Time-Kill Assay and Etest Evaluation for Synergy with Polymyxin B and Fluconazole against Candida glabrata

Abstract

Fluconazole-resistant Candida glabrata is an emerging pathogen that causes fungemia. Polymyxin B, a last-resort antibiotic used to treat multidrug-resistant Gram-negative bacterial infections, has been found to possess in vitro fungicidal activity and showed synergy with fluconazole against a single strain of C. glabrata. Since both agents may be used simultaneously in intensive care unit (ICU) patients, this study was performed to test for possible synergy of this combination against 35 C. glabrata blood isolates, using 2 methods: a time-kill assay and an experimental MIC-MIC Etest method. Thirty-five genetically unique C. glabrata bloodstream isolates were collected from 2009 to 2011, identified using an API 20C system, and genotyped by repetitive sequence-based PCR (rep-PCR). MICs were determined by Etest and broth microdilution methods. Synergy testing was performed using a modified bacterial Etest synergy method and time-kill assay, with final results read at 24 h. The Etest method showed synergy against 19/35 (54%) isolates; the time-kill assay showed synergy against 21/35 (60%) isolates. Isolates not showing drug synergy had an indifferent status. Concordance between methods was 60%. In vitro synergy of polymyxin B and fluconazole against the majority of C. glabrata isolates was demonstrated by both methods. The bacterial Etest synergy method adapted well when used with C. glabrata. Etest was easier to perform than time-kill assay and may be found to be an acceptable alternative to time-kill assay with antifungals.

INTRODUCTION

Candida glabrata has been a rising cause of candidemia in the past few years, second only to Candida albicans (1). According to the Centers for Disease Control and Prevention (CDC), candidemia is the fourth most common hospital-acquired infection in the United States (2). C. glabrata has been known to develop resistance to the azole family of antifungals, the most inexpensive and easily accessible medications to treat candidemia. Further data from the CDC have shown that 7% of Candida species have become fluconazole resistant (2). This development of resistance is most likely due to the routine use of azoles in treatment or prophylaxis. Other treatment options for Candida infections include amphotericin B and the echinocandin family (to which resistance by C. glabrata has occurred) (1). A U.S. surveillance study evaluating Candida isolates from San Francisco, CA, Atlanta, GA, Baltimore, MD, and Connecticut (from 1998 to 2010) determined that C. glabrata isolates are able to adapt to new environments to increase their fitness and antifungal resistance (3). Another surveillance study, evaluating developing resistance in fluconazole-susceptible isolates collected from 2008 to 2011, showed that among 2,329 Candida isolates, the prevalence of azole resistance was unchanged (7%), 32 of the isolates were echinocandin resistant, and 8 C. glabrata isolates were resistant to both fluconazole and echinocandins (4).

Polymyxin B is a last-resort antibiotic used to treat multidrug-resistant Gram-negative bacterial infections. It has been found to possess in vitro fungicidal activity (5, 6). In addition, polymyxin B showed synergy with fluconazole against a single strain of C. glabrata (7) and a single strain of C. albicans (8), although it is not routinely used in the treatment of fungal infections. Because of polymyxin B's reported fungicidal activity and synergy when used in combination with fluconazole, we chose it to further evaluate this finding, using 35 genetically unique C. glabrata bloodstream isolates. The goal of our study was 2-fold: (i) to determine if in vitro synergy was present with the combination of fluconazole and polymyxin B against C. glabrata and (ii) to evaluate a rapid Etest method and an antifungal time-kill assay (TKA) for the detection of synergy.

(Some of the data reported here were presented at the 52nd ICAAC, San Francisco, CA, September 2012 [9], and at the 12th ASM Conference on Candida and Candidiasis, New Orleans, LA, March 2014 [10].)

MATERIALS AND METHODS

Microorganisms, media, and antimicrobial agents.

