Biofilm Formation by the Emerging Fungal Pathogen Trichosporon asahii: Development, Architecture, and Antifungal Resistance

Abstract

Trichosporon asahii is the most common cause of fatal disseminated trichosporonosis, frequently associated with indwelling medical devices. Despite the use of antifungal drugs to treat trichosporonosis, infection is often persistent and is associated with high mortality. This drove our interest in evaluating the capability of T. asahii to form a biofilm on biomaterial-representative polystyrene surfaces through the development and optimization of a reproducible T. asahii-associated biofilm model. Time course analyses of viable counts and a formazan salt reduction assay, as well as microscopy studies, revealed that biofilm formation by T. asahii occurred in an organized fashion through four distinct developmental phases: initial adherence of yeast cells (0 to 2 h), germination and microcolony formation (2 to 4 h), filamentation (4 to 6 h), and proliferation and maturation (24 to 72 h). Scanning electron microscopy and confocal scanning laser microscopy revealed that mature T. asahii biofilms (72-h) displayed a complex, heterogeneous three-dimensional structure, consisting of a dense network of metabolically active yeast cells and hyphal elements completely embedded within exopolymeric material. Antifungal susceptibility testing demonstrated a remarkable rise in the MICs of sessile T. asahii cells against clinically used amphotericin B, caspofungin, voriconazole, and fluconazole compared to their planktonic counterparts. In particular, T. asahii biofilms were up to 16,000 times more resistant to voriconazole, the most active agent against planktonic cells (MIC, 0.06 μg/ml). Our results suggest that the ability of T. asahii to form a biofilm may be a major factor in determining persistence of the infection in spite of in vitro susceptibility of clinical isolates.

Candida species are the most common cause of disseminated nosocomial fungal infections (34). Invasive infections by rarer opportunistic fungal pathogens, however, have recently emerged as a significant problem in treatment of immunocompromised patients (8, 34).

In particular, disseminated life-threatening Trichosporon infection is becoming increasingly common in patients with underlying hematological malignancies, extensive burns, solid tumors, and AIDS, accounting for approximately 10% of all confirmed cases of disseminated fungal infections (8, 12, 14, 32-34). Likewise, nonimmunosuppressed patients have suffered from Trichosporon infections associated with ophthalmologic surgery, infections of prosthetic devices, intravenous drug abuse, and peritoneal dialysis (1, 17, 21).

Trichosporon beigelii was formerly considered the causative agent in trichosporonosis, but recent taxonomic findings based on partial sequences of large-subunit rRNA and DNA relatedness (9, 29, 30) revealed that T. beigelii actually consists of six distinct pathogenic Trichosporon species.

In particular, Trichosporon asahii is the major cause of disseminated or deep-seated trichosporonosis (12, 30). Although most of the reported cases of hematogenous T. asahii infections occurred in patients with leukemia during a neutropenic phase, another relevant predisposing factor for infection is represented by the presence of invasive devices, such as intravenous or urinary catheters (13, 22, 26, 28), endoscopic forceps (18), and arteriovenous graft (15). These findings suggest that prosthetic devices could act as substrates for adhesion and, possibly, growth as biofilms, structured microbial communities embedded in an extracellular polymeric substance (EPS). As for bacterial and Candida biofilms, consequences of T. asahii biofilm growth could be twofold: markedly enhanced resistance to antimicrobial agents and protection from host defenses (2, 4, 7). Therefore, demonstration of the biofilm lifestyle could be critical in relation to therapeutic strategies engaged to treat deep-seated T. asahii infections. In spite of these considerations, no studies have so far dealt with the ability of T. asahii to produce biofilms.

The purpose of our study was threefold. The first objective was to verify whether T. asahii can produce biofilms on abiotic (polystyrene) surfaces, and we addressed this point by developing and optimizing a reproducible T. asahii-associated biofilm model. The second objective was to gain informations about the architecture of possible T. asahii biofilm by means of scanning electron and confocal scanning laser microscopic analyses. The third objective was to evaluate the in vitro susceptibility of T. asahii biofilms to four currently used antifungal agents (amphotericin B, caspofungin, fluconazole, and voriconazole).

