Polyhexamethylene Biguanide and CalcineurinInhibitors as Novel Antifungal Treatments forAspergillus Keratitis

Rachelle A. Rebong,1 Ricardo M. Santaella,1 Brian E. Goldhagen,1 Christopher P. Majka,1

John R. Perfect,2 William J. Steinbach,2 and Natalie A. Afshari1

PURPOSE. To establish polyhexamethylene biguanide (PHMB) asan effective treatment for Aspergillus keratitis in a novel mu-rine model. To determine the ability of the calcineurin inhibi-tors tacrolimus (FK506) and cyclosporine A (CSA) to enhancethe activity of PHMB, amphotericin B (AMB), and voriconazole(VCZ) against Aspergillus keratitis.

METHODS. In vitro studies: Broth antifungal susceptibility testswere performed with PHMB, AMB, VCZ, and FK506, individu-ally and in combination against Aspergillus fumigatus. Mini-mum inhibitory concentrations (MIC) and fractional inhibitoryconcentration index (FICI) values were used to analyze anti-fungal activity. In vivo studies: A novel murine model wascreated to establish Aspergillus keratitis. Infected mice wererandomly assigned to treatment groups receiving saline, CSA,AMB, VCZ, PHMB, AMBCSA, VCZCSA, or PHMBCSA. Anophthalmologist blinded to the treatment groups assessed dis-ease severity daily based on a grading scale. The mean endchange in disease score was compared between groups.

RESULTS. In vitro studies: FK506 in combination with PHMB,VCZ, or AMB enhanced fungal growth inhibition. FICI valuesshowed an additive effect between FK506 and PHMB, AMB, orVCZ. PHMB monotherapy eliminated Aspergillus growth start-ing at 4 g/mL. In vivo studies: All treatment groups showed asignificant improvement in disease score compared to thecontrol group. CSA significantly worsened VCZ activity againstAspergillus keratitis.

CONCLUSIONS. PHMB is an effective inhibitor of Aspergillusgrowth. Further investigation of the role of calcineurin inhib-itors in the treatment for Aspergillus keratitis is warranted.(Invest Ophthalmol Vis Sci. 2011;52:73097315) DOI:10.1167/iovs.11-7739

Fungal keratitis is an important cause of ocular morbidityworldwide, leading to corneal ulceration and scarring, cor-neal transplantation, and blindness.1,2 Whereas yeastlike Can-dida species predominate in more temperate climates, in trop-ical and subtropical areas such as Asia, Africa, and the southern

United States, filamentous fungi such as Aspergillus species arethe most common causes of fungal keratitis.37

Currently available medical treatments for Aspergillus ker-atitis are not only largely ineffective but are also expensive. Asthe fungus penetrates deeper into the cornea, it becomes moredifficult for the currently available treatments to reach anderadicate it. Although 15% to 27% of all patients with fungalkeratitis ultimately require surgical intervention, Aspergillusspecies cause a particularly severe keratitis that necessitateskeratoplasty in 42% to 60% of cases.8 Given the prevalence ofAspergillus keratitis in many areas, particularly agriculturalareas in developing nations, the development of new treatmentoptions to increase effectiveness and to decrease cost is imper-ative.

Despite the high incidence of keratomycoses globally, first-line topical polyene antifungal antibiotics, including amphoter-icin B (AMB) and natamycin, are not effective in completelyeliminating severe keratomycoses.7 Polyene macrolides workby binding to ergosterol, a sterol unique to fungal cell mem-branes, to form pores that allow the leakage of electrolytes andlead to cell death.9 Natamycin is the only U.S. Food and DrugAdministration (FDA)approved topical ocular antifungal, butit has poor corneal penetration, is cost-prohibitive, degradeseasily, and is not commercially available in many regions of theworld.8,10 AMB is easier to handle and has been used to treatkeratomycoses since 1959; however, it is not available as anophthalmic preparation and needs to be compounded as atopical solution (0.15%0.30%).8,11 Topical AMB appears to bewell tolerated at lower concentrations but causes conjunctivalirritation at higher concentrations (0.50%), limiting its use atthese concentrations.2

