Caspase-1 Inhibitor Reduces Severity of Pseudomonas aeruginosa Keratitis in Mice

Abstract

purpose. To test an inhibitor of IL-1β converting enzyme (ICE), with or without ciprofloxacin, in a C57BL/6 mouse model of keratitis induced by Pseudomonas aeruginosa in which corneal perforation is expected.

methods. Clinical score, histopathology, myeloperoxidase (MPO) activity, bacterial counts, and ELISA analysis were used to assess the efficacy of treatment initiated at 18 hours postinfection (p.i.) with ICE inhibitor versus placebo; and with ICE inhibitor plus ciprofloxacin versus placebo plus ciprofloxacin. Efficacy of the ICE inhibitor was also tested and evaluated for clinical score in experimental corneal infection induced by a clinical isolate and a ciprofloxacin-resistant bacterial strain.

results. Clinical scores were reduced at 3, 5, and 7 days p.i. in ICE inhibitor versus placebo-treated mice; reduced scores also were observed with a combined treatment (ICE inhibitor and ciprofloxacin). Further testing (MPO assay) revealed reduced PMN number, particularly striking in ICE inhibitor and ciprofloxacin versus placebo and ciprofloxacin-treated mice. Corneal protein levels for IL-1β and MIP-2 also were reduced in mice treated with the ICE inhibitor versus placebo and in ICE inhibitor and ciprofloxacin versus ciprofloxacin and placebo-treated mice. Treatment with ICE inhibitor also reduced clinical scores after corneal infection with a clinical isolate, KEI-1025, and with a ciprofloxacin-resistant P. aeruginosa strain.

conclusions. Downregulation of IL-1β by ICE together with ciprofloxacin to kill bacteria may provide alternate therapy to current treatment.

Asight-threatening corneal disease, Pseudomonas aeruginosa (P. aeruginosa) keratitis accounts for approximately three quarters of reported cases of contact lens-associated microbial infection. ¹ Disease progresses rapidly to cause ulceration of the cornea and can potentially lead to permanent loss of vision from corneal scarring if not treated aggressively. ² Tissue damage during Pseudomonas keratitis can occur from multiple microbial ^{3 4} and host-associated factors. ^{5 6} In this regard, host inflammatory responses have been shown experimentally to play a critical role in the outcome of ocular infection with P. aeruginosa. ^{7 8 9 10 11 12 13 14}  

Management of bacterial keratitis is intended to eliminate infectious organisms and to suppress the host’s destructive inflammatory reaction. Conventionally, dual therapy ^{15 16 17} with aminoglycosides and third-generation cephalosporins in fortified ophthalmic solutions, or monotherapy ^{18 19} with fluoroquinolones is prescribed for treating Pseudomonas keratitis. The recent increased incidence of refractory bacterial keratitis resulting from antibiotic (especially fluoroquinolone) resistant P. aeruginosa strains, ^{20 21 22 23} is of great concern and also may limit future therapeutic choices. 

Corticosteroids are a standard anti-inflammatory medication to treat residual inflammation with antibacterial therapy. At present, only corticosteroids are available in ophthalmic solutions to suppress the ongoing inflammatory response after bacterial corneal infection. However, identification of the causative organism and response to antibacterial therapy (or antibiotic sensitivity) are the key restrictive factors that must be considered before initiating corticosteroid therapy. The effect (beneficial or detrimental) of corticosteroids in reducing host-mediated tissue damage has not been proven conclusively in bacterial keratitis. ^{24 25 26 27 28} Therefore, the controversial role of corticosteroids and emerging resistance of P. aeruginosa to antibiotics warrant development of new adjunctive therapeutic modalities. 