MATERIALS AND METHODS

Microorganisms.

Three clinical strains of T. asahii (TA309, TA310, and TA311), isolated from blood samples of patients admitted to the “Santo Spirito” Hospital of Pescara for hematological malignancies, and T. asahii ATCC 201110 were used in the present study. Strains were stored at −80°C until use. When needed, strains were thawed and subcultured (37°C, 24 h) once in yeast peptone dextrose (Oxoid, Garbagnate M.se, Italy) broth and then twice on Sabouraud dextrose agar (Oxoid). Identification of the investigated isolates tested was performed by PCR amplification of sequences of the internal transcribed spacer regions, according to the method of Sugita et al. (31), with minor modifications. Briefly, total DNA isolation was performed with the DNeasy tissue kit (QIAGEN, Milan, Italy)—for animal tissues and cells, yeasts, or bacteria—according to the manufacturer's instructions. The primers (Invitrogen, Milan, Italy) were TAAF (forward; 5′-GGATCATTAGTGATTGCCTTAATA-3′) and pITS4 (reverse; 5′-TCCTCCGCTTATTGATATG-3′). PCR was carried out in a total volume of 50 μl containing the following: 10 ng template DNA, 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1 mM MgCl₂, 25 mM (each) dATP, dCTP, dGTP, and dTTP, 2 mM (each) primers, and 0.5 U of Taq DNA polymerase. Amplifications were carried out in a Perkin-Elmer 9700 thermal cycler, using the following program: 95°C for 5 min, followed by 35 cycles consisting of 94°C for 30 s, 58°C for 30 s, and 72°C for 40 s, with a final extension period at 72°C for 10 min. After thermal cycling, 5 μl of the amplified product was separated electrophoretically in 1.5% (wt/vol) agarose gels in Tris-borate-EDTA buffer at a 100-V constant voltage for 50 min, stained with ethidium bromide, and photographed under transilluminated UV light. Trichosporon cutaneum ATCC 62965 and T. asahii ATCC 201110 were used in parallel to test the specificities of oligonucleotide primers. The primers proved quite specific for DNA of T. asahii only, producing approximately 500-bp fragments.

Biofilm production assay.

Forty milliliters of yeast peptone dextrose broth, prepared in a 50-ml Falcon tube, were inoculated with T. asahii isolates grown overnight on Sabouraud dextrose agar plates and incubated overnight at 37°C in an orbital shaker at 130 rpm. Cells were harvested, washed once with sterile phosphate-buffered saline (PBS), and then resuspended in RPMI 1640 (Sigma-Aldrich Italia, Milan, Italy) adjusted to pH 7.0 with 0.165 M morpholinepropanesulfonic acid (MOPS) (Sigma-Aldrich Italia) to a concentration of 10⁵ CFU/ml. Biofilms were formed by pipetting 2 ml of the standardized cell suspension (10⁵ CFU/ml) into commercially available presterilized, polystyrene, 35-mm-diameter tissue culture petri dishes (Iwaki, Bibby Sterilin, Milan, Italy) and incubating them at 37°C (t₀). After 60 min of incubation (adhesion phase), dishes were washed twice with sterile PBS to remove nonadherent cells and then refilled with 2 ml of fresh RPMI 1640-MOPS medium. Samples were then incubated up to 48 h at 37°C, replacing the medium each 24 h of incubation (biofilm formation phase). At the end of the incubation time, the medium was aspirated and nonadherent cells were removed by thoroughly washing the biofilms two times in 2 ml of sterile PBS. As a control, additional dishes were processed in an identical fashion, except that no T. asahii cells were added.

Quantitation of biofilm.

