Confirmation of the role of pneumolysin in ocular infections with Streptococcus pneumoniae

The Role of Pneumococcal Virulence Factors in Ocular Infectious Diseases

Streptococcus pneumoniae is a gram-positive, facultatively anaerobic pathogen that can cause severe infections such as pneumonia, meningitis, septicemia, and middle ear infections. It is also one of the top pathogens contributing to bacterial keratitis and conjunctivitis. Though two pneumococcal vaccines exist for the prevention of nonocular diseases, they do little to fully prevent ocular infections. This pathogen has several virulence factors that wreak havoc on the conjunctiva, cornea, and intraocular system. Polysaccharide capsule aids in the evasion of host complement system. Pneumolysin (PLY) is a cholesterol-dependent cytolysin that acts as pore-forming toxin. Neuraminidases assist in adherence and colonization by exposing cell surface receptors to the pneumococcus. Zinc metalloproteinases contribute to evasion of the immune system and disease severity. The main purpose of this review is to consolidate the multiple studies that have been conducted on several pneumococcal virulence factors and the role each plays in conjunctivitis, keratitis, and endophthalmitis.

Host response and bacterial virulence factor expression in Pseudomonas aeruginosa and Streptococcus pneumoniae corneal ulcers

P. aeruginosa and S. pneumoniae are major bacterial causes of corneal ulcers in industrialized and in developing countries. The current study examined host innate immune responses at the site of infection, and also expression of bacterial virulence factors in clinical isolates from patients in south India. Corneal ulcer material was obtained from 49 patients with confirmed P. aeruginosa and 27 patients with S. pneumoniae, and gene expression of Toll Like Receptors (TLR), cytokines and inflammasome proteins was measured by quantitative PCR. Expression of P. aeruginosa type III secretion exotoxins and S. pneumoniae pneumolysin was detected by western blot analysis. We found that neutrophils comprised >90% cells in corneal ulcers, and that there was elevated expression of TLR2, TLR4, TLR5 and TLR9, the NLRP3 and NLRC4 inflammasomes and the ASC adaptor molecule. IL-1α IL-1β and IFN-γ expression was also elevated; however, there was no significant difference in expression of any of these genes between corneal ulcers from P. aeruginosa and S. pneumoniae infected patients. We also show that 41/49 (84%) of P. aeruginosa clinical isolates expressed ExoS and ExoT, whereas 5/49 (10%) of isolates expressed ExoS, ExoT and ExoU with only 2/49 isolates expressing ExoT and ExoU. In contrast, all 27 S. pneumoniae clinical isolates produced pneumolysin. Taken together, these findings demonstrate that ExoS/T expressing P. aeruginosa and pneumolysin expressing S. pneumoniae predominate in bacterial keratitis. While P. aeruginosa strains expressing both ExoU and ExoS are usually rare, these strains actually outnumbered strains expressing only ExoU in the current study. Further, as neutrophils are the predominant cell type in these corneal ulcers, they are the likely source of cytokines and of the increased TLR and inflammasome expression.

Active Immunization with Pneumolysin versus 23-Valent Polysaccharide Vaccine for Streptococcus pneumoniae Keratitis

PURPOSE

The purpose of this study was to determine whether active immunization against pneumolysin (PLY), or polysaccharide capsule, protects against the corneal damage associated with Streptococcus pneumoniae keratitis.

METHODS

New Zealand White rabbits were actively immunized with Freund's adjuvant mixed with pneumolysin toxoid (ψPLY), Pneumovax 23 (PPSV23; Merck, Whitehouse Station, NJ), or phosphate-buffered saline (PBS), before corneal infection with 10⁵ colony-forming units (CFU) of S. pneumoniae. Serotype-specific rabbit polyclonal antisera or mock antisera were passively administered to rabbits before either intravenous infection with 10¹¹ CFU S. pneumoniae or corneal infection with 10⁵ CFU of S. pneumoniae.

