Activities of Triazole-Echinocandin Combinations against Candida Species in Biofilms and as Planktonic Cells^,

Abstract

Biofilm formation complicates the treatment of various infections caused by Candida species. We investigated the effects of simultaneous or sequential combinations of two triazoles, voriconazole (VRC) and posaconazole (PSC), with two echinocandins, anidulafungin (AND) and caspofungin (CAS), against Candida albicans and Candida parapsilosis biofilms in comparison to their planktonic counterparts. Antifungal activity was assessed by the 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]2H-tetrazolium-5-carboxanilide (XTT) metabolic assay. Antifungal-agent interactions were analyzed by the Bliss independence model in the simultaneous-treatment studies and by analysis of variance (ANOVA) in the sequential-treatment studies. Against C. albicans planktonic cells, the simultaneous combination of PSC (32 to 128 mg/liter) and CAS (0.008 to 0.25 mg/liter) was synergistic; the combinations of PSC (128 to 1,024 mg/liter) with AND (0.03 to 0.5 mg/liter) and VRC (32 to 512 mg/liter) with AND (0.008 to 0.03 mg/liter) were antagonistic. Against C. parapsilosis planktonic cells, the interaction between VRC (32 to 1,024 mg/liter) and CAS (1 to 16 mg/liter) was antagonistic. All simultaneous antifungal combinations demonstrated indifferent interactions against biofilms of both Candida species. Damage to biofilms of both species increased (P < 0.01) in the presence of subinhibitory concentrations of echinocandins (0.008 to 0.064 mg/liter), followed by the addition of PSC (512 mg/liter for C. albicans and 64 to 512 mg/liter for C. parapsilosis) or VRC (256 to 512 mg/liter for C. albicans and 512 mg/liter for C. parapsilosis). Triazole-echinocandin combinations do not appear to produce antagonistic effects against Candida sp. biofilms, while various significant interactions occur with their planktonic counterparts.

INTRODUCTION

Candida bloodstream infections cause significant morbidity and mortality in critically ill patients (30). Candida albicans and Candida parapsilosis are the species most frequently implicated in vascular-catheter-related candidemia (19, 32, 36). The development of candidemia has been associated with the use of central venous catheters or other implantable prosthetic devices, which are highly susceptible to colonization and infection by yeast cells (7, 17). Guidelines recently published by the Infectious Diseases Society of America (IDSA) (28) recommend prompt removal of the foreign body; however, since antifungal agents, on many occasions, are insufficient to cure biofilm-related infections and catheter removal is not always feasible (24), other approaches, such as combination therapy or lock therapy, have been suggested (25, 29, 33).

Biofilm formation is an important virulence factor of Candida spp. in such infections. Yeast cells embedded in biofilms demonstrate phenotypic traits distinct from those of their planktonic counterparts (8, 22, 30). In particular, biofilms exhibit reduced susceptibility to common antimicrobial agents and host defense mechanisms and have survival advantages over planktonic cells (14, 15).

Triazoles and echinocandins, two classes of antifungal agents with distinct mechanisms of action (5), are used as standard therapy for Candida infections, according to IDSA guidelines (10, 28). In refractory cases of invasive fungal infections, combinations of triazoles and echinocandins have been studied as promising therapies to reduce high attributable mortality rates (18). Little is known, however, about the combined effects of the newer antifungal triazoles and echinocandins against biofilms of C. albicans and C. parapsilosis (1, 31).

In the present study, we investigated the combined effects of voriconazole (VRC) or posaconazole (PSC) with anidulafungin (AND) or caspofungin (CAS) against C. albicans and C. parapsilosis biofilms or planktonic cells. We studied triazole-echinocandin interactions in two models of simultaneous and sequential treatment.

(This study was presented in part at the 48th Annual Meeting of the Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) and the 46th Annual Meeting of the Infectious Diseases Society of America (IDSA), Washington, DC, 25 to 28 October 2008; the 19th European Congress of Clinical Microbiology and Infectious Diseases (ECCMID), Helsinki, Finland, 16 to 19 May 2009; and the 4th Trends in Medical Mycology (TIMM), Athens, Greece, 18 to 21 October 2009.)

MATERIALS AND METHODS

Strains.