Azoles are another class of antifungals used in the treatmentof Aspergillus keratitis. They inhibit the cytochrome P-450enzyme 14-demethylase needed to convert lanosterol toergosterol and lead to increased permeability of fungal cellmembranes.12 Voriconazole (VCZ) is a triazole antifungal agentwith a broad spectrum of activity against Aspergillus speciesand is FDA-approved for the treatment of invasive aspergillo-sis.10 The use of oral and topical VCZ, compounded as a 1%solution, has been reported in cases of fungal keratitis.1315 Ithas been effective in treating even Aspergillus infections resis-tant to AMB and has less severe side effects than AMB. Adverseeffects of systemic VCZ include skin rashes and visual distur-bances, which are usually mild and transient.10 Elevation ofhepatic enzymes rarely occurs.12 However, VCZ is very costlyand is not readily available in most parts of the world; like AMB,it must be compounded for use as an ophthalmic solution.Moreover, Aspergillus resistance to VCZ is a growing prob-lem.1618

Polyhexamethylene biguanide (PHMB) is an inexpensive, read-ily available general biocide used as a swimming pool and contactlens disinfectant.19 It is thought to work by disrupting the cell

From the 1Duke University Eye Center and the 2Division of Infec-tious Diseases, Duke University Medical Center, Durham, North Caro-lina.

Supported by an Allergan Horizon Grant and Research to PreventBlindness.

Submitted for publication April 14, 2011; revised June 24, 2011;accepted July 17, 2011.

Disclosure: R.A. Rebong, None; R.M. Santaella, None; B.E.Goldhagen, None; C.P. Majka, None; J.R. Perfect, None; W.J. Steinbach,None; N.A. Afshari, None

Corresponding author: Natalie A. Afshari, Cornea and RefractiveSurgery Service, Duke University Eye Center, DUMC Box 3802, Dur-ham, NC 27710; natalie.afshari@duke.edu.

Cornea

Investigative Ophthalmology & Visual Science, September 2011, Vol. 52, No. 10Copyright 2011 The Association for Research in Vision and Ophthalmology, Inc. 7309

membranes of microorganisms, causing leakage of intracellularcomponents and inhibiting the function of respiratory enzymes.2

In vitro studies demonstrated PHMBs activity against Candidaalbicans, Fusarium solani, and Aspergillus niger.19 In a rabbitmodel of Fusarium keratitis, 0.02% topical PHMB significantlyreduced fungal growth compared to placebo.20 Topical PHMB atconcentrations of 0.02% to 0.053% has been used in the treatmentof Acanthamoeba keratitis, but only one published animal studyand no human studies have evaluated the use of PHMB againstAspergillus keratitis.21 In the study evaluating PHMB treatment forAspergillus keratitis, topical PHMB 0.02% was shown to be mod-erately effective against the infection in a rabbit model, showingsignificantly improved ulcer healing time compared with thecontrol.21 However, only six animals were treated in that study.

Given the limited effectiveness, growing fungal resistance, andhigh cost associated with current treatment options for Aspergil-lus keratitis, the development of new treatments becomes impor-tant. Tacrolimus (FK506) and cyclosporine A (CSA) are immuno-suppressive agents that target the calcineurin pathway to preventrejection after organ transplantation.22 Calcineurin is a serine-threonine phosphatase in the calcium signaling pathway of theconserved cell stress response. In fungi, this signaling pathwayhas been found to mediate growth, morphology, stress responses,and pathogenicity; it has also been shown to have a role inregulating cell wall formation.23 Thus, calcineurin inhibitors mayplay a role not only in immunosuppression but also in inhibitingfungal growth. Evidence to support this hypothesis comes fromprevious studies which showed that calcineurin inhibitors inhibitthe growth of A. fumigatus and that deletion of a gene within thecalcineurin pathway of A. fumigatus reduces hyphal growth andattenuates virulence.2426 In addition, in a murine model of C.albicans keratitis, we have shown that calcineurin promotescorneal fungal infection and that calcineurin inhibitors can worksynergistically with fluconazole in the treatment of Candida ker-atitis.27 In fact, calcineurin inhibitors have already been shown toenhance the activity of caspofungin, an echinocandin that inhibitsfungal cell wall synthesis, against Aspergillus species.28

Our hypothesis is that similar synergy occurs between thecalcineurin inhibitor FK506 and PHMB, AMB, and VCZ, respec-tively, against A. fumigatus. Using calcineurin inhibitors incombination with more conventional antifungals may providea more effective and ultimately less costly treatment option forAspergillus keratitis.