Cytokines (especially IL-1α/β and TNF-α) are optimum therapeutic targets as they can initiate and sustain inflammation in many diseases. Various strategies such as soluble receptors, antibodies, and receptor antagonists or inhibitors are used to block cytokines. These specific anti-cytokine–based therapies reduce inflammation in many chronic inflammatory or autoimmune diseases and are approved by the FDA for human use. ^{29 30 31 32} The importance of IL-1β in the pathogenesis of Pseudomonas keratitis was demonstrated in previous studies. ^{8 9 10} Persistent elevated levels of IL-1β expression are associated with the severity of corneal disease, while reduced levels (after antibody neutralization or inhibition of IL-1β receptors) result in reduced disease severity. These studies prompted us to test an inhibitor of IL-1β converting enzyme (ICE) for its therapeutic effect in bacterial keratitis. ICE, also known as caspase-1, is an intracellular protease that cleaves the precursors of IL-1β and IL-18 into active cytokines. ^{33 34} Although other proteases (including bacterial and host proteases) can process pro-IL-1β, ICE-deficient (ICE^−/−) mice are incapable of releasing mature IL-1β in response to endotoxin. ^{35 36}  

Therefore, this study tested the hypothesis that administration of an ICE inhibitor after onset of experimentally induced Pseudomonas keratitis will result in improved corneal disease by reducing the levels of IL-1β in cornea. The efficacy of an ICE inhibitor in the presence or absence of ciprofloxacin was tested in C57BL/6 (B6) mice in which perforation is the expected outcome and using a standard laboratory strain of P. aeruginosa. The efficacy of the inhibitor was further tested in bacterial corneal infection induced by a clinical isolate (KEI-1025) ³⁷ and a ciprofloxacin-resistant P. aeruginosa (19660CR) strain. 

Materials and Methods

Bacterial Strains

P. aeruginosa strain 19660, used as a standard laboratory strain, produces reproducible corneal pathology in the B6 mouse model. ^{7 8} P. aeruginosa strain 1025 (KEI-1025) was isolated in 1999 from a human microbial keratitis case at the Kresge Eye Institute, Detroit, MI ³⁷ . The laboratory-derived ciprofloxacin-resistant mutant was developed by serially passaging the wild-type P. aeruginosa strain 19660 on ciprofloxacin-containing Luria-Bertani (LB) broth to obtain ciprofloxacin resistance. ³⁸ The ciprofloxacin-resistant P. aeruginosa strain (19660CR) when compared with the parent strain exhibited a 100-fold increase in the minimum inhibitory concentration (MIC) of ciprofloxacin (0.25 mg/mL vs. 25 mg/mL) required for in vitro killing of the bacteria. 

Animal Infection

Eight-week-old female B6 mice (The Jackson Laboratory, Bar Harbor, ME) were used in these experiments because after infection, the cornea of these mice perforate within 5 to 7 days. ³⁹ The left cornea of each anesthetized mouse was scarified with three parallel 1 mm incisions using a sterile 25 5/8 gauge needle under a stereoscopic microscope. Scarified corneas were challenged topically with 1.0 × 10⁶ CFU/μL of P. aeruginosa ATCC strain 19660, a clinical isolate KEI-1025, or the 19660CR strain (developed in our laboratory) in a 5 μL dose as described previously. ³⁹ Eyes were examined macroscopically at 1 day postinfection (p.i.) and at times described below to ensure that all mice were similarly infected and to monitor the course of disease visually. All animals were treated humanely and in full compliance with the Association for Research in Vision and Ophthalmology resolution on usage and treatment of animals in research. 

ICE Inhibitor Formulations

The ICE inhibitor used in these experiments was VRT-043198, which displays potent inhibition of ICE (Ki = 0.8 nM) and selectivity >100-fold versus other non-ICE caspases. ^{40 41} VRT-043198 is the active metabolite of VX-765, which is orally active in models of skin and joint inflammation. ^{40 41} Four coded formulations (phosphate buffered saline, pH 7.2) with or without ICE inhibitor were provided by Vertex Pharmaceuticals Inc. (Cambridge, MA), for subconjunctival and topical (eye drop) administration. 