Thirty-five C. glabrata bloodstream infection isolates were collected in 2009 to 2011 from individual patients at Ochsner Health System, New Orleans, LA. Approval for the study was granted by the Institutional Review Board of the Ochsner Clinic Foundation. Isolates were identified using the API 20C yeast identification system (bioMérieux, Inc., Durham, NC). Sabouraud dextrose agar plates (Becton-Dickinson Microbiology Systems, Sparks, MD) were used for initial subculture of isolates and spiral plating during the TKA; RPMI 1640 agar plates with morpholinepropanesulfonic acid (MOPS) and 2% glucose (Remel, Lenexa, KS) were used for the Etest MIC determinations and synergy tests. Etest strips for fluconazole and polymyxin B (bioMérieux, Inc.) were utilized. Quality control testing was performed with Candida parapsilosis ATCC 22019, C. albicans ATCC 90028, and Pseudomonas aeruginosa ATCC 27853 (11, 12). Standard laboratory powders of fluconazole and polymyxin B (Sigma-Aldrich, St. Louis, MO) were used for determination of broth microdilution (BMD) MICs and for TKA. A stock solution of each agent was prepared in RPMI 1640 buffered with 0.165 mol/liter MOPS (Sigma-Aldrich), with the pH adjusted to 7.0 by use of 1 mol/liter sodium hydroxide. Dimethyl sulfoxide (DMSO) was used as the solvent for fluconazole, and the final DMSO concentration in the test solution was 1%. DMSO at this concentration (without drug) was used in the test as a dilution control (13). RPMI medium was sterilized using a 0.22-μm polystyrene filter. Polymyxin B was prepared in deionized water. Stock solutions of both agents were also sterilized using 0.22-μm filters, aliquoted, and stored at −70°C until used.

Antimicrobial susceptibility testing. (i) Broth microdilution MICs.

The MIC of each antimicrobial was determined following BMD guidelines as described by the Clinical and Laboratory Standards Institute (CLSI) (13). CLSI breakpoints used for interpretation of fluconazole MICs are ≤32 μg/ml for susceptibility-dose dependence (SDD) and ≥64 μg/ml for resistance (R) (11). For the SDD category, susceptibility is dependent on achieving the maximum possible blood level. For fluconazole, doses higher than the standard dosing (6 mg/kg of body weight/day) amount may be required in adults with normal renal function and body habitus. There are no CLSI interpretive guidelines for testing polymyxin B against C. glabrata. The inoculum was prepared by picking 3 to 5 colonies from a 24-h culture on Sabouraud dextrose agar, preparing a suspension in 5 ml of sterile water, and vortexing the suspension for 15 s. The cell density was adjusted to the turbidity of a 0.5 McFarland standard. The stock solution was diluted, and the final inoculum was 0.5 × 10³ to 2.5 × 10³ CFU/ml. Microtiter trays were inoculated and incubated at 35°C. MICs were read and recorded at 24 h. The fluconazole MIC was read as the lowest concentration at which a prominent decrease in turbidity (approximately 50% as determined visually) was observed. The MIC of polymyxin B was read as the lowest concentration to show complete inhibition of visual growth.

(ii) Etest MICs.

MICs were also determined by Etest (bioMérieux, Inc.) in triplicate, following the manufacturer's guidelines for Candida species; mean values were used. The inoculum was prepared by picking 3 to 5 colonies from a 24-h culture on Sabouraud dextrose agar, preparing a suspension in 5 ml of sterile water, and vortexing the suspension for 15 s. The cell density was adjusted to the turbidity of a 0.5 McFarland standard. RPMI plates were inoculated twice with the suspension. RPMI agar plates were incubated at 35°C in a loosely folded plastic bag to maintain moisture. MICs were read and recorded at 24 h. The fluconazole MIC was read as the first point of significant inhibition of growth or 80% inhibition of visual growth. The MIC of polymyxin B was read as the lowest concentration to show complete inhibition of visual growth.

Synergy testing. (i) Etest synergy method.