Quantitation of T. asahii biofilm was performed by using a colorimetric assay based on XTT {sodium 3′-[1-(phenylamino-carbonyl)-3,4-tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate} reduction (cell proliferation kit II; Roche Diagnostics SpA, Milan, Italy). The assay is based on the cleavage of the yellow tetrazolium salt XTT to form an orange formazan dye by metabolically active cells. Briefly, XTT was prepared in RPMI 1640-MOPS at 0.5 mg/ml, and then the electron-coupling agent N-methyl dibenzopyrazine methyl sulfate (PMS) was added to a final concentration of 12.5 μM. A 1.5 ml-aliquot of the XTT-PMS solution was added to each prewashed biofilm and to control wells (for the measurement of background XTT reduction levels). The plates were then incubated in the dark for 2 h at 37°C. Supernatants were centrifuged and then divided into aliquots in a 96-well microplate for reading. A colorimetric change in the XTT reduction assay, which is a direct correlation of the metabolic activity of cells within the biofilm, was then measured in a microtiter plate reader (Sunrise; Tecan Italia srl; Cologno Monzese, Milan, Italy) at 492/620 (read/reference) nm.

Kinetics of biofilm formation.

T. asahii ATCC 201110 was used for kinetic studies of biofilm formation. Biofilms were allowed to develop, as described above, over a series of time intervals (0.5, 1, 2, 4, 8, 24, 48, and 72 h). At each time interval, biofilm formation was assessed by XTT reduction assay and concurrently by total viable cell counting in the same sample in order to avoid a possible bias due to a different sampling. Sessile cells were removed by scraping with a sterile scraper (Iwaki) and then resuspended in sterile PBS by vortexing at maximum speed for 1 min. Serial 10-fold dilutions in sterile PBS were performed on each sample. One hundred microliters of each dilution were plated on Sabouraud dextrose agar and incubated at 37°C for 24 h. Colonies were counted to estimate the total viable cell counts from each sample.

SEM.

For scanning electron microscopy (SEM) analysis, biofilms were preformed on 35-mm-diameter tissue culture polystyrene dishes (Iwaki) and then immersed for 16 h in a mixture of 2% paraformaldehyde (vol/vol) and 2% glutaraldehyde (vol/vol) in 0.15 M sodium cacodylate buffer, pH 7.4, with or without the cationic dye 0.1% alcian blue (Polysciences Europe; Eppelheim, Germany). The samples were washed in 0.15 M cacodylate buffer for 5 min and postfixed for 90 min in 1% OsO₄ (vol/vol) in 0.15 M cacodylate buffer. Samples were rinsed for 5 min in 0.15 M cacodylate buffer and dehydrated in an ascending ethanol series (50, 70, 80, 95, and 100% [twice]; 10 min each) before 30 min of drying with hexamethyldisilazane (Polysciences Europe) and finally were air dried overnight. The specimens were coated with gold-palladium by Polaron E5100 II (Polaron Instruments, Inc.). After processing, samples were observed with a Philips XL30CP scanning electron microscope in the high-vacuum mode at 15 kV. Images were processed for display using Photoshop (Adobe Systems, Inc., San Jose, Calif.) software.

CSLM.

