Biofilm formation and virulence properties of Candida spp. isolated from hospitalised patients in Tunisia

Abstract

Control of opportunistic yeast is one of the factors reducing nosocomial infection in hospital. In this study, we investigated the presence of virulence factors including antifungal susceptibility, haemolytic potency, proteinase and phospholipase production. In addition biofilm formation was assessed using the XTT assay. A total of 24 Candida strains were isolated and identified. The most common species were Candida albicans and Candida glabrata (40.7%), followed by Candida dubliniensis (8.3%), Candida kefyr (4.2%) and Candida krusei (4.2%). In addition, 62.5% of strains were resistant to itraconazole and 16.7% of strains were resistant to fluconazole. All the strains were susceptible to amphotericin B. Phospholipase activity showed that 8.3% of the total strains were very strong producers and 41.7% were strong producers. Furthermore, three strains (12.5%) were found to have very strong proteinase activity. All the tested strains revealed beta haemolytic potency. All Candida strains were adhesive to polystyrene 96-well microtitre plates to differing degrees. The OD values of XTT reduction at 492 nm ranged from 0.025 (ca12) to 0.927 (ca2). The majority of C. albicans strains were highly metabolically active in comparison with the other species. Control of fungal infections in hospitalised patients is important to prevent the emergence of multi-resistant strains to commonly used antifungal agents.

Introduction

In recent years, the widespread use of immunosuppressants has led to the development of opportunistic infections. The human pathogen Candida albicans is the fourth leading cause of nosocomial bloodstream infections (Edmond et al. 1999). Invasive fungal infections remain a major cause of morbidity and mortality, particularly in immunocompromised patients (Segal et al. 2002). Recently, the spectrum of fungal infections has changed significantly in favour of non-Candida albicans species, such as Candida tropicalis, Candida krusei, Candida dubliniensis and Candida glabrata (Flanagan and Barnes 1998). In Brazilian hospitals, it has been shown that 37.5% and 24.3% of candidaemia in children was associated with C. albicans and Candida parapsilosis infection, respectively (Rodero et al. 2002). Fluconazole is an appropriate agent for fungal prophylaxis in patients with prolonged neutropenia. Additionally, amphotericin B is an antifungal with a greater spectrum of activity including Aspergillus species and yeast resistant to fluconazole (Wingard et al. 1991). Lipases are ubiquitous enzymes found in plants and microorganisms; these enzymes might help a microorganism to grow in environments where lipids are the sole carbon source. Previous reports described high lipase activities in microorganisms belonging to Candida rugosa and Candida parapsilosis (Neugnot et al. 2002).

Secretion of specific hydrolytic enzymes such as proteinases by C. albicans as possible virulence factors has been suggested (Neugnot et al. 2002). When C. albicans is phagocytosed by endothelial cells, it converts to the hyphal form and cause damages to the invaded cells (Zink et al. 1996). In addition, the same authors demonstrated that C. albicans with increased phospholipase activity were associated with higher virulence in murine models (Ibrahim et al. 1995). Candida species have the ability to produce a variety of enzymes facilitating hyphal invasion (Cutler 1991). It has been shown that mutants of C. albicans with deficient proteinase secretion were less virulent for mice than the parental organisms (Kwon-Chung et al. 1985). Secretory aspartyl proteinases (Sap) constitute a family of enzymes that are able to degrade several physiologically important substrates such as albumin, immunoglobulin and skin proteins (Cassone et al. 1987). Haemolysins are known to be putative virulence factors facilitating hyphal invasion in disseminated candidiasis. Many Candida species, including C. glabrata, exhibit varying ability to produce α or β haemolysins.

The adherence of C. albicans to host cells is an essential and necessary first step in successful colonisation and the development of pathogenesis and infection (Gristina et al. 1993).

Biofilm formation has been associated with the enhanced capacity of C. parapsilosis to colonise indwelling central venous catheters in individuals receiving intravenous hyper-alimentation (Branchini et al. 1994). Almost total resistance of biofilm-associated organisms to antifungal agents has been reported (Chandra et al. 2001). Candida biofilms are especially widespread and have been observed in most medical devices, such as stents, shunts, implants, endotracheal tubes, pacemakers, and various types of catheter (Ramage et al. 2006). The XTT reduction assay has been used as a routine tool for the quantification of Candida biofilms (Jahn et al. 1995) because it measures cell activity. The intracellular reduction of XTT releases a formozan compound that can be quantified by colorimetric estimation (Roehm et al. 1991).

