Comparison of Two Probes for Testing Susceptibilities of Pathogenic Yeasts to Voriconazole, Itraconazole, and Caspofungin by Flow Cytometry

A cytometric approach to determine the susceptibilities of Candidaspp. and Cryptococcus neoformans to voriconazole, itraconazole,and caspofungin is described. A total of 63 clinical isolateswith different susceptibility patterns were exposed for 1, 2,4, and 6 h to serial concentrations of each antifungal agent,followed by staining with two fluorescent probes: propidiumiodide (PI) and FUN-1. FUN-1 was able to identify the susceptibilitypatterns of the assayed strains to the three agents after 1h. PI penetrated a maximum of 50% of the cells treated withPI, at the highest concentration of caspofungin, 16 µg/ml,after 6 h of incubation (this percentage varied with the strainand was drug concentration and time of incubation dependent)and did not stain cells treated with high concentrations ofeither azole after 6 h. The use of FUN-1 appears to be an excellentfast and reliable alternative to the classical dilution methodfor determining the susceptibility of Candida spp. and C. neoformansto these three antifungal agents.

INTRODUCTION

Top

Introduction

Prior to the introduction of the azoles, few therapeutic choiceswere available for fungemia, which limited the role of antifungalsusceptibility testing (AFST). AFST could be useful in the selectionof empirical treatment and for testing isolates from blood,from deep-seated infections, or from recurrent mucosal infections(8). The role of AFST is also important in light of the increasein the number of severe non-albicans Candida infections andthe introduction of new broad-spectrum azoles and echinocandins.The use of standardized antifungal susceptibility methods hasimproved interlaboratory reproducibility. However, the abilityof antifungal susceptibility testing to predict clinical outcomeis still being elucidated for the new agent caspofungin; tentativebreakpoints are still under discussion for voriconazole by theClinical Laboratory Standards Institute (CLSI; formerly NationalCommittee for Clinical Laboratory Standards). The CLSI microdilutionmethod is labor and time-consuming and difficult to use in smallclinical laboratories. Some alternatives have emerged, includingcolorimetric or fluorescent methods, which could improve definitionof susceptibility patterns and or automation. Commercial antifungalmethods also have been developed (6, 11, 13) that are easy toperform but also require a minimum of 24 to 48 h.

Flow cytometry has been described as an excellent tool for studyingthe susceptibility of different microorganisms, including fungi(9, 16, 20, 23). This methodology allows the timely determinationof susceptibility patterns and can provide additional informationregarding drug action and resistance mechanisms (17-19). Anoptimized cytometric assay to study the susceptibility of Candidaspp. with incubation for 2 h with amphotericin B and 4 h withfluconazole has been described recently and shows a good correlationwith the CLSI method (2).

We evaluated here a cytometric protocol to determine the susceptibilityof Candida spp. and Cryptococcus neoformans to the triazolesitraconazole and voriconazole and to the echinocandin caspofungin.The performance of the following two fluorescent markers wasevaluated: propidium iodide (PI), a marker of cell death (itonly penetrates cells with severe membrane lesions), and FUN-1,a probe that is converted by metabolically active fungi froma diffuse cytosolic pool (green-yellow fluorescence) to redcytoplasmic intravacuolar structures (10). Metabolically disturbedcells show an increase in the intensity of green-yellow fluorescence.

MATERIALS AND METHODS

Top

Materials and Methods

Isolates. Sixty-three clinical strains of Candida spp. (26 Candida albicans,4 C. glabrata, 5 C. guilliermondii, 7 C. parapsilosis, 16 C.krusei, and 5 C. tropicalis) and 3 Cryptococcus neoformans strainswere isolated from blood, urine, lower respiratory tract, andspinal fluid specimens. C. albicans ATCC 90028, C. parapsilosisATCC 22019, and C. krusei ATCC 6258 were also studied. All clinicalisolates were identified by using Vitek system (bioMérieux,Vercieux, France). Until testing, the yeasts were kept at –70°Cin brain heart infusion broth (Difco Laboratories, Detroit,MI) with 5% glycerol. Each strain was subcultured twice on Sabouraudagar (Difco) prior to testing to confirm its purity and viability.