A modified Etest synergy method (14) previously reported for testing antibacterial combinations was performed in triplicate to test combinations for antifungal activity (Fig. 1 and 2). Inoculum preparation and the test medium were the same as those previously described for Etest MIC determination. Fluconazole and polymyxin B Etest strips were placed on different sections of an RPMI agar plate. The agar was marked adjacent to the previously determined MIC value on each strip (for fluconazole, the MIC; and for polymyxin B, 1/2 MIC). For isolates whose fluconazole MIC exceeded the concentration on the Etest strip, the highest concentration (256 μg/ml) was marked on the agar. The strips were removed after incubation for 1 h at room temperature. A new fluconazole strip was placed on the area of the previously removed polymyxin B strip so that the fluconazole MIC corresponded with the mark of the 1/2 MIC for polymyxin B. The polymyxin B Etest strip was applied in reciprocal fashion to the area of the previous fluconazole strip so that the respective MIC values were aligned. RPMI agar plates were incubated at 35°C in a loosely folded plastic bag to maintain moisture. The resulting combination MICs were read and recorded at 24 h. The fluconazole MIC was read as the first point of significant inhibition of growth or 80% inhibition of visual growth. The MIC of polymyxin B was read as the lowest concentration to show complete inhibition of visual growth.

FIG 1.

Etest MICs. Etest MICs for isolate 4 were as follows: polymyxin B (PO) MIC, 512 μg/ml (left); and fluconazole (FL) MIC, >256 μg/ml (right).

FIG 2.

Etest synergy test. MICs for isolate 4 were as follows: original polymyxin B 1/2 MIC, 256 μg/ml; polymyxin B 1/2 MIC after combination with fluconazole, 64 μg/ml; original fluconazole MIC, >256 μg/ml; fluconazole MIC after combination with polymyxin B, 64 μg/ml. ΣFIC = 0.3 (synergy).

To evaluate the effect of the combination in the Etest method, the fractional inhibitory concentration (FIC) was calculated for each antibiotic in each combination, as follows: FIC of fluconazole = MIC of fluconazole in combination/MIC of fluconazole alone, and FIC of polymyxin B = (1/2 MIC of polymyxin B in combination)/(1/2 MIC of polymyxin B alone). The total (summation) fractional inhibitory concentration (ΣFIC) for each isolate was calculated according to the following formula: ΣFIC = FIC of fluconazole + FIC of polymyxin B. To calculate ΣFICs, high, off-scale MICs (>256 μg/ml) were converted to the nearest 2-fold dilution (512 μg/ml), and final ΣFIC values were rounded up to the nearest tenth (e.g., 0.26 was rounded to 0.3). The mean ΣFIC was used to interpret results of the Etest synergy method. “Synergy” was defined as having a ΣFIC of ≤0.5, “indifference” as having a ΣFIC of >0.5 to 4, and “antagonism” as having a ΣFIC of >4 (15).

(ii) Time-kill assay.

There are no CLSI standardized methods available for antifungal time-kill testing. Therefore, antifungal time-kill studies were carried out by a method previously described and evaluated by Klepser et al. (16). Each isolate was tested against fluconazole (MIC) and polymyxin B (1/2 MIC), alone and in combination. MIC values obtained by broth microdilution were used in the time-kill studies. A starting inoculum was prepared by inoculating a tube of sterile water with growth from a 24-h Sabouraud dextrose agar plate and adjusting it to the turbidity of a 0.5 McFarland standard. After vortexing, an aliquot was added to each tube of RPMI 1640 medium with MOPS (containing one or both drugs) to give an inoculum of approximately 10⁵ CFU/ml. One tube without drug was included as a growth control for each isolate. A negative control (growth medium without drug or organism) was also included on each day of testing. Tubes were incubated at 35°C on an orbital shaker and vortexed prior to removing a sample for the determination of colony counts. At predetermined time points (0, 24, and 48 h), samples were aseptically removed, serially diluted in saline, if necessary, and plated onto Sabouraud agar plates in duplicate, using a spiral plater (Spiral Biotech, Norwood, MA). Plates were incubated at 35°C for 24 to 48 h and then scanned and counted using a QCount automated colony counter (Spiral Biotech) that detected colony counts as low as 20 CFU/ml. The mean colony count (CFU/ml) from duplicate samples was used to determine synergy. Sampling methods were evaluated for any antifungal carryover, following the procedure previously described by Klepser et al. (17). No antifungal carryover was observed for fluconazole at the MIC. Any TKA results found to be discordant with the Etest synergy method results were repeated and confirmed the initial time-kill results. “Synergy” was defined as showing a ≥2-log₁₀ decrease in colony count after 24 h with the combination compared to that with the most active single agent alone, “indifference” as showing a <2-log₁₀ increase or decrease in colony count at 24 h with the combination compared to that with the most active single agent alone, and “antagonism” as showing a ≥2-log₁₀ increase in colony count after 24 h with the combination compared to that with the most active single agent alone (15).