RESULTS

After active immunization, clinical scores of corneas of the rabbits immunized with ψPLY and Freund's adjuvant were significantly lower than scores of the rabbits that were mock immunized with PBS and Freund's adjuvant or with PPSV23 and Freund's adjuvant at 48 hours after infection (P ≤ 0.0010), whereas rabbits immunized with PPSV23 and Freund's adjuvant failed to show differences in clinical scores compared with those in mock-immunized rabbits (P = 1.00) at 24 and 48 hours after infection. Antisera from rabbits actively immunized with PPSV23 and Freund's adjuvant were nonopsonizing. Bacterial loads recovered from infected corneas were higher for the ψPLY- and PPSV23-immunized rabbits after infection with WU2, when compared with the mock-immunized rabbits (P ≤ 0.007). Conversely, after infection with K1443, the ψPLY-immunized rabbits had lower bacterial loads than the control rabbits (P = 0.0008). Quantitation of IgG, IgA, and IgM in the sera of ψPLY-immunized rabbits showed high concentrations of PLY-specific IgG. Furthermore, anti-PLY IgG purified from ψPLY-immunized rabbits neutralized the cytolytic effects of PLY on human corneal epithelial cells. Passive administration of serotype-specific antisera capable of opsonizing and killing S. pneumoniae protected against pneumococcal bacteremia (P ≤ 0.05), but not against keratitis (P ≥ 0.476).

CONCLUSIONS

Active immunization with pneumococcal capsular polysaccharide and Freund's adjuvant fails to produce opsonizing antibodies, and passive administration of serotype specific opsonizing antibodies offers no protection against pneumococcal keratitis in the rabbit, whereas active immunization with the conserved protein virulence factor PLY and Freund's adjuvant is able to reduce corneal inflammation associated with pneumococcal keratitis, but has variable effects on bacterial loads in the cornea.

Identification and potency of cyclodextrin-lipid inhibitors of Staphylococcus aureus α-toxin

PURPOSE

Staphylococcus aureus, a leading cause of bacterial keratitis, secretes α-toxin, a cytotoxin active on the corneal epithelium. This study describes the production and testing of chemical inhibitors of α-toxin action.

METHODS

Purified α-toxin was titered by its ability to lyse rabbit erythrocytes in buffered saline (PBS). To prepare potential toxin inhibitors, each of 18 lipids was incorporated into a complex with methyl-β-cyclodextrin (MβCD) or hydroxypropyl-β-cyclodextrin (HPβCD). Serial dilutions of each lipid-cyclodextrin (CD-lipid) complex were mixed with α-toxin prior to the addition of rabbit erythrocytes. Select CD-lipid complexes were mixed with 12 hemolytic units (HU) α-toxin and injected into the rabbit corneal stroma so the resulting corneal erosions could be measured at 4 and 8 hours post-injection (PI). Eyes injected with toxin alone, MβCD, or HPβCD alone served as controls.

RESULTS

Neither form of CD alone inhibited α-toxin. Of the 36 complexes prepared, 6 lipid-CD complexes were found to inhibit >100 HU of α-toxin. Four lipid complexes able to inhibit >200 HU of α-toxin were tested in toxin-injected corneas; at 4 and 8 hours PI, the complexes of cholesterol or lanosterol with MβCD and squalene or desmosterol with HPβCD caused a significant reduction in the corneal erosion size as compared to eyes injected with α-toxin alone (P ≤ 0.05).

CONCLUSIONS

Specific lipid inclusion complexes with either MβCD or HPβCD demonstrated a significant inhibition of α-toxin in both in vitro and in vivo assays. Changes in either the cyclodextrin or lipid of a complex affected the inhibitory activity.

Pathogenesis of A Clinical Ocular Strain of Streptococcus pneumoniae and the Interaction of Pneumolysin with Corneal Cells

Streptococcus pneumoniae is an important cause of bacterial keratitis, an infectious disease of the cornea. This study aimed to determine the importance of pneumolysin (PLY), a pneumococcal virulence factor, in keratitis using a clinical keratitis isolate (K1263) and its isogenic mutant deficient in PLY (K1263ΔPLY) and determine the effect of these strains on primary rabbit corneal epithelial (RCE) cells. Each strain was injected into the corneal stromas of rabbits, clinical examinations were performed, and the recovered bacterial loads were determined. Bacterial extracts were exposed to RCE cells, and morphology and viability were assessed. The mutant strain deficient in PLY, K1263ΔPLY, caused significantly lower ocular disease scores than the parent strain (K1263), although a higher bacterial load was recovered from corneas infected with the mutant strain. Histological examination showed increased inflammatory cells in the anterior chamber and increased edema in eyes infected with the parent strain. RCE cells exposed to the parent strain had significantly decreased cell viability and showed increased evidence of cellular damage. This study confirms that in a strain that can cause clinical keratitis, PLY is a significant cause of the damage associated with pneumococcal keratitis. It also shows for the first time that the results from an in vitro model using RCE cells correlates with in vivo results thereby establishing a less invasive way to study the mechanisms of pneumococcal keratitis.