Determination of MICs. MICs of voriconazole (Pfizer, Groton, CT) and itraconazole (Jansen,Beerse, Belgium) were determined according to the NCCLS M27-A2microdilution method guidelines (12); for caspofungin (Merck,Rahway, NJ), the same conditions described for azoles were used,but the MIC was defined as the lowest concentration in whicha score of 0 (optically clear) was observed.

Flow cytometry analysis. (i) Optimization of the cytometric protocol. Yeast cells were cultured with shaking (200 rpm) at 35°Cin Sabouraud broth until late log phase (as determined by agrowth curve constructed from absorbance readings at an opticaldensity of 600 nm). The cells were centrifuged, suspended inphosphate-buffered saline (Sigma) supplemented with 2% glucose(pH = 7.0), and counted in a Newbauer chamber. Suspensions containing10⁶ cells/ml were incubated with shaking at 35°C with eachantifungal agent for 1, 2, 4, and 6 h; the itraconazole andvoriconazole concentrations were 0.125, 0.5, 1, 2, 4, and 8µg/ml, and the caspofungin concentrations were 0.25, 1,2, 4, 8, and 16 µg/ml.

(ii) PI staining. After the antifungal treatment, yeast cells (10⁶) were washedto avoid quenching of fluorescence and stained with 1 µgof PI/ml (Sigma, St. Louis, MO) in the dark at room temperaturefor 30 min in HEPES solution (pH 7.2) supplemented with 2% glucose(GH solution) (16). Suspensions of untreated (drug-free) andkilled cells with 70% ethanol were stained using the same conditionsused with PI and used as controls.

(iii) FUN-1 staining. Antifungal treated and control cells (10⁶) were washed and stainedwith 0.5 µM FUN-1 [2-choro-4-(2,3-dihydro-3-methyl-(benzo-1,3-thiazol-2-yl)-methylidene)-1phenylquinoliniumiodide;Molecular Probes, Europe BV, Leiden, The Netherlands] in HEPESsolution (pH 7.2) supplemented with 2% glucose (GH solution)in the dark at room temperature for 30 min (16). Suspensionsof untreated (drug free) and cells treated with 1 mM sodiumazide (Sigma) for 1 h were stained under the same conditionswith FUN-1 and used as controls.

(iv) Cytometry analysis. After staining, ca. 30, 000 cells were analyzed (Beckman CoulterXL-MCL flow cytometer with an argon 15-mV laser) at least twicefor each cytometric reading. The cell scattergram, the autofluorescence,and the intensity of fluorescence at FL2 (yellow-green fluorescence,575 nm) and at FL3 (red fluorescence, 620 nm) were recordedby using a logarithmic scale. For PI, the results were expressedas the percentage of positive cells showing high fluorescenceat FL3; for FUN-1, the results were expressed as a stainingindex (SI), defined as the ratio between the mean fluorescenceof treated cell suspensions and the corresponding value forthe nontreated cells at FL2. The susceptibility phenotype wasdefined according to the lowest concentration of the antifungalagent that resulted in an SI above that of the control, thatis an SI of >1, after drug treatment (20). As described foritraconazole (12), MICs of $≤$ 0.125 µg/ml were consideredsusceptible, MICs between 0.25 and 0.5 µg/ml were consideredsusceptible-dose dependent, and MICs of $≥$ 1 µg/ml were consideredresistant. For voriconazole, MICs of $≤$ 1 µg/ml were consideredsusceptible, MICs of 2 µg/ml were considered susceptible-dosedependent, and MICs of $≥$ 4 µg/ml were considered resistant(newly established tentative breakpoints). Although, susceptibilitybreakpoints have not yet been established for caspofungin, strainsthat are inhibited by $≤$ 1 µg/ml could be considered susceptible(14). Whenever the SI was >1 after incubation with concentrations $≥$ 1 µg/ml for itraconazole, >1 µg/ml for caspofungin,and $≥$ 4 µg/ml for voriconazole, the strains were consideredresistant.

Viability assessment. The number of viable cells, both for treated and untreated suspensions,was determined by enumeration of CFU/ml on Sabouraud agar platesafter plating 100 µl of serial dilutions and incubationat 35°C for 48 h.

Agreement between both methods. The categorical agreement between CLSI standards and determinationby flow cytometry, after staining with FUN-1, was determinedby comparing results for each isolate evaluated.