Two well-characterized biofilm-producing Candida strains were used. Candida albicans strain M-61 was obtained from an infected intravascular catheter, and C. parapsilosis strain P/A71 was isolated from sputum. The two strains were kindly donated by M. A. Ghannoum (University Hospital of Cleveland and Case Western Reserve University, Cleveland, OH). Stock cultures were divided into small portions and stored at −35°C in 25% glycerol and 75% peptone.

Growth medium and conditions.

All Candida strains were grown in yeast nitrogen base (YNB) medium (Difco Laboratories, Detroit, MI) supplemented with 50 mM glucose. Twenty milliliters of YNB medium was inoculated with a loopful of Candida from a freshly inoculated Sabouraud glucose agar plate and incubated on a rocker at 37°C overnight. Cells were harvested and washed twice with 0.15 M phosphate-buffered saline (PBS) solution (pH 7.2; Ca²⁺ and Mg²⁺ free; Biochrom KG, Berlin, Germany). Yeast cells were resuspended in 10 ml of PBS, counted after serial dilutions using a hemocytometer, standardized at 5 × 10⁵ blastoconidia/ml, and used immediately.

Biofilm formation.

Biofilms were formed on preconditioned sterile silicone elastomer discs (Bioplexus Corp., Ventura, CA) that had been preincubated with fetal bovine serum (FBS) in 96-well plates (Corning Inc., New York, NY) under constant linear shaking for 24 h. The discs were then washed with PBS to remove residual FBS. For mature-biofilm formation, 5 × 10⁵ blastoconidia/ml of C. albicans M-61 or C. parapsilosis PA/71 were added to the above-mentioned 96-well plates and incubated at 37°C under constant shaking in RPMI 1640 (Sigma-Aldrich, St. Louis, MO) for 48 h or 72 h, respectively. After biofilm formation, the 96-well plates were centrifuged at 2,230 × g for 30 min. The medium was then aspirated, and nonadhering cells were removed by washing them once with sterile PBS (19).

Antifungal agents.

VRC and AND were provided by Pfizer Inc. (New York, NY), PSC by Schering-Plough (Brussels, Belgium), and CAS by Merck and Co. Inc. (Whitehouse Station, NJ). VRC and CAS were obtained in powder form and reconstituted in sterile water to make stock solutions of 10 mg/ml and 5 mg/ml, respectively. AND was obtained in a lyophilized form and reconstituted with 20% alcohol to give a final stock solution of 10 mg/ml. PSC was obtained as a suspension of 40 mg/ml. Stock solutions were stored at −35°C and used within a period of 30 days. Working solutions were prepared in RPMI 1640 buffered to a pH of 7.4 with 0.165 M morpholinepropanesulfonic acid (MOPS) (Sigma-Aldrich) buffer. Given that echinocandins, unlike azoles, retain their activity against C. albicans biofilms (14), we selected concentrations of echinocandins ranging around clinically relevant concentrations (0.008 to 0.064 mg/liter for the sequential model and 0.008 to 16 mg/liter for the simultaneous model). The use of higher concentrations of echinocandins was not considered due to the anticipated in vitro paradoxical effect (3). On the other hand, we used much higher concentrations of triazoles (64 to 512 mg/liter for the sequential model and 16 to 1024 mg/liter for the simultaneous model). Our rationale for using such high concentrations of PSC and VRC was that they could be achieved by the lock therapy approach.

MIC determination.

The planktonic cell MICs for both isolates were determined by the broth microdilution reference method, as described by the Clinical and Laboratory Standards Institute (CLSI [formerly the National Committee for Clinical Standards]) (6). MICs were determined after incubation of 2 × 10³ CFU/ml in RPMI 1640 plus MOPS medium for 48 h and were defined as the lowest antifungal concentrations that caused ≥50% reduction in visible growth. The planktonic MICs of VRC, PSC, CAS, and AND for C. albicans M61, as tested by the reference method M27-A (CLSI), were 0.01, <0.001, 0.06, and 0.003 mg/liter, respectively. The planktonic MICs of VRC, PSC, CAS, and AND for C. parapsilosis PA/71 were 0.03, 0.01, 0.06, and 0.12 mg/liter, respectively.

Biofilm activity was also determined after 48 h or 72 h of incubation of 5 × 10⁵ CFU/ml for C. albicans or C. parapsilosis, respectively, in RPMI 1640 plus MOPS medium using a 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]2H-tetrazolium-5-carboxanilide (XTT) assay. The antifungal concentration that caused a ≥50% reduction in the metabolic activity of the biofilm compared with the control (incubated in the absence of the antifungal agent) was determined, and the lowest concentration resulting in ≥50% reduction in the metabolic activity was considered the biofilm MIC (11, 16, 19).