Biofilms were allowed to form as described above for SEM analysis. After incubation at 37°C for 72 h, samples were gently washed with PBS and incubated for 60 min at 37°C in 4 ml of PBS containing the fluorescent stains FUN-1 (20 μM) and concanavalin A-Alexa Fluor 594 conjugate (ConA) (10 μM), both from Molecular Probes (Eugene, Ore.). FUN-1 is converted to orange-red by metabolically active cells, while ConA binds to glucose and mannose residues of cell wall polysaccharides with green fluorescence. The confocal scanning laser microscopy (CSLM) analysis was performed with an LSM 510 META laser scanning microscope (Zeiss, Germany) attached to an Axioplan II microscope (Zeiss). Biofilms were observed using a ×100 oil immersion objective (α-Plan-FLUAR). The excitation wavelengths were 488 (Argon laser) and 543 (He-Ne laser), and emission wavelengths were 540 and 615 nm for FUN-1 and ConA, respectively. Depth measurements were taken at regular intervals across the width of the device. To determine biofilm structure, a series of horizontal (x-y) optical sections with a thickness of 0.18 μm at 0.51-μm intervals was taken throughout the length of the biofilm. Confocal images of green (ConA) and red (FUN-1) fluorescence were conceived simultaneously using a track mode. Yellow areas represent dual staining. Three-dimensional reconstructions of imaged biofilms were obtained by Amira 3.1.1 (Mercury Computer Systems; Chelmsford, MA) software. The images were captured and processed for display using Adobe Photoshop (Adobe Systems, Inc.) software.

Antifungal susceptibility testing of planktonic and biofilm cells.

Amphotericin B (Sigma-Aldrich Srl), caspofungin-acetate (Merck-Sharp & Dohme Italia SpA), fluconazole (Pfizer Italia, Rome, Italy), and voriconazole (Pfizer Italia) were used in this study. A stock solution was prepared in RPMI 1640-MOPS for each antifungal agent and stored at −80°C until use. The antifungal susceptibility of planktonic cells was assessed in two ways: (i) susceptibility testing was performed with 96-well round-bottom tissue culture plates, as described in the M27-A2 procedure for the susceptibility testing of yeasts (24). (ii) The second assay of susceptibility testing consisted of the XTT-based colorimetric method previously described by Hawser et al. (11), with minor modifications. Briefly, the susceptibility plates were prepared as for the CLSI (formerly NCCLS) method. Two hours prior to the endpoint reading, plates were agitated, and 50 μl of a mixture of XTT (0.5 mg/ml) plus the electron-coupling agent PMS (12.5 μM) were added to all wells. Plates were incubated for 2 h at 35°C to allow color development. MICs in the XTT assay (MIC_XTT) were determined spectrophotometrically (492 nm/620 nm [read/reference]) as the lowest concentration of antifungal agent causing a 100% reduction in metabolic activity (no color change occurring). We chose this endpoint since Hawser et al. previously showed that the 100% reduction in metabolic activity is equivalent to the MIC as determined by the CLSI M27-A2 method for planktonic yeast cells (11). Candida parapsilosis ATCC 22019 and Candida krusei ATCC 6258 were included in each run of experiments for quality control.

For antifungal susceptibility testing of biofilm cells, antifungal agents were tested against 48-h biofilms preformed on 35-mm-diameter tissue culture polystyrene dishes (Iwaki). Following incubation, biofilms were washed twice with sterile PBS and then exposed to each tested antifungal agent for a further 24 h at 37°C, at concentrations ranging from 16 to 1,024 μg/ml, prepared in fresh RPMI 1640-MOPS. A series of antifungal agent-free dishes was also included to serve as controls. The effect of antifungal agents against biofilms was measured using the XTT reduction assay described above for quantitation of biofilm. The antifungal concentrations which caused a 100% reduction in metabolic activity (MIC_XTT) of biofilms were determined.

Statistical analysis.

Each experiment was carried out in triplicate and repeated on two different occasions. Statistical analysis was performed by calculating the Pearson correlation coefficient r, the one-way analysis-of-variance test, and the Newman-Keuls multiple-comparison posttest. The analyses were performed by using Prism version 4.0 for Windows (GraphPad Software, San Diego, Calif.). P values of <0.05 were considered statistically significant.

RESULTS

Standardization and optimization of T. asahii biofilm growth on polystyrene.