METHODS

Aspergillus Strain, Media, andInoculum PreparationWild-type A. fumigatus strain AF293 was used in all experiments.RPMI 1640, prepared according to CLSI (Clinical and Laboratory Stan-dards Institute) standards, was used in all in vitro assays.29 The inocu-lum was prepared from AF293 grown and harvested according to CLSIguideline for antifungal susceptibility testing.29

Drug PreparationFor in vitro assays, VCZ (Vfend; Pfizer, New York, NY) and AMB (Fungi-zone; Bristol-Myers Squibb, New York, NY) were prepared from powdersaccording to CLSI guidelines.29 FK506 was obtained from Astellas PharmaUS, Inc. (Deerfield, IL). Polyhexamylene biguanide was obtained as a0.02% ophthalmic solution from the Duke Pharmacy Compounding Facil-ity (Durham, NC). It was prepared by adding 0.1 mL of a commerciallyavailable 20% PHMB solution (Baquacil; Arch Chemicals, Inc., Norwalk,CT) to 100 mL of normal saline. CSA was obtained as a 2% ophthalmicsolution from the Duke Pharmacy Compounding Facility (Durham, NC).CSA was prepared by adding 0.6 mL of cyclosporine 100 mg/mL to 2.4 mLof corn oil.

In Vitro Study DesignBroth susceptibility testing of A. fumigatus was performed usingFK506, PHMB, AMB, and VCZ individually and in combination (FK506with PHMB, AMB, or VCZ) against wild-type A. fumigatus, accordingto a modified version of the CLSI guidelines.29 A checkerboard titrationwas used, testing two drugs to compare the effects of the drugs aloneand in combination. FK506 was used at concentrations between0.01 and 0.08 g/mL, PHMB between 0.3 and 30 g/mL, AMB between2 and 64 g/mL, and VCZ between 0.125 and 1 g/mL. Samples wereincubated for 48 hours at 37C. At 48 hours, fungal growth wascompared against that in the controls and across the various drugconcentrations and combinations. Elimination of growth (MIC) andinhibition of growth, seen as the blunting of hyphal tips compared tocontrol wells and known as the minimum inhibitory concentration(MIC), were assessed (Fig. 1).

Quantitative Analysis of In Vitro DataThe MIC and fractional inhibitory concentration index (FICI) wereused to analyze in vitro antifungal activity and interactions against A.

FIGURE 1. Examples of morphologicchanges in wild-type A. fumigatus asseen in vitro. (A) Growth of untreatedwild-type A. fumigatus. Note the ex-tensively branching filaments. (B) Inhi-bition of growth with VCZ (0.25 g/mL) and FK506 (40 ng/mL). (C)Inhibition of growth with PHMB (0.03g/mL) and FK506 (2.25 g/mL). Notethe blunted hyphal tips in (B) and (C)(arrows). (D) Elimination of growthwith PHMB (3 g/mL). VCZ, voricona-zole; FK506, tacrolimus; PHMB, poly-hexamethylene biguanide. Bar, 200 m.

7310 Rebong et al. IOVS, September 2011, Vol. 52, No. 10

fumigatus. The FICI is used to analyze drug interactions in combina-tion and is calculated with the following formula: FICI [(MIC A incombination)/MIC A] [MIC B in combination)/MIC B]. Interpretationof the FICI can vary. In our study, interpretation was determinedaccording to accepted standards by the following: FICI 0.5, syner-gistic effect; 0.5 but 1, additive effect; 1 but 4, indifferenteffect; and 4, antagonistic effect.30

In Vivo Study Design. A prospective randomized control studywas performed according to a modified version of a previous proto-col.27 This protocol was approved by the Duke Institutional AnimalCare and Use Committee. All study animals were treated according tothe ARVO Statement for the Use of Animals in Ophthalmic and VisionResearch.

A total of 158 adult male BALB/c mice, each weighing 20 to 25 g,were immunosuppressed with methylprednisolone (100 mg/kg) ondays 5, 1, and 1 of inoculation, to rapidly establish infection. Anintramuscular injection of a ketamine (10 mg/mL)-xylazine (1 mg/mL)mixture and a drop (30 L) of proparacaine 0.5% ophthalmic solu-tion (Alcon, Fort Worth, TX) on the cornea were given for anesthesia.A drop of moxifloxacin 0.5% ophthalmic solution (Vigamox; Alcon)was then administered to the cornea to avoid bacterial contamination.The right corneas of 158 mice were de-epithelialized with a 30-gaugeneedle. Two 5-L suspensions, each containing 106 A. fumigatusspores, were evenly distributed in succession on the scratched corneaof each mouse.

A previously described disease grading scale from 0 (no disease) to4 (severe disease) was modified, and animals were graded as follows byan ophthalmologist who was blinded to the drug treatments (Fig. 2): 0,no sign of epithelial defect or infection; 1, signs of epithelial scratchesbut no infiltrate; 2, corneal infiltrate covering 25% to 50% of thecorneal surface; 3, corneal infiltrate covering 50% to 75% of the cornealsurface; and 4, corneal infiltrate covering 75% of the corneal sur-face.27 Eyes that perforated received an automatic score of 4.

