|Home | About | Journals | Submit | Contact Us | Français|
The most common form of oral candidiasis, denture-associated stomatitis, involves biofilm growth on an oral prosthetic surface. Cells in this unique environment are equipped to withstand host defenses and survive antifungal therapy. Studies of the biofilm process on dentures have primarily been limited to in vitro models. We developed a rodent acrylic denture model and characterized the Candida albicans and mixed oral bacterial flora biofilm formation, architecture, and drug resistance in vivo, using time course quantitative culture experiments, confocal microscopy, scanning electron microscopy, and antifungal susceptibility assays. We also examined the utility of the model for measurement of C. albicans gene expression and tested the impact of a specific gene product (Bcr1p) on biofilm formation. Finally, we assessed the mucosal host response to the denture biofilm and found the mucosal histopathology to be consistent with that of acute human denture stomatitis, demonstrating fungal invasion and neutrophil infiltration. This current oral denture model mimics human denture stomatitis and should be useful for testing the impact of gene disruption on biofilm formation, studying the impact of anti-infectives, examining the biology of mixed Candida-oral bacterial flora biofilm infections, and characterizing the host immunologic response to this disease process.
Candida spp. frequently establish biofilm lifestyles, proliferating as multicellular communities attached to surfaces, such as a medical device (21, 31). Adherent and encased in a polymeric extracellular matrix, biofilm cells are phenotypically distinct from free-floating planktonic cells. Not only are these organisms capable of proliferating in healthy hosts by surviving immune factors, they demonstrate increased resistance to commonly used antifungal drug therapies (1, 35, 39, 55). The majority of Candida biofilm infections are associated with placement of medical devices, such as a dentures, venous catheters, or urinary catheters (29). Treatment of these infections remains difficult due to the extraordinary drug resistance exhibited during biofilm growth. Although often costly and inconvenient, device removal is typically required for eradication of infection (48).
Oral candidiasis commonly occurs in those who are colonized with Candida and have risk factors for development of disease, such as HIV infection, chemotherapy treatment, or antibiotic exposure (63, 67). Local breaks in mucosa caused by mucositis, radiation, and trauma are also predisposing factors. Oral candidiasis may occur as an acute or chronic illness and can involve the tongue, buccal mucosa, or palate. The most common form of oral Candida infection, denture stomatitis, is prevalent in the elderly population and affects up to 70% of denture wearers (58, 66, 70). The condition begins with Candida biofilm growth on the denture-mucosa interface. Biofilm growth progresses over the denture surface, leading to inflammation of the denture-exposed palatal mucosa (54, 67). Clinical manifestations range from no symptoms to severe pain and difficulty swallowing (70). While Candida albicans is responsible for the majority of oral candidiasis (70 to 80%), infection by Candida glabrata, Candida tropicalis, and Candida parapsilosis has been reported less frequently (20, 63).
In vitro biofilm models have the advantage of lower cost and are ideal for high-throughput screening assays and antifungal testing (12, 56, 61, 71). These models have provided important phenotypic information regarding drug resistance in Candida biofilms. The models have also been valuable for characterizing the role of surface adherence properties, such as topography and hydrophobicity. However, it is difficult for in vitro models to account for all the factors which likely play a role during biofilm infection (38). For example, cells in the in vitro systems are not exposed to salivary immune components and proteins which may precondition, or coat, the surface and promote adherence. Although several of the host components which play a key role in promoting Candida adherence have been defined, the interaction is likely a complex phenomenon (42). In the denture biofilm niche, additional components hypothesized to impact this growth state include companion oral bacterial flora, contact between the device biofilm and the host mucosal surface, and salivary flow dynamics. Each of these conditions would be challenging to mimic in vitro.
Several in vivo models have been developed for study of oral candidiasis and denture stomatitis (58). The earliest studies utilizing acrylic plates fitted to the palates of primates found that Candida infection of these prostheses mimicked the histopathology of human denture stomatitis (8, 9). However, the cost and difficulty of maintaining primates prevented widespread use. An alternative model utilized rats with molded acrylic plates (47). Advantages of the rat model include reduced size and cost. The rat mucosal surface is also relatively easy to colonize with Candida, although maintenance of colonization often requires antibiotic administration (58). The rat models have been integral for describing the basic pathologies of oral candidiasis and denture stomatitis. The investigations have primarily been directed toward the host response, and in-depth examination of the Candida biofilm phenotype has not been undertaken.
Here we adapt and test a rat model for Candida-associated denture stomatitis, with a focus on Candida biofilm formation. We chose to use a common acrylic denture material to form an oral device directly in the animal. One major advantage of the current device placement is the ability to form the denture without the costly and time-consuming ex vivo mold process. The model mimics denture stomatitis with respect to Candida biofilm formation and histology. The experimental system proved useful for testing of the impact of gene products and measurement of the response to treatment. It also demonstrated the value for study of mixed microbial biofilm interactions, comparisons with other models of biofilm niche infection, such as a venous catheter model, and examination of the host response to biofilm.
Specific-pathogen-free male Sprague-Dawley rats at weights of 350 g (Harlan Sprague-Dawley, Indianapolis, IN) were used for all studies. Animals were maintained in accordance with the American Association for Accreditation of Laboratory Care criteria, and all studies were approved by the Institutional animal Care Committee. Rats were housed individually in an environmentally controlled room in metabolic cages to avoid oral contamination with bedding and animal waste. Animals were fed an Ensure liquid diet while dentures were in place to avoid contamination of biofilms with food particles. The animals were examined for signs of distress at 6 h after surgery and then every 12 h throughout the study. The oral mucosal area was examined twice daily for signs of inflammation or purulence.