RESULTS

Top

Results

Untreated cells stained with PI showed a very low intensityof fluorescence and, in contrast, ethanol-treated cells showeda high red fluorescence. The two histograms are so dissimilarthat the results were expressed as positive (dead cells, Fig.1A) or negative (viable cells, Fig. 1B) for PI. Candida cellstreated with both triazoles did not stain with PI, a findingwhich agrees with cell viability results; viability was notaffected in these experimental conditions, even with the highesttriazole concentrations after 6 h of incubation. Treatment withfour times MIC of caspofungin during 1 h did not result in PIstaining of yeast cells. Conversely, a growing percentage ofcells become PI positive as the drug concentration and timeof incubation increased (Fig. 2A). A maximum of 20% PI-positivecells were observed after 6 h (Fig. 1C and D), varying betweenstrains (10 to 20%) and, when incubated with the highest concentration(16 µg/ml), this percentage could be a maximum of 50%after 6 h, varying between susceptible strains (30 to 50%).These effects were not observed on resistant strains to thedrug. Nevertheless, after plating, the number of nonviable cellssurpassed the number of PI stained cells on susceptible strains(Fig. 2B). After 6 h with the highest concentration of caspofunginused (16 µg/ml) and despite the fact that more than 90%of the cells were nonviable, approximately 50% of the cellswere stained by PI (Fig. 2). This effect was also variable withthe strains.

View larger version (22K):
[in this window]
[in a new window]
FIG. 1. Three-dimensional plots illustrating yeast cells of a typical susceptible strain to caspofungin stained with PI. (A) Ethanol-treated cells, all of the population stained with PI. (B) Viable cells, PI negative. (C) Cells treated with 1 µg of caspofungin/ml (four times the MIC) during 1 h (most of the population did not stain with PI). (D) Cells treated with 1 µg of caspofungin/ml for 6 h (26.6% of the cells were PI positive).

View larger version (31K):
[in this window]
[in a new window]
FIG. 2. (A and B) Percentage of PI-positive (PI+) cells (A) and nonviable cells (B) after exposure of C. albicans, an example of a susceptible strain, to different concentrations of caspofungin for 6 h.

Yeast cells exposed to 1 mM sodium azide (used as a controlof metabolic impairment) and stained with FUN-1 showed an increasein fluorescence intensity compared to nontreated cells correspondingto an SI of >1 (Fig. 3A). The effect was similar for susceptiblestrains after incubation with low drug concentrations (0.125for itraconazole; 1 µg/ml for the other two), whereasthis effect only occurred among resistant strains after incubationwith the highest drug concentrations (Fig. 3). A dose-dependentand time-dependent increase in SI (SI > 1) was seen withthe three tested antifungals. These differences were evidentafter 1 h of incubation for the three antifungal agents, withan excellent correlation with NCCLS microdilution MIC results.The susceptibility phenotypes of the Candida and Cryptococcusspecies evaluated to itraconazole, voriconazole, and caspofunginby NCCLS and flow cytometry methods after FUN-1 staining areshown in Table 1.

View larger version (34K):
[in this window]
[in a new window]
FIG. 3. Histograms determined by flow cytometry analysis at FL2 log of typical Candida strains after different treatments for 1 h and stained with FUN-1. (A) Treatment with sodium azide as a control of the staining. (B) Treatment with voriconazole, representative of the azoles. (C) Treatment with caspofungin.

View this table:
[in this window]
[in a new window]
TABLE 1. Agreement between itraconazole, voriconazole, and caspofungin phenotypes of Candida spp. and C. neoformans as determined by NCCLS M27-A2 and flow cytometry (FC: FUN-1 staining) methods