RESULTS

Antimicrobial susceptibility testing.

Etest MICs (μg/ml) were as follows: for fluconazole, 8 to >256 (77% SDD and 23% R); and for polymyxin B, 32 to 1,024 (no interpretive guidelines are available for testing of Candida spp.). Broth microdilution MICs (μg/ml) were as follows: for fluconazole, 0.5 to 256 (80% SDD and 20% R); and for polymyxin B, 32 to 256 (no interpretive guidelines are available for testing of Candida spp.). Fluconazole and polymyxin B MIC values were considered to be in essential agreement (EA) between the broth microdilution and Etest methods when MICs were within two 2-fold dilutions. EA of fluconazole MICs obtained by the broth microdilution and Etest methods was high (91%). EA between the broth microdilution and Etest methods for polymyxin B MIC values was excellent (100%) (Table 1).

TABLE 1.

MIC (μg/ml) results by BMD and Etest

Antimicrobial agent	Method	MIC range	Median MIC	Agreement between BMD and Etest (no. of isolates with agreement/total no. of isolates [%])
				Essential agreement	Categorical agreement
Fluconazole	BMD	0.5-256	16	32/35 (91)	32/35 (91)
	Etest	8–>256	16
Polymyxin B	BMD	32-256	256	35/35 (100)	—a
	Etest	32-1,024	192

^a—, not calculated because interpretive guidelines for polymyxin B with Candida spp. are not available.

Categorical agreement (CA) was also evaluated for fluconazole MICs. CA referred to fluconazole MICs that fell within the same interpretive category (susceptible-dose dependent or resistant). CA of fluconazole MICs determined by BMD and Etest was high (91%). Three isolates did not show CA. Two were SDD by BMD but R by Etest, and one isolate was R by BMD but SDD by Etest (Table 2). This finding is similar to fluconazole MIC results for C. glabrata previously reported by Borghi et al. (18). There is a trailing growth that occurs in broth microdilution assays with fluconazole and some Candida isolates, causing isolates to appear susceptible after 24 h but completely resistant at 48 h (13). In vivo studies using murine models have shown that this type of isolate should be categorized as susceptible rather than resistant (19, 20). Therefore, results from the 24-h reading (for both methods) were used for our study, as recommended by CLSI and Etest guidelines (13, 21).

TABLE 2.

Fluconazole MICs by BMD and Etest that were not in categorical agreement

C. glabrata isolate (n = 3)	Fluconazole MIC (μg/ml)
	BMD	Etest
Isolate 2	8	48
Isolate 19	64	16
Isolate 7	32	48

Synergy testing.

Results determined by both synergy methods are presented in Table 3. The Etest method showed synergy of the combination of fluconazole (MIC) and polymyxin B (1/2 MIC) against 19/35 (54%) C. glabrata isolates (ΣFIC values, 0.3 to 0.5), including 6/8 (75%) fluconazole-resistant isolates. In comparison, the TKA showed synergy against 21/35 (60%) isolates (−2.0- to −4.0-log₁₀ change after 48 h), including 7/8 (88%) fluconazole-resistant isolates. Fungicidal activity (≥3-log₁₀ reduction in CFU/ml) of the combination was demonstrated against 10/21 (48%) isolates by the TKA. Isolates that did not exhibit synergy by either method were termed indifferent. No antagonism was seen. Agreement between the Etest method and TKA was found with 21/35 (60%) isolates (Table 4).

TABLE 3.