Mice that developed clinically significant infection, indicated byreaching a grade of 2, were included in the study and started onantifungal treatment. Eligible mice were randomly assigned to one ofeight treatment groups: balanced salt solution (control; Akorn, LakeForest, IL), CSA, AMB, VCZ, PHMB, AMBCSA, VCZCSA, orPHMBCSA.

Study mice received four doses (5 L/dose) of the treatment drugeach day for 10 consecutive days. For combination therapy, the drugswere administered in succession with at least 2 minutes betweendoses. Grading was performed daily with an operating microscope(Carl Zeiss Meditec, Dublin, CA). The results from two independentexperiments were combined and analyzed.

Statistical Analysis of In Vivo Data

The mean changes in disease score between groups were comparedusing ANOVA, in which all pairwise comparisons between the eightgroups were performed by using Students t-test. We included all micethat developed a disease score of 2 and excluded mice that pro-gressed to perforation of the cornea before treatment could beinitiated. Data from mice whose corneas perforated after treatmentwas initiated were included (statistical analyses by JMP 8 software;SAS Institute, Cary, NC).

RESULTS

FK506 Enhanced PHMB, AMB, and VCZ Activityagainst A. fumigatus In Vitro

With FK506 alone, inhibition of A. fumigatus growth occurredstarting at 0.02 g/mL; fungal growth elimination did notoccur, even at the highest concentration (Fig. 3). When FK506was added to PHMB, AMB, or VCZ, fungal growth inhibitionwas enhanced and/or occurred at lower concentrations thanwhen PHMB, AMB, or VCZ was used individually. PHMB elim-inated growth starting at 4 g/mL but showed no effects atlower concentrations. With the addition of at least 0.01 g/mLFK506, PHMB still eliminated growth at 4 g/mL but alsoinhibited fungal growth at 0.5 g/mL PHMB (Fig. 3). AMBinhibited growth at the highest concentration 64 g/mL. Withthe addition of 0.08 g/mL FK506, fungal growth inhibitionoccurred at 8 g/mL AMB (Fig. 3). VCZ alone inhibited growthstarting at 0.25 g/mL and eliminated growth starting at 1g/mL. Fungal growth inhibition was enhanced at 0.25 g/mLVCZ with the addition of at least 0.02 g/mL FK506 (Fig. 3).

FICI values indicated an in vitro additive effect when FK506was combined with any of the study medications. The FICIvalues calculated from the individual and combined MICs ofthe drugs are 0.625, 0.625, and 1 for PHMBFK506,AMBFK506, and VCZFK506, respectively (Table 1).

All Murine Model Treatment Groups ShowedSignificant Improvement in Disease ScoreCompared to the Control Group

A total of 103 of 158 mice achieved grade 2 infection andwere included in the study, with 12 to 14 mice per group.The control group showed an increase or worsening ofdisease score after 10 days of treatment (1.08 1.08). All

FIGURE 2. Disease grading scale forevaluating mice corneas. The righteyes of BALB/c mice were scrapedand inoculated with wild-type A. fu-migatus. Study eyes were viewedand digitally photographed throughan operating microscope. (A) A scoreof 0 indicated no sign of epithelialdefect or infection. (B) A score of 1showed signs of epithelial scratchesbut no infiltrate. (C) A score of 2showed a corneal infiltrate covering25% to 50% of the corneal surface.(D) A score of three showed a cor-neal infiltrate covering 50% to 75% ofthe corneal surface. (E) A score of 4showed a corneal infiltrate covering75% of the corneal surface or anysigns of perforation, as shown in thisimage.

IOVS, September 2011, Vol. 52, No. 10 Novel Treatments for Aspergillus Keratitis 7311

other treatment groups showed an improvement in diseasescore after 10 days of treatment, which can be seen as adecrease in the average daily disease score per group over

the complete course of treatment (Fig. 4) or calculated as anegative mean end change in disease score (Table 2). Of thetreatment groups, AMB alone led to the largest mean enddecrease in disease score from baseline (1.84 1.28),whereas VCZCSA led to the smallest mean end decrease indisease score from baseline (0.42 1.22). P values for thecomparison of the mean end change in disease score be-tween the control group and the other study groups were allsignificant (P 0.05; Table 3).

CSA significantly worsened VCZ activity against A. fumiga-tus infection in study mice. Significant differences were foundbetween VCZCSA and VCZ alone, AMB alone, andPHMBCSA (P 0.0158, 0.0023, and 0.0320, respectively) instudy mice. Mice receiving VCZCSA showed significantly lessimprovement in disease score than did the mice receiving VCZalone, AMB alone, or PHMBCSA. The rest of the comparisonsbetween noncontrol treatment groups were not significant(Table 3).

