In Vitro Fungicidal Activities of Anidulafungin, Caspofungin, and Micafungin against Candida glabrata, Candida bracarensis, and Candida nivariensis Evaluated by Time-Kill Studies

Abstract

Anidulafungin, caspofungin, and micafungin killing activities against Candida glabrata, Candida bracarensis, and Candida nivariensis were evaluated by the time-kill methodology. The concentrations assayed were 0.06, 0.125, and 0.5 μg/ml, which are achieved in serum. Anidulafungin and micafungin required between 13 and 26 h to reach the fungicidal endpoint (99.9% killing) against C. glabrata and C. bracarensis. All echinocandins were less active against C. nivariensis.

TEXT

Candida glabrata follows Candida albicans as the second or third most prevalent cause of candidemia worldwide (1). C. glabrata presents decreased antifungal susceptibility to fluconazole and other current antifungal drugs and can rapidly acquire resistance (2). Candida bracarensis and Candida nivariensis, two species closely related to C. glabrata, have been recently described (3, 4), but there is scarce information on the prevalence, antifungal susceptibility patterns, and clinical significance of these cryptic species (1, 5–7). The present study aimed to determine the killing activities of echinocandins against C. glabrata, C. bracarensis, and C. nivariensis (Table 1). Strains were identified by metabolic properties (ATB ID 32C; bioMérieux, Marcy l'Étoile, France) and molecular methods, as previously described (5, 7).

TABLE 1.

MIC of the indicated drug

Strain or isolate	MIC (μg/ml) of:
	Anidulafungin	Caspofungin	Micafungin
C. glabrata
UPV/EHU 03-282a	0.06	0.25	0.12
UPV/EHU 03-285a	0.06	0.5	0.12
UPV/EHU 03-287a	0.06	0.25	0.12
ATCC 90030	0.06	0.25	0.25
NCPF 3203	0.06	0.25	0.12
GMb (median)	0.06 (0.06)	0.29 (0.25)	0.14 (0.12)
C. bracarensis
NCYC 3133	0.06	0.25	0.12
NCYC 3397	0.06	0.12	0.12
GM (median)	0.06 (0.06)	0.17 (0.19)	0.12 (0.12)
C. nivariensis
UPV/EHU 11-284a	0.12	0.25	0.12
CBS 9984	0.06	0.25	0.12
CECT 11998T	0.06	0.25	0.12
GM (median)	0.08 (0.06)	0.25 (0.25)	0.12 (0.12)

^aBloodstream isolate.

^bGM, geometric mean of MIC.

Caspofungin (Merck Sharp & Dohme, Madrid, Spain), micafungin (Astellas Pharma, Madrid, Spain), and anidulafungin (Pfizer SLU, Madrid, Spain) were dissolved in dimethyl sulfoxide. Further dilutions were prepared in standard RPMI 1640 medium (Sigma-Aldrich, Madrid, Spain). MICs (defined as minimum concentrations that produce ≥50 growth reduction) were determined following the M27-A3 document (8). Time-kill studies were carried out on microtiter plates for the computer-controlled microbiological incubator BioScreen C MBR (LabSystems, Vantaa, Finland) in RPMI 1640 (final volume, 200 μl; inoculum, 1 × 10⁵ to 5 × 10⁵ CFU/ml). The echinocandin concentrations assayed were 0.06, 0.125, and 0.5 μg/ml, which are achieved in serum after standard doses (9). Plates were incubated at 36 ± 1°C without agitation. At 0, 2, 4, 6, 8, 24, and 48 h, aliquots of 6 or 10 μl were removed from both the control and each test solution well, serially diluted in phosphate-buffered saline (PBS), and plated onto Sabouraud agar to determine the number of CFU per milliliter. Each experiment was performed twice for each isolate (10–13). The antifungal carryover effect was determined as formerly reported (12).

Time-kill data were fitted to the exponential equation N_t = N₀ × e^kt, where t is incubation time, N_t represents viable yeast cells at time t, N₀ is the starting inoculum, and k is the killing or growing rate. The goodness of fit for each isolate/drug was assessed by the r² value (>0.8). The times needed to achieve 50, 90, 99 and 99.9% reductions in growth (t₅₀, t₉₀, t₉₉, and t_99.9, respectively) were calculated from the k value, as described previously (12). Analysis of variance was performed to determine significant differences in killing kinetics. A P value of <0.05 was considered significant.

Micafungin and anidulafungin MICs were lower than those of caspofungin against the three species evaluated (Table 1) and were similar to those reported in previous studies: conversely, caspofungin MICs tended to be higher against C. glabrata (6, 7, 14–18), although the in vitro caspofungin MIC results should be cautiously interpreted since a recent study has shown interlaboratory variability in modal MICs (19). Based on this unresolved technical problem related to caspofungin susceptibility testing, we primarily discuss in this paper the results that we obtained for anidulafungin and micafungin.

Other authors have reported head-to-head or comparative data regarding the fungicidal activities of the echinocandins for other Candida species, such as C. albicans, Candida krusei, Candida parapsilosis, or Candida lusitaniae, using time-kill methods (10, 11, 15–17, 20, 21), but to our knowledge, this is the first study that has made a head-to-head comparison of the killing kinetics of these three echinocandins using blood isolates and strains of C. bracariensis and C. nivariensis. The lack of similar reports in the literature has precluded comparisons.