Assessment of Streptococcus pneumoniae capsule in conjunctivitis and keratitis in vivo neuraminidase activity increases in nonencapsulated pneumococci following conjunctival infection

PURPOSE

The pneumococcal capsule is required for pathogenesis in systemic infections, yet reports show most conjunctivitis outbreaks are caused by nonencapsulated pneumococci, while keratitis infections are caused by encapsulated strains. This study aims to determine the effect of capsule in pneumococcal keratitis and conjunctivitis in rabbit models of infection.

METHODS

A capsule-deficient isogenic mutant was created using homologous transformation. Parent and mutant strains were injected within the upper bulbar conjunctiva (conjunctivitis) or into the corneal stroma (keratitis) of New Zealand white rabbits. Clinical examinations were performed 24 and 48 hr post-infection at which time corneas or conjunctivae were removed, homogenized, and plated to determine the recovered bacterial load. Whole eyes were removed for histological examination. The neuraminidase activity was determined following in vitro and in vivo growth.

RESULTS

There were no significant differences in clinical scores between the eyes infected with the parent or mutant for either infection, nor was there a difference in the amount of bacteria recovered from the cornea. In the conjunctivae, however, the mutant strain was cleared by the host faster than the parent strain. Histological examination showed slightly more infiltrating polymorphonuclear leukocytes (PMN) and macrophages in the conjunctivae infected with the parent strain. The neuraminidase activity of both strains was not significantly different when the strains were grown in vitro. However, the neuraminidase activity of the parent was significantly less than that of the mutant at 3 and 12 hr post conjunctival infection.

CONCLUSIONS

Although more outbreaks of pneumococcal conjunctivitis are tied to nonencapsulated S. pneumoniae strains, this study showed that an encapsulated strain was capable of establishing conjunctivitis in a rabbit injection model and survive attack by the host immune system longer than its nonencapsulated isogenic mutant. Nonetheless, the nonencapsulated pneumococci had an increased neuraminidase activity level in vivo when compared to the parent strain.

Cirrhosis-induced defects in innate pulmonary defenses against Streptococcus pneumoniae

Background

The risk of mortality from pneumonia caused by Streptococcus pneumoniae is increased in patients with cirrhosis. However, the specific pneumococcal virulence factors and host immune defects responsible for this finding have not been clearly established. This study used a cirrhotic rat model of pneumococcal pneumonia to identify defect(s) in innate pulmonary defenses in the cirrhotic host and to determine the impact of the pneumococcal toxin pneumolysin on these defenses in the setting of severe cirrhosis.

Results

No cirrhosis-associated defects in mucociliary clearance of pneumococci were found in these studies, but early intrapulmonary killing of the organisms before the arrival of neutrophils was significantly impaired. This defect was exacerbated by pneumolysin production in cirrhotic but not in control rats. Neutrophil-mediated killing of a particularly virulent type 3 pneumococcal strain also was significantly diminished within the lungs of cirrhotic rats with ascites. Levels of lysozyme and complement component C3 were both significantly reduced in bronchoalveolar lavage fluid from cirrhotic rats. Finally, complement deposition was reduced on the surface of pneumococci recovered from the lungs of cirrhotic rats in comparison to organisms recovered from the lungs of control animals.

Conclusion

Increased mortality from pneumococcal pneumonia in this cirrhotic host is related to defects in both early pre-neutrophil- and later neutrophil-mediated pulmonary killing of the organisms. The fact that pneumolysin production impaired pre-neutrophil-mediated pneumococcal killing in cirrhotic but not control rats suggests that pneumolysin may be particularly detrimental to this defense mechanism in the severely cirrhotic host. The decrease in neutrophil-mediated killing of pneumococci within the lungs of the cirrhotic host is related to insufficient deposition of host proteins such as complement C3 on their surfaces. Pneumolysin likely plays a role in complement consumption within the lungs. Our studies, however, were unable to determine whether pneumolysin more negatively impacted this defense mechanism in cirrhotic than in control rats. These findings contribute to our understanding of the defects in innate pulmonary defenses that lead to increased mortality from pneumococcal pneumonia in the severely cirrhotic host. They also suggest that pneumolysin may be a particularly potent pneumococcal virulence factor in the setting of cirrhosis.

