Efficacy of Gemifloxacin for the Treatmentof Experimental Staphylococcus aureus Keratitis

Xianggen Wu, Haoran Jiang, Yuanyuan Xu, Wenjie Yue, Lingling Yang,Zicheng Song, Hao Chen, and Ting Liu

Abstract

Purpose: The objective of the present study was to evaluate the effectiveness of topically applied gemifloxacin forthe treatment of experimental Staphylococcus aureus keratitis in a rabbit model.Methods: Rabbit corneas were intrastromally injected with *100 colony-forming units (CFU) of S. aureusATCC25923. Eight hours (early treatment) or 16 h (late treatment) after the injection, 1 topical drop of balancedsalt solution (BSS), gemifloxacin ophthalmic solution (0.5%), levofloxacin ophthalmic solution (0.5%), or gati-floxacin eye gel (0.3%) was applied to each eye every 15min for 5 doses and then, every 30min for 14 doses. Theeyes were examined both before and after treatment. The corneas were harvested from treated and untreatedrabbits for the quantitation of bacteria and histological observation.Results: In the early-treatment groups, all 3 fluoroquinolones significantly lowered the clinical severity ofinfection and the median erosion area of the cornea compared with the BSS control (P= 0.000). In the late-treatment groups, gemifloxacin and levofloxacin did not cause a significant reduction in clinical scores comparedwith the BSS control (P = 0.107 and 0.531, respectively), but the gatifloxacin caused a significant reduction inclinical scores compared with the BSS control (P= 0.011). The median erosion area significantly decreased withtreatment with gemifloxacin, gatifloxacin, and levofloxacin in both early- and late-treatment groups, whencompared with the control group (P 0.022). In the early-treatment groups, the gemifloxacin, gatifloxacin, andlevofloxacin groups had significantly lower CFU recovered from the corneas compared with the control group(P < 0.01), while in the late-treatment groups, levofloxacin failed to reduce the CFU recovered from the corneascompared with the control group (P = 0.695). The minimal inhibitory concentrations for gemifloxacin, gati-floxacin, and levofloxacin against S. aureus ATCC25923 were 0.0625, 0.0625, and 0.125mg/L, respectively.Conclusions: Gemifloxacin, similar to gatifloxacin and levofloxacin, can significantly lower the clinical severityand CFU per cornea observed in S. aureus keratitis when early treatment is implemented. Significantly, gemi-floxacin showed a significant efficacy improvement in reducing the bacterial load recovered from the corneas inthe late-treatment experiment.

Introduction

Staphylococcus aureus is a major cause of bacterialkeratitis, a sight-threatening condition.13 S. aureus ocularinfections can cause severe inflammation, pain, corneal per-foration, scarring, and loss of visual acuity.1,4 Treatment forthese ocular bacterial infections is primarily empiric withbroad-spectrum antibiotics that are effective against the mostcommon bacteria associated with these ocular infections.5

However, with the increasing and widespread use of topicalantibiotics, there is justifiable concern over the emergence ofresistant organisms.6 Strategies for the prevention of an in-

crease in ocular pathogen resistance should be developedand implemented. In addition, new antibiotics and new an-tibiotic formulations are needed to manage future cases of S.aureusinduced keratitis.4

Fluoroquinolones have been investigated for a variety ofinfections caused by S. aureus.7 They possess concentration-dependent bactericidal activity and rapidly penetrate to sitesof infection.8 However, the rapid emergence of resistance toolder generations of these compounds has raised concerns.8

Newer-generation fluoroquinolones have improved activityat topoisomerase IV, which is the primary target site in S.aureus.9 Gemifloxacin is a fluoronaphthyridone with a

State Key Laboratory Cultivation Base, Shandong Provincial Key Laboratory of Ophthalmology, Shandong Eye Institute, ShandongAcademy of Medical Sciences, Qingdao, P.R. China.

