|Home | About | Journals | Submit | Contact Us | Français|
Candida albicans biofilms form on indwelling medical devices (e.g., denture acrylic or intravenous catheters) and are associated with both oral and invasive candidiasis. Here, we determined whether surface modifications of polyetherurethane (Elasthane 80A [E80A]), polycarbonateurethane, and poly(ethyleneterephthalate) (PET) can influence fungal biofilm formation. Polyurethanes were modified by adding 6% polyethylene oxide (6PEO), 6% fluorocarbon, or silicone, while the PET surface was modified to generate hydrophilic, hydrophobic, cationic, or anionic surfaces. Formation of biofilm was quantified by determining metabolic activity and total biomass (dry weight), while its architecture was analyzed by confocal scanning laser microscopy (CSLM). The metabolic activity of biofilm formed by C. albicans on 6PEO-E80A was significantly reduced (by 78%) compared to that of biofilm formed on the nonmodified E80A (optical densities of 0.054 ± 0.020 and 0.24 ± 0.10, respectively; P = 0.037). The total biomass of Candida biofilm formed on 6PEO-E80A was 74% lower than that on the nonmodified E80A surface (0.46 ± 0.15 versus 1.76 ± 0.32 mg, respectively; P = 0.003). Fungal cells were easily detached from the 6PEO-E80A surface, and we were unable to detect C. albicans biofilm on this surface by CSLM. All other surface modifications allowed formation of C. albicans biofilm, with some differences in thearchitecture. Correlation between contact angle and biofilm formation was observed for polyetherurethane substrates (r = 0.88) but not for PET biomaterials (r = −0.40). This study illustrates that surface modification is a viable approach for identifying surfaces that have antibiofilm characteristics. Investigations into the clinical utility of the identified surfaces are warranted.
Indwelling medical devices, including dental materials and central venous catheters, are commonly used in the care of patients. These devices include dentures, dental implants, intravenous catheters, stents, pacemaker valves, contact lenses, etc., and provide a diverse range of surfaces where microbial infection can become established (5, 21). The increasing use of implants and medical devices has led to a concomitant increase in the incidence of device-related infections (23), with the most common fungal infections being due to Candida albicans. Management of device-related Candida infections can be challenging, and removal of the infected device is generally needed to establish cure. However, removal of the device can be associated with complications, resulting in significant medical and economic impact. Therefore, device-associated infections due to biofilms represent a significant problem.
One approach to prevent or reduce biofilm formation on biomaterials is to modify their surface chemistries, e.g., by adding surface-modifying end groups (SMEs) or by altering the chemical composition of substrates. Such modifications have been shown to prevent or reduce bacterial adhesion and biofilm formation (9, 10, 18, 20). However, the effect of different surface chemistries on the ability of C. albicans to form biofilms on bioprosthetic surfaces has not been investigated.
In this study, we determine whether the ability of C. albicans to form biofilm in vitro is affected by (i) the presence of SMEs (6% [wt/wt] polyethylene oxide [6PEO], 6% [wt/wt] fluorocarbon [6FC], or silicone) on polyurethane (polyetherurethane and polycarbonateurethane) or (ii) chemical modifications (hydrophobic, hydrophilic, anionic, or cationic) of poly(ethyleneterephthalate) (PET) biomaterial. Our analyses revealed that different SMEs variably influence the ability of C. albicans to form biofilm. Moreover, these investigations identified an SME (6PEO) which, when added to Elasthane 80A (E80A) biomaterial, significantly reduced the ability of C. albicans to form biofilm. Our data demonstrate that surface modifications of biomaterials can be a potentially useful approach to reduce or manage biofilm formation on medical devices. The identified surfaces may have utility in manufacturing new indwelling medical devices that prevent biofilm formation, thereby reducing associated infections.
C. albicans strain SC5314, a clinical isolate obtained from a candidiasis patient, was a generous gift from William Fonzi (Georgetown University). This fungal strain was grown overnight at 37°C in yeast nitrogen base medium (Difco Laboratories, Detroit, MI) supplemented with 50 mM glucose. Cells were harvested, washed with phosphate-buffered saline (PBS), and standardized to 1 × 107 cells/ml.