Simultaneous combination antifungal treatment of biofilms or planktonic cells.

Mature biofilms or planktonic cells were coincubated for 24 h at 37°C in RPMI medium (control) or with serially 2-fold-diluted concentrations of PSC or VRC ranging from 16 to 1,024 mg/liter and of CAS or AND ranging from 0.008 to 16 mg/liter alone or simultaneously—PSC plus CAS, PSC plus AND, VRC plus CAS, and VRC plus AND—in a checkerboard format (20). Biofilm or planktonic cell metabolic activities were then assessed by the XTT assay. Biofilm MICs in the presence of different concentrations of antifungal agents alone or in combination were determined as described above.

Sequential combination antifungal treatment of biofilms.

Mature biofilms were formed in the presence of subinhibitory concentrations of echinocandins. An inoculum of 5 × 10⁵ blastoconidia/ml of C. albicans or C. parapsilosis in 96-well plates containing silicone discs was incubated in RPMI medium (control) or in RPMI containing 2-fold-diluted concentrations of CAS or AND ranging from 0.008 to 0.064 mg/liter for 48 h and 72 h at 37°C, respectively. Biofilms were subsequently incubated with VRC or PSC (64 to 512 mg/liter), respectively, for another 24 h at 37°C. Biofilm metabolic activity was then assessed by XTT assay. Biofilm MICs in the presence of different concentrations of antifungal agents, alone or in sequential combinations, were determined as described above. For sequential treatments, only CAS followed by PSC and AND followed by VRC were examined.

Assessment of biofilm and planktonic cell metabolic activities.

Measurement of biofilm or planktonic cell metabolic activities was performed using the XTT metabolic-reduction assay, as previously described (14). Briefly, after incubation for 48 or 72 h, the plates were centrifuged at 4,000 rpm for 30 min and washed twice in sterile water. After the washes, PBS containing 0.25 mg/ml XTT and 40 μg/ml coenzyme Q₀ was added. After incubation at 37°C for 1 h, 100 μl was transferred to a new plate and the optical density (OD) was assessed spectrophotometrically. Absorbance at 450 nm with a reference wavelength of 690 nm was measured using an automated plate reader (Anthos 2010; Labtech Instruments, Salzburg, Austria). Percent metabolic activity was calculated with the following equation: (1 − X/C) × 100, where X is the OD of antifungal-agent-containing wells and C is the OD of control wells with fungi only.

Analysis of antifungal-agent interactions. (i) Simultaneous treatment.

Assays were carried out in 5 to 12 replicates on different days. Antifungal-agent interactions were analyzed using the Bliss independence model (20). According to this model, the expected theoretical percentage of growth (E_ind) (compared to an antifungal-agent-free control) describing the effect of the combination of two antifungal agents is calculated with the following equation: E_ind = E_A × E_B, where E_A and E_B are the experimental percentages of growth when each antifungal agent acts alone. For each combination of x mg/liter of antifungal agent A with y mg/liter of antifungal agent B in each of the independent replicate experiments, the experimentally observed percentage of growth, E_obs, was subtracted from E_ind. When the mean ΔΕ (ΔΕ = E_ind − E_obs) was positive and its 95% confidence interval (CI) did not include 0, significant synergy was claimed for that specific combination of x mg/liter of antifungal agent A with y mg/liter of antifungal agent B. When the mean ΔΕ was negative without its CI overlapping 0, statistically significant antagonism was claimed. In any other case, indifference was concluded.

(ii) Sequential treatment.

Assays were carried out in 10 replicates on different days. The average value of these replicates was then used in the data analysis to calculate the mean ± standard error (SE) of all the experiments performed under the same conditions. Differences between mean values were assessed by repeated-measures analysis of variance (ANOVA). A Dunnett multiple-comparison test was used for statistical comparisons between untreated cells and cells treated with antifungal agents alone or in combination. InStat (GraphPad Inc., San Diego, CA) was used for data analysis. A two-tailed P value of <0.05 indicated statistical significance.

RESULTS

Biofilm MICs.