Cholesterol as treatment for pneumococcal keratitis: cholesterol-specific inhibition of pneumolysin in the cornea

PURPOSE

The purpose of this study was to determine whether cholesterol, the host cell receptor for pneumolysin of Streptococcus pneumoniae, could effectively treat pneumococcal keratitis.

METHODS

New Zealand White rabbits were intrastromally injected with 10(5) colony-forming units (CFUs) of S. pneumoniae D39. Corneas were treated with topical drops of 1% cholesterol every 2 hours beginning 25 hours after infection and were examined by slit lamp microscopy 24, 36, and 48 hours after infection. Rabbits were killed, and CFUs were recovered from the corneas after the final slit lamp examination (SLE). Minimal inhibitory concentration (MIC) assays of cholesterol against bacteria were performed. Specific inhibition of pneumolysin by cholesterol in the rabbit cornea was tested by intrastromal injection of pneumolysin with or without cholesterol and was compared with cholesterol inhibition of pneumolysin in vitro using hemolysis assays with rabbit erythrocytes.

RESULTS

Corneas treated with cholesterol had significantly lower SLE scores 48 hours after infection than corneas treated with vehicle (P = 0.0015). Treated corneas also had significantly less log(10) CFUs than vehicle-treated corneas (P = 0.0006). Cholesterol at a 1% concentration was bactericidal to bacteria in vitro, and lower concentrations of cholesterol were partially inhibitory in a concentration-dependent manner. Cholesterol also specifically inhibited 1 mug pneumolysin in vivo and up to 50 ng pneumolysin in vitro.

CONCLUSIONS

Topical cholesterol is an effective treatment for S. pneumoniae keratitis. Cholesterol not only inhibits pneumolysin, it is also bactericidal.

Pneumococcal keratitis: a clinical profile

AIM

To study the clinical features of pneumococcal keratitis and response to ciprofloxacin therapy.

METHODS

A retrospective analysis was undertaken of 58 patients with culture-proven pneumococcal keratitis seen over a period of 2 years.

RESULTS

Pneumococcal keratitis accounted for 33.3% of bacterial keratitis. Most cases presented with non-severe keratitis (77.5%). Co-existing sac pathology was more frequent in pneumococcal ulcers as compared to non-pneumococcal bacterial ulcers (50%vs 9%, P < 0.001). Characteristic clinical features enabling an accurate clinical diagnosis were found in 27.5% and lanceolate diplococci on Gram's stain were identified in 76% of cases. In vitro testing showed a high susceptibility to cephazolin and ciprofloxacin. All patients received ciprofloxacin as first-line therapy. Eighty per cent responded well with complete healing of the ulcer. A second drug was required in 8.5%.

CONCLUSION

Ciprofloxacin therapy can be effective in the treatment of pneumococcal keratitis.

Contribution of pneumolysin and autolysin to the pathogenesis of experimental pneumococcal endophthalmitis

PURPOSE

To determine the contribution of pneumolysin and autolysin, two putative pneumococcal virulence proteins, to the pathogenesis of Streptococcus pneumoniae endophthalmitis.

METHODS

Endophthalmitis was established in Lewis rats by intravitreal injection of pneumococcal strains at an inoculum of 10 organisms. The virulence of three closely related type 2 S. pneumoniae strains were compared: a pneumolysin-deficient derivative (PLN-A), an autolysin-deficient derivative (AL-6), and their isogenic wild-type parent (D 39). Clinical and histologic inflammation scores were compared 24 hours and 48 hours after inoculation.

RESULTS

Eyes infected with PLN-A and AL-6 strains showed less anterior segment inflammation clinically at 24 hours than did eyes infected with the wild-type strain. Histologic examination at 24 hours showed significantly less corneal infiltration and vitritis and more relative preservation of retinal tissue in eyes infected with PLN-A and AL-6 strains than in eyes infected with the wild-type strain. At 48 hours, no observable differences between PLN-A and wild-type strains were present clinically or histologically. Histologically, however, the AL-6 strain caused less retinal damage than did the wild-type strain.

CONCLUSIONS

Intraocular infection with pneumolysin-deficient S. pneumoniae results in less severe tissue damage in the first 24 hours of disease compared with infection with pneumolysin-producing S. pneumoniae. Autolysin-deficient S. pneumoniae shows a similar degree of attenuated virulence. Pneumolysin and autolysin may contribute to the early pathogenesis of pneumococcal endophthalmitis.