Treatment Schedule

Various concentrations, ranging from 30 to 1000 μM, of ICE inhibitor formulations as well as treatment initiation times (2, 6, 18, and 24 hours p.i.) were tested (data not shown). The optimal efficacy of the ICE inhibitor was observed at 300 μM concentration when treatment was initiated at 18 hours p.i. All formulations were nontoxic to the eye in otherwise untreated mice and had no direct (in vitro) ability to kill bacteria (Table 1) . B6 mice were injected subconjunctivally at 18 hours p.i. with 10 μL of 300 μM concentration of ICE inhibitor or placebo, followed by topical application of ICE inhibitor or placebo with or without ciprofloxacin (Ciloxan, a 0.3% solution of ciprofloxacin; Alcon, Ft. Worth, TX) at 18 hours p.i. and then three times per day for 7 days. Topical delivery of ICE inhibitor/placebo and ciprofloxacin had at least a 30-minute interval between them, to provide each sufficient time for contact with the ocular surface. Initiation of ICE inhibitor therapy at 18 hours p.i. was chosen to provide more clinically relevant data. In addition, by this time point, it would be expected that a patient would notice ocular symptoms such as haziness, discomfort, or pain, and would seek care. 

Treatment Groups and Animal Numbers

Four treatment groups included: ICE inhibitor (Group 1), placebo (Group 2), ICE inhibitor plus ciprofloxacin (Group 3), and placebo plus ciprofloxacin (Group 4). The total number of mice used (pooled from each of two separate but similar experiments) was: for clinical scores, n = 10/group/time (for each of three bacterial strains tested); histopathology, n = 6/group (only strain 19660); MPO, n = 10/group (only strain 19660); bacterial counts, n = 10/group (only strain 19660); bacterial counts after in vitro treatment with ICE inhibitor versus PBS, n = 10/group/bacterial strain; ELISA, n = 10/group (only strain 19660). 

Clinical Examination

For clinical score assessment, mice were color coded and examined in masked fashion by two independent observers at 1, 3, 5, and 7 days p.i. to grade the severity of disease visually after P. aeruginosa infection. Ocular disease was graded and clinical scores were expressed using the following scale: ⁴² 0, clear or slight opacity partially covering the pupil; +1, slight opacity fully covering the anterior segment; +2, dense opacity partially or fully covering the pupil; +3, dense opacity covering the anterior segment; and +4, corneal perforation. 

Histopathology

For histopathological examination, eyes (n = 6/group) from Groups 1 to 4 treated mice infected with strain 19660 were enucleated at 7 days p.i. Eyes were immersed in PBS, rinsed and placed in a fixative containing 1% osmium tetroxide, 2.5% glutaraldehyde, and 0.2 M Sorenson’s phosphate buffer (pH 7.4) in 1:1:1 ratio at 4°C for 3 hours. Eyes were transferred into a fresh fixative after 1.5 hours and then dehydrated in graded ethanols, embedded in Epon-araldite, sections cut, stained with a modified Richardson’s stain, and photographed as previously described. ⁴³  

Measurement of Myeloperoxidase (MPO) Activity

Samples were assayed for MPO activity as described previously. ⁴⁴ Corneas (n = 10/group) from Groups 1 to 4 treated mice infected with strain 19660 were collected at 7 days p.i. and homogenized in 1 mL of hexadecyl trimethylammonium bromide (HTAB) buffer (0.5% HTAB in 50 mM phosphate buffer, pH 6.0). The samples were subjected to three freeze–thaw cycles and then centrifuged at 16,000g for 20 minutes The supernatant was mixed with 50 mM phosphate buffer (pH 6.0), containing 16.7 mg/mL O-dianisidine hydrochloride and 0.0005% hydrogen peroxide at a 1:30 ratio in a total 3 mL volume. The change in absorbance at 460 nm was continuously monitored for 5 minutes. The results were expressed as units of MPO/cornea. One unit of MPO activity corresponds to approximately 2.0 × 10⁵ PMN. ⁴⁴  

Quantitation of Viable Bacteria in Cornea

At 7 days p.i., corneas (n = 10/group) from Group 1 to 4 treated mice infected with strain 19660 were collected and the number of viable bacteria determined. Individual corneas were homogenized in sterile PBS and aliquots (100 μL) of serial dilutions were plated onto Pseudomonas isolation agar (Difco, Detroit, MI) plates in triplicate. Plates were incubated for 24 hours at 37°C. Results were expressed as log₁₀ number of CFU/cornea ± SEM. Bacterial colony enumeration also was determined for all three bacterial strains tested in vitro in the presence of 300 μM inhibitor versus PBS (placebo). 