Since conditions for T. asahii biofilm formation on polystyrene surfaces have not been established, preliminary experiments were carried out to standardize and optimize biofilm formation. Basic parameters for biofilm growth—such as inoculum size (10⁴, 10⁵, or 10⁶ CFU/ml), adhesion time (30, 60, or 120 min), and biofilm formation time (24, 48, or 72 h)—were considered for optimization. Figure 1 illustrates T. asahii ATCC 201110 biofilm growth, assessed by using XTT and viable-count assays, to determine optimal adhesion time, inoculum concentration, and duration of growth. By using an inoculum size of 10⁴ or 10⁵ CFU/ml, biofilm production significantly increased with incubation and adhesion times. On the contrary, biofilm produced with an inoculum size of 10⁶ CFU/ml and adhesion time of 30 or 120 min reached a plateau by 48 h. Maximum biofilm formation was achieved after 72 h of incubation, with an adhesion time of 60 min and an inoculum size of 10⁶ CFU/ml, an adhesion time of 30 min and inoculum size of 10⁵ CFU/ml, and an adhesion time of 60 min and inoculum size of 10⁵ CFU/ml, without any statistically significant difference. However, the last experimental conditions gave the best correlation with the XTT assay (Pearson r = 0.993). Therefore, we selected an inoculum size of 10⁵ CFU/ml, an adhesion time of 60 min, and a biofilm formation time of 72 h as optimal experimental conditions for growing T. asahii biofilm on polystyrene surfaces.

FIG.1.

Optimization of Trichosporon asahii biofilm growth on polystyrene surface. Effect of adhesion time (30 [▪], 60 [▴], or 120 [▾] min), incubation time (24, 48, or 72 h), and inoculum concentration (A, B: 10⁴ CFU/ml; C, D: 10⁵ CFU/ml; E, F: 10⁶ CFU/ml) on T. asahii ATCC 201110 biofilm formation assessed by viable count (A, C, and E) and XTT (B, D, and F) assays. Values were plotted as means ± standard deviations (bars).

In vitro biofilm formation by T. asahii.

The kinetic of biofilm formation by the T. asahii ATCC 201110 strain on the surfaces of polystyrene wells over 72 h is illustrated in Fig. 2, as revealed by both the colorimetric XTT formazan salt reduction assay and total viable counts. The biofilm growth curve showed that T. asahii rapidly adhered to polystyrene after a 30-min incubation, exhibiting low metabolic activity in the first 4 h. Thereafter, the biofilm matured, and its complexity exponentially increased after 6 h up to 72 h, probably reflecting the increased number of cells that constituted the mature biofilm. The excellent correlation between results of XTT and viable-count assays (Pearson correlation coefficient r = 0.936; P < 0.0001) clearly supports that the XTT assay absorbance readings were proportional to the cellular density in the biofilm (Fig. 2). Thus, in the present study we used the more rapid and simpler XTT-based method to allow high-throughout analysis of antifungal efficacy.

FIG. 2.

Kinetic of biofilm formation by T. asahii ATCC 201110 as determined by both XTT (optical density at 492 nm/optical density at 620 nm [OD_492/620]) and viable-count (CFU/plate) assays. Values were plotted as means ± standard deviations (bars). Linear regression analysis revealed a regression coefficient of 0.936 (P < 0.0001) between viable-count and XTT assay results.

T. asahii biofilm imaging.

(i) SEM. SEM analysis was used to monitor biofilm formation and to analyze the morphological characteristics of adhered cells (Fig. 3 and 4). After 30 min, adherent budding yeast cells stuck to the wells in a random manner (Fig. 3A) (adhesion phase). Filamentation was initiated within 1 h, and after 4 h, early biofilm appeared as small microcolonies, consisting mainly of filamentous forms (microcolony formation phase) (Fig. 3B). After 6 h, hyphal forms from neighboring microcolonies merged into an intricate network of spatially dispersed filamentous forms that, after 8 h, intertwined to form a coherent monolayer of woven-like structures (Fig. 3C). During the maturation phase (24 to 72 h), the complexity of the biofilm increased from a monolayer into a multilayered-structure, involving all fungal morphologies (yeast cells, hyphae, and pseudohyphae), resulting in a compact structure covering the entire surface of the well (Fig. 3D to G). The kinetics of biofilm formation by different clinical isolates showed a pattern similar to that of the type strain, although minor variations in individual ability to adhere to polystyrene were detected (data not shown).