Tables Tables11 and and22 summarize the biomaterials (15-mm-diameter disks each) used in this study: (i) polyetherurethane and polycarbonateurethane materials with their different SMEs (provided by the Polymer Technology Group, Berkeley, CA) and (ii) poly(ethyleneterephthalate) and its surface modifications (referred to here as Matsuda materials; provided by T. Matsuda, Kyushu University, Japan).
The materials supplied by the Polymer Technology Group consisted of polyetherurethanes and polycarbonateurethanes complemented with different SMEs. All polyurethane biomaterials used in this study were copolymers of two or more chemically distinct polymer blocks. These block copolymers contained a hard segment and a soft segment (see Table Table1for1for details). E80A is a polyetherurethane composed of a 4,4′-methylene bisphenyl diisocyanate (MDI) hard segment and a polytetramethyleneoxide (PTMO) soft segment. The E80A surface was modified by addition of (i) 6PEO, (ii) 6FC, or (iii) silicone with an MDI hard segment and a soft segment of PTMO and polydimethylsiloxane (PDMS) (Pursil20 80A). Bionate 90A is a polycarbonateurethane with an MDI hard segment and a poly(1,6-hexyl-1,2-ethyl carbonate) diol (PHECD) soft segment; Carbosil 90A is a silicone polycarbonateurethane made of an MDI hard segment and a soft segment of PDMS and an aliphatic, hydroxyl-terminated polycarbonate. Both Bionate 90A and Carbosil 90A have silicone SMEs covalently bonded to the base polymer. Polymer films were made by drying the resins for 16 to 18 h in a forced-air dryer at 65.6°C (150°F) prior to extrusion. All materials were extruded as described previously (4, 20). The extruded films were cut into disks, cleaned with ethanol, and used for biofilm formation.
As shown in Table Table2,2, the Matsuda materials consisted of poly(ethyleneterephthalate) (PET) and its chemical modifications (hydrophobic, hydrophilic, anionic, and cationic surfaces). The hydrophobic surface was PET coated with poly(benzyl-N,N-diethyldithiocarbamate-costyrene) (BDEDTC),and the hydrophilic surface was BDEDTC-coated PET polymerized with polyacrylamide (PAAm). The cationic and anionic surfaces were BDEDTC-coated PET polymerized with the methiodide of poly(dimethylaminopropyl-acrylamide) (DMAPAAmMeI) or the sodium salt of poly(acrylic acid) (PAANa), respectively.
The contact angle, determined using a goniometer, is commonly used to evaluate the hydrophilicity (wettability) of substrates (4, 20) and is determined by measuring the angle between the tangent to the surface of a liquid droplet made with the surface of the solid sample.
Biofilms were formed on different substrates by using the denture-based in vitro model described previously by our group (7). Briefly, biomaterial disks (diameter, 15 mm) were precoated with pooled, unstimulated whole saliva from healthy volunteers (collected according to a protocol approved by the Institutional Review Board, Case Western Reserve University [IRB protocol no. 12-04-28]) for 90 min, and a standardized C. albicans cell suspension (1 × 107 cells/ml) was applied to the surfaces placed in a tissue culture plate. The cells were allowed to adhere for 90 min at 37°C and then submerged in yeast nitrogen base medium supplemented with 50 mM glucose for 48 h to allow biofilm formation.
The ability of C. albicans to form biofilm on different substrates was quantitated by determining the metabolic activity and dry weight of the formed biofilm, as described previously (7). Briefly, for metabolic activity determination, biomaterial disks with biofilm were transferred to fresh tissue culture plates containing XTT [2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide; 12.5 μg/ml] and menadione (1 μM) in 4 ml PBS, and the XTT formazan product in the supernatant was measured spectrophotometrically. For dry weight measurement, biofilms were scraped from each disk in 4 ml PBS and transferred to a centrifuge tube. The scraped biofilm suspension was filtered through a preweighed filter (0.22-μm-diameter pore size), washed with PBS, dried at 37°C for 24 to 48 h, and weighed to determine the total biomass. Candida biofilms were visualized by confocal scanning laser microscopy (CSLM) as described previously (6), using a combination of the fluorescent dyes FUN-1 and concanavalin A-Alexa Fluor 488 conjugate (Molecular Probes, Inc., Eugene, OR).