Pneumolysin-induced complement depletion during experimental pneumococcal bacteremia

To quantify complement depletion by pneumolysin during Streptococcus pneumoniae bacteremia, cirrhotic and control rats were infected intravenously with one of three isogenic mutant strains of S. pneumoniae expressing different forms of pneumolysin. Outcome measures included clearance of the organisms from the bloodstream, alterations in 50% serum hemolytic complement (CH(50)) activity and complement C3 levels during infection, and serum opsonic capacity at 18 h postinfection. Cirrhotic rats had significantly lower CH(50) and C3 levels than control rats, both before and after infection. However, initial complement levels did not predict bacterial load after 18 h of infection. Changes in CH(50) and C3 levels over the 18-h period correlated with numbers of H+C+ but not H+C- or PLY- organisms in the bloodstream at 18 h postinfection. The sera of cirrhotic rats infected with the H+C+ strain had significantly decreased levels of C3 and showed significantly lower opsonizing activity for S. pneumoniae than sera from H+C+-infected control rats. These studies suggest that under limiting concentrations of complement, the expression of pneumolysin by pneumococci has a significant, negative effect on serum complement levels and reduces serum opsonic activity.

Streptococcal competence for genetic transformation: regulation by peptide pheromones

Although the capacity for genetic transformation is perhaps the most famous attribute of pneumococcus, use of this genetic phenomenon as a tool for study of the biology of the organism and of its pathogenicity has been largely restricted to a few favored unencapsulated strains, both by the delicacy of the conditions required for development of competence, and by experience that encapsulated strains transformed poorly. We discuss here the recent discovery of a small stable inexpensive peptide pheromone that acts as a quorum-sensing signal and that induces competence under a wide variety of conditions and in encapsulated strains. Its use circumvents some if not all limitations to the expression of transformability in pneumococcus and therefore expands opportunities for application of tools molecular genetics to many strains of pneumococcus without prior genetic manipulation.

Role of Pneumolysin's complement-activating activity during pneumococcal bacteremia in cirrhotic rats

We investigated the role of pneumolysin's complement-activating activity during Streptococcus pneumoniae bacteremia in a hypocomplementemic, cirrhotic host. Isogenic mutant pneumococcal strains, in which pneumolysin was expressed from a plasmid, were used. These strains included H+C+, expressing wild-type pneumolysin with both cytolytic and complement-activating activity; PLY-, carrying the plasmid without the pneumolysin gene; and, H+C-, expressing pneumolysin with cytolytic activity only. In control rats, intravenous infection with 2.0 x 10(7) CFU of H+C+ per ml of blood resulted in a decrease in bacteremia of 3.5 log units by 18 h postinfection and 55% mortality. By contrast, cirrhotic rats infected similarly with the H+C+ strain demonstrated a 0.2-log-unit increase in bacteremia by 18 h postinfection and 100% mortality. Both control and cirrhotic rats cleared the PLY- strain more effectively from their bloodstreams by 18 h postinfection (6.2 and 5. 6 log unit decreases, respectively). Infection with the PLY- strain also resulted in low mortality (0 and 14%, respectively) for control and cirrhotic rats. When infected with the H+C- strain (without complement-activating activity), both groups cleared the organism from their bloodstreams nearly as well as they did the PLY- strain. Furthermore, the mortality rate for control and cirrhotic rats was identical after infection with the H+C- strain. These studies suggest that pneumolysin production contributes to decreased pneumococcal clearance from the bloodstream and higher mortality in both control and cirrhotic rats. However, pneumolysin's complement-activating activity may uniquely enhance pneumococcal virulence in the hypocomplementemic, cirrhotic host.

Roles of autolysin and pneumolysin in middle ear inflammation caused by a type 3 Streptococcus pneumoniae strain in the chinchilla otitis media model