Fluconazole and polymyxin B MICs by BMD and Etest and synergy test results for Etest and time-kill assay

C. glabrata blood isolate (n = 35)	Fluconazole MIC (μg/ml)			Polymyxin B MIC (μg/ml)			Synergy test result (value, description)b
	BMD	Original Etesta	Etest with polymyxin B in combinationa	BMD	Original Etesta	Etest with fluconazole in combinationa	Time-kill assay log10 change after 48 h with fluconazole (MIC) + polymyxin B (1/2 MIC)	Etest result with fluconazole (MIC) + polymyxin B (1/2 MIC) (mean ΣFIC)a
1	128	>256	96	256	128	24	−3.0, SYN	0.5, SYN
2	8	48	8	256	128	16	−3.0, SYN	0.5, SYN
3	64	>256	64	256	192	48	−2.3, SYN	0.4, SYN
4	128	>256	48	256	512	64	−2.6, SYN	0.3, SYN
5	32	32	6	256	256	16	−2.4, SYN	0.4, SYN
6	128	>256	48	256	384	48	−2.0, SYN	0.3, SYN
7	32	48	12	256	384	12	−2.3, SYN	0.5, SYN
8	8	16	3	128	192	24	−3.4, SYN	0.4, SYN
9	32	16	6	256	256	24	−2.7, SYN	0.5, SYN
10	16	16	4	256	192	24	−4.0, SYN	0.5, SYN
11	8	24	4	256	384	32	−3.0, SYN	0.4, SYN
12	32	32	4	256	1,024	96	−2.1, SYN	0.4, SYN
13	8	16	4	256	256	24	−2.0, SYN	0.5, SYN
14	16	16	6	256	128	32	−4.0, SYN	0.7, IND
15	16	12	3	256	96	32	−2.0, SYN	0.9, IND
16	8	12	6	256	192	32	−4.0, SYN	0.6, IND
17	32	24	16	128	32	12	−3.1, SYN	1.4, IND
18	4	16	4	256	192	32	−2.0, SYN	0.7, IND
19	64	16	3	128	128	16	−3.2, SYN	0.6, IND
20	128	192	48	256	96	16	−2.0, SYN	0.7, IND
21	16	16	6	128	192	32	−3.0, SYN	0.7, IND
22	16	24	2	32	48	3	−0.5, IND	0.5, SYN
23	16	24	6	256	192	24	−1.2, IND	0.5, SYN
24	16	16	4	256	256	24	−1.6, IND	0.4, SYN
25	8	16	4	256	256	32	−1.2, IND	0.5, SYN
26	32	16	4	128	256	32	−0.8, IND	0.5, SYN
27	16	12	3	128	256	32	−1.4, IND	0.5, SYN
28	256	>256	192	128	64	32	−1.0, IND	1.4, IND
29	8	16	6	128	96	24	+1.1, IND	1.1, IND
30	8	12	6	256	64	24	+1.3, IND	1.4, IND
31	8	8	4	64	48	12	−0.8, IND	1.1, IND
32	16	16	6	128	128	24	−0.4, IND	0.8, IND
33	8	16	8	128	48	16	−0.5, IND	1.4, IND
34	0.5	16	4	128	128	32	+1.0, IND	0.9, IND
35	16	24	8	128	128	16	−1.1, IND	0.6, IND

^aPerformed in triplicate.

^bSYN, synergy; IND, indifference; ΣFIC, summation fractional inhibitory concentration.

TABLE 4.

Agreement between Etest method and time-kill assay for determination of synergy with combination of fluconazole (MIC) and polymyxin B (1/2 MIC) against C. glabrata

Etest result	No. (%) of isolates with time-kill assay result		Total (%)
	Synergy	Indifference
Synergy	13	6	19 (54)
Indifference	8	8	16 (46)
Total	21 (60)	14 (40)	35

DISCUSSION

There were two goals of this study: (i) to determine if synergy occurred with the combination of fluconazole and polymyxin B against C. glabrata and (ii) to compare a rapid Etest method with an antifungal TKA for detection of synergy. A factor that prompted this investigation was the comparable results of the Etest synergy method and TKA for testing antibacterial combinations, with the Etest method being simpler to perform (14, 22).