JOURNAL OF OCULAR PHARMACOLOGY AND THERAPEUTICSVolume 28, Number 4, 2012 Mary Ann Liebert, Inc.DOI: 10.1089/jop.2011.0229

420420420

pyrrolidine group at the C7 position.10 It has a broad spec-trum of activity associated with improved Gram-positivepotency and retention of the good Gram-negative activityshown by older fluoroquinolones.11 Gemifloxacin was one ofthe most effective antibiotics against conjunctival bacteriaisolated from refractive surgery patients.12 It is generallythought that topical ocular administration of antibiotics willproduce higher antibiotic concentrations in ocular tissue thanthose achieved in the serum after systemic administration.9

However, until now, there have been no ophthalmic for-mulations of gemifloxacin available in clinics. In our previ-ous study, the prophylactic effect of topical gemifloxacin forexperimental endophthalmitis was evaluated, and the resultsshowed that gemifloxacin might be as effective as moxi-floxacin and levofloxacin for topical prophylaxis and for thetreatment of S. aureusinduced endophthalmitis in the rabbitmodel.13

The purpose of this study was to compare levofloxacin,gatifloxacin, and gemifloxacin for their in vivo effectivenessin treating experimental S. aureus keratitis. The stock strain ofS. aureus, ATCC25923, used in this study was selected, be-cause this pathogen is usually the first or second mostcommonly isolated cause of bacterial keratitis in manycountries in the world. The treatment times were chosen tocompare the effectiveness of fluoroquinolones during both aperiod of active bacterial replication (early treatment) and asubsequent phase with reduced bacterial replication (latetreatment) in the cornea.4,14,15 The results demonstrate thepotency of these antibiotics and reveal the value of gemi-floxacin as a potential new ocular therapy.

Methods

Bacterial cultures

A stock strain of S. aureus ATCC25923 (Qingdao EyeHospital, Qingdao, China) was selected as the test organismand was incubated overnight at 37C. Several isolated colo-nies were inoculated into 10-mL tryptic soy broth (TSB) andincubated for 1822 h at 37C and 200 rpm in a shaking airincubator. This culture was diluted 1:100 in TSB and incu-bated for*2h at 37C and 200 rpm in a shaking air incubatoruntil the optical density of the culture was A600= 0.2. Serial 10-fold dilutions of this logarithmic-phase culture were made inTSB. The 10- 4 dilution, equivalent to 100 colony-forming units(CFU) per 5mL, was used for the corneal infection of rabbits.Accuracy of the inoculum was verified by plating 0.1mL ofeach serial 10-fold dilution in triplicate on TSA.

Corneal infection of rabbits

Male and female New Zealand albino rabbits were ob-tained from Qingdao Kangda Foodstuffs Co., Ltd. (LicenseNo. SCXK [Lu] 20070023). Animal care and procedures wereconducted according to the Principles of Laboratory AnimalCare. The experimental animals weighed between 2 and2.5 kg and were housed individually in an air-conditionedand light-controlled room at 25C 2C and at 70% 5%relative humidity. The rabbits were given standard pelletfeed and provided with water ad libitum. All animals werehealthy and free of clinically observable ocular abnormali-ties. The animal study was approved by the Shandong EyeInstitute Ethics Committee for Animal Experimentation,Qingdao, Shandong, China.

Each rabbit was systemically anesthetized by an intra-muscular injection with 25mg/kg ketamine hydrochlorideand 25mg/kg chlorpromazine. The left eye of each rabbitwas designated as the experimental eye. Topical anesthesiawas achieved with the administration of 1 drop of propar-acaine ophthalmic solution to the experimental eye. Eachexperimental eye was held steady with clamping forceps,and 5mL of S. aureus ATCC25923 (containing 100 CFU) wasinjected directly into the corneal stroma with a 33-gaugeHamilton needle (30 bevel, 13mm) on a 10-mL Hamiltonsyringe (Hamilton Company).

Treatment regimen

Early treatment. Eight hours postinfection (PI), everyrabbit was examined to collect animals with similar severitiesof infection. The rabbits were then randomly assigned togroups to verify that each group had similar scores beforetreatment. They were then divided into 4 treatment groupsand 1 baseline group, with 12 rabbits in each group. Thetreatment groups were as follows: sterile balanced salt so-lution (BSS; Qingdao Huaren Pharmaceutical Co., Ltd.) ascontrol group, gatifloxacin eye gel (Diyou, 3mg/g, con-taining carbomer and hyaluronic acid as eye gel base,chlorobutanol as preservatives; Shenyang Sinqi Pharmaceu-tical Co., Ltd.), levofloxacin ophthalmic solution (Cravit,5mg/mL; Santan Pharmaceutical Co., Ltd.), and gemi-floxacin ophthalmic solution (sterilized water solution con-taining gemifloxacin mesylate 6.23mg/mL, equivalent togemifloxacin 5mg/mL, with sodium hydroxide to adjust pHto *5.30, and with sodium chloride to adjust osmolality to*300mOsm, the gemifloxacin powder was provided byDalian Meilun Biology Technology Co., Ltd.). Gatifloxacineye gel and levofloxacin ophthalmic solution were used intheir commercially available formulations. One drop wasplaced on each experimental eye every 15min for 5 dosesand then, every 30min for 14 doses (19 total doses over8 h).16 One hour after the last drop, the rabbit experimentaleyes were examined and scored.