FIGURE 3. In vitro growth of A. fu-migatus after treatment with FK506.In vitro growth inhibition occurredat lower concentrations of PHMBwith the addition of FK506 and wasenhanced at 0.25 g/mL VCZ withthe addition of at least 0.02 g/mLFK506. In vitro growth of A. fumiga-tus after treatment with increasingdoses of (A) 0.01, (B) 0.04 g/mLFK506 alone. Inhibition occurred atlower concentrations of PHMB andAMB, with the addition of FK506; (C)0.5 g/mL PHMB0.01 g/mLFK506. Note the increased area ofclearing (arrow); (D) 0.5 g/mLPHMB0.02 g/mL FK506; (E) 16g/mL AMB0.08 g/mL FK506;and (F) 64 g/mL AMB0.08 g/mLFK506. In vitro growth inhibition wasenhanced at 0.25 g/mL VCZ with theaddition of at least 0.02 g/mL FK506;(G) 0.25 g/mL VCZ alone; and (H)0.25 g/mL VCZ0.08 g/mL FK506.Bar, 200 m.

TABLE 1. In Vitro Additive Effects of FK506 and Antifungals againstWild-Type A. fumigatus by Microdilution Checkerboard Interaction

DrugMIC

(g/mL)MICs of Antifungal Agent

Combination Drug A/FK506 (g/mL) FICI

FK506 0.02 AMB 64.00 8.000/0.01 0.625VCZ 0.25 0.125/0.01 1.000PHMB 4.00 0.500/0.01 0.625

FICI is used to analyze drug interactions in combination. FICI [(MIC A in combination)/MIC A] [(MIC B in combination)/MIC B]. FICIvalues were interpreted as follows: 0.5, synergistic effect; 0.5 but 1,additive effect; 1 but 4, indifferent effect; and 4, antagonistic effect.

7312 Rebong et al. IOVS, September 2011, Vol. 52, No. 10

Alopecia Was a Notable Side Effect in Study Mice

All animals treated with CSA, either alone or in combination,developed alopecia around the treated eye (Fig. 5).

DISCUSSION

In this study, we sought to determine whether PHMB mono-therapy is a viable treatment for Aspergillus keratitis and toascertain the ability of calcineurin inhibitors to enhance theantifungal activity of PHMB, AMB, and VCZ against A. fumiga-tus, with the ultimate goal of providing better and less expen-sive treatment options for Aspergillus keratitis. To test ourhypotheses, we developed a novel murine model for Aspergil-lus keratitis. In addition to establishing the efficacy of usingcalcineurin inhibitors for Aspergillus keratitis, our studies pro-vided additional evidence to demonstrate the effectiveness ofPHMB against Aspergillus. Each of the drugs tested showedantifungal activity individually against A. fumigatus. This resultwas expected for AMB and VCZ; even the calcineurin inhibitorFK506 had been shown in previous studies to have an effectagainst A. fumigatus.24 However, data on PHMBs effective-ness against A. fumigatus were limited to a single study in arabbit model, showing that PHMB was moderately effective butnot better than natamycin in clearing Aspergillus keratitis.21 Inour study, PHMB was effective, not only in inhibiting but alsoin eliminating Aspergillus growth. Interestingly, unlike theother drugs tested, PHMB did not show a progression of fungalgrowth inhibition with increasing doses of the drug; rather, thetransition between dilutions abruptly changed from no effectto complete elimination of fungal growth. Further in vitro andin vivo studies are warranted to determine PHMBs activityagainst Aspergillus. If shown to be effective, PHMBessen-tially, diluted pool cleanercould become a revolutionary andvery affordable option for the treatment of Aspergillus keratitisfor patients in agricultural areas of developing countries,where it is a significant problem.

Based on our initial in vitro studies, FICI values indicate thatthe calcineurin inhibitor FK506 displays additive effects whencombined with PHMB, AMB, or VCZ against A. fumigatus. Anadditive, as opposed to a synergistic, effect observed betweenthese drugs may be explained in several ways. First, comparedwith Candida, against which CSA and fluconazole appear toshow in vivo synergy, Aspergillus causes a more virulent cor-neal infection.27 In Candida keratitis, even deep corneal le-sions treated with medical therapy alone generally carry afavorable prognosis; in Aspergillus keratitis, deep corneal le-sions generally do not respond to medical therapy and requiresurgical intervention.2

The role of the calcineurin pathway may differ betweenCandida and Aspergillus species. More information on cal-cineurin-dependent genes in Candida species versus in Asper-gillus species are available and differences between the two

FIGURE 4. Mean daily disease scoresover 10 days of treatment of the studymice. The control group showed anincrease or worsening of mean dailydisease score, whereas all othergroups showed a decrease or im-provement of mean daily diseasescore over the treatment course.BSS, balanced salt solution.