The biofilm MIC of VRC was 512 mg/liter, that of PSC was 256 mg/liter, that of AND was 0.125 mg/liter, and that of CAS was 0.0625 mg/liter for C. albicans M61. The biofilm MICs of VRC, PSC, CAS, and AND were 256, 256, 1, and 2 mg/liter, respectively, for C. parapsilosis PA/71.

Simultaneous combination antifungal treatment of biofilms or planktonic cells.

Simultaneous treatment of C. albicans planktonic cells with CAS (0.008 to 0.25 mg/liter) and PSC (32 to 128 mg/liter) resulted in synergistic interaction (mean ΔΕ value of significant interactions, 39% [range, 6% to 54%]; mean SE, 7% [range, 1% to 9%]) (Fig. 1A). The combination of CAS (0.008 to 16 mg/liter) with VRC (16 to 1,024 mg/liter) resulted in indifferent interaction (see the supplemental material). In comparison, antagonistic interactions against C. albicans planktonic cells were observed with (i) simultaneous incubation of AND (0.008 to 0.03 mg/liter) and VRC (32 to 512 mg/liter) (mean ΔΕ value of significant interactions, −12% [range, −19% to −4%]; mean SE, 2% [range, 1% to 3%]) (Fig. 1B) and (ii) simultaneous incubation of AND (0.03 to 0.5 mg/liter) and PSC (128 to 1,024 mg/liter) (mean ΔΕ value of significant interactions, −49% [range, −84% to −22%]; mean SE, 8% [range, 5% to 12%]) (Fig. 1C). In contrast, all combinations of antifungal agents studied, CAS with POS or VRC and AND with POS or VRC, exhibited indifferent interactions against biofilms (see the supplemental material).

Fig. 1.

Interaction surface plots obtained from analysis with the Bliss independence model of antifungal drug interactions against planktonic cells of C. albicans (A to C) and C. parapsilosis (D). The plots represent combinations of CAS and PSC (A), AND and VRC (B), AND and PSC (C), and CAS and PSC (D). The zero plane (ΔΕ = 0) represents indifferent interactions, whereas volumes above (ΔΕ > 0) and below (ΔΕ < 0) the zero plane represent synergistic and antagonistic interactions, respectively.

With the exception of antagonistic interactions observed after simultaneous incubation of planktonic cells of C. parapsilosis with CAS (1 to 16 mg/liter) and VRC (32 to 1,024 mg/liter) (mean ΔΕ value of significant interactions, −38% [range, −52% to −22%]; mean SE, 6% [range, 3% to 9%]) (Fig. 1D), neither synergy nor antagonism was found for the remaining antifungal agent combinations tested against planktonic cells or mature biofilms of C. parapsilosis (see the supplemental material).

Sequential antifungal treatment.

After formation of C. albicans biofilms in the presence of CAS and further incubation with no triazole, we observed minimal fungal damage across all concentrations of CAS used (0 to 5.9%) (Fig. 2). In comparison, we found significant increases in biofilm damage to C. albicans when CAS at concentrations of 0.008 to 0.064 mg/liter was followed by 128 mg/liter of PSC (32.7% for PSC alone versus 39.2 to 47.7% for CAS plus PSC; P < 0.01) or CAS at 0.064 mg/liter with 256 mg/liter of PSC (35.8% for PSC alone versus 45.1% for CAS plus PSC; P < 0.01). In addition, significant increases in fungal damage were found when CAS at 0.016 to 0.064 mg/liter was combined with 512 mg/liter of PSC (41.2% for PSC alone versus 51.9 to 58.7% for CAS plus PSC [Fig. 2]; P < 0.01).

Fig. 2.

Sequential antifungal treatment of C. albicans biofilms with CAS and PSC. Shown is fungal damage to C. albicans biofilm as evidenced by decrease of metabolic activity in an XTT assay after formation of a biofilm in the presence of CAS and sequential incubation with PSC. The values on the x axis represent CAS concentrations. The values on the y axis represent percent biofilm damage to C. albicans compared to untreated controls (0% fungal damage). The bars represent the means ± SE of values derived from 8 experiments performed on different days. The horizontal line denotes the defined biofilm MIC (see the text). a, P < 0.01 for concentrations of CAS (0.008, 0.032, and 0.064 mg/liter) combined with 128 mg/liter of PSC; b, P < 0.01 for 0.064 mg/liter of CAS combined with 256 mg/liter of PSC; c, P < 0.01 for increasing concentrations of CAS (0.016 to 0.064 mg/liter) combined with 512 mg/liter of PSC.