Late treatment. Sixteen hours PI, the rabbits were exam-ined and divided into 5 groups that were similar to the early-treatment groups. The same treatment schedule as wasfollowed in the case of the early-treatment groups was car-ried out.

Examination and scoring of rabbit eyes

The experimental eyes were examined by a masked ob-server (Wenjie Yue, a trained ophthalmologist) at 8 h (pre-treatment) and 17 h (post-treatment) PI in the early-treatmentgroups and at 16 h (pre-treatment) and 25 h (post-treatment)PI in the late-treatment groups. Five parameters were as-sessed to determine the severity of the infection: conjunctivalinjection, conjunctival chemosis, corneal infiltrate, cornealedema, and hypopyon. Each parameter was given a grade of0 (normal) to 4 (maximally severe) by the observer, and the 5parameters were added together to achieve a total score witha theoretical maximum of 20. The first 2 parameters, con-junctival injection and chemosis, were graded without theuse of a slit-lamp microscope, while the other 3 parameterswere observed and scored under a slit-lamp microscope.After observation of the 5 parameters just mentioned, each

GEMIFLOXACIN TREATMENT KERATITIS 421

experimental eye was stained with fluorescein to aid inmeasuring corneal epithelial erosions.

Euthanasia and tissue harvest

For the quantitation of baseline CFU per experimentalcornea, to form the baseline group, 9 rabbits were euthanized(at 8 h PI in the early-treatment groups or at 16 h PI in thelate-treatment groups). In addition, the animals were eu-thanized at 17 h PI in the early-treatment groups or at 25 h PIin the late-treatment groups to evaluate the effects of thetreatment and to conduct the final examinations. Nine ex-perimental corneas in each group were harvested, homoge-nized, serially diluted in BSS, and plated on TSA. The plateswere incubated overnight at 37C, and the colonies werecounted. The other 3 experimental eyes of each treatmentgroup were removed for histology. The whole eyes wereharvested and fixed in 10% formaldehyde solution for at least24h and dehydrated with gradient alcohol. The eyes wereprepared for hematoxylineosin staining using routine meth-ods and analyzed with light microscopy to examine the con-junctiva, the cornea, the limbus, the chamber angles, and thesclera. The presence of neutrophils, lymphocytes, macro-phages, fibroblasts, and giant cells was regarded as evidenceof tissue response. Every sample was treated simultaneouslyto reduce variations in the fixation procedure.

Minimal inhibitory concentration assays

The minimal inhibitory concentrations (MICs) of gati-floxacin, levofloxacin, and gemifloxacin against S. aureusATCC25923 were determined using the macrodilution brothmethod.14,15 The MIC for each antibiotic was determined tobe the lowest concentration at which no turbidity was ob-served, taking into account the final 2-fold dilution of eachantibiotic when the bacterial suspension was added.

Statistical analysis

Data were analyzed by using SPSS 10.0 software. The totalclinical scores in early or late treatment were analyzed using

the KruskalWallis test, and comparisons between early andlate treatment were determined by using the MannWhitneytest. The average corneal epithelial erosion area in early orlate treatment was analyzed using the multiple comparisonin analysis of variance (ANOVA), and comparisons betweenearly and late treatment were determined using the inde-pendent-samples t-test. For quantitative analysis of the effi-cacy of the experimental keratitis treatment, bacterial loadsper cornea were transformed to logarithmic values; themultiple comparison in ANOVA was used to compare theefficacy in early or late treatment; and comparisons betweenearly and late treatment were determined using the inde-pendent-samples t-test. A P value < 0.05 was consideredsignificant.