Quantitation of Cytokine Proteins in Corneal Homogenate

Protein levels for IL-1β and macrophage inflammatory protein (MIP)-2 were tested in Group 1 to 4 treated mice infected with strain 19660 using ELISA kits (R & D Systems, Minneapolis, MN) per instructions of the manufacturer. Corneas (n = 10/group) were removed at 7 days p.i. and immediately stored at −70°C. Before analysis, individual corneas were homogenized in 250 μL of 0.1% Tween 20-PBS with a glass Kontes pestle (Fisher, Itasca, IL) centrifuged at 5000g for 10 minutes at 4°C, and supernatants were used to quantify IL-1β and MIP-2 proteins. Results are reported as pg/mL/cornea. 

Statistical Analysis

The difference between conditions at individual time points was tested (Groups 1 to 4) by the Mann–Whitney U test (GraphPad Prism, San Diego, CA). An unpaired, two-tailed Student’s t-test was used to determine statistical significance for data from MPO assay, bacterial counts, and ELISA analyses between treated and control groups. Mean differences were considered significant at P ≤ 0.05. Experiments were repeated at least twice to ensure reproducibility; pooled data from two separate experiments are shown. 

Results

Efficacy Testing of ICE Inhibitor, VRT-043198, in Corneal Infection Induced by Three Strains of P. aeruginosa

P. aeruginosa Strain 19660.

By visual observation, ICE inhibitor-treated mice showed a significant decrease in disease severity at 3 (P = 0.05), 5 (P = 0.001), and 7 (P = 0.0001) days p.i. compared to placebo-treated mice (Fig. 1A) . Combined therapy with the ICE inhibitor and ciprofloxacin resulted in significantly lower clinical scores at 3 (P = 0.007) and 5 (P = 0.014), but not 7 (P = 0.28) days p.i. compared to the placebo and ciprofloxacin-treated group (Fig. 1B) . Comparisons also were made between ICE inhibitor only versus ciprofloxacin only and ICE inhibitor only versus ICE inhibitor plus ciprofloxacin treated mice at 7 days p.i. (Table 2) . A significant difference (P = 0.003) between ciprofloxacin only and ICE inhibitor only treated groups and between ICE inhibitor plus ciprofloxacin versus ICE inhibitor treated mice (P = 0.0007) was detected. 

P. aeruginosa Strain KEI-1025.

In the ICE inhibitor versus placebo-treated KEI-1025 infected mice, a significant decrease in disease severity (clinical score) was seen at 3 (P = 0.004), 5 (P = 0.015), and 7 (P = 0.007) days p.i. (Fig. 1C) . Combined treatment with the ICE inhibitor and ciprofloxacin versus placebo and ciprofloxacin also showed significantly decreased clinical scores at 3 (P = 0.007) and 5 (P = 0.019) but not at 7 (P = 0.09) days p.i. (Fig. 1D) . Comparisons also were made between ICE inhibitor only versus ciprofloxacin only and between ICE inhibitor only versus ICE inhibitor plus ciprofloxacin-treated mice at 7 days p.i. (Table 2) . Significant difference (P = 0.04) was observed between ciprofloxacin only and ICE inhibitor only treated groups and between ICE inhibitor plus ciprofloxacin-treated versus ICE inhibitor only treated mice (P = 0.001). 