Similarly, when C. albicans biofilms were formed in the presence of the same range of concentrations of AND and further incubated with no triazole, the observed biofilm damage was less than 15% (Fig. 3). When VRC was added, significant increases in fungal damage were observed. Thus, when AND at a concentration of 0.032 or 0.064 mg/liter was combined with 64 mg/liter of VRC, biofilm damage increased from 33.8% (VRC) to 46.0% and 45.4% (AND plus VRC), respectively (Fig. 3) (P < 0.01). Similarly, when AND at concentrations ranging from 0.016 to 0.064 mg/liter was combined with 128 mg/liter of VRC, biofilm was damaged by 39.2 to 47.3%, compared to 31.9% with VRC alone (Fig. 3) (P < 0.01). Additional antifungal combinations that exhibited significant increases were 0.032 mg/liter of AND with 256 mg/liter of VRC (39.0% for VRC versus 49.1% for AND plus VRC) and 0.032 or 0.064 mg/liter of AND with 512 mg/liter of VRC (52.6% for VRC versus 60.4% and 64.8% for AND plus VRC, respectively [Fig. 3]; P < 0.01).

Fig. 3.

Sequential antifungal treatment of C. albicans biofilms with AND and VRC. Shown is fungal damage to C. albicans biofilm as evidenced by decrease of metabolic activity in an XTT assay after formation of a biofilm in the presence of AND and sequential incubation with VRC. The values on the x axis represent AND concentrations. The values on the y axis represent percent biofilm damage to C. albicans compared to untreated controls (0% fungal damage). The bars represent the means ± SE of values derived from 8 experiments performed on different days. The horizontal line denotes the defined biofilm MIC (see the text). a, P < 0.01 for concentrations of AND (0.032 and 0.064 mg/liter) combined with 64 mg/liter of VRC; b, P < 0.01 for concentrations of AND (0.016, 0.032, and 0.064 mg/liter) combined with 128 mg/liter of VRC; c, P < 0.01 for 0.032 mg/liter of AND combined with 256 mg/liter of VRC; d, P < 0.01 for concentrations of AND (0.032 and 0.064 mg/liter) combined with 512 mg/liter of VRC.

Significant increases in biofilm damage were also found when C. parapsilosis biofilms were formed in the presence of subinhibitory concentrations of echinocandins and subsequently exposed to PSC or VRC. For example, while biofilm formed in the presence of CAS (0.008 to 0.064 mg/liter) and further incubated with no triazole was not damaged more than 8%, when 0.008 to 0.064 mg/liter of CAS was combined with 512 mg/liter of PSC, there was a significant increase in biofilm damage, from 40.2% (for PSC alone) to 49.0 to 58.5% (for CAS plus PSC) [Fig. 4]; P < 0.01). Likewise, when CAS at a concentration of 0.032 mg/liter was combined with 256 mg/liter of PSC, biofilm damage was increased from 35.8% by PSC to 44.4% by the combined activity of CAS plus PSC (Fig. 4) (P < 0.01).

Fig. 4.

Sequential antifungal treatment of C. parapsilosis biofilms with CAS and PSC. Fungal damage to C. parapsilosis biofilm as evidenced by decrease of metabolic activity in an XTT assay after formation of a biofilm in the presence of CAS and then incubation with PSC. The values on the x axis represent CAS concentrations. The values on the y axis represent percent biofilm damage to C. parapsilosis compared to untreated controls (0% fungal damage). The bars represent the mean ± SE of values derived from 8 experiments performed on different days. The horizontal line denotes the defined biofilm MIC (see the text). a, P < 0.01 for 0.016 and 0.064 mg/liter of CAS combined with 128 mg/liter of PSC; b, P < 0.01 for increasing concentrations of 0.032 mg/liter of CAS combined with 256 mg/liter of PSC; c, P < 0.01 for increasing concentrations of CAS (0.008 and 0.032 to 0.064 mg/liter) combined with 512 mg/liter of PSC.