Each experiment was performed three times, with triplicates each time. All statistical analyses (unpaired t test and correlation coefficient) were performed using StatView (version 4.5; Abacus Concepts Inc., Berkeley, CA), and a P value of <0.05 was considered statistically significant.
In this study, we determined whether the metabolic activity and total biomass of C. albicans biofilms, as determined by reduction of XTT dye and dry weight measurement, respectively, are affected by different surface modifications.
We found that metabolic activity of C. albicans biofilm formed on E80A-6FC or E80A-silicone did not differ significantly from that of biofilm formed on the nonmodified E80A surface (mean optical densities [ODs] ± standard deviations [SDs] of 0.416 ± 0.148 and 0.315 ± 0.184, respectively; P ≥ 0.571) (Table (Table3).3). However, we observed clusters of C. albicans cells on the 6PEO-E80A surface which did not mature into biofilm. These clusters quickly detached from the 6PEO-E80A surface while the XTT and biomass assays were performed. XTT analyses revealed a minimal metabolic activity compared to that of biofilm formed on the nonmodified E80A surface (mean ODs ± SDs of 0.054 ± 0.02 and 0.240 ± 0.103 for E80A-6PEO and E80A, respectively; P = 0.037) (Table (Table3),3), suggesting that only remnant C. albicans cells were present on the PEO surface. For the polycarbonateurethane materials, we did not detect significant differences in metabolic activity between biofilm formed on nonmodified (Bionate 90A) and silicone-modified (Carbosil 90A) substrate surfaces (mean ODs ±SDs of 0.476 ± 0.032 and 0.456 ± 0.128, respectively; P = 0.799) (Table (Table33).
To investigate whether hydrophilic and anionic modifications also affect the ability of C. albicans to form biofilms, we determined the metabolic activities of biofilms formed on four modified PET materials. Our data showed that the metabolic activities of C. albicans biofilms formed on these surfaces did not differ significantly from those of biofilms formed on the nonmodified PET substrate (P ≥ 0.143 for all comparisons with respect to the nonmodified PET surface) (Table (Table33).
To determine whether a correlation between the observed increases in metabolic activity and biofilm mass exists among biofilms formed by C. albicans on different surface-modified substrates, we determined the dry weights of biofilms formed on these surfaces. Minimal dry weight was observed for C.albicans grown on 6PEO-modified E80A. The dry weight of C.albicans grown on 6PEO-E80A was 74% less than that of biofilm formed on the nonmodified surface (Table (Table3).3). The dry weights of biofilms formed on E80A-6FC, the silicone-modified E80A surface, silicone-modified polycarbonateurethane (Carbosil 90A), or anionic and hydrophobic PET surfaces were not statistically significantly different from the dry weight of biofilm formed on the nonmodified surface (P < 0.05, Table Table3).3). In contrast, the dry weight of C. albicans biofilms formed on cationic, anionic, and hydrophilic PET surfaces were significantly higher than that of biofilm formed on the nonmodified PET surface (Table (Table33).
The metabolic activities of biofilms formed by C. albicans on E80A and polycarbonateurethane surfaces correlated well with the corresponding dry weight values (Pearson correlation coefficients of 0.679 and 1.0, respectively). A similar correlation was observed between XTT metabolism and dry weight for biofilms formed on nonmodified, anionic, and hydrophobic PET surfaces (Pearson correlation coefficient of 0.867) but not for those formed on the cationic and hydrophilic PET surfaces. Taken together, our data showed that the only substrate modification that inhibited biofilm formation by C. albicans was the addition of the 6PEO SME.