P. aeruginosa Strain 19660CR.

Corneas infected with a ciprofloxacin-resistant strain of P. aeruginosa (19660CR) and treated with the ICE inhibitor showed significantly lower clinical scores at 3 (P = 0.02) and 7 (P = 0.004) days p.i. compared to the placebo-treated mice (Fig. 1E) . Likewise, significantly lower clinical scores at 5 (P = 0.02) and 7 (P = 0.004) days p.i. were observed in the ICE inhibitor and ciprofloxacin treated compared to the placebo and ciprofloxacin-treated group (Fig. 1F) . Comparisons also were made between ICE inhibitor only versus ciprofloxacin only and between ICE inhibitor only versus ICE inhibitor plus ciprofloxacin-treated mice at 7 days p.i. (Table 2) . A significant difference (P = 0.004) was observed between ciprofloxacin only and ICE inhibitor only treated groups. No significant difference (P = 1.0) between ICE inhibitor only and ICE inhibitor plus ciprofloxacin-treated mice was observed, confirming the in vivo resistance of this strain to ciprofloxacin. 

Slit Lamp Microscopy (P. aeruginosa Strain 19660)

Slit lamp microscopy (Figs. 2A and 2B) of mice infected with strain 19660 and treated with the ICE inhibitor versus placebo at day 7 p.i. provided photographic documentation of the ocular disease response. In the ICE inhibitor-treated cornea (Fig. 2A) , noticeably less cellular infiltration was observed, mainly localized in the central cornea over the pupil. Opacity seen in the inferior corneal region was due to infiltrating cells that had gravity settled in the anterior chamber. In contrast, all mice treated with the placebo exhibited corneal perforation (Fig. 2B) . Slit lamp examination in corneas treated with the ICE inhibitor and ciprofloxacin (Fig. 2C) showed only slight corneal opacity compared to placebo and ciprofloxacin-treated corneas (Fig. 2D) , which showed more cellular infiltrate in the cornea and the anterior chamber. 

Histopathology (P. aeruginosa Strain 19660)

Since clinical scoring is a visual observation, histopathology also was used to provide further confirmation of disease progress at 7 days after infection with strain 19660 and ICE inhibitor treatment. ICE inhibitor-treated eyes showed markedly reduced infiltrating cells in the corneal stroma with a minimal anterior chamber inflammatory cell response (Fig. 3A) . In contrast, the placebo-treated B6 mice showed a heavy cellular infiltrate in the cornea with complete denudation of the corneal epithelium, central stromal degradation, severe edema, severe anterior chamber inflammation, and perforation (Fig. 3B) . Corneas treated with the ICE inhibitor and ciprofloxacin showed few inflammatory cells along the corneal endothelium and in the anterior chamber (Fig. 3C) compared to a typical eye treated with the placebo and ciprofloxacin (Fig. 3D) which showed a heavier inflammatory cell infiltrate in the anterior chamber and adherent to the corneal endothelium. 

Polymorphonuclear (PMN) Infiltration (P. aeruginosa Strain 19660)

Since PMN infiltration into cornea after bacterial infection is associated with bacterial clearance, as well as corneal destruction, if they persist, ^{7 14} MPO activity was assayed to quantify PMN infiltration in the cornea of the ICE inhibitor versus placebo-treated mice infected with strain 19660 at 7 days p.i. ICE inhibitor-treated mice showed a significantly lower (P = 0.04) number of PMN compared to the placebo-treated mice. Similarly in the ICE inhibitor plus ciprofloxacin versus placebo plus ciprofloxacin-treated mice, MPO activity was significantly reduced (P = 0.0024) in cornea at 7 days p.i. The units of MPO activity per cornea (± SEM) from data pooled from two separate, similar experiments are shown in Figure 4 . 

Bacterial Clearance (P. aeruginosa Strain 19660)

Viable bacterial plate counts were determined in cornea (n = 10/group) from ICE inhibitor (3.291 + 1.347 log₁₀ CFU) versus placebo-treated (7.213 + 0.2149 log₁₀ CFU) groups infected with strain 19660 at 7 days p.i. A significantly decreased (P = 0.02) number of viable bacteria in the corneas of ICE inhibitor-treated when compared to the placebo-treated group was detected. No bacterial colonies were isolated from plate counts in corneas from ICE inhibitor and ciprofloxacin or placebo and ciprofloxacin-treated groups at 7 days p.i. (data not shown). 

