage to selection of an alternative agent. For anaerobic organisms, including Gram-positive cocci or Gram-negative rods, penicillin remains

the preferred agent. Aerobic and anaerobic Gram-positive rods may also respond to penicillin therapy. A combination of trimethoprim (16 mg/ml) and sulphamethoxazole (80 mg/ml) eyedrops is an

effective treatment for _Nocardia_ keratitis.104 Sulphonamides, tetracyclines, erythromycin, or amikacin may be alternate agents. The recommended initial therapy of acid-fast-stained

positive corneal scrapings in suspected microbial keratitis is topical amikacin 10–20 mg/ml, one drop every half hour as indicated.105 Subconjunctival amikacin 20 mg in 0.5 ml may be

adjunctively administered. Systemic amikacin is not routinely utilized, but may be added in selected cases with corneal perforation or extension of infection to involve the sclera. Other

antibiotics reported to be effective against a significant number of isolates of _M. fortiutum_ and _M. chelonae_ include cefoxitine, ciprofloxacin, doxycycline, erythromycin, imipenem,

kanamycin, netilmicin, ofloxacin, and tobramycin. Sulphamethoxazole may sometimes have significant activity. The fluoroquinolones have been shown to be active against the most important

nontuberculous mycobacteria causing keratitis, including some species that are highly resistant to standard antituberculous drugs. _M. fortuitum_ is highly susceptible to both ciprofloxacin

and ofloxacin.106,107 The minimal inhibitory concentrations of ciprofloxacin have varied from 0.01 to 12.5 _μ_g/ml. The MICs of ofloxacin have ranged from 0.03 to 1.2 _μ_g/ml and no

resistant strains have been identified. Experimental animal models and isolated clinical reports108 have suggested that topical ciprofloxacin alone or in combination with topical amikacin

may be useful in the treatment of _M. fortuitum_ keratitis. Experimental studies assessing the _in vivo_ efficacy of newer macrolides, including clarithromycin and azithromycin,109,110

suggest a potential role in the therapy of _M. fortuitum_ keratitis. The therapy of nontuberculous mycobacterial keratitis is complicated by the relatively poor correlation between _in

vitro_ susceptibility profiles and clinical behaviour. Following initial response, the infection may worsen, even if the pathogen remains ‘sensitive’ by _in vitro_ susceptibility testing.

The response of bacterial keratitis to antibiotic therapy must be monitored with frequent clinical observation. Slit-lamp biomicroscopic examination twice daily on hospitalized patients

should be performed. The response to therapy may be difficult to appreciate with assessment within the first few days owing to organism and host factors increasing inflammation, and the

reaction to corneal scrapings, as well as frequent local antibiotics. Within 48–72 h of effective bactericidal therapy, the progression of keratitis is halted. The stromal infiltrate

consolidates, the anterior chamber inflammation subsides, and the epithelial and stromal healing begins. Clinical response varies depending on the responsible pathogen, duration of

infection, antecedent factors, and host response. After 36–48 h, the frequency of antibiotic administration can be tapered. Caution with cessation of therapy is recommended for organisms

that may persist in corneal tissue, such as _P. aeruginosa_, and prolonged therapy may be required. The principal end points of treatment are re-epithelialization and nonprogression of

stromal infiltrates. The patient should be monitored for tolerance of antimicrobial agents, as toxicity may retard epithelial and stromal healing. In addition, resistant organisms may fail

to respond appropriately to topical therapy. ADJUNCTIVE THERAPY Owing to the rich innervation of the cornea, ulcerative keratitis is frequently accompanied by significant pain. Pain control

with acetinominophen or other analgesics may result in improved patient comfort and more effective delivery of the treatment regimen. Topical cycloplegic agents should be administered to

relieve ciliary spasm, alleviate pain, and to prevent the formation of synechiae. Topical 0.25% scopolamine or homatropine 5% used t.i.d. to q.i.d. is usually adequate. Significant

intraocular inflammation may result in a secondary glaucoma. Elevated intraocular pressure should be monitored and treated with topical beta blocker or topical or oral carbonic anhydrase

inhibitors as required for control. Most patients can be effectively managed as outpatients, if there is an adequate support system to allow compliance with the treatment plan. Patching

should be avoided in the initial therapy of bacterial keratitis, as this may result in a microenvironment favourable to accelerated organism replication. Following eradication of the

causative bacteria, patching may be applied to assist re-epithelialization. Therapeutic soft contact lenses may be a useful adjunct to assist epithelial healing. Antibiotic administration

should continue over the therapeutic soft contact lens. Caution should be exercised as infection may occasionally complicate therapeutic soft contact lens usage.111 The therapeutic lens may

provide some tectonic support with impending or microscopic corneal perforation. One of the principal objectives of therapy for bacterial keratitis is to prevent tissue destruction and

irreversible structural alterations. A number of adjunctive modalities have been suggested to reduce the destructive effects of various enzymes released with progressive bacterial keratitis.

Injured corneal epithelium, infiltrating neutrophils, and some bacterial organisms elaborate various enzymes that contribute to stromal keratolysis. Collagenase is also produced by host

corneal tissue. Enzyme inhibitors, including disodium edetate (EDTA 0.05 M), or acetylcysteine (Mucomyst, 20%) or Heparin 2% have been shown to be effective experimentally.112,113,114 Matrix

metalloproteinases (MMPs) play an important role in corneal wound healing and in pathologic conditions. The activity of these proteases is regulated by the presence of tissue inhibitors of

metalloproteinases. Infectious ulcerative keratitis may be modified by controlling the MMP activity in the tissue. MMP-2 is a normal constituent of the corneal stroma, but MMP-9 is found

under pathologic conditions and after corneal wounding. Both MMP-2 and MMP-9 have the capacity to degrade basement membrane collagens (types IV and VII). TGF-beta augments the expression of

MMP-9 in corneal tissue. Synthetic inhibitors of matrix metalloproteinase have been shown to inhibit proteases from _P. aeruginosa_ _in vitro_ and to prevent experimental _Pseudomonas_

keratitis.115 In addition, synthetic metalloproteinase inhibition may potentiate the antiangiogenic effect of high-dose prednisolone for inhibition of endotoxin-induced corneal

neovascularization in rabbits.116 Clinical trials assessing the potential adjunctive role of synthetic matrix metalloproteinase inhibitors are nearing completion. The precise role and the

timing of adjunctive topical corticosteroid usage in the therapy of bacterial keratitis has remained controversial. Reduction in the host inflammatory response which may contribute to

corneal destruction provides the rationale for corticosteroids usage. Corticosteroids effectively decrease the host inflammatory response initiated by bacterial exo- or endotoxins and lytic

enzymes released from polymorphonuclear neutrophils. Several experimental studies have failed to demonstrate any deleterious effect from the addition of topical corticosteroids with

concomitant bactericidal therapy for bacterial keratitis.117,118,119 Microbial cultures from patients with keratitis after 3 days of gentamicin therapy grew more bacteria in individuals

receiving topical corticosteroids, with a longer treatment course required to eradicate the infection using this combined treatment.120 Combined antibiotic therapy with adjunctive

corticosteroids did not improve the clinical course of experimental _Pseudomonas_ keratitis compared with treatment with antibiotic alone regardless of the timing or strength of