Results

Rabbit keratitis model

Early treatment. Eight hours after the inoculation ofS. aureus, obvious symptoms of infection, such as photo-phobia, conjunctival congestion, edema, purulent discharge,and even hypopyon, were observed. There was no obviouscorneal infiltrate or edema. The median total clinical score 8 hafter inoculation was 4. If effective treatment had not beencarried out, such as within the control group, then the con-junctival congestion, edema, and purulent discharge becamemuch more severe. At this point, corneal infiltrate and edemacould be observed in all eyes in the control group, and hy-popyon was also found in 6 eyes. The median clinical scorefor the control group at the end of the observation periodwas 13 (P = 0.000 when compared with the baseline group).After treatment (17 h PI), gemifloxacin, gatifloxacin, and le-vofloxacin caused a significant reduction in clinical scorescompared with the BSS control (Table 1; P = 0.001 to levo-floxacin, and P= 0.000 to gemifloxacin and gatifloxacin, re-spectively). The clinical score from the gemifloxacin groupwas not significantly different compared with the levo-floxacin group (P = 0.727), but the gatifloxacin group causeda greater reduction in clinical scores compared with the

Table 1. Median Clinical Scores (Range) of Rabbit Eyes After the Earlyand Late Treatments (n = 12 Per Group)

Conjunctivalinjection

Conjunctivalchemosis

Cornealinfiltrate

Cornealedema Hypopyon Total

Early treatmentBaseline 1 (12) 2 (12) 1 (01) 0 (0) 0 (0) 4 (25)a

Control 4 (34) 4 (34) 2 (24) 2.75 (24) 0.5 (02) 13 (1019)Gemifloxacin 2 (13) 1.5 (13) 2 (12.5) 1.5 (12) 0 (0) 6 (49)a

Gatifloxacin 0.5 (0.52) 0.5 (0.51) 1 (0.52) 0.5 (02.5) 0 (0) 2.75 (1.565)a,b

Levofloxacin 2 (0.53) 2 (0.53) 1 (0.53) 1.5 (0.54) 0 (03) 6.5 (216)a

Late treatmentBaseline 3.0 (34) 3.75 (34) 3 (24) 3 (33.5) 2 (03) 14 (1317)Control 4 (4) 4 (24) 3 (24) 3 (24) 2 (14) 15.5 (1120)Gemifloxacin 4 (24) 4 (04) 2 (03) 3 (04) 1 (04) 14 (219)Gatifloxacin 3.5 (24) 3 (14) 2 (13) 3 (04) 1.5 (04) 12 (419)c

Levofloxacin 4 (24) 4 (24) 3 (14) 3 (14) 2 (04) 15 (620)

aSignificant reduction in clinical scores compared with the BSS control (P= 0.001 to levofloxacin, and P = 0.000 to baseline, gemifloxacin, andgatifloxacin, respectively).

bSignificant reduction in clinical scores compared with gemifloxacin and levofloxacin groups (P= 0.000 to gemifloxacin and 0.003 tolevofloxacin, respectively).

cSignificant reduction in clinical scores compared with the BSS control (P= 0.041).BSS, balanced salt solution.

422 WU ET AL.

gemifloxacin and levofloxacin groups (P = 0.000 and 0.003,respectively). The average erosion area for each group at thispoint of time was as follows: gemifloxacin, 3.36mm2; gati-floxacin, 0.36mm2; levofloxacin, 4.98mm2; control group,15.11mm2; and baseline group, 1.95mm2. The average ero-sion area significantly decreased with treatment withgemifloxacin, gatifloxacin, and levofloxacin by 17 h PI,when compared with the control group (all 3 P = 0.000).Figure 1 shows representative eyes from each treatmentgroup at 17 h PI.

Late treatment. Sixteen hours after the inoculation ofS. aureus, severe symptoms of infection, such as inability toopen the eyes, purulent discharge, severe conjunctival con-gestion and edema, hypopyon, and corneal infiltrate andedema, were observed. The median clinical score 16 h afterinoculation was 14. If effective treatment had not been car-ried out, such as within the control group, then the infectionwas ongoing, and the median clinical score for the controlgroup at the end of the observation period was 15.5 (P= 0.212when compared with the baseline group). After treatment(25 h PI), gemifloxacin and levofloxacin did not cause a sig-nificant reduction in clinical scores compared with the BSScontrol (P = 0.115 and 0.658, respectively), but gatifloxacincaused a significant reduction in clinical scores compared