FIG. 3.

(A to F) SEM images of T. asahii ATCC 201110 biofilm formed on a polystyrene surface after 30 min or 4, 8, 24, 48, or 72 h, respectively. (G) Magnification of F. Magnification, ×100 (A to F) or ×500 (G).

FIG. 4.

SEM images of mature (72-h-old) T. asahii TA309 biofilms on polystyrene. Biofilm cells stained without (A) or with (B) 0.1% alcian blue. Fixation for 22 h in aldehyde containing 0.1% alcian blue visualizes an extensive network of EPS filaments surrounding cells and bridging cell surfaces. Magnification, ×7,000.

In order to better visualize EPS in biofilm samples, we used the cationic dye alcian blue (Fig. 4). Immersion fixation of T. asahii biofilms in the paraformaldehyde-glutaraldehyde cocktail without alcian blue revealed a complete lack of detectable EPS on the cell surface (Fig. 4A). On the contrary, addition of 0.1% alcian blue to the aldehyde fixation and 22 h of fixation induced an entirely different appearance (Fig. 4B). During the biofilm formation, the amount of EPS increased in parallel with incubation time. After 4 h of incubation, the surfaces of fungal cells were covered by a predominantly noncellular material with an amorphous granular appearance (data not shown) until T. asahii sessile communities in the mature biofilm were completely embedded in an EPS that resulted in an extensive network of filaments bridging surfaces of individual cells (Fig. 4B). Close examination of biofilm structures showed coaggregating T. asahii cells mediating the adhesion onto the polystyrene surface.

(ii) CSLM.

For CSLM examination of a 72-h-old T. asahii TA311 biofilm, we used a combination of the fluorescent dyes FUN-1 (staining metabolically active cells) and ConA (staining cell wall glucose and mannose residues). Vertical (x-z) sections or side views of the three-dimensional (3-D) reconstructed images were used to determine biofilm thickness and architecture (Fig. 5A). Further, Fig. 5B to D shows three-dimensional representations of the biofilm resulting from the compilation of a series of individual x-y sections taken across the z axis. The thickness of the biofilm at maturity varied from 25 to nearly 40 μm. Mature T. asahii biofilm consisted of a highly organized structure, displaying significant channelling and porosity. This heterogeneity was in terms of the distribution of fungal cells (indicated by the red color due to FUN-1 staining localized in dense aggregates in the cytoplasm of metabolically active cells) and EPS (green coloration resulting from ConA binding to mannose and glucose residues present in cell wall polysaccharides). In fact, mature T. asahii biofilms revealed internal regions of a dense network of metabolically active cells (yeast cells and hyphal elements) completely encased within EPS, with a basal layer consisting of both densely packed yeast and hyphal structures encased in and often obscured by a surrounding extracellular material. EPS increased through the entire thickness, from the basal to the upper layers (Fig. 5C and D).

FIG. 5.

CSLM examination of T. asahii TA311 biofilm with both FUN-1 and ConA stains after 72 h of development. (A) Orthogonal images of the basal layer show that mature biofilm consisted of mostly metabolically active (red, FUN1-stained) cells embedded in the polysaccharide extracellular material (green, ConA stained). (B to D) Three-dimensional representations of T. asahii biofilm: (B) top view; (C) basal layer, lateral view; (D) upper layer, lateral view. Image capture was set for simultaneous visualization of both green and red fluorescence (A, C, D) or for visualization of red fluorescence only (B). Magnification, ×100.

Susceptibility of planktonic and sessile (biofilm) cells to antifungal agents.