Significant biofilm damage was also noted with the combinations of AND and VRC (Fig. 5). Biofilm damage to C. parapsilosis reached at most 5% after biofilm formation in the presence of AND and no further incubation of VRC. However, AND (0.008 to 0.064 mg/liter) in combination with VRC (256 or 512 mg/liter) increased the percent damage to the biofilm from 48.6% (256 mg/liter of VRC alone) to 66.9 to 74.6% (for AND plus VRC) or from 67.3% (512 mg/liter of VRC alone) to 77.6 to 91.1% for the combination AND and VRC (Fig. 5) (P < 0.01).

Fig. 5.

Sequential antifungal treatment of C. parapsilosis biofilms with AND and VRC. Fungal damage to C. parapsilosis biofilm as evidenced by decrease of metabolic activity in an XTT assay after formation of a biofilm in the presence of AND and then incubation with VRC. The values on the x axis represent AND concentrations. The values on the y axis represent percent biofilm damage to C. parapsilosis compared to untreated controls (0% fungal damage). The bars represent the mean ± SE of values derived from 8 experiments performed on different days. The horizontal line denotes the defined biofilm MIC (see the text). a, P < 0.01 for 0.016 mg/liter of AND combined with 64 mg/liter of VRC; b, P < 0.01 for 0.064 mg/liter of AND combined with 128 mg/liter of VRC; c, P < 0.01 for increasing concentrations of AND (0.008 to 0.064 mg/liter) combined with 256 mg/liter of VRC; d, P < 0.01 for increasing concentrations of AND (0.008 to 0.064 mg/liter) combined with 512 mg/liter of VRC.

DISCUSSION

In this study, we evaluated the effect of triazole-echinocandin combinations used either simultaneously or sequentially against C. albicans and C. parapsilosis biofilms and the triazole-echinocandin combinations used only simultaneously against planktonic cells. To our knowledge, this is the first study to examine the sequential interactions of CAS or AND and VRC or PSC, as well as the simultaneous effect of AND and PSC or VRC, against Candida biofilms. Pretreatment with echinocandins (CAS and AND) at subinhibitory concentrations significantly increased the efficacy of subsequent triazoles (VRC and PSC) against Candida biofilms. When used simultaneously, all the combinations against Candida biofilms exhibited indifferent interactions. In the case of planktonic cells, differential interactions of simultaneous combinations were observed. In particular, the following interactions were observed against planktonic cells: synergy for the combination of PSC (32 to 128 mg/liter) and CAS (0.008 to 0.25 mg/liter) at low concentrations against C. albicans, antagonism for the combination of AND (0.03 to 0.25 mg/liter) with either VRC (16 to 512 mg/liter) or PSC (128 to 1,024 mg/liter) against C. albicans, and antagonism for VRC (32 to 1,024 mg/liter) and CAS (0.5 to 8 mg/liter) against C. parapsilosis. For the remaining concentrations and combinations tested in planktonic cells, indifferent interactions were observed.

While the guidelines of the IDSA recommend catheter removal in cases of existing vascular catheters and candidemia (28), this is not always feasible and sometimes could be a life-threatening procedure (24). Almost invariably, a biofilm can be detected on the surfaces of indwelling medical devices, constituting a critical virulence determinant in catheter-related infections (17). From the clinical perspective, the most important feature of biofilms is their resistance to traditional antifungal agents (4). As the development of potent antibiofilm agents is exceedingly slow, combination therapy appears to be a very promising solution for biofilm-related infections. The objectives of antifungal combination therapy are to (i) increase efficacy by means of a potential synergistic effect, (ii) broaden the spectrum of antifungal activity, and (iii) prevent the development of resistance (12).

Using an in vitro biofilm model, we sought to determine the combination effect of two newer antifungal agents belonging to different antifungal classes with distinct mechanisms of action. Our experiments were based on the hypothesis that echinocandins and triazoles do not antagonize each other, as they act at different sites in Candida. Echinocandins, which inhibit cell wall synthesis by disrupting 1,3-β-d-glucan synthesis, may enhance the activity of triazoles by increasing the rate or degree of their access to the cell membrane (21, 27). This theoretical approach could be particularly applicable in the case of biofilm development, given that the biofilm matrix contains excessive amounts of glucan (2).