Cytokine/Chemokine Protein Expression Using P. aeruginosa Strain 19660

Protein levels for IL-1β and MIP-2 in the ICE inhibitor versus placebo-treated groups were determined at 7 days p.i. using ELISA analysis (Fig. 5) . Significantly lower protein levels for IL-1β (P = 0.023; Fig. 5A ) and MIP-2 (P = 0.012; Fig. 5C ) were detected in the ICE inhibitor compared to the placebo-treated group. Ciprofloxacin treatment markedly reduced the levels of both IL-1β and MIP-2, presumably due to a lessening of the pro-inflammatory stimulus achieved by bacterial killing (Figs. 5B and 5D) . Nevertheless, in the ICE inhibitor and ciprofloxacin-treated group, protein levels for IL-1β (P = 0.036; Fig. 5B ) and MIP-2 (P = 0.04; Fig. 5D ) also were significantly reduced compared to the placebo and ciprofloxacin-treated group. 

Discussion

Bacterial keratitis remains a major cause of sight-limiting scarring and visual impairment, especially in extended-wear contact lens users, ^{45 46 47} despite the efficacy of broad spectrum antibacterial agents. Over 30 million people use contact lenses in the United States alone, ⁴⁸ and 1 in 2500 daily wear contact lens users and 1 in 500 extended-wear contact lens users develop bacterial keratitis each year. ⁴⁷ Traditionally, broad spectrum antibiotics (often ciprofloxacin) therapy is promptly instituted in keratitis cases, before obtaining results from culture specimens to identify a causative organism. Although antimicrobial treatment is often able to render a sterile cornea, it does not guarantee a clear visual axis, due to residual host-derived inflammation. The latter often may necessitate the use of corticosteroids to restore corneal clarity. In some cases, use of corticosteroids may have potential adverse effects, including delayed corneal wound healing. ^{49 50}  

In this study we tested and found that use of an ICE inhibitor only could reduce corneal disease when compared with placebo treatment. Although use of the ICE inhibitor alone did not fully stop inflammation or reduce its levels as much as ciprofloxacin, ICE inhibitor versus placebo-treated eyes did not perforate. In this regard, treatment with the ICE inhibitor was most efficacious in adjunctive therapy to complement the bacterial killing effects of ciprofloxacin and, together with the antibiotic, synergistically appeared to downregulate the host inflammatory response better than use of either of the agents alone. Treatment with the ICE inhibitor, VRT-043198, also contributed, indirectly, to lessening bacterial growth (as it had no ability to kill bacteria), perhaps because bacteria are not able to disseminate in a cornea in which damage is reduced. In addition, there may be other as yet untested mechanisms that contribute to bacterial clearance, for example, in ICE^−/− mice that endogenously lack caspase-1 versus wild-type mice, enhanced corneal epithelial cell apoptosis has been reported ⁵¹ . This too could serve to contain the bacterial infection and prevent stromal damage. In addition, ICE inhibitor (as well as ciprofloxacin) treatment reduced levels of IL-1β that may in turn decrease inducible nitric oxide synthase (iNOS). ⁵² This effect could then lead to a decrease in sustained nitric oxide (nitrite) production which is potentially toxic, dependent on its concentration and the microenvironment in which it is produced. ⁵³ ICE inhibitor-treated mice showed significantly reduced levels of not only IL-1β, but MIP-2 (chemoattractants for PMN); reduced PMN infiltration and bacterial load compared to the placebo-treated group. Ciprofloxacin treatment significantly reduced levels of these cytokines with even greater effect, but most striking was the significant reduction of both cytokine levels and PMN number when the combination therapy of ICE inhibitor and ciprofloxacin was used. In addition, histopathological examination of the ICE inhibitor-treated group showed markedly reduced infiltrating cells with intact corneal epithelium, whereas all the corneas of the placebo-treated mice had perforated. Addition of the ICE inhibitor with topical antibiotic (ciprofloxacin) produced further improvement of corneal disease outcome, most evidenced by the reduction in inflammatory infiltrate in the anterior chamber and associated with the corneal endothelium (compare Figs. 3C and 3D ). 