The ultimate aim of our study was to investigate the antifungal susceptibility of clinical Candida species isolated from Tunisian hospitalised patients and to study their proteinase production, phospholipase activity, haemolytic potency and biofilm formation on an abiotic surface.

Materials and methods

Strain isolation and identification

The human pathogenic yeast used in this study was isolated from patients suffering from condidosis (14 strains from the paediatric service, 10 from other services) in the Ibn El JAZZAR hospital (Kairouan, centre of Tunisia).

The study included specimens that contained fungal strains isolated from blood culture, catheter, urine, stool, wounds, etc. The strains were cultured using standard mycological procedures. Cultures were incubated at 30°C for 24 h on Sabouraud chloramphenicol agar plates. Species identification was performed using the Api ID32 C system (BioMerieux, Marcy l’Etoile, France) according to the manufacturer’s specification, and the results were read using an automated microbiological mini-Api (BioMerieux).

In vitro antifungal susceptibility

Candida strains isolated from hospitalised patients were evaluated in vitro for their antifungal susceptibility toward the four antifungal agents using the commercial kit ATB Fungus2 (BioMerieux) according to the manufacture’s specification. Susceptibility to 5-fluorocytosin (5-FC), amphotericin B (AMB), fluconazol (FCA) and itraconazol (ITR) was tested.

ATB fungus2 was found a reliable and reproducible method with a repeatability of 96.6%, and a reproducibility of 95.4% and showed an excellent correlation of 91.7% with reference minimum inhibitory concentrations (MICs; Luo et al. 2001). The MICs of fluconazole, itraconazole, amphotericin B and 5-fluorocytosin were also determined.

Phospholipase and proteinase assay

Phospholipase activity was assayed according to a method described previously (Price et al. 1982). Briefly, a yeast suspension of each strain was made in 5 ml yeast peptone dextrose medium (YPD: 10 g yeast extract, 10 g peptone, 10 g glucose and distilled water up to 1,000 ml) and incubated at 37°C for 18 h. Following incubation, 1.5 ml of the yeast culture was centrifuged at 3,000 rpm for 5 min. The pellets obtained were washed twice in phosphate buffered saline (PBS) and centrifuged. After standardising the suspensions at 5 MacFarland with a Densimat (BioMerieux), 1 μl was plated in duplicate, at equidistant points, in phospholipase agar medium made by mixing 10 g peptone; 30 g glucose; 57.3 g NaCl; 0.55 g CaCl₂, 100 mL 50 % egg yolk enrichment and 900 ml distilled water. The inoculated plates were cultured at 37°C for 4 days. Phospholipase activity (Pz) was measured by dividing the colony diameter by the diameter of the precipitation zone around the colony formed on the plate. According to this system, a lower Pz ratio corresponds to a higher enzyme activity (higher virulence). A Pz of 1.0 was evaluated as negative, 0.99 > Pz > 0.9 as weak, 0.89 > Pz > 0.8 as mild, 0.79 > Pz > 0.7 as strong and Pz < 0.69 as very strong enzyme activity (Price et al. 1982).

The proteinase assay was studied in proteinase agar as previously studied (Aoki et al. 1990). The test medium contains 0.5 g bovine serum albumin (BSA), 0.2 g yeast extract, 60 ml of a solution containing 0.04 g MgSO₄⋅7H₂O, 0.5 g K₂HPO₄, 1 g NaCl, and 4 g glucose; the pH was adjusted to 3.5. The presence of proteinase activity was determined by the formation of a transparent halo around the yeast colonies. Proteinase activity (PZ) was measured and calculated in terms of the ratio of the diameter of the colony and that of the colony plus the precipitation zone (Price et al. 1982). Assays were repeated three times for each strain.