with the BSS control (P = 0.041). There were no significantdifferences in scores when gemifloxacin, gatifloxacin, andlevofloxacin were compared (P 0.093). When comparedwith early treatment, gemifloxacin, gatifloxacin, and levo-floxacin were not as effective in the reduction of clinicalscores during late treatment as they had been during earlytreatment (P = 0.001, 0.000, and 0.000, respectively). The av-erage erosion area for each group at this point of time was asfollows: gemifloxacin, 5.17mm2; gatifloxacin, 8.50mm2; le-vofloxacin, 7.96mm2; control group, 16.23mm2; and baselinegroup, 16.94mm2. The average erosion area also significantlydecreased with treatment with gemifloxacin, gatifloxacin,and levofloxacin at 25 h PI, when compared with the controlgroup (P = 0.001, 0.022, and 0.018, respectively), and gemi-floxacin and levofloxacin were similarly effective, but gati-floxacin was less effective in the reduction of the cornealepithelial erosion area during late treatment as they had beenduring early treatment (P = 0.590, 0.216 and 0.016, respec-tively). Figure 1 also shows representative eyes from eachtreatment group at 25 h PI.

Corneal CFU recovery

Early treatment. The gemifloxacin-, gatifloxacin-, and le-vofloxacin-treatment groups had significantly lower CFUrecovered from the corneas compared with the control group(P < 0.01). Gemifloxacin and gatifloxacin significantly re-duced the corneal CFU compared with levofloxacin(P = 0.000 and 0.005, respectively), and the gatifloxacin-treatment group had significantly reduced CFU recoveredfrom the corneas compared with the gemifloxacin groups(P = 0.000). (Fig. 2).

Late treatment. The gemifloxacin- and gatifloxacin-treatment groups had significantly lower CFU recoveredfrom corneas compared with the control group (bothP = 0.000), while the levofloxacin group failed to reduce theCFU recovered from corneas compared with the controlgroup (P = 0.695). The gemifloxacin group had a lower CFUrecovered but was not significantly different when comparedwith gatifloxacin (P = 0.078). When compared with earlytreatment, gemifloxacin, gatifloxacin, and levofloxacin were

FIG. 1. The typical clinical presentation of eyes infectedafter early and late treatment.

FIG. 2. The number of corneal colony counts [mean standard deviation (SD)] for each treatment in the Staphylo-coccus aureus keratitis model with early treatment (n = 9).*Significantly lower colony-forming units (CFU) recoveredfrom the corneas compared with the control group (P < 0.01).#Significantly reduced corneal CFU compared with levo-floxacin (P = 0.000 to gatifloxacin and 0.005 to gemifloxacin,respectively). Significantly reduced CFU compared with thegemifloxacin groups (P = 0.000).

GEMIFLOXACIN TREATMENT KERATITIS 423

not effective in lower CFU during late treatment as they hadbeen during early treatment (3 P = 0.000). (Fig. 3).

Histopathology

Early treatment. Within the baseline group, the cornealepithelium was intact, and the corneal fibers were arrangedin an orderly fashion with no significant infiltration of theinflammatory cells. Additionally, the corneal endothelial

cells could be seen, and there was little serous effusion in theanterior chamber. Finally, the anterior angle was open withno significant inflammatory cell infiltration. Within the con-trol group, minor edema of the corneal epithelium andstroma could be seen, but it contained no significant in-flammatory cell infiltration. Additionally, vasodilatation ofthe iris was observed, more inflammatory exudate could beseen in the anterior chamber, and more inflammatory cellinfiltration could be seen in the anterior angle than in thebaseline group. Mild corneal stromal edema, slight iris vas-cular expansion, and inflammatory cell infiltration in theanterior angle could be found in all 3 treatment groups.Additionally, inflammatory exudate was observed in theanterior chamber in the levofloxacin group corneas (see inFig. 4).

Late treatment. Within the baseline group, the cornealepithelium was intact, but the corneal stroma had significantedema with significant infiltration of inflammatory cells. Thisgroup also had serious serous effusion in the anteriorchamber and inflammatory cells and exudate in the anteriorangle. Additionally, severe vasodilation was observed in theiris. Within the control group, local corneal ulceration, per-foration, and corneal epithelial defects could be observed.