DISCUSSION

Top

Discussion

Yeast represents at present the fourth leading organism causingsepticemia in the United States, Europe, and Australia (7, 15,22), in particular in intensive care units (4). The availabilityof rapid and reliable tools to determine the susceptibilityof yeasts, namely, Candida and Cryptococcus spp., to antifungalagents is mandatory. In the last few years, the susceptibilityto antifungal agents has been evaluated by several flow cytometrymethods (16, 17, 18, 19, 20, 21, 23); these methods can savetime and provide additional information regarding mechanismsof action and resistance of the drugs tested. We have alreadydeveloped cytometric protocols to study amphotericin B, flucytosine,and fluconazole (16, 19), but each antifungal class may requirea different protocol that takes into consideration its molecularcharacteristics and mechanisms of action. Cytometric methodologyhas also been evaluated for testing itraconazole with acridineorange, ethidium bromide, and fluorescein diacetate, but thesemethods require 8 to 24 h to yield results (9). A flow cytometryassay has recently been described to study Aspergillus fumigatusversus voriconazole using PI as the marker of fungicidal activity(21). However, although voriconazole may be fungicidal for Aspergillusspp., it is only fungistatic against yeasts. A single studyhas used cytometric methodology with caspofungin, using confocalmicroscopy after staining Candida cells with FUN-1, to studyits effect on planktonic Candida cells versus cells on biofilms(1).

Due to the lack of fast and reliable cytometric assays, we optimizedour cytometric methodology to study the susceptibility of pathogenicyeasts to these three antifungal agents. Two fluorescent probes,previously optimized by us (16), PI and FUN-1, were compared.

In the present study, the performance of the PI probe was limitedfor evaluating the two triazoles in vitro activities, similarto our results for fluconazole and flucytosine (16). This effectwas expected since they are only fungistatic drugs. Ramani etal. were able to stain with PI yeast cells treated with fluconazole,but only after using a detergent of the membrane, sodium deoxycholate,that facilitates the diffusion of PI into the cells (20). Althoughit was possible to appreciate the fungicidal effect of caspofunginagainst some Candida spp., the percentage was variable and lowwith the PI probe, mainly after a short incubation time. Caspofunginhas in vitro fungicidal activity against certain strains ofCandida spp. (5). Since PI did not penetrate the yeast cellsquickly, we could conclude that the fungicidal effect of caspofunginwas not initiated by producing primary cytoplasmic membranelesions but perhaps by inhibiting the synthesis of fungal cellwall (3). In contrast, when permeability occurs after a shortincubation time (15 min), the primary effect could be due tomembrane lesions (17, 18). However, it has been demonstratedthat although cells treated with amphotericin B were dead, therewas no diffusion of PI into these dead cells (16).

After uptake and accumulation in the cell cytoplasm, FUN-1 isconverted only by metabolically active yeasts. Sodium azide(1 mM) inhibits the respiratory metabolism, rendering cellsunable to process the probe (10). The intensity of fluorescenceat FL2 of azide-treated cells was significantly higher thanthat of nontreated cells, corresponding to an SI of >1. Accordingly,an SI of >1 after antifungal treatment indicated a seriousimpairment of yeast metabolism, having the advantage of beingan earlier marker of cell damage, before cell death occurs bya fungicidal agent. For susceptible strains, the fluorescenceintensity of cells treated with low drug concentrations increasedcompared to control cells (SI > 1). This effect only occurredin the resistant strains when challenged with the highest drugconcentrations (4 to 16 µg/ml); the effect was similarto that observed previously with fluconazole (14) and flucytosine(19, 14). We were able to identify itraconazole categoricalendpoints after 1 h of incubation with this agent; the overalltime in other studies for this incubation period is 4 to 6 hfor yeasts (20, 2) and 2 to 3 h for A. fumigatus (21).

Using the two fluorescent probes, we were then able to achievea better understanding of what is happening to the yeast cellsafter incubation with serial drug concentrations and have obtaineda kinetic perspective of the activity of caspofungin againstyeasts. In conclusion, flow cytometry analysis with FUN-1 provedto be an excellent tool and superior to the PI probe for studyingthe susceptibility of Candida spp. and C. neoformans to thesethree drugs. Once again, the cytometric approach to susceptibilitytesting has proven to have potential as an important tool inresearch and clinical laboratories to provide in vitro resultsfor clinical isolates in a timely manner. The reproducibilityand clinical utility of this methodology, especially for caspofungin,is yet to be determined.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbiology, Porto School of Medicine, Alameda Prof. Hernani Monteiro, 4200-Porto, Portugal. Phone: 00351-919078512. Fax: 00351-229962096. E-mail: cpinavaz{at}yahoo.com.

REFERENCES

Top