corticosteroid therapy.121 Enhancement or recurrence of _Pseudomonas_ keratitis has been observed with concomitant corticosteroid and antibiotic treatment.122,123 Prior use of

corticosteroids may mask the clinical signs of stromal and intraocular inflammation, such that bacterial invasion may not be detected. Antecedent use establishes the potential for severe

rebound stromal inflammation with keratolysis after withdrawal of corticosteroids during the initial management of infection. Severe keratitis with marked stromal thinning may accelerate to

corneal perforation with corticosteroid usage. Corticosteroids may have a limited role in the therapy of bacterial keratitis to suppress the deleterious effects of inflammation once

effective bactericidal therapy has eliminated or reduced the pathogen(s). With Gram-positive keratitis, judicious topical application of corticosteroid may be initiated after several days of

intensive specific antimicrobial therapy. In confirmed Gram-negative infection, or if there is doubt regarding a possible Gram-negative coinfection, corticosteroid therapy should be

deferred for a longer period of aggressive specific topical antibiotic therapy. Some clinicians recommend withholding steroids if clinical improvement is progressing without steroids, using

steroids only after several days of antibiotics if there is persistent inflammation that does not seem to improve, concomitant antibiotic coverage when steroids are used, avoidance of

steroids in thin corneas with threatening perforation, and use of steroids for 24 h prior to a therapeutic penetrating keratoplasty for active bacterial keratitis.124 Topical corticosteroids

may be administered in low concentrations and infrequent dosing for 1–2 days on a trial basis to monitor for adverse effects and increased as indicated. Long-acting periocular

corticosteroids and oral corticosteroids should be avoided in bacterial keratitis. Cryotherapy has been applied in experimental animal models with demonstration of bactericidal effects.125

Cryotherapy may be useful in select cases of focal peripheral corneal ulcerations or in _Pseudomonas_ sclerokeratitis.126,127 Caution should be exercised with the administration of

cryotherapy, as severe toxic effects may be additive to the keratitis process. The precise role for adjunctive topical nonsteroidal therapy for bacterial keratitis is not determined.

Diclofenac sodium 0.1% therapy did not adversely effect the results of antibiotic therapy with gentamicin 0.3% in the treatment of experimental _Pseudomonas_ keratitis.128 In an experimental

model of _Pseudomonas_ keratitis in rabbits, recurrence was observed in 85% of steroid-treated rabbits _vs_ 12.5% of flurbiprofen-treated rabbits.129 Conversely, in a _S. pneumoniae_ model,

there were no recurrences experienced, either with steroids or nonsteroidal topical treatment.130 The application of tissue adhesive (isobutyl cyanoacrylate) or other analogues has been

recommended in progressive stromal keratolysis, with thinned Descemetoceles, or small perforated infectious ulcerations. The tissue adhesive may have some inherent antibacterial activity.130

It is toxic to corneal endothelium and thus should be applied only for small perforations. Tissue glue is useful to restore the integrity of the anterior segment and to postpone the need

for surgery until antibiotic and anti-inflammatory therapy have reduced the ocular inflammation. The edges and bed of the ulceration should be debrided and dried with methylcellulose sponges

prior to the application of tissue adhesive. A therapeutic soft contact lens should be placed over the tissue adhesive to prevent irritation and to protect the glue from mechanical effects

of the eyelids. Tissue toxicity, microbial colonization, use of therapeutic soft contact lenses, and long-term broad-spectrum antibiotics may precipitate tissue adhesive-related microbial

keratitis.131 Masking of the underlying infection and the development of resistant organisms should be considered when using tissue adhesive. Excimer laser photoablative treatment of

microbial keratitis has been investigated in experimental animal models.132,133 Results of these investigations indicate that advanced stromal keratitis with deep suppuration cannot be

eradicated using the excimer laser. As corneas may be perforated inadvertently during treatment, excimer laser therapy of infectious keratitis should be approached with caution and used only

for very select superficial and well-circumscribed lesions. Carbon dioxide laser therapy has also been investigated in experimental _P. aeruginosa_ keratitis.134 Corneal patch grafting may

be an alternative to the application of tissue adhesives for small corneal perforations resulting from bacterial keratitis. Small partial conjunctival flaps may be used in peripheral

ulcerations to assist with healing, but are not recommended for use in impending or perforated central bacterial keratitis. If there is a large perforation or a residual necrotic cornea, a

therapeutic penetrating keratoplasty may be indicated.135 Maximal antibiotic therapy to eradicate the corneal pathogen(s) and to reduce inflammation are recommended prior to surgery. In

addition to topical intensive antibiotic treatment, parenteral antibiotic therapy should be instituted in the perioperative period. The surgeon must select a large trephine to completely

excise the area of infection. A free-hand dissection may be required in limbal to limbal suppurative keratitis. Sclerocorneal grafting may be performed, although with a limited postoperative

success because of glaucoma, rejection, and other problems. An oversized graft (0.50–0.75 mm) is recommended. A portion of the corneal button should be excised and placed in tissue media

for grinding and culture of the tissue homogenate. The remainder of the specimen should be submitted for histopathologic study, including special stains for light microscopy and electron

microscopy where indicated. Placement of interrupted sutures is recommended over continuous running sutures, given the intense inflammatory reaction. Aggressive topical corticosteroids

should be applied in the postoperative period along with concomitant antibiotic therapy as indicated. Parenteral therapy should be continued if there is suspicion for intraocular

dissemination. Therapeutic penetrating keratoplasty was successful in restoring anatomic and visual results in 75% of grafts performed for bacterial keratitis in one study.136 Complicated

cataract may result from severe bacterial keratitis.137 Cataract may result from bacterial toxins, iridocyclitis, and treatment toxicity. Cataract formation may result from severe bacterial

keratitis alone, but is probably enhanced by concurrent treatment with high-dose topical corticosteroids. Surgical rehabilitation may require combined cataract extraction with penetrating

keratoplasty, depending on the degree of corneal scarring and opacification with cataract formation. THE NEW FRONTIER: MICROBIAL GENOMICS AND PROTEOMICS While frequent administration of

topically applied antimicrobial agents has served patients and ophthalmic clinicians well in therapy of severe infectious keratitis, newer tools are required to enhance our understanding of

the molecular pathogenesis of infections involving the cornea. Advances in microbial genomics and proteomics offer the exciting potential to develop new therapeutic targets involved in the

cascade of events that is triggered by onset of bacterial keratitis. Such advances shall hopefully elevate treating clinicians beyond the ‘exorcist approach’ to a more sophisticated strategy

of addressing not only the invading organism(s) but the host response as well. The availability of complete genome sequences has dramatically changed the opportunities for developing novel

and improved therapies as well as preventive strategies, including vaccines.138 Complete genomic databases provide an inclusive catalogue of all potential candidate vaccines for any

bacterial pathogen. In conjunction with adjunct technologies, including bioinformatics, random mutagenesis, microarrays, and proteomics, a systematic and comprehensive approach to

identifying vaccine discovery can be undertaken. Genomics must be used in conjunction with population biology to ensure that the vaccine can target all pathogenic strains of a species. A

proof in principle of the utility of genomics is provided by the recent exploitation of the complete genome sequence of _N. meningitidis_ group B.138 Compounding the threat from newer