FIG. 3. The number of corneal colony counts (mean SD)for each treatment in the S. aureus keratitis model with latetreatment (n = 9). *Significantly lower CFU recovered fromcorneas compared with the control group (P = 0.000).

FIG. 4. Histology of eyesections stained with hema-toxylin and eosin in the early-treatment groups. Small arrowsin figures indicate the inflam-matory cell infiltration, andlarge arrows indicate the in-flammatory exudates ( 100).

424 WU ET AL.

The corneal stroma contained edema, fibrosis, and a largeamount of inflammatory cell infiltration. Severe vasodilationwas observed in the iris, and large amounts of inflammatoryexudate could be seen in the anterior chamber and on thesurface of the iris. Finally, serious inflammatory cell infiltrationcould be found in the anterior angle. In the gemifloxacingroup, there was mild corneal stromal edema but no signifi-cant inflammatory cell infiltration. A small amount of in-flammatory exudate was seen in the anterior chamber as wellas a small amount of inflammatory cell infiltration was ob-served in the anterior angle. In the gemifloxacin group, noexpansion of the iris blood vessels was observed. In the gati-floxacin and levofloxacin groups, there was mild corneal epi-thelial edema and severe corneal stromal edema withinflammatory cell infiltration as well as large amounts of in-flammatory exudate in the anterior chamber, vasodilation withinflammatory cell infiltration in the iris, and a large number ofinflammatory cells in the anterior angle (see in Fig. 5).

Minimum inhibitory concentrations

The MICs for gemifloxacin and gatifloxacin against S.aureus ATCC25923 were 0.0625mg/L. The MIC for levo-floxacin was 0.125mg/L.

Discussion

Gemifloxacin is a fluoroquinolone antibacterial compoundwith enhanced affinity for bacterial topoisomerase IV and abroad spectrum of activity against Gram-positive and Gram-negative bacteria.11 It has shown potent antibacterial activityagainst clinical isolates and reference strains in both in vitrostudies and experimental models of infection in animals.Gemifloxacin is now widely used for the treatment of re-spiratory and urinary tract infections.11 However, to ourknowledge, there are no ophthalmic formulations of gemi-floxacin available in the clinic. An antibiotic susceptibilitystudy showed that gemifloxacin was one of the most activeagents against conjunctival bacteria from refractive surgerypatients.12 Our previous study also showed that gemi-floxacin might be as effective as moxifloxacin and levo-floxacin for topical prophylaxis and for the treatment ofS. aureusinduced endophthalmitis in the rabbit model.13 Allthese results indicate that gemifloxacin, similar to othernewer fluoroquinolones such as moxifloxacin and gati-floxacin, might be effective in an ocular topical use for ocularinfections.

Bacterial keratitis can ultimately lead to a loss of vision ifnot properly treated, and S. aureus is a common cause of this

FIG. 5. Histology of eye sec-tions stained with hematoxylinand eosin in the late-treatmentgroups. Small arrows in figuresindicate the inflammatory cellinfiltration, and large arrowsindicate the inflammatoryexudates ( 100).

GEMIFLOXACIN TREATMENT KERATITIS 425

infection. The present study demonstrates the effectivenessof gemifloxacin in treating experimental keratitis caused byS. aureus and compares it with 0.5% levofloxacin ophthalmicsolution (Cravit) and 0.3% gatifloxacin eye gel, which are the2 most popular fluoroquinolone ophthalmic formulationscurrently used in mainland China. The levofloxacin oph-thalmic solution used in this study does not contain anypreservatives or mucoadhesive polymers. To compare effi-cacy, the gemifloxacin eye drops applied in this research onlyhad pH and osmotic regulation and did not contain anyadded pharmaceutical excipients such as preservatives, sur-face active agents, or mucoadhesive polymers that couldimprove the corneal absorption of the gemifloxacin or exhibitantibacterial activity. A 0.5% concentration was selected,which is a popular concentration for fluoroquinolones. Withgatifloxacin, the only fourth-generation fluoroquinolone withan ophthalmic topical formulation available in mainlandChina is a 0.3% concentration with preservatives. One ofthese formulations was a 0.3% eye drop, and the other was a0.3% eye gel; a 0.3% gatifloxacin eye gel was selected andused for comparison in this study.

