INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

New antifungal agents are urgently needed to combat an increasing number of life-threatening, systemic fungal infections. A fungicidal mode of action is highly desirable since many patients requiring antifungal therapy are immunocompromised (28). Enzymes required for fungal cell wall synthesis are excellent targets for development of nontoxic fungicidal drugs. Lipopeptide antifungals of the echinocandin class of compounds are noncompetitive inhibitors of (1,3)--D-glucan synthase of Candida spp. (24, 25) and Aspergillus spp. (2) and are also effective against other pathogenic fungi, such as Pneumocystis carinii (1, 12).

Previous studies demonstrated the utility of flow cytometric technology in detecting the rapid (3 h) fungicidal activity of cilofungin, a narrow-spectrum semisynthetic echinocandin B (ECB) analog. Those studies provided a positive correlation between flow cytometric detection of fluorescent-dye uptake and classical microbiological techniques (10) for measurement of fungicidal activity. In the present study, we used the same methods and a single-cell-deposition technique to compare the fungicidal activities of amphotericin B (AMB) and cilofungin to that of a new, more potent ECB analog, LY303366. We also varied the in vitro drug treatment conditions to examine relationships between compound exposure time and subsequent killing of target yeasts.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Yeast cultures. Candida albicans A26 (ATCC 90234) and Saccharomyces cerevisiae YPH499 were obtained from Eli Lilly & Co. (Indianapolis, Ind.) Infectious Disease Research stock cultures and maintained on yeast-peptone-dextrose (YPD; Difco, Detroit, Mich.) agar at 4°C. For experiments, organisms were grown overnight in YPD broth in 32°C shaker bath and diluted to 2 × 10⁶ cells/ml in fresh YPD broth just before use.

Antifungal agents. Antifungal agents were diluted in their appropriate solvents (listed below) to prepare 1.0-mg/ml stock solutions. Cilofungin (Eli Lilly & Co.) was diluted in ethanol, LY303366 (Eli Lilly & Co.) was dissolved in reagent-grade methanol, and crystalline AMB (Sigma, St. Louis, Mo.) was dissolved in dimethyl formamide. Fresh dilutions of compound in YPD broth were prepared for each experiment.

Compound treatments. Active cultures and experimental compounds were both diluted in YPD broth (1 ml each), combined in sterile 12- by 75-mm polypropylene test tubes, and placed in a 32°C shaker bath for defined time intervals. For pulsed-compound-exposure experiments, 2 ml of phosphate-buffered saline (PBS) was added to tubes after pulsed exposure, the tubes were centrifuged for 5 min at 400 × g, and the supernatant was aspirated from these pulsed cultures. The pulsed-exposure cells were then resuspended in 2.0 ml of fresh YPD broth containing a drug vehicle, whereas the continuous-exposure cells were resuspended in the existing broth containing the drug and vehicle. Tubes were returned to the 32°C shaker bath to complete a 3- or 5-h incubation period. In experiments where cultures were exposed to drug for the last 5 min of the experiment, 20 µl of a 100× drug solution was added to the 2-ml culture, and further processing for propidium iodide (PI) staining, sorting, or CFU assay began after the 5-min exposure.

PI staining and flow cytometric analysis. For PI staining of cultures, cells were pelleted by centrifugation for 5 min at 400 × g, the supernatant was discarded, and the cells were resuspended in PI (Sigma) solution (25 µg/ml in PBS) for 20 min at room temperature before analysis on an EPICS Elite flow cytometer (Coulter Corp., Hialeah, Fla.). During flow cytometric analysis, individual cells were detected and categorized by their forward-angle and right-angle scatter of incident 488-nm laser light as displayed in Fig. 1A. This light scatter data plot was used to establish a gated region that excluded cell clusters (>2 cells) from the fluorescence analysis. The light scatter gated PI fluorescence of individual cells was acquired by using a 630-nm band-pass filter (FL3) and displayed in single-parameter histograms (Fig. 1B through E). By using the cytometer's onboard software, the fluorescence histograms were integrated to determine the percentage of PI (living) and PI⁺ (dead) yeast cells.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Flow cytometric evaluation of treated C. albicans cells that were stained with PI. Actively growing yeast cultures were treated with fungicidal agents as described in the text, mixed with PI, and analyzed on the flow cytometer. (A) Light scatter gating region used for the analysis of red fluorescence of each sample. FS, forward scatter; SS, side scatter. (B through E) Red fluorescence histograms displaying the integrated percentage of PI+ cells. The treatments for the panels were medium only (control) (B), 70% ethanol (C), LY303366 for 3 h (D), and LY303366 for a 5-min pulse (E).

