In vivo challenging of polymyxins and levofloxacin eye drop against multidrug-resistant Pseudomonas aeruginosa keratitis.

The purposes of this study were to establish a rabbit multidrug-resistant Pseudomonas aeruginosa (MDRP) keratitis model, and test the efficacy of levofloxacin, colistin methanesulfate (CL-M), colistin sulfate (CL-S) and polymyxin B (PL-B) against MDRP infection. In a rabbit eye, making a 2-mm circular corneal excision, and MDRP strain #601 or representative P. aeruginosa strain IID1210 were instilled into the corneal concavity. IID1210 was used to confirm this model developed P. aeruginosa keratitis. After MDRP keratitis developed, we treated the eyes with levofloxacin, CL-M, CL-S or PL-B eye drops. The infected eyes were evaluated by clinical score, histopathological examination and viable bacterial count (CFU). Rabbits developed MDRP keratitis reproducibly after instilled the bacteria into the corneal lesion. MDRP produced severe keratitis similarly with IID1210, as shown by slit lamp examination and clinical score. In MDRP keratitis models, clinical scores and viable bacterial counts were significantly lower in levofloxacin- and CL-M-treated groups compared with PBS-treated group, but the magnitudes of reduction were not remarkable. However, clinical scores were dramatically lowered in CL-S- and PL-B-treated groups compared with PBS-treated group. CL-S- and PL-B-treated group were kept corneal translucency and little influx of polymorphonuclear neutrophils in histopathological examination. In addition, both CL-S- and PL-B-treated groups were not detected viable bacteria in infected cornea. Using our MDRP keratitis model, we showed that topical levofloxacin and CL-M are not adequately effective, while CL-S and PL-B are efficacious in controlling MDRP keratitis. Especially, PL-B, which is commercially available eye drop, might be most effective against MDRP.