pathogens is the rapid evolution of drug resistance by microorganisms that is rendering many existing antimicrobial agents obsolete. Thus, there is an urgent need for the development of new

classes of antimicrobial agents and the identification of new drug targets. In the last two decades, the search for completely novel antibacterial agents has acquired a new sense of urgency

because of the remarkable rise of antibiotic resistance among key bacterial pathogens. More recently, the advent of bacterial genomics has provided investigators with the data and

bioinformatic tools to identify rationally novel antibacterial targets and the genome-scaled methodologies to validate them.139 Less than a decade has elapsed since the publication of the

first complete bacterial genome sequence, but more than 50 complete microbial genome sequences are now available. Advances in high-throughput automated DNA sequencing have delivered a wealth

of genetic information in the form of whole-genome sequences of microbial pathogens. Coupled with this advancement has been the development of new genetic tools and computational advances

capable of selecting genes of particular interest as well as testing for the effects of candidate drugs. The existing bacterial genome data set provides both advantages and limitations for

the rational identification of novel antibacterial targets.139 Since the ability to identify rapidly essential genes where loss of function is coincident with loss of viability is the most

important task of genomics-based target validation, essential testing methodologies (in which molecular genetic techniques are used to determine whether or not a gene product is required for

viability of the parent cell) are critical and their amenability to genome-scaled analysis exceptionally important. For example, the emergence and dissemination of drug-resistant

pneumococcal strains, coupled to changing patterns of virulence and the inadequacy of available vaccines, calls for an aggressive search for novel targets for antibiotic and vaccine

development.140 Microbial genomics techniques allow genetic and biochemical tools to be employed to tackle discovery, design, and development of new anti-infective agents based on the

identification of hundreds of new targets. Novel approaches have been employed to identify potential antibiotic and vaccine targets in _S. pneumoniae_. Recently identified virulence factors,

as well as molecules essential for bacterial viability, cell wall integrity, and infectivity, have been explored as new potential targets in the battle against pneumococci.140 Structural

genomics stands out among the emerging fields of proteomics since it influences the drug discovery process at so many points. Recent developments in protein expression technologies, X-ray

crystallography, and NMR spectroscopy provide the essential elements for high-throughput structure determination platforms.141,142 Bioinformatics methods to interrogate the resulting data

will provide comprehensive, genome-wide databases of protein structure. Genomic sequencing and methods for high-throughput expression and protein purification are furthest advanced for

microbial genes and so these have been the early targets for structural genomics initiatives.142 The information will be invaluable in understanding gene function, designing broad-spectrum

small molecule inhibitors and in better understanding drug–host interactions. The process of prokaryotic drug discovery has been highly successful for over more than half a century, yet the

number of exploited bacterial targets is a mere fraction, less than 0.1% of the potential targets (based on total number of bacterial genes identified by gene sequence projects).143 To

better understand the potential for drug intervention, multiple paradigms have been established in the pharmaceutical industry, all with some semblance of commonality and uniqueness to

provide proprietary positioning, yet no company has been successful to date in taking a genomics approach to the finish line of having a genomics-based drug on the market. The whole

bacterial genome sequence data in itself is merely a starting point for drug discovery of novel antibacterial targets and, eventually, drugs. In order to leverage this large amount of data,

it is necessary to match an understanding of the microbial physiology of pathogenic bacteria to disease processes and identify the gene products for which intervention may reduce or

eliminate the infectious state. Many new molecular-based technologies (proteomics, transcriptional profiling, studies of gene expression _in vivo_) have originated or have expanded into

wider use, and have been made possible by the availability of complete bacterial genome sequence information and subsequent bioinformatic analytic tools. Proteomics has begun to provide

insight into the biology of microorganisms. The combination of proteomics with genetics, molecular biology, protein biochemistry, and biophysics is particularly powerful, resulting in novel

methods to analyse complex protein mixtures. Emerging proteomic technologies promise to increase the throughput of protein identifications from complex mixtures and allow for the

quantification of protein expression levels. The techniques of proteomics (high-resolution two-dimensional electrophoresis and protein characterization) are widely used for microbiological

research to analyse global protein synthesis as an indicator of gene expression.144 The rapid progress in microbial proteomics has been achieved through the wide availability of whole-genome

sequences for a number of bacterial groups. Beyond providing a basic understanding of microbial gene expression, proteomics has also played a role in medical areas of microbiology. Progress

has been made in the use of the techniques for investigating the epidemiology and taxonomy of human microbial pathogens, the identification of novel pathogenic mechanisms, and the analysis

of drug resistance. In each of these areas, proteomics has provided new insights that complement genomic-based investigations. A major challenge ahead in proteomics is the analysis of

genetically heterogeneous bacterial populations and the integration of the proteomic and genomic data for these bacteria. The characterization of the proteomes of bacterial pathogens growing

in their natural hosts remains a future challenge. As genomics is highly promising for vaccine development, the postgenomic era holds the even greater possibility of rational design of

novel vaccines for important human pathogens.145,146 The discovery and development of these new vaccines is likely to be accomplished through integrated proteomic strategies. Although most

proteomic studies are based on two-dimensional gel electrophoresis (2D-PAGE) as a separation technique, new methods have been developed within the past 2 years that provide complementary

information concerning microbial protein expression. The 2D-PAGE technique in combination with Western blotting has been successfully applied in the discovery of antigens from _Helicobacter

pylori_, _Chlamydia trachomatis_, and _Borrelia garinii_.146 Two-dimensional semipreparative electrophoresis has provided complementary information regarding membrane protein expression in a

strain of _H. pylori_. Through two-dimensional liquid chromatography–tandem mass spectrometry, the most comprehensive information to date regarding protein expression in yeast was obtained.

This technique may become an important tool in vaccinology. In the field of infectious diseases, there is an urgent need for global approaches that can efficiently, precisely, and

integratively study structural and functional genomics and proteomics of microbial infections (infectomics).147 The combination of new (eg DNA and protein microarrays) and traditional

approaches (eg cloning, PCR, gene knockout and knockin, and antisense) will help overcome the challenges we are facing today. It is possible that the global phenotypic changes (infectomes)

in microbes and their host during infections are encoded by the genomes of microbial pathogens and their hosts.147 These phenotypical changes are expressed in certain environmental

conditions, such as the corneal and ocular surface milieu, devoted to specific microbe–host interactions. Global drug responses (pharmacomes) in microbes and their host can be detected by

genomic and proteomic approaches. Genome-wide approaches to genotyping and phenotyping or expression profiling will eventually lead to global dissection of microbial pathogenesis, efficient

and rapid diagnosis of infectious diseases, and the development of novel strategies to control infections. The key fundamental issue of infectious diseases is how to understand the

interactions between microbial pathogens and their hosts by using infectomics globally and integratively. The elucidation of the function of new genes is not a simple matter and human

genetic studies have been among the most useful in identifying genes implicated in disease, although homologues, or orthologues, of human genes in mouse, worm, fly, and yeast genomes have

also been useful in deciphering gene function.148 The advent of genomics has led into the study of protein structure and function under the rubric of proteomics. Proteins function in a cell

mostly by interacting with other proteins, and protein interaction maps of cellular circuits are now available. Screening strategies to address protein–protein interactions are being

developed, and many drug targets in the future will be directed towards these interactions. In addition to using genetic and analytical approaches for finding new drug targets, chemical

libraries can be used to inhibit the activity of new proteins, and thus reveal function. The combination of high-throughput screening with testing compounds for ADME (absorption,

distribution, metabolism, and excretion) and toxicity will help in the early clarification of clinical utility of both new drug targets and drug candidates. Bioinformatics has, out of

necessity, become a key aspect of drug discovery in the genomic revolution, contributing to both target discovery and target validation.149 Bioinformatics will play an increasingly important

role in response to the waves of genome-wide data sources that have become available to the industry, including expressed sequence tags, microbial genome sequences, model organism

sequences, polymorphisms, gene expression data, and proteomics. However, these knowledge sources must be intelligently integrated. The future for application of advances in microbial

genomics and proteomics towards more sophisticated management strategies for bacterial infections of the cornea beyond the ‘exorcist approach’ appears very bright indeed. REFERENCES *