TABLE 2. Sample Size and Mean End Change in Disease Score after10 Days of Treatment of Study Mice

Treatment Group Sample SizeEnd Change(Mean SD)

Balanced salt solution (control) 12 1.08 1.08CSA 12 1.08 1.26AMB 13 1.84 1.28AMBCSA 13 1.11 1.12VCZ 13 1.53 0.78VCZCSA 13 0.42 1.22PHMB 13 0.96 1.27PHMBCSA 14 1.39 1.16

TABLE 3. Comparison of the Mean Changes in Disease Scorebetween Two Groups of Study Mice

Group A Group B P

Balanced salt solution CSA

genera with regard to gene targets as well as phenotypesattributed to calcineurin signaling have been identified.31

Whether these differences can account for the two generasdifferent responses to calcineurin inhibitors is a topic thatshould be studied further.

The additive as opposed to synergistic effects betweenFK506 and PHMB, AMB, or VCZ may reflect the different targetsites for the drugs. Caspofungin inhibits the activity of 1,3--D-glucan synthase, an enzyme critical for fungal cell wall synthe-sis.28 AMB and VCZ, meanwhile, inhibit fungal cell membranesynthesis by interfering with ergosterol. The synergistic activ-ity of calcineurin inhibitors with caspofungin and their additiveactivity with AMB and VCZ may be explained by the possibilitythat the calcineurin pathway plays a role in cell wall synthesisof Aspergillus species. Incidentally, although the antifungalactivity of PHMB is not as well understood, its additive effectwith calcineurin inhibitors supports the idea that PHMB acts atthe level of the fungal cell membrane as opposed to the cellwall.

Although the calculated FICI values did not meet the criteriafor synergy, the observed morphologic changes clearly showedthat FK506 enhanced the ability of PHMB, AMB, and VCZ toinhibit Aspergillus growth in vitro. Our in vitro studies indicatethat calcineurin inhibitors like FK506 can have a role in im-proving the antifungal treatment of Aspergillus keratitis.

Our in vivo studies showed little statistical significancewhen comparing noncontrol treatment groups, probably be-cause of the low sample size. A smaller number of animals wasdesirable for this study, as it allowed us to establish and eval-uate a novel mouse model of Aspergillus keratitis in addition totesting the effects of different drug combinations. Previousmodels of Aspergillus keratitis used rabbits, whose larger cor-neas make experimentation easier.32,33 A mouse model ofAspergillus keratitis was created by Zhong et al.,34 who used amore complicated procedure similar to epikeratophakia com-pared with our method involving corneal de-epithelialization.Our studies showed that a procedure involving scratching ofthe cornea, as opposed to a more technically complicatedprocedure similar to epikeratophakia, can lead to successfulinfection of over 65% of inoculated mice.34 This mouse modelprovides a more practical and economical option which can beused in future research on Aspergillus keratitis.

We elected to use the calcineurin inhibitor CSA in our invivo experiments rather than using FK506, as we did in our invitro experiments. We initially chose FK506 for the in vitroexperiments because it has been shown to be superior to CSAin its traditional role as an immunosuppressant after organtransplantation.35 We then chose CSA for the in vivo experi-ments, because CSA is already commercially available and usedto treat other ocular diseases such as dry eye syndrome. Al-though FK506 and CSA act on the same pathway by inhibitingcalcineurin phosphatase, they bind to different members of theimmunophilin familyFK506 to FKBP12 and CSA to cyclophi-linwhich, in cases of organ transplantation, leads to differentrisk and benefit profiles.35 Whether these differences translateto the calcineurin pathway in fungi and whether these differ-ences are significant are questions that have yet to be an-swered. If FK506 and CSA in fact behave differently in Asper-gillus, it would help explain the discrepancy between ourresults.

Key differences between in vitro studies and in vivo studiesusing calcineurin inhibitors may stem from the fact that cal-cineurin inhibitors in vivo not only exhibit an antifungal effectagainst Aspergillus but may also exert immunosuppressiveeffects on the host. During in vivo experiments, since calcineu-rin inhibitors are traditionally used as immunosuppressantagents, they may limit the hosts ability to combat infection.Alternatively, calcineurin inhibitors may help suppress thehosts inflammatory response, which at times can cause asmuch or more damage to the cornea than does the infection.To better define this issue, the balance between the antifungaland immunosuppressive effects of calcineurin inhibitors inhost species requires further investigation.