Streptococcus pneumoniae cell wall and pneumolysin are important contributors to pneumococcal pathogenicity in some animal models. To further explore these factors in middle ear inflammation caused by pneumococci, penicillin-induced inflammatory acceleration was studied by using three closely related pneumococcal strains: a wild-type 3 strain (WT3), its pneumolysin-negative derivative (P-1), and into autolysin-negative derivative (A-1). Both middle ears of chinchillas were inoculated with one of the three pneumococcal strains. During the first 12 h, all three strains grew in vivo at the same rate, and all three strains induced similar inflammatory cell responses in middle ear fluid (MEF). Procaine penicillin G was given as 12 h to one-half of the animals in each group, and all treated chinchillas had sterile MEF at 24 h. Penicillin significantly accelerated MEF inflammatory cell influx into WT3-and P-1-infected ears at 18 and 24 h in comparison with the rate for penicillin-treated A-1-infected ears. Inflammatory cell influx was slightly, but not significantly, greater after treatment of WT3 infection than after treatment of P-1 infection. Interleukin (IL)-1beta and IL-6, but not IL-8, concentrations in MEF at 24 h reflected the penicillin effect on MEF inflammatory cells; however, differences between treatment groups were not significant. Results suggest that pneumococcal otitis media pathogenesis is triggered principally by the inflammatory effects of intact and lytic cell wall products in the middle ear, with at most a modes additional pneumolysin effect. Investigation strategies that limit the release of these products or neutralize them warrant further investigation.

Growth and virulence of a complement-activation-negative mutant of Streptococcus pneumoniae in the rabbit cornea

Our previous work has demonstrated the importance of pneumolysin in the virulence of S. pneumoniae in a rabbit intracorneal model. This was accomplished by showing that deletion of the gene encoding pneumolysin resulted in reduced virulence, whereas restoration of the wild-type gene resulted in restoration of the virulent phenotype. To assess the importance of a particular domain in the pneumolysin molecule, we have now constructed a strain which produces a pneumolysin molecule which is hemolytic but which bears a site-specific mutation in the domain known to be associated with the complement-activating properties of this molecule. Comparison of the virulence of this strain with that of a strain bearing the wild-type gene showed statistically significantly lower total slit lamp examination (SLE) scores at 12, 18, 24, and 36 h (particularly with respect to fibrin formation), but no difference at 48 h. Determination of colony forming units (CFU) in eyes infected with the two strains showed approximately 10(6) bacteria per cornea until 36 h. Between 36 and 48 h, the bacteria were almost completely cleared with very few bacteria recoverable at the later time point. The loss of virulence observed with this mutation in the complement-activation domain of pneumolysin, though less than that observed with the gene deletion mutant, suggests that complement activation by pneumolysin has a significant role in the pathology observed in this model of corneal infection.

Pharmacokinetic considerations in the treatment of bacterial keratitis

The eye is relatively impermeable to micro-organisms and other environmental elements. However, if corneal integrity is breached by trauma, a sight-threatening bacterial infection can result. Staphylococcus aureus, Pseudomonas aeruginosa, and Streptococcus pneumoniae are the most common bacterial pathogens associated with infection of compromised corneas. Bacterial enzymes and toxins, as well as factors associated with the host immune response, can lead to tissue destruction during corneal infection. For successful therapy, an antibacterial agent must be active against the pathogen and must be able to overcome the permeability barrier of the cornea. Topical application of antibacterial agents adequately delivers drugs to the cornea and aqueous humour. However, drug concentrations at the site of infection are not always sufficient to rapidly kill infective organisms. Infections with antibiotic-resistant strains present an even greater therapeutic challenge. In addition, sterilisation of the cornea by antibacterial agents does not eliminate inflammation and corneal scarring that accompany infection. Steroidal and non-steroidal antiinflammatory agents limit corneal scarring during experimentally induced bacterial keratitis. However, although promising, concomitant use of these drugs with antibacterial agents remains controversial. Two ocular drug delivery systems that provide high and sustained concentrations of drug to ocular tissues are corneal collagen shields and transcorneal iontophoresis. The collagen shield, originally designed as a bandage lens, prolongs drug contact with the cornea. Chemotherapeutic studies of experimental bacterial keratitis demonstrate that shields hydrated with antibacterial agents reduce bacteria in the cornea as well as or better than frequent applications of fortified antibacterial drops. Transcorneal iontophoresis uses an electric current to drive charged drugs into the cornea. In experimentally induced bacterial keratitis, transcorneal iontophoresis of antibacterial agents is superior to topically administered ocular drops for reducing the numbers of bacteria in the cornea. Although both drug delivery systems appear to be well tolerated and nontoxic in animal models, clinical trials in patients are required to determine the usefulness of these drug delivery systems in clinical trials. Based on present experimental results, future therapy of bacterial keratitis will involve efficient drug delivery devices, the use of new antibacterial agents or combinations of presently available antibacterial agents, and careful use of adjuvant anti-inflammatory agents.