A limited number of in vitro studies thus far have evaluated the effects of antifungal combinations against C. albicans biofilms. Concerning the triazole-echinocandin combinations, Bachmann et al. identified, by time-kill studies, that the simultaneous combination of caspofungin with high fluconazole concentrations was antagonistic (1). Similarly, the effect of another triazole-echinocandin combination, namely, voriconazole-micafungin, against C. albicans biofilms appeared to be antagonistic (13). The mechanism of this antagonism was attributed to upregulation of Hsp90-related stress responses, which enhanced cell wall integrity (13). A subsequent study, including 30 clinical C. albicans isolates, showed that the combination of caspofungin, at high concentrations, and voriconazole against biofilms did not exhibit improved in vitro activity compared to caspofungin alone (34). The antifungal combinations of fluconazole, amphotericin B, and caspofungin were found to be indifferent by checkerboard testing, whereas some trend toward additivity was observed for an amphotericin B and caspofungin combination (1). Notably, synergism was reported in the in vitro interaction of amphotericin B with posaconazole against C. albicans biofilms, while the amphotericin B and caspofungin combination yielded indifferent effects (35).

The effects of serial antifungal sequential combinations against C. albicans biofilms have been evaluated for micafungin followed by voriconazole and vice versa. The results of this study showed that treatment with micafungin, at a concentration 16 times higher than the MIC, followed by voriconazole was as effective as micafungin alone, while the opposite sequence (voriconazole first and then micafungin) was less effective than micafungin alone (13). Our sequential-treatment model showed that substantial damage was caused to the growing biofilm when preexposure to subinhibitory concentrations of CAS or AND (0.008 to 0.064 mg/liter) was followed by VRC (256 to 512 mg/liter for C. albicans and 512 mg/liter for C. parapsilosis) and PSC (512 mg/liter for C. albicans and 64 to 512 mg/liter for C. parapsilosis). In spite of their differences, both studies support the serial antifungal combination of an echinocandin followed by a triazole.

The findings of our sequential-treatment model provide insights into biofilm resistance underlying the role of the extracellular matrix, and particularly β-glucan. Nett et al., using a similar in vitro biofilm model, examined the role of the β-1,3-glucan synthase gene Fks1p in triazole susceptibility, matrix production, and drug sequestration during C. albicans biofilm growth (23). They demonstrated that glucan synthesis by Fks1p is critical for biofilm drug resistance, suggesting that azoles and echinocandins, which target Fks1p, may be synergistic in C. albicans biofilms. In aggregate, the findings of our study and that of Nett et al. are complementary, since the latter provides the pharmacological basis of our findings. Whether our findings could be applicable in the management of biofilm-related infections remains to be verified in future animal studies and clinical trials.

While antifungal combinations have been studied primarily against Candida planktonic cells, data from in vitro studies have been somewhat inconclusive. The interactions appeared to be strain specific, and correlations with clinical outcome were not determined (21). The discrepancies noted between different in vitro studies could be attributed to different methodologies used for testing and to the way the results were analyzed and interpreted. One intriguing finding of our study concerns the azole-echinocandin combinations against planktonic Candida cells. While azole-echinocandin combinations are expected to produce indifferent or synergistic effects (34), we found that VRC combined with AND or CAS against C. albicans or C. parapsilosis planktonic cells, respectively, produced antagonistic effects. Notably, PSC combined with CAS consistently produced synergistic effects against C. albicans or C. parapsilosis planktonic cells. The exact reason for this phenomenon is not known; however, it could be attributed to the different properties of azole binding to different Candida cells (37).

The traditional technique for assessing antifungal combinations is the calculation of the fractional inhibitory concentration (FIC) index. However, the use of the FIC index has many limitations, which have been reviewed elsewhere, while the main problem, which particularly concerns biofilm experiments, is the index's reproducibility problems (26). The development of new models based on the response surface has been advocated to be more robust than the FIC index, resolving many problems of drug interaction modeling. In an attempt to better characterize antifungal agents' interactions, we analyzed our data with a Bliss independence model. Bliss independence theory is based on the assumption that two drugs act independently, with the probabilistic sense of independence. This model summarizes drug interactions with a single parameter, includes the statistical-significance levels of the interactions, and gives results that are reproducible and less sensitive to experimental error (9, 20).

In conclusion, our study has shown that triazole-echinocandin combinations are marginally effective when used serially, starting with the echinocandin, while their simultaneous use demonstrates indifferent effects against Candida biofilms. While moderately effective, our sequential-treatment model provides some supplementary evidence for intrinsic Candida biofilm resistance to antifungal agents. Not surprisingly, variable interactive activities are exerted against their planktonic counterparts.