In vitro susceptibilities of biofilm versus planktonic cells of T. asahii to amphotericin B, caspofungin, voriconazole, and fluconazole are summarized in Table 1. MICs determined by the CLSI method (MIC_CLSIs) revealed that voriconazole was the most active agent against planktonic T. asahii cells of strains tested (MIC, 0.06 μg/ml). On the contrary, caspofungin (MIC, 32 μg/ml), fluconazole (MIC range, 2 to 4 μg/ml), and amphotericin B (MIC range, 4 to 16 μg/ml) were revealed to be poorly active. The same trend in susceptibility was confirmed by the results of the XTT assay: the overall agreement between XTT and CLSI readings was excellent (93.7%), as determined for isolates whose MIC_XTT was either the same as or within 1 dilution of the MIC_CLSI.

TABLE 1.

In vitro activities of antifungal agents against different T. asahii strains under planktonic or biofilm growth conditions

Antifungal agent	Isolate	Susceptibility (μg/ml) ofa:
		Planktonic cells		Biofilm (MICXTT)
		MICCLSI	MICXTT
Amphotericin B	TA309	8	8	>1,024
	TA310	8	8	>1,024
	TA311	16	8	512
	ATCC 201110	4	4	256
Caspofungin	TA309	32	32	>1,024
	TA310	32	32	512
	TA311	32	32	128
	ATCC 201110	32	32	512
Voriconazole	TA309	0.06	0.12	>1,024
	TA310	0.06	0.12	>1,024
	TA311	0.06	0.12	>1,024
	ATCC 201110	0.06	0.12	>1,024
Fluconazole	TA309	2	4	>1,024
	TA310	2	4	>1,024
	TA311	4	>128	>1,024
	ATCC 201110	4	4	>1,024

^aSusceptibility is expressed as the MIC of the antifungal agent for the isolate, determined by the CLSI method or the XTT assay.

The in vitro activities of caspofungin, amphotericin B, fluconazole, and voriconazole against preformed T. asahii biofilms were assessed using the XTT assay (Table 1). All of the antifungal agents tested showed decreased activity against preformed biofilms of all T. asahii strains, since the biofilm MIC_XTTs were generally much greater than the concentration of antifungal agent required to inhibit planktonic cells. Data revealed that T. asahii biofilms were intrinsically resistant to all antifungal agents tested, although at different levels: as indicated by MIC_XTTs, the activity against biofilms was reduced by 6 to 8 log₂ times for amphotericin B, 2 to 6 log₂ times for caspofungin, 14 log₂ times for voriconazole, and 3 to 9 log₂ times for fluconazole, compared with their activities against planktonic cultures. Importantly, although voriconazole showed potent activity against planktonically grown T. asahii, this antifungal never eradicated biofilms, as reflected by residual metabolic activity of biofilms at concentrations up to 1,024 μg/ml.

DISCUSSION

Trichosporon asahii is the most common cause of fatal disseminated trichosporonosis (12, 30), frequently associated with indwelling medical devices (13, 15, 18, 22, 26, 28). In spite of antifungal drugs administered, trichosporonosis is often persistent or reestablished soon after treatment (12, 34). These clinical observations induced us to evaluate the ability of T. asahii to form biofilms on polystyrene surfaces as a means of efficient host infection.

Our results demonstrate that T. asahii, like many other fungal and bacterial species (6), can produce biofilms through a discrete sequence of events, including fungal surface adhesion, microcolony formation, and biofilm maturation. Within 30 min of incubation, T. asahii rapidly adhered to polystyrene, probably because of the intrinsic high-level cell surface hydrophobicity displayed by all strains tested (data not shown). Within the first 4 h, although on a low metabolic profile, adhered cells organized themselves as a microcolony, probably representing an early microbial adaptative response to the new environment. Thereafter, biofilm complexity exponentially increased after 6 h up to 72 h, in parallel with the increased number of cells in the mature biofilm. Planktonic T. asahii can exist as yeast cells, pseudohyphae, or hyphae (5). SEM analysis revealed that mature T. asahii biofilms consist of a mixture of yeast and filamentous forms, suggesting that biofilm formation in T. asahii is not morphology specific. Our findings are consistent with those of other authors reporting on Candida spp. biofilms (3, 27).