O'Brien TP, Green WR . Keratitis. In: Mandell GL _et al_ (eds). _Principles and Practice of Infectious Diseases_, Chap 93, 4th ed. Churchill Livingstone: New York, 1995, pp 1110–1118.

Google Scholar  * Jones DB . A plan for antimicrobial therapy in bacterial keratitis. _Trans Am Acad Ophthalmol Otolaryngol_ 1975; 79: 95. Google Scholar  * Jones DB . Initial therapy of

suspected microbial corneal ulcers II. Specific antibiotic therapy based on corneal smears. _Surv Ophthalmol_ 1979; 24: 97. CAS  PubMed  Google Scholar  * O'Brien TP, Hazlett LD .

Pathogenesis of ocular infection. In: Pepose JS, Holland GN, Wilhelmus KR (eds). _Ocular Infection and Immunity_, Chap 15. Mosby: St Louis, MO, 1995. Google Scholar  * Jones DB .

Pathogenesis of bacterial and fungal keratitis. _Trans Ophthalmol Soc UK_ 1978; 98: 367. CAS  Google Scholar  * Baker NR . _Pseudomonas aeruginosa_ exoenzyme S is an adhesin. _Infect Immun_

1991; 59(9): 2859. CAS  PubMed  PubMed Central  Google Scholar  * Hoepelman AIM, Tuomanen EL . Consequences of microbial attachment: directing host cell functions with adhesins. _Infect

Immun_ 1992; 60: 1729. CAS  PubMed  PubMed Central  Google Scholar  * Reichert R, Stern GA . Quantitative adherence of bacteria to human corneal epithelial cells. _Arch Ophthalmol_ 1984;

102: 1394. CAS  PubMed  Google Scholar  * Doig P _et al_. Role of pili in the adhesion of _Pseudomonas aeruginosa_ to human respiratory epithelial cells. _Infect Immun_ 1988; 56: 1641. CAS 

PubMed  PubMed Central  Google Scholar  * Holder IA, Naglich JG . Experimental studies of the pathogenesis of infection due to _pseudomonas aeruginosa_ infection. _J Trauma_ 1986; 26: 118.

CAS  PubMed  Google Scholar  * Nicas TI, Iglewski BH . Toxins and virulence factors of _Pseudomonas aeruginosa_. In: Sokatch JR (ed). _The Bacteria. Vol X. The Biology of Pseudomonas_.

Academic Press: New York, 1986. Google Scholar  * Iglewski BH, Liu PV, Kabat D . Mechanism of action of _Pseudomonas aeruginosa_ exotoxin A: adenosine diphosphate-ribosylation of mammalian

elongation factor II _in vitro_ and _in vivo_. _Infect Immun_ 1977; 15: 138. CAS  PubMed  PubMed Central  Google Scholar  * Cryz SJ _et al_. Role of lipopolysaccharide and virulence of

_Pseudomonas aeruginosa_. _Infect Immun_ 1984; 44: 508. CAS  PubMed  PubMed Central  Google Scholar  * Costerton JW . _Pseudomonas aeruginosa_ in nature and diseases. In: Sabath LD (ed).

_Pseudomonas aeruginosa: The Organisms, The Disease it Causes and their treatment_. Hans Huber: Bern, 1980. Google Scholar  * Patti JM, Allen PL, McGavin MJ _et al_. MSCRAMM-mediated

adherence of microorganisms to host tissues. _Annu Rev Microbiol_ 1994; 48: 585. CAS  PubMed  Google Scholar  * Woods DE _et al_. Role of fibronectin in the prevention of adherence of

_Pseudomonas aeruginosa_ to buccal cells. _J Infect Dis_ 1981; 143: 784. CAS  PubMed  Google Scholar  * Hazlett LD _et al_. Evidence for _N_-acetylmanosomine as an ocular receptor for _P.

aeruginosa_ adherence to scarified cornea. _Invest Ophthalmol Vis Sci_ 1987; 28: 1978. CAS  PubMed  Google Scholar  * Stern GA, Weitzenkorn D, Valenti J . Adherence of _Pseudomonas

aeruginosa_ to the mouse cornea: epithelial v. stromal adherence. _Arch Ophthalmol_ 1982; 100: 1956. CAS  PubMed  Google Scholar  * Hazlett LD, Moon MM, Berk R . _In vivo_ identification of

sialic acid as the ocular receptor for _Pseudomonas aeruginosa_. _Infect Immun_ 1986; 51: 687. CAS  PubMed  PubMed Central  Google Scholar  * Hazlett LD _et al_. _Pseudomonas aeruginosa_

pili mediate binding to human corneal epithelial glycoproteins. _10th International Congress Eye Research_, Stresa, Italy, 1992. Google Scholar  * Rudner XL _et al_. Corneal epithelial

glycoproteins exhibit _Pseudomonas aeruginosa_ pilus binding activity. _Invest Ophthalmol Vis Sci_ 1992; 33: 2185. CAS  PubMed  Google Scholar  * Liu X, Hazlett LD, Berk RS . Systemic and

topical protection studies using _Pseudomonas_ flagella or pili. _Invest Ophthalmol Vis Sci_ 1990; 31: 449. Google Scholar  * Rudner XL _et al_. _Pseudomonas aeruginosa_ pili interact

differently with glycosylated proteins of immature cornea. _Invest Ophthalmol Vis Sci_ 1992; 33: 844. Google Scholar  * Lee KK _et al_. Mapping the surface regions of _Pseudomonas

aeruginosa_ PAK pilin: the importance of the C-terminal region of adherence to human buccal epithelial cells. _Mol Biol_ 1989; 3: 1493–1499. CAS  Google Scholar  * Saiman L, Ishimoto K, Lory

S, Prince A . The effect of piliation and exoproduct expression on the adherence of _Pseudomonas aeruginosa_ to receptor monolayers. _J Infect Dis_ 1990; 161: 541. CAS  PubMed  Google

Scholar  * Baum J, Panjwani NJ . Adherence of _Pseudomonas_ to soft contact lenses and cornea: mechanisms and prophylaxis. In: Cavanaugh HD (ed). _The Cornea: Transaction of the World

Congress on the Cornea_, Vol III. Raven Press: New York, 1988. Google Scholar  * Drake D, Montie TC . Flagella, motility, and invasion virulence of _Pseudomonas aeruginosa_. _J Gen

Microbiol_ 1988; 134: 43–52. CAS  PubMed  Google Scholar  * Ansorg RA _et al_. Differentiation of the major flagellar antigens of _Pseudomonas aeruginosa_ by the slide coagulation technique.