Since biofilm formation involves adhesion to a surface as the first step and because contact angle is an important factor in adhesion to surfaces, we hypothesized that the ability of these surfaces to promote biofilm formation is mediated by changes in contact angle. To test this hypothesis, we determined the correlation between the contact angles of surfaces and the amounts of biofilm formed on them. Water contact angles of the biomaterials tested in our studies are provided in Tables Tables11 and and2.2. We found that while increasing contact angle correlated with metabolic activity (r = 0.88) for biofilms formed on polyetherurethane surfaces, there was an inverse correlation between the two variables for biofilm formed on polycarbonate surfaces (r = −1). In contrast, our analyses did not reveal a significant correlation between contact angle and metabolic activity for biofilm formed on PET biomaterials (r = −0.19). Low water contact angle values indicate a hydrophilic surface, whereas high water contact angles are indicative of a hydrophobic surface. Our studies showed that the correlation between contact angle and biofilm formation was surface dependent. The biomaterials tested in our studies are soft materials and not hard dental material, where surface roughness is a factor. Therefore, contact angle in this case is not a reflection of surface smoothness.
Figure 1A to D shows three-dimensional reconstructions of images of C. albicans biofilm formed on the polyetherurethane (E80A) surface and its modifications. The biofilm formed by C. albicans on nonmodified E80A surface was ~310 μm thick (Fig. (Fig.1A),1A), while there was no detectable biofilm formation by C. albicans on the E80A-6PEO surface (Fig. (Fig.1B).1B). Biofilms formed on E80A-6FC were 70 μm thick (Fig. (Fig.1C),1C), while those formed on E80A-silicone were 220 μm (Fig. (Fig.1D).1D). The thickness of biofilm formed by C. albicans on Bionate 90A was similar to that of biofilm formed on Carbosil 90A surface (data not shown).
CSLM analysis revealed no substantial difference in the thicknesses of C. albicans biofilms formed on the different PET surfaces, although some differences in architecture were noted. In this regard, biofilms formed by C. albicans on the nonmodified PET surface had a thickness (90 μm) similar to those of biofilms formed on chemically modified PET surfaces (Fig. 1E to I). The thicknesses of the basal layer of biofilm formed on anionic and cationic surfaces (85 μm and 123 μm, respectively) were also similar to that of biofilm formed on the nonmodified PET surface (90 μm). Interestingly, C. albicans biofilms formed on the nonmodified and nonionic (hydrophilic or hydrophobic) surfaces contained numerous hyphal elements throughout their structure (Fig. 1F and G), while biofilms formed on ionic (anionic and cationic) surfaces had protruding hyphae that were observed in side views of these biofilms (data not shown). These studies showed that C. albicans biofilms were formed on all the PET surfaces tested, with some differences in architecture.
Overall, the different surface modifications studied influenced the ability of C. albicans to form biofilm to various extents. However, the addition of the 6PEO SME was the only surface modification that resulted in no biofilm formation.
The purpose of this study was to identify surface modifications that can prevent or inhibit Candida biofilm formation on medical devices. We compared the ability of C. albicans to form biofilms on surfaces modified by different SMEs or chemical modifications. We found that of all the surface modifications tested, the 6PEO SME was the only one that significantly inhibited Candida biofilm formation.
Surface modifications, including addition of SMEs and chemical modifications, have been previously tested for their ability to allow bacterial adhesion and biofilm formation on dental and catheter materials (10, 11, 17, 18). Olsson et al. (18) investigated protein adsorption and saliva-mediated bacterial adherence on untreated, hydrophobic surfaces and on polyethylene oxide-coated glass and ceramic crown surfaces. These investigators showed that hydrophobic and PEO-treated surfaces exhibited much lower (or no) colonization and pellicle and plaque formation. Dijk et al. (8) reported similar results, demonstrating that biofilm formation by C. albicans and Candida tropicalis on a silicone rubber voice prosthesis treated with a colloidal palladium/tin solution (resulting in a thin metal coat) was significantly less than that on untreated prostheses. Nikawa et al. (17) showed that fluoric and heat-cured silicone denture lining materials promote the lowest colonization by C.albicans. In this study, we identified surface modifications that can eliminate or reduce the ability of C. albicans to form biofilm on biomaterial surfaces.