Flow cytometric cell sorting. Vehicle control or compound-treated yeast cells were single-cell sorted with a Coulter EPICS Elite ESP cell sorter as described below. For each test sample, the light scatter data were acquired and used to establish a sorting region that included only single cells and budding cells. By using the cell sorter's capabilities, single cells or budding cells were electrostatically deflected into individual microculture wells containing 200 µl of YPD broth in 96-well tissue culture plates (Becton Dickinson Labware, Franklin Lakes, N.J., item 3072). These plates were incubated at 30°C without shaking for 48 h and then visually inspected for yeast growth. The percentage of single cells that formed visible colonies in wells (percent single-cell growth [SCG%]) was determined as follows: (no. of wells with a visible yeast colony after 48 h/total no. of wells inoculated by single-cell deposition) × 100.

CFU assay. Aliquots of cultures were removed just prior to PI staining or sorting and serially diluted in PBS. Selected dilutions were plated in duplicate on YPD agar. The plates were incubated for a minimum of 48 h at 30°C, colonies were counted, and results were expressed as the percent reduction in CFU per milliliter compared to the growth in nontreated control cultures.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of LY303366 on PI staining of yeast. Antifungal effects can be detected by flow cytometric analysis of PI-treated yeast cells (10). Certain antifungal agents increase the PI staining of yeast. The effects of two fungicidal agents on C. albicans from a representative experiment are displayed in Fig. 1. Figure 1A shows the light scatter gating region used for analysis of red fluorescence. In Fig. 1B, the PI fluorescence for untreated C. albicans cultures illustrates the very dim autofluorescence profile attained. Figures 1C and D display the PI fluorescence for other treatments of C. albicans. All cells that developed a PI fluorescence intensity greater than that noted for the untreated control cells (Fig. 1B) were defined as PI⁺. Treatment of cells with 70% ethanol (Fig. 1C) gave a >100-fold increase in PI fluorescence intensity, resulting in virtually all of the cells becoming PI⁺. Continuous treatment of cultures with the minimal fungicidal concentration (MFC), the lowest concentration of drug resulting in more than 99% killing of cells, of LY303366 (for C. albicans, 0.01 µg/ml) for 3 h (Fig. 1D) resulted in 87.7% of the C. albicans cells becoming PI⁺. Finally, cultures that were pulse treated with LY303366 at the MFC for only 5 min (Fig. 1E), followed by one wash and further incubation in compound-free broth for the remainder of the 3-h incubation, displayed a similar percentage of PI⁺ cells (85.7%) as did those that were continuously treated during the 3-h incubation. Virtually identical results were obtained when S. cerevisiae cells were similarly tested (data not shown).

Determination of SCG% of PI⁺ and PI cultures. S. cerevisiae cultures were treated with a moderately potent concentration of LY303366 (0.02 µg/ml) for 3 h, stained with PI, and analyzed on the cytometer. Data from a representative experiment are summarized in Fig. 2. From this PI fluorescence profile, cell-sorting regions were established around PI and PI⁺ regions (Fig. 2). Single cells from these two regions were deposited into microculture wells of 96-well plates and incubated at 30°C for 48 h. The number of wells (out of 48 from each condition per plate) that displayed positive yeast growth are shown in Fig. 2. These data demonstrate that after treatment with LY303366, virtually all the Saccharomyces cells that stained with PI were nonviable (SCG% = 4.2), while the PI cells exhibited a SCG% (95.8) equal to that of the untreated controls (96.0). These data also indicate that PI itself (in the absence of fungicidal agents) does not alter yeast SCG%. Similar results were obtained with C. albicans (data not shown).

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Single-cell sorting of cells treated with LY303366 and stained with PI. Active cultures of S. cerevisiae were treated with LY303366 (0.02 µg/ml) for 3 h and mixed with PI. The cells were analyzed on the cytometer, and individual cells were sorted into microwell cultures based on their red fluorescence intensity as defined by regions M2 (PI) and M1 (PI+). After 2 days of static incubation at 30°C, the microwells were visually inspected for determination of cell growth. The number of wells (out of 48 used) with S. cerevisiae growth is displayed for each of three plates in this representative experiment.