_J Clin Microbiol_ 1984; 20: 84. CAS  PubMed  PubMed Central  Google Scholar  * Montie TC _et al_. Motility, virulence, and protection with the flagella vaccine against _Pseudomonas

aeruginosa_ infection. _Antibiot Chemother_ 1987; 39: 233. CAS  PubMed  Google Scholar  * Rudner XL, Hazlett LD, Berk RS . Systemic and topical protection studies using _Pseudomonas

aeruginosa_ flagella in an ocular model of infection. _Curr Eye Res_ 1992; 11: 727. CAS  PubMed  Google Scholar  * Hyndiuk RA . Experimental _Pseudomonas_ keratitis: I. Sequential electron

microscopy. II. Comparative therapy trials. _Trans Am Ophthalmol Soc_ 1981; 79: 541. CAS  PubMed  Google Scholar  * Koch JM _et al_. Experimental _Pseudomonas aeruginosa_ keratitis from

extended wear of soft contact lenses. _Arch Ophthalmol_ 1990; 108: 1453. CAS  PubMed  Google Scholar  * Cotran RS . New roles for the endothelium and inflammation and immunity. _Am J Pathol_

1987; 29: 407. Google Scholar  * Fajardo LF . The complexity of endothelial cells. A review. _Am J Clin Pathol_ 1989; 92: 241. CAS  PubMed  Google Scholar  * Staugas REM _et al_. Induction

of tumor necrosis factor (TNF) and interleukin-I (IL-I) by _Pseudomonas aeruginosa_ and exotoxin A-induced suppression of lymphoproliferation and TNF, lymphotoxin, gammainterferon, and IL-I

production in human leukocytes. _Infect Immun_ 1992; 60: 3162. CAS  PubMed  PubMed Central  Google Scholar  * Berk RS, Beisel K, Hazlett LD . Genetic studies on the murine corneal response

to _P. Aeruginosa_. _Proc Soc Exp Biol Med_ 1983; 172: 488. PubMed  Google Scholar  * Hazlett LD, Rosen D, Berk RS . Experimental _Pseudomonas_ eye infection in cyclophosphamide-treated

mice. _Invest Ophthalmol Vis Sci_ 1977; 16: 649. CAS  PubMed  Google Scholar  * Twining S, Lohr KM, Moulder JE . The immune system in the experimental keratitis. A model and early effects.

_Invest Ophthalmol Vis Sci_ 1986; 27: 507. CAS  PubMed  Google Scholar  * Cleveland RP _et al_. The role of complement in _Pseudomonas_ ocular infections. _Invest Ophthalmol Vis Sci_ 1983;

24: 237. CAS  PubMed  Google Scholar  * Hazlett LD . Pathogenic mechanisms of _P. aeruginosa_ keratitis: a review of the role of T cells, Langerhans cells, PMN, and cytokines. _DNA Cell

Biol_ 2002; 21(5–6): 383–390. CAS  PubMed  Google Scholar  * Hazlett LD, McClellan SA, Rudner XL, Barrett RP . The role of Langerhans cells in _Pseudomonas aeruginosa_ infection. _Invest

Ophthalmol Vis Sci_ 2002; 43(1): 189–197. PubMed  Google Scholar  * Kernacki KA, Barrett RP, McClellan S, Hazlett LD . MIP-1alpha regulates CD4+T cell chemotaxis and indirectly enhances PMN

persistence in _Pseudomonas aeruginosa_ corneal infection. _J Leukoc Biol_ 2001; 70(6): 911–919. CAS  PubMed  Google Scholar  * Thakur A, Xue ML, Wang W, Lloyd A, Wakefield D, Willcox MD .

Expression of macrophage migration inhibitory factor during _Pseudomonas_ keratitis. _Clin Exp Ophthalmol_ 2001; 29(3): 179–182. CAS  PubMed  Google Scholar  * Hazlett LD, McClellan S,

Barrett R, Rudner X . B7/CD28 costimulation is critical in susceptibility to _Pseudomonas aeruginosa_ corneal infection: a comparative study using monoclonal antibody blockade and

CD28-deficient mice._J Immunol_ 2001; 166(2): 1292–1299. CAS  PubMed  Google Scholar  * Hazlett LD, Rudner XL, McClellan SA, Barrett RP, Lighvani S . Role of IL-12 and IFN-gamma in

_Pseudomonas aeruginosa_ corneal infection. _Invest Ophthalmol Vis Sci_ 2002; 43(2): 419–424. PubMed  Google Scholar  * Steuhl KP _et al_. The effect of immunization on corneal infection by

_Pseudomonas aeruginosa_. _Invest Ophthalmol Vis Sci_ 1987; 28: 1559. CAS  PubMed  Google Scholar  * Hibbs MS _et al_. Gelatinase (type IV collagenase) immunolocalization in cells and

tissues: use of an anti-serum to rabbit bone gelatinase that identifies high and low Mr forms. _J Cell Sci_ 1989; 92: 487. Google Scholar  * Fini ME, Jirard MT . Expression of

collagenolytic/gelatinolytic metalloproteinase by normal cornea. _Invest Ophthalmol Vis Sci_ 1990; 32: 12779. Google Scholar  * Kirschner SE, Twining SS . The effect of _Pseudomonas

aeruginosa_ exoproducts on corneal proteases. _Invest Ophthalmol Vis Sci_ 1990; 31 (Suppl): 487. Google Scholar  * Callegan MC, Engel LS, Hill JM, O'Callaghan RJ . Corneal virulence of

_Staphylococcus aureus_: roles of alpha-toxin and protein A in pathogenesis. _Infect Immun_ 1994; 62(6): 2478. CAS  PubMed  PubMed Central  Google Scholar  * Howe TR, Iglewski BH : Isolation

and characterization of alkaline protease-deficient mutants of _Pseudomonas aeruginosa_ _in vitro_ and _in vivo_. _Infect Immun_ 1984; 3: 1058. Google Scholar  * Ohman DE, Barns RP,

Iglewski BH . Corneal infection in mice with toxin-A and elastase mutants of _Pseudomonas aeruginosa_. _J Infect Dis_ 1980; 142: 547. CAS  PubMed  Google Scholar  * Burns FR, Gray RP,

Patterson CA . Inhibition of alkali-induced corneal ulceration and perforation by a thio peptide. _Invest Ophthalmol Vis Sci_ 1990; 31: 107. CAS  PubMed  Google Scholar  * Heck LW, Morihara

K, Abrahamson DR . Degradation of soluble laminin and depletion of tissue-associated basement laminin by _Pseudomonas aeruginosa_ elastase and alkaline protease. _Infect Immun_ 1986; 54:

149. CAS  PubMed  PubMed Central  Google Scholar  * Twining SS, Davis SD, Hyndiuk RA . Relationship between proteases and descemetocele formation in experimental _Pseudomonas_ keratitis.