Haemolysin potency

Haemolysin potency was evaluated using a modification technique of the plate assay described previously (Quindos et al. 1994). In brief, 10 μl of a yeast suspension (10⁸cells/ml) was spotted onto sheep blood agar made by mixing 70 ml fresh sheep blood with 1,000 ml Sabouraud dextrose agar (SDA) supplemented with 3% glucose (w/v). The plates were incubated at 37°C in 5% CO₂ for 48 h. Positive haemolytic potency was recorded by the presence of a distinct translucent halo around the inoculum site. The diameters of the zones of lysis and the colony were measured and the ratio (equal to or larger than 1) was used as a haemolytic index to represent the intensity of haemolysin production by the tested strains. The assay was conducted in duplicate.

Biofilm formation

Strains were cultured aerobically at 30°C for 18 h on SDA and a loopful of growth was inoculated in Yeast Nitrogen Base (YNB) broth (Difco, Detroit, MI). After 18 h of incubation, the cells were harvested, washed twice with PBS (pH 7.2) and resuspended in YNB supplemented with 100 mM glucose. Standard Candida suspensions were adjusted to 0.5 McFarland (10⁷cells/ml) using a Densimat (BioMerieux).

A 100 μl aliquot of a standardised cell suspension was transferred into each well of a polystyrene flat-bottomed 96-well microtitre plate (Iwaki, Tokyo, Japan), and the plate was incubated for 1.5 h at 30°C in a rotary shaker at 75 rpm to allow the yeast to adhere to the surfaces of the wells (Kuhn et al. 2002).

Following the adhesion phase, cells suspensions were aspirated, and each well was washed twice with PBS to remove non-adherent cells. A total of 200 µl yeast nitrogen base medium supplemented with 100 mM glucose was transferred into each of the washed wells, and the plates were incubated at 37°C in a shaker at 75 rpm. The biofilms were allowed to develop for up to 48 h, and the yeasts were then quantified by the XTT (Sigma-Aldrich, Steinheim, Germany) reduction assay (Melo et al. 2006). The medium was changed daily by the aspiration of the spent medium and the addition of fresh medium. All assays were repeated five times on two separate occasions.

XTT reduction assay

The XTT reduction assay was performed according to methods described previously (Kuhn et al. 2002). Briefly, XTT solution (1 mg/ml) was prepared in PBS, filter-sterilised and stored at −70°C. Menadione (Sigma-Aldrich, Buchs, Switzerland) solution (0.4 mM) was prepared in acetone and sterilised immediately before each assay. The biofilms were first washed five times with PBS, and then 200 µl PBS and 12 µl XTT-menadione solution (12.5:1 v/v) were added to each of the prewashed wells and the control wells. The microtitre plate was then incubated for 2 h in the dark at 37°C. Following incubation, 100 µl of the solution was transferred to fresh wells, and the colour change in the solution was measured with a multiskan reader (GIO. DE VITA E C, Rome, Italy) at 492 nm. The absorbance values for the controls were then subtracted from the values of the tested wells to eliminate spurious results due to background interference.

Statistical analysis

Statistical analysis was performed on SPSS v.17.0 statistics software. Statistical differences and significance were assessed by one-way ANOVA test and Wilcoxon signed ranks test, as appropriate, to evaluate the phospholipase, protease activity (Pz values) according the type of strains and the origin of isolated strains. A P value <0.05 was considered significant.

Results

Biochemical characterisation of Candida strains

In total, five different species of yeast were identified among the 24 isolated strains. The most common species were Candida albicans (41.7 %) and Candida glabrata (41.7 %); followed by Candida dubliniensis, 2 (8.3 %); Candida kefyr, 1 (4.2 %) and Candida krusei, 1 (4.2 %).

In vitro antifungal susceptibility

The results of the antifungal susceptibility test showed that 16 out of 24 strains were resistant to itraconazole (62.5%) and 16.7 % were resistant to fluconazole (4 out of 24). All the tested strains were susceptible to amphotericin B. Furthermore, an intermediate phenotype was observed as follows: 3 strains out of 24 susceptible to fluconazole (16.7%), and one strain to 5-fluorocytosine (4.2%) as presented in Table 1.