We demonstrated that the ability of C. albicans to form biofilm is influenced by the surface chemistries of the biomaterials used. Although all the surface modifications affected the ability of C. albicans to form biofilms, only one modification, the addition of the 6PEO SME, reduced the ability of C.albicans to form biofilm on the E80A surface. The metabolic activity and dry weight of C. albicans cells adhered to E80A-6PEO were reduced by 78% and 74%, respectively (compared to those of biofilm formed on nonmodified E80A). The low dry weight of C. albicans biofilm formed on 6PEO correlates well with the reduction in metabolic activity of biofilm formed on the same material. Moreover, when using CSLM, we were unable to detect any biofilm formation by C.albicans on the E80A-6PEO surface. Since very low, yet noticeable, metabolic activity and dry weight were observed for biofilm formed on E80A-6PEO, the low XTT values as well as inability of CSLM to detect biofilm on this surface may be because fungal cells did not adhere strongly to the modified biomaterial and were likely detached during manipulations performed for confocal analysis.
We observed statistically significant increases in the dry weights of biofilms formed on cationic and hydrophilic PET surfaces compared to those of biofilm formed on the nonmodified PET surface. In a different study, Brodbeck et al. (4) showed that adhesion of monocytes/macrophages to hydrophilic and anionic PET substrates is greatly reduced compared to that with nonmodified PET surfaces. Since the cell surface hydrophobicity of Candida cells is known to be an important factor in its adherence to acrylic surfaces (19), this factor may also play a role in the formation of biofilm on PET surfaces.
We employed XTT assay and total biomass determination to show that PEO-modified polyetheruerthane significantly inhibited biofilm formation by C. albicans. Using these established methods, it is possible to demonstrate a correlation between XTT and dry weight analyses. We found that metabolic activity and total biomass correlated well for Candida biofilms formed on the polyurethanes and two (anionic and hydrophobic) PET surfaces but not for those formed on cationic and hydrophilic PET surfaces. In a previous publication, Kuhn et al. (16) showed that the relationship between XTT colorimetric signal and organism number is not always linear. This observation can explain the apparent lack of correlation between XTT activity and dry weight of biofilms formed on some surfaces, since the number of cells contributes significantly to the total biomass of biofilms. Kuhn et al. (16) also showed that while the XTT formazan product readily appears in solution, there can be a significant amount of retained intracellular product, which becomes soluble only after cell treatment with dimethyl sulfoxide. Moreover, the retention of residual formazan product varies with different species of Candida; such variation may also be induced by the different substrates tested in this study. Since total biomass values are independent of formazan product formation, a lack of correlation between the metabolic activity and total biomass of biofilms formed on some surfaces is not entirely unexpected.
The mechanisms underlying the inability of C. albicans to form biofilm on 6PEO-modified surfaces is unknown. The role of different surface chemistries in microbial adhesion and biofilm formation has been proposed to be a complex interplay between different microbial and host factors, including thermal cycling (17), protein coating (saliva or serum) (17, 18), hydrophobicity (18), and nonspecific physiochemical forces or specific ligand-receptor interactions (20).
Our studies revealed that while more biofilm was formed on polyetherurethane surfaces with higher contact angles, no correlation was seen for PET biomaterials, indicating that the influence of contact angles on biofilm formation is surface dependent. Our results are in agreement with previous studies showing that the correlation between contact angle and biofilm formation is dependent on the surface tested (1-3, 12-14, 22). Jansen and Kohnen (12) investigated the influence of contact angles of modified polymers on adherence of Staphylococcus epidermidis and showed a correlation between contact angle measurements and adherence. Similar correlations between contact angle and biofilm formation were noted for amalgam and resin composites (22), poly(vinyl chloride) (PVC) catheters (25), and glass (1). However, other investigators have demonstrated no correlation between contact angle and biofilm formation. In this regard, Jones et al. (15) showed that no correlation existed between contact angle and biofilm formation on PET and high-density polyethylene bottles. Balazs et al. (2) reported complete inhibition of adhesion and colonization by Pseudomonas aeruginosa of PVC with an ultrahydrophobic surface (contact angle o >120 degrees). In a separate study, Webb et al. (24) demonstrated a positive correlation between adhesion of the fungus Aureobasidium pullulans and the contact angle of unplasticized PVC. However, plasticized PVC, with a relative reduction of 13° in contact angle (compared to that of unplasticized PVC), showed enhanced adhesion and biofilm formation. Taken together, these studies clearly demonstrate that the correlation between contact angle and biofilm formation is dependent on the substrate surface.