_Curr Eye Res_ 1986; 5: 503. CAS  PubMed  Google Scholar  * Bejarano PA _et al_. Degradation of basement membranes by _Pseudomonas aeruginosa elastase_. _Infect Immun_ 1989; 57: 3783. CAS 

PubMed  PubMed Central  Google Scholar  * Kharazami A . Mechanisms involved in the evasion of host defense by _Pseudomonas aeruginosa_. _Immunol Lett_ 1991; 30: 201. Google Scholar  *

Preston M _et al_. Kinetics of serum, tear, and corneal antibody responses in resistant and susceptible mice intracorneally infected with _Pseudomonas aeruginosa_. _Infect Immun_ 1992; 60:

885. CAS  PubMed  PubMed Central  Google Scholar  * Kreger AS _et al_. Immunization against experimental _Pseudomonas aeruginosa_ and _Serratia marcescens_ keratitis vaccination with

lipopolysaccharide endotoxin and proteases. _Invest Ophthalmol Vis Sci_ 1986; 27: 932. CAS  PubMed  Google Scholar  * Valenton MJ, Okumoto M . Toxin producing strains of _Staphylococcus

epidermidis_. _Arch Ophthalmol_ 1973; 89: 187. Google Scholar  * Jelyaszewicz J, Wadstrom T (eds). _Bacterial Toxins in Cell Membranes_. Academic Press: London, 1978. Google Scholar  *

Mondino BJ, Rabin BS, Kessler E _et al_. Corneal rings with Gram negative bacteria. _Arch Ophthalmol_ 1977; 95: 2222. CAS  PubMed  Google Scholar  * Belmont JB, Ostler HB, Chandler RD,

Schwab I . Non-infectious ring-shaped keratitis associated with _Pseudomonas aeruginosa_. _Am J Ophthalmol_ 1982; 93: 338. CAS  PubMed  Google Scholar  * Jones DB . Polymicrobial Keratitis.

_Trans Am Ophthalmol Soc_ 1981; 79: 153. CAS  PubMed  PubMed Central  Google Scholar  * Baum JL, Jones DB : Initial therapy of suspected microbial corneal ulcers. _Surv Ophthalmol_ 1979; 24:

97. CAS  PubMed  Google Scholar  * Wilhelmus KR . Bacterial corneal ulcer. _Int Ophthalmol Clin_ 1984; 24: 1. CAS  PubMed  Google Scholar  * Gomez JT, Robinson NN, Osato MS _et al_.

Comparison of acridine orange and Gram stains in bacterial keratitis. _Am J Ophthalmol_ 1988; 106: 735. CAS  PubMed  Google Scholar  * Groden LR, Rodnite J, Brinser JH, Genvert GI . Acridine

orange and Gram stains in infectious keratitis. _Cornea_ 1990; 9: 122. CAS  PubMed  Google Scholar  * Groden LR, Brinser JH : Out-patient treatment of microbial corneal ulcers. _Arch

Ophthalmol_ 1986; 104: 84. CAS  PubMed  Google Scholar  * O'Brien TP, Reynolds LA . Basic ocular pharmacotherapy. _J Ophthalmol Nurs Tech_ 1995; 14(4): 160. CAS  Google Scholar  * Shell

JW : Pharmacokinetics of topically applied ophthalmic drugs. _Surv Ophthalmol_ 1982; 26: 207. CAS  PubMed  Google Scholar  * Lesar TS, Fiscella RG . Antimicrobial drug delivery to the eye.

_Drug Intel Clin Pharm_ 1985; 19: 642. CAS  Google Scholar  * Barza M . Antibacterial agents in the treatment of ocular infections. _Infect Dis Clin N Am_ 1989; 3: 533. CAS  Google Scholar 

* Davis SD, Sarff LD, Hyndiuk RA . Topical tobramycin therapy of experimental _Pseudomonas_ keratitis: an evaluation of some factors that potentially enhance efficacy. _Arch Ophthalmol_

1978; 96: 123. CAS  PubMed  Google Scholar  * Kupferman A, Leibowitz HM . Topical antibiotic therapy of _Pseudomonas aeruginosa_ keratitis in _guinea pigs_. _Arch Ophthalmol_ 1979; 97: 1699.

CAS  PubMed  Google Scholar  * Davis SD, Sarff LD, Hyndiuk RA . Antibiotic therapy of experimental _Pseudomonas_ keratitis in guinea pigs. _Arch Ophthalmol_ 1977; 95: 1638. CAS  PubMed 

Google Scholar  * Glasser DB, Gardner S, Ellis JG, Pettit TH . Loading doses and extended dosing intervals in topical gentamicin therapy. _Am J Ophthalmol_ 1985; 99: 329. CAS  PubMed  Google

Scholar  * Robin JS, Ellis PP . Ophthalmic ointments. _Surv Ophthalmol_ 1978; 22: 335. CAS  PubMed  Google Scholar  * Hanna C, Hof HC, Smith WG . Influence of drug vehicle on ocular contact

time of sulfacetamide sodium. _Ann Ophthalmol_ 1985; 17: 560. CAS  PubMed  Google Scholar  * Osborn E, Baum JL, Ernst C, Koch P . The stability of ten antibiotics in artificial tears. _Am J

Ophthalmol_ 1976; 82: 775. CAS  PubMed  Google Scholar  * Baum J, Barza M, Hsushan D, Weinstein L . Concentration of gentamicin in experimental corneal ulcers. _Arch Ophthalmol_ 1974; 92:

315. CAS  PubMed  Google Scholar  * Kupferman A, Leibowitz HM . Antibiotic therapy of bacterial keratitis: topical application or a periocular injection? _Invest Ophthalmol Vis Sci_ 1980;

19: 112. Google Scholar  * Baum JL, Barza M . Topical _versus_ subconjunctival treatment of bacterial corneal ulcers. _Ophthalmology_ 1983; 90: 162. CAS  PubMed  Google Scholar  * Stern GA,

Driebe WT : The effect of fortified antibiotic therapy on the visual outcome of severe bacterial corneal ulcers. _Cornea_ 1982; 1: 341. Google Scholar  * Liebowitz HM, Ryan WJ, Kupferman A :

Route of antibiotic administration in bacterial keratitis. _Arch Ophthalmol_ 1981; 99: 1420. Google Scholar  * Baum J . Treatment of bacterial ulcers of the cornea in the rabbit: a

comparison of administration by eyedrops and subconjunctival injections. _Trans Am Ophthalmol Soc_ 1982; 80: 369. CAS  PubMed  PubMed Central  Google Scholar  * Hessburg PC . Treatment of

_Pseudomonas_ keratitis in humans. _Am J Ophthalmol_ 1966; 61: 896. CAS  PubMed  Google Scholar  * Burris TE, Newsom DI, Rowsey JJ . Hessburg subpalpebral antibiotic lavage of _Pseudomonas_

corneal and corneoscleral ulcers. _Cornea_ 1982; 1: 347. Google Scholar  * Matoba AY, McCulley JP . The effect of therapeutic soft contact lenses on antibiotic delivery to the cornea.