Another issue to consider with regard to our in vivo studiesis that the preparation of the CSA ophthalmic solutions mayhave influenced their ability to be delivered effectively to thesite of infection. Compared to the balanced salt solution,VCZ, AMB, and PHMB ophthalmic solutions, the CSA oph-thalmic solution was more viscous due to being prepared incorn oil. CSA was more difficult to administer, not onlybecause the viscous solution was difficult to apply but alsobecause the animals more vigorously wiped it off afterapplication. The alopecia resulting from CSA administrationmay be due to this vigorous wiping rather than or in additionto the properties of the drug itself. A viscous substance likeCSA may be better applied and therefore absorbed across alarger surface area, like the human cornea, as opposed to thesmaller corneas of our study animals. Considering alternativepreparations for CSA may improve drug administration to andpenetration of the cornea in future experiments.

Future in vivo studies are needed to assess the effects ofusing calcineurin inhibitors to enhance the antifungal treat-ment of Aspergillus keratitis. If proven effective, the use ofcalcineurin inhibitors in conjunction with current antifungaltherapies may lower the concentrations needed for the indi-vidual drugs to exert their desired effects. Using calcineurininhibitors to enhance the treatment of Aspergillus keratitiswould help avoid potential side effects at higher doses of theindividual drugs and would increase the effectiveness of cur-rent treatment regimens against Aspergillus keratitis. Perhapsmore importantly, the use of calcineurin inhibitors may help inavoiding the problem of growing fungal resistance when indi-vidual drugs are used alone.

References

1. Klotz SA, Penn CC, Negvesky GJ, Butrus SI. Fungal and parasiticinfections of the eye. Clin Microbiol Rev. 2000;13:662685.

2. Thomas PA. Current perspectives on ophthalmic mycoses. ClinMicrobiol Rev. 2003;16:730797.

FIGURE 5. All mice treated with CSA developed localized alopecia.

7314 Rebong et al. IOVS, September 2011, Vol. 52, No. 10

3. Basak SK, Basak S, Mohanta A, Bhowmick A. Epidemiological andmicrobiological diagnosis of suppurative keratitis in Gangetic WestBengal, eastern India. Indian J Ophthalmol. 2005;53:1722.

4. Bharathi MJ, Ramakrishnan R, Vasu S, Palaniappan R. Aetiologicaldiagnosis of microbial keratitis in South India: a study of 1618cases. Indian J Med Microbiol. 2002;20:1924.

5. Hagan M, Wright E, Newman M, Dolin P, Johnson G. Causes ofsuppurative keratitis in Ghana. Br J Ophthalmol. 1995;79:10241028.

6. Leck AK, Thomas PA, Hagan M, et al. Aetiology of suppurativecorneal ulcers in Ghana and south India, and epidemiology offungal keratitis. Br J Ophthalmol. 2002;86:12111215.

7. Srinivasan M. Fungal keratitis. Curr Opin Ophthalmol. 2004;15:321327.

8. Thomas PA. Fungal infections of the cornea. Eye (Lond). 2003;17:852862.

9. Gallis HA, Drew RH, Pickard WW. Amphotericin B: 30 years ofclinical experience. Rev Infect Dis. 1990;12:308329.

10. Hariprasad SM, Mieler WF, Lin TK, Sponsel WE, Graybill JR. Vori-conazole in the treatment of fungal eye infections: a review ofcurrent literature. Br J Ophthalmol. 2008;92:871878.

11. Wood TO, Tuberville AW, Monnett R. Keratomycosis and ampho-tericin B. Trans Am Ophthalmol Soc. 1985;83:397409.

12. Kaur IP, Rana C, Singh H. Development of effective ocular prep-arations of antifungal agents. J Ocul Pharmacol Ther. 2008;24:481493.

13. Mehta H, Mehta HB, Garg P, Kodial H. Voriconazole for thetreatment of refractory Aspergillus fumigatus keratitis. Indian JOphthalmol. 2008;56:243245.

14. Jones A, Muhtaseb M. Use of voriconazole in fungal keratitis. JCataract Refract Surg. 2008;34:183184.

15. Jhanji V, Sharma N, Mannan R, Titiyal JS, Vajpayee RB. Manage-ment of tunnel fungal infection with voriconazole. J CataractRefract Surg. 2007;33:915917.

16. Snelders E, van der Lee HA, Kuijpers J, et al. Emergence of azoleresistance in Aspergillus fumigatus and spread of a single resis-tance mechanism. PLoS Med. 2008;5:e219.