Since the ICE inhibitor reduced the inflammatory infiltrate and prevented perforation (when compared with placebo) after bacterial corneal infection using a P. aeruginosa ATCC strain 19660, the next logical step was to begin to test at least one more P. aeruginosa strain, preferably a clinical isolate. Our results demonstrate that the ICE inhibitor versus placebo treatment lessens disease (no perforation) not only against a standard ATCC laboratory strain (19660), but also against a clinical isolate (KEI-1025). Obviously, it would be advisable to test additional clinical isolates as treatment outcome may differ, dependent on the strain of bacteria. 

To further test the ICE inhibitor, a ciprofloxacin-resistant strain (19660CR) derived from the parent 19660 strain was produced and tested. The strain appeared normal in growth characteristics, and antibiotic resistance was maintained at least through three passages (unpublished results). Initial testing showed that clinical scores were significantly reduced in the ICE inhibitor versus placebo-treated corneas after infection with strain 19660CR. However, virulence of this mutant was decreased compared to the parent strain during in vitro generation. While this is not unusual and has been reported previously, ³⁹ use of a more virulent ciprofloxacin-resistant strain (clinical isolate) and rigorous study of cytokine and chemokine levels will be needed to confirm our initial observations. Various studies ^{20 21 22} have shown a link between in vitro antibiotic resistance and clinical failure to respond to antibiotic in keratitis patients. Garg et al. ²¹ reported that of 141 culture-proven cases of Pseudomonas keratitis, 22 cases were caused by isolates resistant to ciprofloxacin (mean MIC 43 mg/mL). Of the 19 (of 22) cases treated initially with ciprofloxacin, 15 (76.7%) worsened or showed no clinical improvement after 3 days of intensive therapy and required modification of antibiotic therapy, corneal grafting, or evisceration. Increasing incidence of antibiotic resistance of Pseudomonas and failure to respond to antibacterial therapy leading to adverse outcomes provide strong reasons to search for new therapeutic strategies. An ICE inhibitor could be a novel therapeutic strategy for antibiotic-resistant Pseudomonas keratitis cases. 

Caspase-1 (ICE) also is required for processing of IL-18, as well as IL-1. ⁵⁴ The former cytokine has been shown to be protective (via induction and tight regulation of IFN-γ) in a model of bacterial keratitis in the BALB/c mouse (whose cornea heals after similar bacterial infection). ^{12 55} Levels of IL-18 have not been determined in the B6 mouse cornea, but after infection, sustained levels of IL-12 (never detected in the BALB/c mouse cornea) are detected in the B6 mouse and it is IL-12 that regulates sustained IFN-γ production, contributing to corneal perforation in this model. ^{43 55} Because the IL-18 receptor and IL-1 receptor are very similar and signal through closely related, if not identical, pathways, ⁵⁴ future experiments may be required to determine whether treatment with an ICE-inhibitor interferes with IL-18 induction of IFN-γ and how this is affected by combined treatment with the ICE inhibitor and ciprofloxacin. 

In summary, these findings demonstrate that ICE inhibitor therapy can reduce P. aeruginosa-induced keratitis when compared with placebo treatment, that disease does progress, but corneas do not perforate. Evidence is also provided that combination treatment using an antibiotic such as ciprofloxacin together with the ICE inhibitor provides better disease outcome than antibiotic alone. 

 Supported by Vertex Pharmaceuticals, Inc., and in part by NIH Grants P30EY04068 and R01EY02986.

 Submitted for publication January 14, 2004; revised May 19, 2004; accepted May 28, 2004.