_Ophthalmology_ 1985; 92: 97. CAS  PubMed  Google Scholar  * Busin M, Goebbels M, Spitznas M . Medicated bandage lenses for sustained gentamicin release. _Ophthalmology_ 1987; 94 (Suppl):

124. Google Scholar  * O'Brien TP, Sawusch MR, Dick JD _et al_. Use of collagen corneal shields _versus_ soft contact lenses to enhance penetration of topical tobramycin. _J Cataract

Refract Surg_ 1988; 14: 505. CAS  PubMed  Google Scholar  * Unterman SR, Rootman DS, Hill JM _et al_. Collagen shield drug delivery: therapeutic concentrations of tobramycin in the rabbit

cornea and aqueous humor. _J Cataract Refract Surg_ 1988; 14: 500. CAS  PubMed  Google Scholar  * Sawusch MR, O'Brien TP, Dick JD _et al_. Use of collagen corneal shields in the

treatment of bacterial keratitis. _Am J Ophthalmol_ 1988; 106: 279. CAS  PubMed  Google Scholar  * Baum JL, Barza M, Weinstein L . Preferred routes of antibiotic administration in treatment

of bacterial ulcers of the cornea. _Int Ophthalmol Clin_ 1973; 13: 31. CAS  PubMed  Google Scholar  * Gilbert ML, Wilhelmus KR, Osato MS . Intracanalicular collagen implants enhance topical

antibiotic bioavailability. _Cornea_ 1986; 5: 167. CAS  PubMed  Google Scholar  * Schaeffer HE, Krohn DL . Liposomes and topical drug delivery. _Invest Ophthalmol Vis Sci_ 1982; 22: 220. CAS

  PubMed  Google Scholar  * Rootman DS, Hobden JA, Jantzen JA _et al_. Iontophoresis of tobramycin for treatment of experimental _Pseudomonas_ keratitis in rabbits. _Arch Ophthalmol_ 1988;

106: 262. CAS  PubMed  Google Scholar  * Callegan MC, O'Callaghan RJ, Hill JM . Pharmacokinetic considerations in the treatment of bacterial keratitis. _Clin Pharmacokinet_ 1994; 27(2):

129. CAS  PubMed  Google Scholar  * Cunha-Vaz J . The blood–ocular barriers. _Surv Ophthalmol_ 1979; 23: 279. CAS  PubMed  Google Scholar  * Burns RP, Rhodes DM . _Pseudomonas_ eye

infection as a cause of death in premature infants. _Arch Ophthalmol_ 1961; 65: 517. CAS  PubMed  Google Scholar  * Gelender H, Rettich C . Gentamicin-resistant _Pseudomonas aeruginosa_

corneal ulcers. _Cornea_ 1984; 3: 21. CAS  PubMed  Google Scholar  * O'Brien TP, Maguire M, Fink N _et al_. Results from the bacterial keratitis study: topical ofloxacin _versus_

cefazolin and aminoglycoside in therapy for acute bacterial keratitis. _Arch Ophthalmol_, 1995. * Hyndiuk RA, Eiferman RA, Caldwell DR _et al_. Comparison of ciprofloxacin 0.3% (Ciloxan)

solution to fortified tobramycin/cefazolin in treatment of bacterial corneal ulcers. _Ophthalmology_, 1996; 103: 1854–1862. CAS  PubMed  Google Scholar  * Donnenfeld ED, Cohen EJ, Barza M,

Baum J . Treatment of _Nocardia_ keratitis with topical trimethoprim-sulfameth-oxazole. _Am J Ophthalmol_ 1985; 99: 601. CAS  Google Scholar  * Swenson JM, Wallace Jr RJ, Silcox VA _et al_.

Antimicrobial sensitivity of five subgroups of _Mycobacterium fortuitum_ and _Mycobacterium chelonae_. _Antimicrob Agents Chemother_ 1985; 28: 807. CAS  PubMed  PubMed Central  Google

Scholar  * Gay JD, Deyoung DR, Roberts GD . _In Vitro_ activity of norfloxacin and ciprofloxacin against _Mycobacterium tuberculosis_, _M. avium_ complex, _M. chelonei_, _M. fortuitum_, _M.

kansasii_. _Antimicrob Agents Chemother_ 1984; 26: 94. CAS  PubMed  PubMed Central  Google Scholar  * Wallace RJ, Bedsole G, Sumter G _et al_. Activities of ciprofloxacin and ofloxacin

against rapidly growing Mycobacteria with demonstration of acquired resistance following single-drug therapy. _Antimicrob Agents Chemother_ 1990; 34: 65. CAS  PubMed  PubMed Central  Google

Scholar  * Hwang DG, Biswell R . Ciprofloxacin therapy of _Mycobacterium chelonae_ keratitis. _Am J Ophthalmol_ 1993; 115: 114. CAS  PubMed  Google Scholar  * Field JH, Backhoff IK, Dick JD,

O'Brien TP . Comparative topical treatment of _Mycobacterium fortuitum_ keratitis in rabbits. _Invest Ophthalmol Vis Sci_ 1993; 34(4): 737. Google Scholar  * Husain SE, Matoba AY,

Husain N, Jones DB . Antimicrobial efficacy of clarithromycin and azithromycin against _Mycobacterium chelonae_ and _Mycobacterium abscessus_. _Invest Ophthalmol Vis Sci_ 1993; 34(4): 729.

Google Scholar  * Brown SI, Bloomfield S, Pearce DB, Tragakis M . Infections with a therapeutic soft lens. _Arch Ophthalmol_ 1974; 91: 275. CAS  PubMed  Google Scholar  * Slansky HH, Dohlman

CH, Berman MB . Prevention of corneal ulcers. _Trans Am Acad Ophthalmol Otolaryngol_ 1971; 75: 1208. CAS  PubMed  Google Scholar  * Ellison A, Poirier R . Therapeutic effects of heparin on

_Pseudomonas_-induced corneal ulceration. _Am J Ophthalmol_ 1976; 82: 619. CAS  PubMed  Google Scholar  * Mehra KS, Singh R, Bhatia RPS . Lysozyme in corneal ulcer. _Ann Ophthalmol_ 1975; 7:

1470. CAS  PubMed  Google Scholar  * Barletta JP, Balch KC, Dimova HG _et al_. Inhibition of _Pseudomonas_ corneal ulcers by a synthetic matrix metalloproteinase inhibitor (Galardin).

_Invest Ophthalmol Vis Sci_ 1993; 34(4): 1749. Google Scholar  * Brown M, Li WW, Dutt S _et al_. A synthetic metalloproteinase inhibitor potentiates the anti-angiogenic effect of

prednisolone in the cornea. _Invest Ophthalmol Vis Sci_. * Davis SD, Sarff LD, Hyndiuk RA . Corticosteroid in experimentally induced _Pseudomonas_ keratitis. _Arch Ophthalmol_ 1978; 96: 126.