The infection of S. aureus in the cornea developed rapidlyif no treatment was available. At 8 h PI, the typical clinicalsymptoms were observed (as in the baseline group of earlytreatment), such as closed eyes, white purulent discharge inthe conjunctival sac, and hypopyon in some anterior cham-bers. At 17 h PI, the clinical symptoms just mentioned weremuch more severe (as in the control group of early treat-ment), and stroma infiltrate and edema could also be ob-served in the cornea. The symptoms were much less severein the 3 treatment groups than within the control group. Thelate therapy was significantly less effective than early ther-apy, and the symptoms just mentioned were further aggra-vated in the late-treatment groups, simultaneous withstromal infiltration, edema, and other symptoms. In the 3treatment groups, the worsening of clinical symptoms wassignificantly slower than in the control group, and there wasalso a significant improvement in clinical symptoms.Among the 3 experimental groups, gatifloxacin was moreeffective in improving the clinical symptoms than gemi-floxacin and levofloxacin, which failed to improve the clini-cal symptoms.

Although the MIC of gatifloxacin was the same as that ofgemifloxacin for S. aureus, the gatifloxacin eye gel performedbetter in the rabbit keratitis model. The hydrogel formulationand the preservative chlorobutanol in the gatifloxacin eye gelmay have contributed to the antibacterial efficacy demon-strated by gatifloxacin in our model. Several articles havereported that mucoadhesive polymers and preservativesmay improve the efficacy of topical eye formulations.1619

There was a major difference in the level of antibiotic ef-fectiveness in reducing CFU in the early-phase versus later-phase therapy. Levofloxacin produced a 3.23-log reductionin CFU per cornea when it was applied early in infection, butit essentially lost all effectiveness during later-phase therapy.Gatifloxacin produced a 5.32-log reduction in CFU per cor-nea during early therapy but only had a 0.99-log reductionper cornea during the later phase of infection. Gemifloxacinproduced a 3.73- and 1.37-log reduction per cornea in early-and later-phase therapy, respectively, which was proposedto offer a more effective means of reducing CFU in the cor-nea. However, bacterial CFU were not completely eliminatedin any of the treatment groups.

The current experimental study demonstrated that gemi-floxacin, similar to gatifloxacin and levofloxacin, can signif-icantly lower the clinical severity and CFU per corneaobserved in S. aureus keratitis when early treatment is im-plemented. Significantly, gemifloxacin showed a remarkablyimproved efficacy in reducing the bacterial load recoveredfrom the corneas in late treatment.

Acknowledgments

This research was supported by the Taishan Scholar Pro-gram, Shandong Province of China (Project No. ts20081148),the National Natural Science Foundation of China (projectNo. 81000369), and the Young and Middle-Aged ScientistsResearch Awards Fund of Shangdong Province, China(project No. BS2009SV053). The authors would like to thankProfessor Weiyun Shi, Shandong Eye Institute Qingdao,China, for reviewing the article.

Author Disclosure Statement

The authors have no proprietary or commercial interest inany of the drugs or companies mentioned.

References

1. Solomon, R., Donnenfeld, E.D., Holland, E.J., et al. Microbialkeratitis trends following refractive surgery: results of theASCRS infectious keratitis survey and comparisons withprior ASCRS surveys of infectious keratitis following keratore-fractive procedures. J. Cataract Refract. Surg. 37:13431350, 2011.

2. Rocha, G.A., Silva, R.F., Lopes, M.F., Pereira, N.C., andSousa, L.B. Main pathogens and in vitro antimicrobial sus-ceptibility in bacterial keratitis: 5-year study, 20052009. Arq.Bras. Oftalmol. 74:2832, 2011.

3. Marquart, M.E. Animal models of bacterial keratitis. J.Biomed. Biotechnol. 2011:680642, 2011.

4. Dajcs, J.J., Thibodeaux, B.A., Marquart, M.E., Girgis, D.O.,Traidej, M., and OCallaghan, R.J. Effectiveness of cipro-floxacin, levofloxacin, or moxifloxacin for treatment of ex-perimental Staphylococcus aureus keratitis. Antimicrob. AgentsChemother. 48:19481952, 2004.

5. Bertino, J.S. Jr. Impact of antibiotic resistance in the man-agement of ocular infections: the role of current and futureantibiotics. Clin. Ophthalmol. 3:507521, 2009.

6. McHugh, S.M., Collins, C.J., Corrigan, M.A., Hill, A.D., andHumphreys, H. The role of topical antibiotics used as pro-phylaxis in surgical site infection prevention. J. Antimicrob.Chemother. 66:693701, 2011.