Patel et al. (20) showed that the adhesion of S. epidermidis to hydrophobic surfaces is greater than the adhesion of this bacterial species to hydrophilic surfaces (e.g., 6PEO). These investigators used a rotating disk model to characterize adhesion of leukocytes and S. epidermidis on polycarbonateurethanes and polyetherurethanes modified with SMEs (6PEO, 6FC, and silicone) under dynamic flow conditions. These studies showed that modification of materials with polydimethylsiloxane and PEO SMEs reduces bacterial adhesion, while fluorocarbon SMEs enhance adhesion. A different effect of 6PEO on bacterial adhesion to glass and ceramic surfaces was demonstrated by Olsson et al. (18). These investigators determined the effect of immobilized PEO on protein adsorption and bacterial adherence in vitro to glass and in vivo to ceramic crown surfaces. In vitro, more protein and bacteria bound to untreated glass than to hydrophobic and PEO-treated glass. On the other hand, in vivo, pellicle and plaque formation was similar on the untreated ceramic and PEO surfaces, but less plaque formed on these surfaces than on adjacent normal tooth surfaces. Almost no plaque accumulated on the hydrophobic crown surface, and it was virtually devoid of stainable pellicle. The different effects of 6PEO SME on bacterial adherence to materials may be due to the different substrates used in these studies and suggested that the effect of the 6PEO SME is dependent on the substrate material to which this SME is applied.
In the only study focusing on fungal adhesion to substrates with modified surfaces, Nikawa et al. (17) investigated the growth of C. albicans on seven saliva-coated, serum-coated, or protein-free (uncoated) thermocycled commercial soft lining materials. For control resilient liners (not thermocycled and uncoated), fungal colonization was found to be dependent on the type of commercial resilient liner used; fluoric and heat-cured silicone materials promoted the lowest colonization, while cold-cured silicone materials and heat-cured acrylic resin exhibited the highest colonization capacity. Thermal cycling and protein coating (saliva or serum) significantly promoted fungal colonization on the materials. These investigators suggested that aging of the materials and the presence of host biological fluids promote yeast colonization on denture lining materials.
Based on the above discussion, it is likely that the mechanism(s) responsible for inhibition of the ability of C. albicans to form intact biofilm on 6PEO-modified surfaces is multifactorial and calls for further investigations which are beyond the scope of this study. It is possible that this inhibition is mediated by the prevention of Candida adhesion to the substrate surface.
In conclusion, we identified the 6PEO SME as a surface-modifying agent that inhibits C. albicans biofilm formation. Examination of the ability of 6PEO to inhibit biofilm formation in vivo is warranted. The results reported in this study may have important clinical implications in the design of novel biomaterials that have antibiofilm properties.
Funding from the National Institutes of Health (grant 1R01 DE13932-01A1) and a Bristol-Myers Squibb Freedom to Discover Award to M.A.G., from the American Heart Association (Scientist Development grant 0335313N) and the NIH/NIAMS/Skin Diseases Research Center (grant P30-AR-39750) to P.K.M., from an Infection Grant (NIH/NIBIB EB 00279) to J.M.A., and from the CSLM core facility (NCI grant P30CA43703-12) is gratefully acknowledged.
We thank T. Matsuda of Kyushu University, Japan, and R. Ward of the Polymer Technology Group, Berkeley, Calif., for providing the materials used in this study.