Table 1 In vitro antifungal susceptibility of isolated strains and the minimum inhibitory concentration (MIC) achieved with ATB Fungus2

We also noted that two C. albicans (Ca6 and Ca7) were resistant to fluconazole and itraconazole with a higher MIC >128 μg/ml and > 4 μg/ml, respectively (Table 1). In addition, nine C. glabrata strains were found to be resistant to itraconazole (MIC>1 μg/ml). Moreover, all C. dubliniensis and C. krusei isolates were resistant to itraconazole. In addition, a significant difference in incidence of resistance to itraconazole between the fluconazole-resistant and fluconazole-susceptible Candida isolates was noted (P < 0.015).

Phospholipase, proteinase assay and haemolytic potency

The results presented in Table 2 showed that 2 out of 24 (8.3%) strains had a very strong phospholipase enzyme activity (Pz < 0.69) and 10 strains (41.7%) were strong (0.7 < Pz < 0.79), whereas 4 strains (16.7%) exhibited medium activity (0.8 < Pz < 0.89) In addition, no activity was observed in eight (33.3%) strains (Pz = 1). The Wilcoxon signed ranks test revealed that there is a statistically significant difference between phospholipase activity and the type of strain (C. albicans vs non-albicans species; P < 0.001).

Table 2 Phospholipase and proteinase activity of Candida spp. strains

The results of the proteinase activity assay showed that three strains had very strong activity (12.5%) and two strains (8.3%) had strong proteinase activity. In addition, 19 strains (79.2%) were found to be proteinase negative (PZ = 1). The paired samples test revealed that there is a statistically significant difference between the type of strains and the proteinase enzyme activity (P < 0.001).

The haemolysin production was evaluated using a sheep blood agar medium. All the tested strains have beta haemolytic activity, although C. glabrata (Ca18) showed a much higher beta-haemolytic activity than all other tested species. The Wilcoxon signed ranks test indicated that there is a statistically significant difference between haemolytic potency and phospholipase activity and between haemolytic potency and proteinase activity in the tested strains (P < 0.001).

In vitro quantitative biofilm formation by Candida strains on polystyrene plates

Biofilm formation by the Candida strains on the surface of polystyrene wells over 24 h as revealed by the colorimetric XTT formazan salt reduction assay is illustrated in Fig. 1. All Candida strains adhered to the polystyrene 96-well microtitre plate to differing degrees. The OD values of XTT reduction at 492 nm ranged from 0.025 (ca12) to 0.927 (ca2) as presented in Fig. 1.

Fig. 1

Biofilm oxidative activity of Candida strains on polystyrene plates after a 48-h incubation period

The majority of Candida albicans strains were highly metabolically active in comparison with the other species. Ca2 display the highest oxidative activity (0.927). Among the isolated Candida glabrata strains, two (ca11 and ca16) have a highly biofilm oxidative activity, at 0.567 and 0.594, respectively. In addition Candida dubliniensis (ca22), Candida krusei (ca23) and Candida kefyr (ca24) also have a high biofilm oxidative activity.

Discussion

This study isolated and identified five different Candida species. The most common species were Candida albicans and Candida glabrata (10/24). It has been reported that non-albicans strains were the causative agents in patients who had received fluconazole prophylaxis (Safdar et al. 2001). The results of the antifungal susceptibility test showed that 15 out of 24 strains were resistant to itraconazole (62.5%). In addition, 16.7% of the tested strains were resistant to fluconazole (4 out of 24). Furthermore, an intermediate phenotype was observed: 3 strains out of 24 were resistant to fluconazole (16.7%) and one strain to 5-fluorocytosine (4.2%) as shown in Table 1. Fluconazole prophylaxis may increase colonisation with azole-resistant Candida species in bone marrow transplant recipients (Marr et al. 2000). Two strains of C. albicans (Ca6 and Ca8) were resistant to both fluconazole and itraconazole (Table 1). Recently, an increase in the isolation of azole-resistant C. albicans strains has been reported (Waltimo et al. 2000). Fluconazole is a triazole agent that is established as a first-line antifungal for the treatment of candidosis. Furthermore, the repeated use of fluconazole to treat candidiasis infection led to the appearance of clinically resistant yeast. Itraconazole is prescribed as an alternative to fluconazole for treating candidosis. In this study, a higher prevalence of itraconazole resistant strains was noted. Amphotericin B is traditionally used in topical formats, although it may be administrated for the treatment of systemic infections in hospitalised patients.