 Disclosure: A. Thakur, Vertex Pharmaceuticals, Inc. (F); R.P. Barrett, Vertex Pharmaceuticals, Inc. (F); J.A. Hobden, Vertex Pharmaceuticals, Inc. (F); L.D. Hazlett, Vertex Pharmaceuticals, Inc. (F)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Effect of ICE Inhibitor on Bacterial Colonization In Vitro in Presence of ICE Inhibitor or PBS

Bacterial Strains Tested	Bacterial Colonies Recovered After 18 Hour Culture
		ICE Inhibitor (300 μM)	PBS
P. aeruginosa 19660	103	97
P. aeruginosa KEI-1025	89	112
P. aeruginosa-19660CR	115	109

Figure 1.

 

Ocular disease grades. (A), (C), and (E): clinical scores in the ICE inhibitor versus placebo-treated mice at 1, 3, 5, and 7 days p.i. with P. aeruginosa 19660, KEI-1025, and 19660CR strains, respectively. (B), (D), and (F): clinical scores in the ICE inhibitor plus ciprofloxacin versus placebo plus ciprofloxacin treated mice at 1, 3, 5, and 7 p.i. with 19660, KEI-1025, and19660CR strains.

Table 2.

 

Comparison of Clinical Scores in ICE Inhibitor Only versus Ciprofloxacin Only, and ICE Inhibitor Only versus ICE Inhibitor Plus Ciprofloxacin-Treated Mice at 7 Days p.i.

Treatment Groups	19660	P value	KEI-1025	P value	19660CR	P value
ICE inhibitor only	2.2 ± 0.13		1.9 ± 0.15		0.95 ± .22
Ciprofloxacin only	1.5 ± 0.18	0.003	1.3 ± 0.15	0.04	1.6 ± 0.14	0.004
ICE inhibitor only	2.2 ± 0.2		1.9 ± 0.15		0.95 ± .22
ICE inhibitor + Ciprofloxacin	1.2 ± .13	0.0007	0.9 ± 0.10	0.001	0.95 ± 0.45	1.0

Figure 2.

 

Slit lamp photographs at 7 days p.i. with P. aeruginosa strain 19660. Representative eyes were photographed: (A) considerably less opacity was observed in cornea treated with the ICE inhibitor; (B) cornea was perforated in the placebo-treated mice; (C) cornea showed slight opacity in the ICE inhibitor plus ciprofloxacin-treated mice; and (D) in the placebo plus ciprofloxacin-treated mice, more opacity was observed in both cornea and anterior chamber. Magnification, ×8.75.

Figure 3.

 

Histopathological examination of P. aeruginosa strain 19660 infected cornea at 7 days p.i. A modest cellular infiltrate in the corneal stroma with an intact corneal epithelium and minimal anterior chamber response in the ICE inhibitor-treated mice (A); a heavy cellular infiltrate in the stroma with complete destruction of the corneal epithelium, severe anterior chamber response, and perforation in the placebo-treated mice (B); few infiltrating cells in the anterior chamber and associated with the endothelium in the ICE inhibitor plus ciprofloxacin-treated eye (C); a heavier cellular infiltrate in the anterior chamber in the placebo plus ciprofloxacin-treated mice when compared with the micrograph in Figure 3C (D). Magnification, ×165.

Figure 4.

 

Corneal MPO activity at 7 days p.i. with P. aeruginosa strain19660. Individual corneas were collected for MPO analysis. Results are presented as units MPO activity/cornea ± SEM (ICE inhibitor versus placebo P = 0.04; ICE inhibitor plus ciprofloxacin versus placebo plus ciprofloxacin, P = 0.0024).

Figure 5.

 

Corneal IL-1β and MIP-2 protein levels at 7 days p.i. with P. aeruginosa strain 19660. Corneas treated with the ICE inhibitor only (A) or together with ciprofloxacin (B) showed significantly reduced (P = 0.023, P = 0.036, respectively) protein levels for IL-1β compared to corneas treated with placebo only or placebo plus ciprofloxacin. Corneas treated with the ICE inhibitor only (C) or together with ciprofloxacin (D) showed significantly decreased (P = 0.012, P = 0.04, respectively) MIP-2 protein levels compared to corneas treated with the placebo or placebo plus ciprofloxacin. Results are presented as pg/mL/cornea.