7. Tungsiripat, T., Sarayba, M.A., Kaufman, M.B., et al. Fluor-oquinolone therapy in multiple-drug resistant staphylococ-cal keratitis after lamellar keratectomy in a rabbit model.Am. J. Ophthalmol. 136:7681, 2003.

8. Leonard, S.N., Kaatz, G.W., Rucker, L.R., and Rybak, M.J.Synergy between gemifloxacin and trimethoprim/sulfa-methoxazole against community-associated methicillin-resistant Staphylococcus aureus. J. Antimicrob. Chemother. 62:13051310, 2008.

9. Romanowski, E.G., Mah, F.S., Yates, K.A., Kowalski, R.P.,and Gordon, Y.J. The successful treatment of gatifloxacin-resistant Staphylococcus aureus keratitis with Zymar (gati-floxacin 0.3%) in a NZW rabbit model. Am. J. Ophthalmol.139:867877, 2005.

10. Morrissey, I., and Tillotson, G. Activity of gemifloxacinagainst Streptococcus pneumoniae and Haemophilus influ-enzae. J. Antimicrob. Chemother. 53:144148, 2004.

426 WU ET AL.

11. Ramji, J.V., Austin, N.E., Boyle, G.W., et al. The dispositionof gemifloxacin, a new fluoroquinolone antibiotic, in ratsand dogs. Drug Metab. Dispos. 29:435442, 2001.

12. Chung, J.L., Seo, K.Y., Yong, D.E., et al. Antibiotic suscep-tibility of conjunctival bacterial isolates from refractive sur-gery patients. Ophthalmology 116:10671074, 2009.

13. Wu, X., Chen, H., Jiang, H., Xu, Y., Liu, T., and Xu, L. Pro-phylactic effect of topical fluoroquinolones in a rabbit modelof Staphylococcus aureus endophthalmitis. J. Ocul. Pharmacol.Ther. 2011 [Epub ahead of print]; DOI: 10.1089/jop.2011.0136.

14. Sanders, M.E., Moore, Q.C., 3rd, Norcross, E.W., Shafiee, A.,and Marquart, M.E. Efficacy of besifloxacin in an earlytreatment model of methicillin-resistant Staphylococcus aure-us keratitis. J. Ocul. Pharmacol. Ther. 26:193198, 2010.

15. Sanders, M.E., Norcross, E.W., Moore, Q.C., 3rd, Shafiee, A.,and Marquart, M.E. Efficacy of besifloxacin in a rabbit modelof methicillin-resistant Staphylococcus aureus keratitis. Cornea28:10551060, 2009.

16. Wu, X.G., Xin, M., Chen, H., Yang, L.N., and Jiang, H.R.Novel mucoadhesive polysaccharide isolated from Bletillastriata improves the intraocular penetration and efficacy oflevofloxacin in the topical treatment of experimental bacte-rial keratitis. J. Pharm. Pharmacol. 62:11521157, 2010.

17. Romanowski, E.G., Mah, F.S., Kowalski, R.P., Yates, K.A.,and Gordon, Y.J. Benzalkonium chloride enhances the anti-

bacterial efficacy of gatifloxacin in an experimental rabbitmodel of intrastromal keratitis. J. Ocul. Pharmacol. Ther. 24:380384, 2008.

18. Parmar, P., Salman, A., Kalavathy, C.M., et al. Comparisonof topical gatifloxacin 0.3% and ciprofloxacin 0.3% for thetreatment of bacterial keratitis. Am. J. Ophthalmol. 141:282286, 2006.

19. Sasaki, H., Nagano, T., Yamamura, K., Nishida, K., andNakamura, J. Ophthalmic preservatives as absorption pro-moters for ocular drug delivery. J. Pharm. Pharmacol. 47, 703707, 1995.

Received: November 23, 2011Accepted: January 25, 2012

Address correspondence to:Dr. Xianggen Wu

State Key Laboratory Cultivation BaseShandong Provincial Key Laboratory of Ophthalmology

Shandong Eye InstituteNo. 5, Yanerdao Road

Shinan DistrictQingdao 266071

P.R. China

E-mail: wuxianggen@126.com

GEMIFLOXACIN TREATMENT KERATITIS 427