CAS  PubMed  Google Scholar  * Smolin G, Okumoto M, Leong-Sit L . Combined gentamicin–tobramycin–corticosteroid treatment II. Effect on gentamicin-resistant _Pseudomonas_ keratitis. _Arch

Ophthalmol_ 1980; 98: 473. CAS  PubMed  Google Scholar  * Badenoch PR, Hay GJ, McDonnell PJ _et al_. A rat model of bacterial keratitis. Effect of antibiotics and corticosteroids. _Arch

Ophthalmol_ 1985; 103: 718. CAS  PubMed  Google Scholar  * Leibowitz HM, Kupferman A . Topically administered corticosteroids: Effect on antibiotic treated bacterial keratitis. _Arch

Ophthalmol_ 1980; 98: 1237. Google Scholar  * Bohigian GM, Foster CS . Treatment of _Pseudomonas_ keratitis in a rabbit with antibiotic steroid combination. _Invest Ophthalmol Vis Sci_ 1977;

16: 553. CAS  PubMed  Google Scholar  * Harbin T . Recurrence of corneal _Pseudomonas_ infection after topical steroid therapy. _Am J Ophthalmol_ 1964; 58: 670. CAS  PubMed  Google Scholar

  * Burns RP . _Pseudomonas_ _aeruginosa_ keratitis: mixed infections of the eye. _Am J Ophthalmol_ 1969; 67: 257. CAS  PubMed  Google Scholar  * Stern GA, Buttross M . Use of

corticosteroids in combination with antimicrobial drugs in the treatment of infectious corneal disease. _Ophthalmology_ 1991; 98: 847. CAS  PubMed  Google Scholar  * Alpren TVP, Hyndiuk RA,

Davis SD _et al_. Cryotherapy for experimental _Pseudomonas_ keratitis. _Arch Ophthalmol_ 1979; 97: 711. CAS  PubMed  Google Scholar  * Codere F, Brownstein S, Jackson B . _Pseudomonas_

_aeruginosa_ scleritis. _Am J Ophthalmol_ 1981; 91: 706. CAS  PubMed  Google Scholar  * Eiferman RA . Cryotherapy of _Pseudomonas_ keratitis and scleritis. _Arch Ophthalmol_ 1979; 97: 1637.

CAS  PubMed  Google Scholar  * Nassaralla BA, Wang XW, Wee WR _et al_. Effect of diclofenac sodium on antibiotic treatment of experimental _Pseudomonas_ keratitis. _Invest Ophthalmol Vis

Sci_ 1994; 36(4): 4721. Google Scholar  * Gritz DZ, Kwitko S, Trousdale MD _et al_. Recurrence of microbial keratitis concomitant with anti-inflammatory treatment in an animal model.

_Cornea_ 1992; 11(5): 404. CAS  PubMed  Google Scholar  * Eiferman RA, Snyder JW . Antibacterial effect of cyanoacrylate glue. _Arch Ophthalmol_ 1983; 101: 958. CAS  PubMed  Google Scholar 

* Cavanaugh TB, Gottsch JD . Infectious keratitis in cyanoacrylate adhesive. _Am J Ophthalmol_ 1991; 111(4): 466. CAS  PubMed  Google Scholar  * Gottsch JD, Gilbert ML, Goodman DF _et al_.

Excimer laser ablative treatment of microbial keratitis. _Ophthalmology_ 1991; 98(2): 146. CAS  PubMed  Google Scholar  * Serdarevic O _et al_. Excimer laser therapy for experimental

_Candida_ keratitis. _Am J Ophthalmol_ 1985; 99: 534. CAS  PubMed  Google Scholar  * Sarno EM _et al_. Carbon dioxide laser therapy of _Pseudomonas_ _aeruginosa_ keratitis. _Am J Ophthalmol_

1984; 97: 791. CAS  PubMed  Google Scholar  * Kirkness CM, Ficker LA, Steele AD, Rice NS . The role of penetrating keratoplasty in the management of microbial keratitis. _Eye_ 1991; 5(4):

425. PubMed  Google Scholar  * Killingsworth DW, Stern GA, Driebe WT _et al_. Results of therapeutic penetrating keratoplasty. _Ophthalmology_ 1993; 100(4): 534. CAS  PubMed  Google Scholar

  * Lotti R, Dart JK . Cataract as a complication of severe microbial keratitis. _Eye_ 1992; 6(4): 400. PubMed  Google Scholar  * Moxon R, Rappuoli R . Bacterial pathogen genomics and

vaccines. _Br Med Bull_ 2002; 62(1): 45–58. PubMed  Google Scholar  * Chalker A, Lunsford R . Rational identification of new antibacterial drug targets that are essential for viability using

a genomics-based approach. _Pharmacol Ther_ 2002; 95(1): 1. CAS  PubMed  Google Scholar  * Di Guilmi AM, Dessen A . New approaches towards the identification of antibiotic and vaccine

targets in _Streptococcus pneumoniae_. _EMBO Rep_ 2002; 3(8): 728–734. CAS  PubMed  PubMed Central  Google Scholar  * Waterman SR . Bacterial genomics as a potential tool for discovering new

antimicrobial agents. _Am J Pharmacogenomics_ 2001; 1(4): 263–269. CAS  PubMed  Google Scholar  * Buchanan SG, Sauder JM, Harris T . The promise of structural genomics in the discovery of

new antimicrobial agents. _Curr Pharm Des_ 2002; 8(13): 1173–1188. CAS  PubMed  Google Scholar  * Dougherty TJ, Barrett JF, Pucci MJ . Microbial genomics and novel antibiotic discovery: new

technology to search for new drugs. _Curr Pharm Des_ 2002; 8(13): 1119–1135. CAS  PubMed  Google Scholar  * Haney SA, Alksne LE, Dunman PM, Murphy E, Projan SJ . Genomics in anti-infective

drug discovery getting to endgame. _Curr Pharm Des_ 2002; 8(13): 1099–1118. CAS  PubMed  Google Scholar  * Klade CS . Proteomics approaches towards antigen discovery and vaccine development.

_Curr Opin Mol Ther_ 2002; 4(3): 216–223. CAS  PubMed  Google Scholar  * Nilsson CL . Bacterial proteomics and vaccine development. _Am J Pharmacogenomics_ 2002; 2(1): 59–65. CAS  PubMed 

Google Scholar  * Huang SH, Triche T, Jong AY . Infectomics: genomics and proteomics of microbial infections. _Funct Integr Genomics_ 2002; 1(6): 331–344. CAS  Google Scholar  * Fernandes PB

. The ATCG of drug discovery. _Curr Opin Mol Ther_ 2000; 2(6): 624–632. CAS  PubMed  Google Scholar  * Searls DB . Using bioinformatics in gene and drug discovery. _Drug Discov Today_ 2000;

5(4): 135–143. CAS  PubMed  Google Scholar  Download references AUTHOR INFORMATION AUTHORS AND AFFILIATIONS * Ocular Infectious Diseases, Ocular Microbiology Laboratory, The Wilmer

Ophthalmological Institute, The Johns Hopkins University School of Medicine, Baltimore, MD, USA T P O'Brien Authors * T P O'Brien View author publications You can also search for

this author inPubMed Google Scholar CORRESPONDING AUTHOR Correspondence to T P O'Brien. ADDITIONAL INFORMATION Presented at the 32nd Cambridge Ophthalmological Symposium, The Cornea,

4–6 September 2002 RIGHTS AND PERMISSIONS Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE O'Brien, T. Management of bacterial keratitis: beyond exorcism towards

consideration of organism and host factors. _Eye_ 17, 957–974 (2003). https://doi.org/10.1038/sj.eye.6700635 Download citation * Received: 28 February 2003 * Accepted: 28 February 2003 *

Published: 20 November 2003 * Issue Date: 01 November 2003 * DOI: https://doi.org/10.1038/sj.eye.6700635 SHARE THIS ARTICLE Anyone you share the following link with will be able to read this

content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to clipboard Provided by the Springer Nature SharedIt content-sharing initiative

KEYWORDS * cornea * bacterial * keratitis * pseudomonas * antibiotics * genomics