The nondestructive CSLM technique enabled in situ visualization of the intact biofilm community. Vertical (x-z) sectioning (side view) of 3-D reconstructed images and 3-D representations suggested that the architecture of mature T. asahii biofilm consisted of the presence of thin areas of metabolically active hyphal and yeast forms interwoven with EPS composed of polysaccharides, such as mannose and glucose residues present in the cell wall, as suggested by the diffuse staining with ConA. Overall, results indicated that mature T. asahii biofilms displayed a typical microcolony/water channel architecture, with extensive spatial heterogeneity. This structure is similar to that seen with bacterial biofilm systems (6) and may represent an optimal arrangement for the influx of nutrients, disposal of waste products, and establishment of microniches throughout the biofilm.

Despite the increased frequency and severity of trichosporonosis, few data on the susceptibility of T. asahii to antifungal agents are available, with no established breakpoints to define antifungal susceptibility for Trichosporon spp. Our results would indicate a better performance of voriconazole, a recent azole compound, than of fluconazole, amphotericin B, or caspofungin against planktonic T. asahii cells, in accordance with previous in vitro studies (20, 25). Susceptibility data generated for planktonic cells, however, do not account for the intrinsic resistance exhibited by their sessile counterparts. In fact, our results showed that biofilm-associated cells of T. asahii were intrinsically resistant to all antifungals tested. In particular, T. asahii biofilms were up to 16,000 times less sensitive to voriconazole, suggesting that this antifungal may be useful for prevention of biofilm formation rather than for treating already-formed biofilm. The intrinsic resistance of T. asahii biofilms to azoles may not be a consequence of the fungistatic nature of these drugs, since both fungicidal caspofungin and amphotericin B were also inactive against T. asahii biofilms. Our results correlate with previous reports for Candida spp., suggesting that the biofilm phenotype may be associated with resistance to antifungal drug therapy (3, 10, 16, 27). Overall, the disparity between MICs for planktonic and sessile cultures from an identical isolate may therefore explain why, even with current antifungal treatment, mortality of patients with disseminated trichosporonosis can be as high as 70% (5); furthermore, our data may partially explain the poor correlation between clinical (in vivo) and mycological (in vitro) resistance.

A better knowledge of the mechanisms of antifungal drug resistance may lead to the development of novel therapies for biofilm-based diseases. Multiple mechanisms have been proposed for the biofilm resistance phenomenon (23). Metabolic quiescence has been proposed as a mechanism of antimicrobial resistance in biofilm bacteria (19) and fungi (23). However, our data show that T. asahii biofilm-embedded cells actively metabolize XTT, in line with data previously reported for Candida albicans and C. parapsilosis (16). It is therefore unlikely that this mechanism is a major factor promoting antifungal resistance of T. asahii biofilms. It has been proposed that EPS, produced by and enveloping sessile communities of cells, could act as a barrier to the diffusion of antibiotics and/or as an ion exchange resin to bind charged antibiotic molecules (23). Our results indicate that EPS increased throughout the incubation time, providing an increasingly organized and extensive network of filaments pervading both blastospores and hyphal elements in the entire biofilm thickness. This mechanism, therefore, seems to be plausibly associated with the lack of activity of antifungals observed for T. asahii biofilms in our model, although different and purposely designed experiments will be needed to further address this point.

The biofilm lifestyle may also serve as a safe reservoir for the release of infecting cells into the surrounding fluid phase to move onto other surfaces. Thus, the ability to form a biofilm may represent a key factor for the survival of T. asahii, which seems to be particularly well adapted to colonization of polystyrene surfaces.