The mean time-kill curves and standard deviations of anidulafungin, caspofungin, and micafungin against the isolates tested are depicted in Fig. 1. The killing activities of the three echinocandins were species dependent, and there is probably variability among strains inside each species. Most echinocandin activities and killing rate studies against several species of Candida have usually included a low number of isolates similar to the number of strains in the present study (10, 11, 15).

FIG 1.

Mean time-kill plots for anidulafungin, caspofungin, and micafungin against five C. glabrata, three C. nivariensis, and two C. bracarensis strains. Each data point represents the mean result ± standard deviation (error bars) from the indicated number of isolates. Broken lines represents ≥99.9% growth reduction compared with that of the initial inoculum (fungicidal effect).

For anidulafungin, the mean maximum log decreases in CFU per milliliter were >3 logs for C. glabrata, C. bracarensis, and C. nivariensis (5.17 ± 0.52, 5.58 ± 0.04, and 3.62 ± 2.78 logs, respectively) at 0.5 μg/ml. The mean maximum log decreases were >3 logs for 0.125 μg/ml of micafungin, against C. glabrata (3.83 ± 1.91 logs) and C. bracariensis (4.43 ± 1.63 logs). Similar results have been observed by Nguyen et al. (15, 16) for C. glabrata, as these authors reported that anidulafungin and micafungin were fungicidal against 3 isolates of this species (0.48 to 0.96 μg/ml and 0.24 to 2 μg/ml, respectively).

There was a paradoxical effect for C. bracarensis, as 0.125 μg/ml of micafungin caused more lethality than 0.5 μg/ml at 24 h. This effect, defined as the ability to grow at high antifungal concentrations, but not at intermediate concentrations, has also been reported by other authors (10, 11). However, this paradoxical growth does not preclude the in vivo response to echinocandin therapy in Candida dubliniensis infections (22). Against C. nivariensis, micafungin did not show lethality as there was not a reduction in CFU/ml with respect to the initial yeast inoculum.

Figure 2 shows the effects of echinocandin concentrations on the killing rates for the three species. Against C. glabrata, the highest killing rate was reached with 0.125 μg/ml of micafungin; however, against C. nivariensis, anidulafungin showed the highest killing rate at 0.5 μg/ml. Even at the lowest concentration (0.06 μg/ml), anidulafungin showed killing activity against C. bracarensis.

FIG 2.

Effect of concentration on killing rates (K) of anidulafungin, caspofungin, and micafungin. Values above the broken lines indicate growth, and values below the broken lines indicate killing.

Table 2 shows the mean times needed to kill 50, 90, 99, and 99.9% of the initial inoculum for each echinocandin concentration. Anidulafungin and micafungin required between 13 and 26 h to reach the fungicidal endpoint (99.9% killing) against C. glabrata (13.23 h at 0.5 μg/ml) and C. bracarensis (22.11 and 22.72 h at 0.125 μg/ml, respectively). Only anidulafungin reached the fungicidal endpoint (36.54 h at 0.5 μg/ml) against C. nivariensis. The repercussions of protein binding on the in vitro killing should be very similar for the three echinocandins as the protein binding is >96% for all of them (23).

TABLE 2.

Time to achieve 50, 90, 99, and 99.9% growth reductions from the initial inoculum at the indicated concentration

Drug and reduction parametera	Time (h) to growth reduction at drug concn (μg/ml) shown
	C. glabrata			C. nivariensis			C. bracarensis
	0.06	0.125	0.5	0.06	0.125	0.5	0.06	0.125	0.5
Anidulafungin
t50	NAb	13.3	1.33	NA	NA	3.67	6.2	2.22	1.4
t90	NA	44.17	4.41	NA	NA	12.18	20.6	7.37	4.66
t99	NA	>48	8.82	NA	NA	24.36	41.19	14.74	9.32
t99.9	NA	>48	13.23	NA	NA	36.54	>48	22.11	13.97
Caspofungin
t50	NA	NA	6.9	NA	NA	8.78	NA	5.62	2.71
t90	NA	NA	22.95	NA	NA	29.18	NA	18.66	9.01
t99	NA	NA	45.9	NA	NA	>48	NA	37.32	18.03
t99.9	NA	NA	>48	NA	NA	>48	NA	>48	27.05
Micafungin
t50	NA	2.36	1.31	NA	NA	NA	NA	2.28	2.62
t90	NA	7.82	4.36	NA	NA	NA	NA	7.58	8.7
t99	NA	15.65	8.72	NA	NA	NA	NA	15.15	17.39
t99.9	NA	23.47	13.08	NA	NA	NA	NA	22.72	26.09

^at₅₀, t₉₀, etc., time (hours) to achieve 50% growth reduction, 90% growth reduction, etc.

^bNA, not achieved.

In conclusion, while MIC values of anidulafungin and micafungin were similar for the three species, differences in fungicidal activities were observed. All echinocandins were more lethal against C. glabrata and C. bracarensis than against C. nivariensis. Anidulafungin was the most active agent against C. bracarensis and C. nivariensis, and micafungin was the most active agent against C. glabrata. However, time-kill data reveal that micafungin did not show killing activity against C. nivariensis. The lower activity of echinocandins against C. nivariensis highlights the importance of correct identification and of knowing the antifungal susceptibility patterns of these new cryptic species for an adequate therapeutic approach to infections caused by the C. glabrata clade.