Concentration response of LY303366 on PI staining and single-cell sorting. We studied the effect of LY303366 concentration on S. cerevisiae viability. In addition to studying the effect of LY303366 on PI uptake, we also utilized a cell sorter to confirm the fungicidal effect. This instrument facilitates automated, selective single-yeast-cell deposition into microculture wells of a 96-well plate. In the first series of cell-sorting experiments, we examined the effect of treatments of various concentrations of LY303366 on S. cerevisiae viability by using both the PI-staining technique and single-cell sorting during the same experiments. Results of these studies are displayed in Fig. 3. After a 3-h incubation with LY303366, the percentage of PI⁺ cells (Fig. 3A) increased with increasing compound concentration. There was a concomitant decrease in the SCG% in the wells, as determined by cell sorting (Fig. 3A), as the concentration of LY303366 increased. Linear-regression analysis of this data (Fig. 3B) demonstrated a very high correlation (r = 0.998) of effect for these two measurements of cell viability. The same correlation was observed in experiments using C. albicans (data not shown).

View larger version (13K): [in this window] [in a new window]   FIG. 3.   PI staining and SCG% for S. cerevisiae treated for 3 h with LY303366. (A) Active yeast cultures were treated with twofold serial dilutions (0.005 to 0.16 µg/ml) of LY303366 and either stained with PI and flow cytometrically analyzed () or single-cell sorted for SCG% determinations (). Each data point represents the mean ± 1 standard error of the mean (SEM) for four experiments. (B) Mean SCG% and mean percentage of PI+ cells for each LY303366 concentration tested. The associated linear-regression line and correlation coefficient are shown.

Time course of LY303366 fungicidal activity. We compared PI staining and SCG% for C. albicans cultures that were incubated with LY303366 throughout a 3-h period. At 30-min intervals, aliquots of LY303366 (0.01 µg/ml)-treated cultures were either analyzed for PI fluorescence or sorted as single cells into the wells of 96-well microculture plates. As shown in Fig. 4, percentages of PI⁺ cells increased throughout the 3-h period with an attendant decrease in SCG%. After 180 min of treatment, 78.1% of the cells were PI⁺ and had an SCG% of 13.8. 

View larger version (17K): [in this window] [in a new window]   FIG. 4.   Time course of the fungicidal activity of LY303366 against C. albicans. An actively growing C. albicans culture was incubated with LY303366 (0.01 µg/ml) for a total of 3 h at 32°C. At 30-min intervals, an aliquot of treated culture was removed and either cell sorted for determination of SCG% () or stained with PI () and flow cytometrically analyzed. Each data point represents the mean percentage of each activity ± 1 standard error of the mean (SEM) for three separate experiments. The mean SCG% and the mean percentage of PI+ cells for 3-h control cultures were 95 and <2, respectively.

Fungicidal effects of an initial 5-min exposure to antifungals. Results from a previous study with ECB analogs (10) demonstrated that ECB analogs kill C. albicans rapidly. In that study, treatment with cilofungin caused 90% of the C. albicans cells to become PI⁺ within 3 h. In the present study, we modified that method to include compound removal (by washing with PBS) in order to investigate the effects of even shorter exposure times on PI fluorescence and reduction in CFU. The data in Table 1 demonstrate that an initial 5-min exposure to LY303366 was equally as potent as a full 3-h exposure period. A 5-min pulse with LY303366 at the MFC (0.01 µg/ml) followed by a 175-min incubation in drug-free broth resulted in a mean of 88.3% PI⁺ cells and an average reduction in CFU per milliliter of 99.7%, compared with a mean of 87.0% PI⁺ cells and an average reduction in CFU per milliliter of 99.6% for the 3-h continuous drug treatment. While the 3-h PI assay correlates well with the CFU and SCG% results for LY303366 (Fig. 2), in this study a 5-h PI assay underestimated the killing activity of AMB as determined by CFU assays. Continuous or initial 5-min pulsed exposure to AMB at 2.5 µg/ml (four times the MFC) produced lower percentages of PI⁺ cells than with the LY303366 treatments. An initial 5-min pulsed exposure to AMB was less effective than a continuous 5-h incubation in AMB, as measured by the PI and CFU reduction assays (Table 1).