This study demonstrated the presence of synergy of the combination of polymyxin B and fluconazole against C. glabrata isolates—even using a lower, subinhibitory concentration of polymyxin B (1/2 MIC). An interesting finding with the Etest method was that the addition of polymyxin B not only decreased the MIC of fluconazole but also showed an increase in activity by reducing the diffuse growth endpoint produced by C. glabrata. When the combination was used, the trailing growth and diffuse (static) growth ellipses were significantly clearer and more concise, indicating synergistic activity.

The Etest synergy method is a potential method for evaluating in vitro antimicrobial synergy that can be performed easily in a clinical microbiology laboratory. Many researchers consider an Etest synergy method to be a measure of inhibitory activity by antibiotics. However, while evaluating Etest MIC results, the reader is able to detect slight hazes of growth and resistant subpopulations that are included when reading the endpoint. For this reason, Etest MICs can be higher than broth microdilution MICs. The Etest synergy test for yeast has a turnaround time of approximately 2 days from the point that the organism is growing on an agar plate, as follows: day 1, setting up the 2 Etest strips for MIC determination and incubating them overnight; day 2, reading the Etest MIC for each Etest strip, setting up the Etest synergy test plate, and incubating it overnight; and day 3, reading the Etest synergy test plate, calculating the ΣFIC, and reporting results. Even though this method has a 2-day turnaround time, the amount of actual hands-on time at the bench is substantially shorter. Sources of error include the alignment of the Etest strips to match the MIC of each strip and subjectivity (depending on level of expertise) in reading the Etest MICs and synergy tests.

Because time-kill studies with bacteria have been used to evaluate the bactericidal activity of the combination being tested, TKA has often been considered more relevant for clinical situations (15, 23). However, time-kill testing is time-consuming and expensive and requires substantial expertise to perform; thus, it is not practical for most clinical microbiology laboratory settings. Furthermore, very few antifungal time-kill studies have been performed, and there are no standardized CLSI guidelines. TKA has a turnaround time of approximately 5 days for yeast (after the MIC is determined), with many hours of dedicated hands-on time at the bench, as follows: day 1, preparing and aliquoting stock drug solutions/dilutions/media and subculturing the isolate to be tested; day 2, inoculating test bottles, removing and plating samples, and incubating both test bottles and plates overnight; days 3 and 4, removing, diluting, plating, and incubating both test bottles and plates overnight; and day 5, determining the colony counts on plates, compiling data, and analyzing the data for the presence of synergy. Sources of error for TKA include variations in measurements of preparations, pipetting error, and subjective readings of the colony counts.

Understanding the mechanisms of action of polymyxin B and fluconazole helps to explain our findings of in vitro synergy. Polymyxins bind lipopolysaccharide and anionic phospholipids in the bacterial cell membrane, disrupting membrane integrity (7). It is possible that polymyxin B acts similarly against the fungal cell membrane. Fluconazole's mechanism of action involves impairing synthesis of ergosterol, the major lipid of fungal cell membranes. Zhang et al. proposed that fluconazole exacerbates intracellular levels of Ca²⁺ and H⁺ ions, interfering with ionic homeostasis of the membrane. Changes in the biosynthetic pathway result in reduced ergosterol levels and altered membrane structure, allowing for antimicrobials to take effect (24).

In conclusion, both methods showed synergy of polymyxin B and fluconazole against the majority of our 35 C. glabrata bloodstream isolates. The discrepancy of results between methods may be strain dependent or due to differences in testing methods, such as the use of liquid medium (time-kill assay) versus solid medium (Etest). We recognize that polymyxin B Etest strips are approved in the United States for investigational use only and that Etest MIC-MIC synergy testing is a research method. In addition, the antifungal TKA is a nonstandardized research method. Based on our findings, the Etest method was more efficient, simpler to use, and reproducible. It can easily be performed in a clinical microbiology laboratory and should be evaluated further.

The polymyxin B concentrations (1/2 MIC) used in this synergy study are high and cannot be achieved in human patients. Additional testing for synergy with lower concentrations of polymyxin B (2 to 5 μg/ml) and fluconazole (6 to 10 μg/ml) (clinically obtainable serum concentrations) in combination should be performed. The results from this in vitro study may or may not correlate clinically. In vivo studies are still needed to evaluate the combination of polymyxin B and fluconazole against C. glabrata in clinical situations.