In this study, we also noted that two strains (Ca22 and Ca23) had MICs = 1 μg/ml for amphotericin B and were resistant to itraconazole. It has been reported that Candida species may develop azole cross-resistance (Muller et al. 2000).

Many virulence attributes may be involved during the pathogenic process, including production of extracellular proteases and phospholipases, as well as haemolytic activity (Costa et al. 2010; Krcmery and Barnes 2002). In this study, the proteinase, phospholipase activities and haemolytic potency of Candida spp. strains isolated from clinical samples were evaluated. Phospholipase positive activity was detected in 66.7% of strains. In addition, two strains (8.3%) were very strong for phospholipase enzyme activity (Pz < 0.69) and ten strains (41.7%) were strong (0.7 < Pz < 0.79). These results agree well with those obtained by others (Bosco et al. 2003). On the other hand, 100% phospholipase-positive isolates were found elsewhere (de Oliveira et al. 1998). This enzyme has been associated with C. albicans and non-albicans species. Statistical analysis revealed a significant difference between phospholipase activity and the type of strain (P < 0.001). It has been reported that phospholipase activity has been detected in species other than C. albicans (Ghannoum 2000). Very strong phospholipase activity was noted in two C. albicans strains (ca1 and ca9) isolated from catheters.

In this study only 20.5% of strains (n = 5) were proteinase producers. This result contradicts that of another study reporting that 95% of C. albicans produced protease (Kantarcioglu and Yucel 2002). We also noted the presence of three very strong producer strains (ca, ca11 and ca20) as presented in Table 3.

Table 3 Phospholipase (Pz), proteinase (PZ) activity and haemolytic potency in the tested strains. SD Standard deviation

Some pathogenic micro-organisms secrete haemolytic factors in order to obtain haemoglobin as a source of iron (Belanger et al. 1995). The quantitative data indicated that all the tested strains have beta-haemolytic activity, although C. glabrata (Ca12) showed a much higher beta-haemolytic activity than all the other tested species. It has been reported that C. albicans has haemolytic activity (Manns et al. 1994) and it has been observed that non-C. albicans species are able to produce one or more types of haemolysin in vitro, with differences among species (Luo et al. 2001).

The ability of Candida isolates to form biofilm on polystyrene plate was tested using the XTT reduction assay. This technique is considered ideal for quantification of Candida biofilm mass (Jin et al. 2004). We noted that adhesion ability differed from one strain to another for the same species. The majority of C. albicans strains were highly metabolically active in comparison with other species.

The OD values in the biofilm reduction assay ranged from 0.025 (ca12) to 0.927 (ca2) as presented in Fig. 1. This result was agreement with a recent study which reported considerable differences in biofilm production among Candida species (Shin et al. 2002). It has been reported that XTT metabolism can be influenced by extrinsic factors (Neugnot et al. 2002).

Ca2 displayed the highest biofilm oxidative activity (0.927). Among the isolated C. glabrata, two strains (ca11 and ca16) have a highly oxidative activity of 0.567 and 0.594, respectively. In addition, Candida dubliniensis (ca22), Candida krusei (ca23) and Candida kefyr (ca24) also have high biofilm oxidative activity.

Nevertheless, the present results clearly show that differences in the adhesion ability among the isolated strains to abiotic surface are strain dependent. Different intra-species adherence ability has also been reported by other authors for non-albicans species (Henriques et al. 2007; Panagoda et al. 2001).

According to the results of the present study, we conclude that phospholipase and proteinase production by Candida strains may indicate the virulence of such isolates. The findings reported here indicate that many Candida species exhibit various capacities to produce haemolysin factor, and their adhesion ability to abiotic surfaces contributes to their persistence and to biomaterial colonisation. Control of nosocomial infections caused by these opportunistic pathogens is important due to their modest susceptibility to common antifungal agents, particularly triazole.