Determination of minimum fungicidal exposure time. We used the pulsed-exposure technique and PI fluorescence analysis in conjunction with CFU assays to compare the potencies and activities of selected antifungals for C. albicans. Cultures were incubated with antifungals for defined periods, washed with PBS, further incubated in compound-free YPD broth (for 180 or 300 min total), and tested by the PI fluorescence and CFU assays. The concentrations tested were multiples of the MFCs of the antifungal compounds for each strain as previously established by a conventional broth microdilution technique (9). After multiple drug concentrations were tested over various exposure periods, the minimum compound exposure times that induced the maximal fungicidal activity of each compound were determined; they are summarized in Table 2. The minimum exposure time required for LY303366 at the MFC (0.01 µg/ml) to exert its maximal fungicidal effect was only 5 min. Cilofungin at a concentration of 1.6 µg/ml (four times the MFC) was not effective in pulsed exposures below the full 180-min period. For AMB, the minimal exposure time determination shown in Table 2 was based on CFU results, since the PI assay underestimates C. albicans killing by this compound (Table 1). A 30-min exposure (2.5 µg/ml, or four times the MFC) was as effective in the CFU assay as a full 5-h exposure.

Effect of timing of pulsed exposure to LY303366 on fungicidal activity. C. albicans cultures were exposed to LY303366 (0.01 µg/ml) by three different treatment regimens in order to help elucidate the importance of an active-growth period following compound exposure for manifestation of fungicidal effects. Four cultures were started, with the first being an untreated control. The second was incubated with LY303366 continuously for 3 h. The third culture was pulse incubated with LY303366 for 5 min, washed in PBS, and returned to a 32°C shaker bath for 175 min. The fourth culture was incubated without compound for 175 min, followed by the addition of LY303366 during the last 5 min. At 180 min, aliquots from each test culture were assayed by the PI fluorescence, CFU reduction, and SCG% assays. Data from three such experiments are summarized in Table 3. Results obtained by all three methods indicate that an initial 5-min pulsed exposure to LY303366, followed by 175 min of incubation without compound, was as effective in killing C. albicans cells as a continuous 3-h exposure. In contrast, a 5-min exposure, which occurred at the end of the 180-min experiment and was analyzed immediately for antifungal activity, displayed a significantly decreased fungicidal effect.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Treatment options for serious fungal infections are currently limited. New antifungal drugs are urgently needed, especially compounds with a fungicidal mode of action. In this study, we used flow cytometric analysis, single-cell-sorting techniques, and standard microbiological methods to investigate the activity of LY303366 against C. albicans and S. cerevisiae. Our results indicate that LY303366 is a highly potent and rapidly acting fungicidal compound. While our methods are novel, our results are consistent with an accumulating body of knowledge regarding the potency and broad-spectrum fungicidal activity of this promising antifungal compound (11, 20, 26).

Advancements in flow cytometry instrumentation have pushed the technology into microbiology laboratories worldwide. A thorough review of flow cytometric applications for microbiology was published recently (4). Flow cytometry offers unique advantages for studying the fungicidal activities of selected compounds against yeasts, providing the ability to examine the effects of compounds on thousands of individual cells within seconds. This type of analysis facilitates the understanding of the heterogeneous nature of yeast populations with respect to drug sensitivity. In a previous study (10), members of our group showed how flow cytometric measurement of PI fluorescence could distinguish the activity of fungicidal agents, such as cilofungin, from that of a fungistatic compound such as fluconazole. The PI method was particularly well suited for determining the extent of fungal cell killing by cilofungin; more than 90% of the cells treated with cilofungin became PI⁺ within 3 h of treatment at the MFC. In the present study, we applied this method to investigate the antifungal effects of a more potent ECB analog, LY303366. Using this approach, we demonstrated an approximately 40-fold increase in potency of LY303366 over cilofungin. Other investigators, using standard in vitro and in vivo microbiology assays, have demonstrated that LY303366 has potent, broad-spectrum fungicidal activity (11, 16, 23, 29-31).

The additional cell-sorting experiments described in this paper supported our hypothesis that LY303366-treated, PI⁺ yeast cells were indeed nonviable (Fig. 2). This confirmation is important because others have shown that certain bacteria may exhibit staining with selected viability dyes and yet be able to recover and grow (4, 7, 13). In addition, other studies have indicated that treatment of yeast with the fungistatic agent fluconazole can increase PI fluorescence slightly, yet these cells resume their growth and exhibit normal PI-staining patterns after removal from compound (19).

Results from the concentration response and time course experiments (Fig. 3 and 4) demonstrate that the PI assay and single-cell sorting for determination of SCG% complement each other. In the LY303366 concentration response experiments, PI staining correlated (r > 0.99) with decreased SCG%, further validating the PI method for measuring fungicidal activity of LY303366 against C. albicans and S. cerevisiae.