17. Bellete B, Raberin H, Morel J, Flori P, Hafid J, Manhsung RT.Acquired resistance to voriconazole and itraconazole in a patientwith pulmonary aspergilloma. Med Mycol. 2010;48:197200.

18. Beernaert LA, Pasmans F, Van Waeyenberghe L, et al. Avian Asper-gillus fumigatus strains resistant to both itraconazole and voricona-zole. Antimicrob Agents Chemother. 2009;53:21992201.

19. Messick CR, Pendland SL, Moshirfar M, et al. In-vitro activity ofpolyhexamethylene biguanide (PHMB) against fungal isolates asso-ciated with infective keratitis. J Antimicrob Chemother. 1999;44:297298.

20. Fiscella RG, Moshifar M, Messick CR, Pendland SL, Chandler JW,Viana M. Polyhexamethylene biguanide (PHMB) in the treatmentof experimental Fusarium keratomycosis. Cornea. 1997;16:447449.

21. Panda A, Ahuja R, Biswas NR, Satpathy G, Khokhar S. Role of 0.02%polyhexamethylene biguanide and 1% povidone iodine in experi-mental Aspergillus keratitis. Cornea. 2003;22:138141.

22. Clipstone NA, Crabtree GR. Identification of calcineurin as a keysignalling enzyme in T-lymphocyte activation. Nature. 1992;357:695697.

23. Fortwendel JR, Juvvadi PR, Pinchai N, et al. Differential effects ofinhibiting chitin and 1,3-{beta}-D-glucan synthesis in ras and cal-cineurin mutants of Aspergillus fumigatus. Antimicrob AgentsChemother. 2009;53:476482.

24. Steinbach WJ, Schell WA, Blankenship JR, Onyewu C, Heitman J,Perfect JR. In vitro interactions between antifungals and immuno-suppressants against Aspergillus fumigatus. Antimicrob AgentsChemother. 2004;48:16641669.

25. Steinbach WJ, Cramer RA Jr, Perfect BZ, et al. Calcineurin controlsgrowth, morphology, and pathogenicity in Aspergillus fumigatus.Eukaryot Cell. 2006;5:10911103.

26. Pinchai N, Perfect BZ, Juvvadi PR, et al. Aspergillus fumigatuscalcipressin CbpA is involved in hyphal growth and calcium ho-meostasis. Eukaryot Cell. 2009;8:511519.

27. Onyewu C, Afshari NA, Heitman J. Calcineurin promotes infectionof the cornea by Candida albicans and can be targeted to enhancefluconazole therapy. Antimicrob Agents Chemother. 2006;50:39633965.

28. Kontoyiannis DP, Lewis RE, Osherov N, Albert ND, May GS. Com-bination of caspofungin with inhibitors of the calcineurin pathwayattenuates growth in vitro in Aspergillus species. J AntimicrobChemother. 2003;51:313316.

29. Clinical and Laboratory Standards Institute. Reference method forbroth dilution antifungal susceptibility testing of filamentous fungi;approved standard. Document M38-A. Clinical and LaboratoryStandards Institute, Wayne, PA; 2002.

30. Afeltra J, Dannaoui E, Meis JF, Rodriguez-Tudela JL, Verweij PE. Invitro synergistic interaction between amphotericin B and pentam-idine against Scedosporium prolificans. Antimicrob Agents Che-mother. 2002;46:33233326.

31. Steinbach WJ, Reedy JL, Cramer RA Jr, Perfect JR, Heitman J.Harnessing calcineurin as a novel anti-infective agent against inva-sive fungal infections. Nat Rev Microbiol. 2007;5:418430.

32. Garcia ML, Herreras JM, Dios E, Argueso P, Almaraz A. Evaluationof lectin staining in the diagnosis of fungal keratitis in an experi-mental rabbit model. Mol Vis. 2002;8:1016.

33. Dong X, Shi W, Zeng Q, Xie L. Roles of adherence and matrixmetalloproteinases in growth patterns of fungal pathogens in cor-nea. Curr Eye Res. 2005;30:613620.

34. Zhong W, Yin H, Xie L. Expression and potential role of majorinflammatory cytokines in experimental keratomycosis. Mol Vis.2009;15:13031311.

35. Haddad EM, McAlister VC, Renouf E, Malthaner R, Kjaer MS, GluudLL. Cyclosporin versus tacrolimus for liver transplanted patients.Cochrane Database Syst Rev. 2006:CD005161.

IOVS, September 2011, Vol. 52, No. 10 Novel Treatments for Aspergillus Keratitis 7315