After confirming the lethal effects of a 3-h treatment with LY303366, we then focused on determining the minimal exposure period sufficient to kill C. albicans by comparing pulsed-compound exposures to continuous 3- or 5-h incubations. Using this approach, we determined (Table 2) that binding of LY303366 to C. albicans is rapid; exposure to the MFC (0.01 µg/ml) for 5 min killed >99% of the cells. These data demonstrate that chemical modifications of the fatty acid side chains of the ECB nucleus (6) incorporated in the synthesis of LY303366 resulted in an agent with much greater potency than cilofungin. The rapid killing of yeast cells by LY303366 after a short exposure period may contribute to its potent efficacy in vivo (3).

The rapid action and enhanced potency of LY303366 compare favorably with those of AMB, a polyene antifungal that has been used clinically since the 1950s (8). AMB has been reported to bind ergosterol in fungal cell membranes within 30 min (27). Cell membrane permeability is increased, first for potassium ions and then for other cell constituents, resulting in cell death within hours after exposure. Binding of AMB was recently shown to decrease yeast plasma membrane potential within 30 min of AMB exposure (18). Flow cytometric analysis of PI staining has been used to study AMB activity against yeasts (22) and to investigate serum effects (15) or synergy (21) or antagonism (14) between antifungals. Our studies also demonstrated that a 30-min incubation with AMB has an effect on C. albicans, an effect measured by increased PI fluorescence (Table 2). Although AMB binds rapidly, our data indicate that AMB does not increase yeast cell membrane permeability to PI as early as LY303366 does. C. albicans cultures treated with AMB for 5 h display only one-fifth of the PI⁺ cells displayed by cultures treated with LY303366 for 3 h (Table 1). Some investigators have used deoxycholate to enhance PI entry into AMB-treated cells (14). AMB is highly effective against most clinical fungal isolates (17); however, nephrotoxicity limits its clinical utility. Among its effects, LY303366 targets a fungus-specific enzyme, glucan synthase (5), which may contribute to an improved toxicity profile.

Because LY303366 binds rapidly to yeast cells and has an inhibitory effect on cell wall synthesis, its antifungal effects would be expected to occur over time in a growing culture. One paradox emerged from the pulsed-exposure experiments: while a 5-min pulsed exposure to LY303366 was shown to be sufficient to eventually kill target cells (Table 1), deposition of single cells into microwells following 30 min of incubation with LY303366 resulted in a smaller-than-expected decrease in SCG% (Fig. 4). We hypothesized that active-growth conditions are important for maximal killing activity of LY303366. To test this idea, we conducted experiments in which cells were treated with LY303366 for 5 min, with or without an active-growth period (incubation in a shaker water bath) for 175 min in compound-free medium (Table 3). The PI staining, CFU, and SCG% assays all indicated that a 5-min pulse with compound without the active-growth period resulted in a significantly reduced fungicidal effect compared to the effect on cultures that were actively grown in a shaking water bath for an additional 175 min. To reconcile these results, we considered the fact that for each of the three fungicidal activity assays, cells were suspended in drug-free medium following the drug pulse. It is possible that when LY303366-treated cells are suspended in drug-free broth, some of the drug will come off of the cells. We hypothesize that if yeast cells remain in log-phase growth, as promoted by the shaker bath, most will begin cell division and, therefore, die before much drug comes off. With static culture conditions following the drug pulse, such as the microtiter well or agar plate, drug release could become a more significant factor because fewer cells will divide while a lethal-threshold level of drug remains bound to the cell. Our results imply that yeasts growing at lower rates may require somewhat longer exposure periods in order to be killed by this agent.

In summary, we have utilized flow cytometric analysis, cell sorting, and standard microbiological plating methods to study the fungicidal activities of selected antifungals against S. cerevisiae and C. albicans. Our data indicate that cell sorting may be a valuable tool for studying and differentiating various antifungal activities. The PI fluorescence and cell-sorting methods provide tools for evaluating certain modes of action that are not easily accomplished by standard microbiological techniques. In addition, our data indicate that LY303366 is a potent, rapidly acting antifungal agent. In our experiments, we confirmed that LY303366-treated PI⁺ yeast cells are nonviable. Our data also suggest that LY303366 works best against actively growing yeasts. The potency and rapid action of LY303366 suggest that it may be useful for antifungal therapy of human disease.