Prior to placement of denture appliances, rats were immunosuppressed with a single dose of cortisone (200 mg/kg subcutaneously) on the day of infection to mimic a protocol commonly utilized for rodent oral candidiasis (49). For a subset of experiments, either ampicillin sodium-sulbactam sodium at 100 mg intraperitoneally (i.p.) twice a day (BID) or aztreonam at 50 mg/kg subcutaneously BID was administered 48 h prior to denture placement to reduce enteric bacterial colonization common to rodents due to coprophagy, or the consumption of feces. Dosing regimens were chosen based upon those previously shown to be effective in the treatment of rodent systemic bacterial infections (D. R. Andes, unpublished data).
Initial studies of the metabolic cages demonstrated that the rats continued to practice coprophagy. To address this, we modified a tube caging procedure that has been utilized in nutritional absorption studies (15). Specifically, an additional group of animals were housed in polyvinyl chloride (PVC) plastic tube caging in which the ability to turn and reach fecal material was eliminated. Animal monitoring during this additional housing restraint was increased to every 6 h throughout the study.
A 32-gauge stainless steel Babcock orthodontic wire (Miltex) was threaded across the hard palate and secured between cheek teeth of an anesthetized rat to serve as an anchor for the denture material (Fig. (Fig.1).1). Cheek teeth were etched using Uni-Etch 32% semigel etchant, with benzalkonium chloride (Bisco, Inc., Schaumburg, IL) for microretention. A metal spatula was placed over the hard palate to create an approximately 3- by 5-mm area across the palate and less than 1 mm of space between the hard palate and the denture material. This space was used for Candida inoculation. Cold-cure acrylic temporary crown and bridge material (Maxitemp HP; Henry Schein, Melville, NY) was applied over the cheek teeth and wire to create a surface approximately 8 by 10 mm in size and approximately 2-mm thick. Material was rapidly shaped with a small spatula to produce a rectangular denture, which was allowed to solidify for 5 min, after which the spatula was removed.
Candida albicans strains DAY185 and CJN702 (bcr1Δ/bcr1Δ) were used for all studies (17, 43, 44). The strains were stored in 15% (vol/vol) glycerol stock at −80°C and maintained on yeast extract-peptone-dextrose (YPD) medium plus uridine (1% yeast extract, 2% peptone, 2% dextrose, and 80 μg/ml uridine) prior to experiments. Prior to denture inoculation, cells were grown at 30°C in YPD plus uridine liquid media with orbital shaking at 200 RPM overnight. To prepare the inoculum, cells were washed with phosphate-buffered saline (PBS), enumerated by hemocytometer counting, and resuspended in saline at 108 cells/ml. The final inoculum concentration was confirmed by microbiologic enumeration.
After removal of the spatula, the hard palate beneath the acrylic device was inoculated with 108 C. albicans cells/ml (0.1 ml). Animals were sacrificed after 6, 24, 48, or 72 h for histopathology or device removal and study. For treatment studies, antifungals were applied to mature biofilms after 48 h of biofilm growth.
To determine the viable burden of Candida and the accompanying oral flora, microbiologic counting was performed on dentures. Devices were placed in 2 ml 0.15 M NaCl, sonicated for 10 min, and vortexed. For Candida microbiology, 1:10 dilutions were plated on Sabouraud dextrose agar (SAB) plus 2.5 μg/ml chloramphenicol. For determination of aerobic bacteria, dilutions were plated on Mueller-Hinton agar plus 2.5 μg/ml amphotericin B. For anaerobic cultures, specimens were passed into an anaerobic chamber (80% N2, 10% H2, and 10% CO2; Coy Laboratory Products, Grass Lake, MI), where dilutions were made in prereduced anaerobically sterilized PY (peptone yeast extract; Anaerobe Systems, Morgan Hill, CA) broth and placed on reduced plates of LKV (laked blood with kanamycin and vancomycin), PEA (phenylethyl alcohol agar), and EYA (egg yolk agar) that were incubated in the chamber at 35°C for 5 days (media from Remel, Lenexa, KS). Dilutions were passed out of the chamber, plated, and incubated in 5 to 7% CO2 for 1 day. Organisms were identified by Gram stain morphology and rapid testing, e.g., catalase, coagulase, PYR (pyrrolidonyl aminopeptidase), and streptococcal grouping (13). Experiments were performed on two or three separate occasions.
Acrylic devices were processed for scanning electron microscopy (SEM) as previously described for venous catheters (3). After removal of each denture, it was washed with phosphate-buffered saline (PBS) and placed in fixative (1% [vol/vol] glutaraldehyde and 4% [vol/vol] formaldehyde in PBS) overnight. The samples were rinsed with PBS, placed in 1% osmium tetroxide for 30 min, and rinsed with PBS. The samples were subsequently dehydrated in a series of ethanol washes (30% for 10 min, 50% for 10 min, 70% for 10 min, 95% for 10 min, and 100% for 10 min). Final desiccation was accomplished by critical-point drying (Tousimis, Rockville, MD). Specimens were mounted on aluminum stubs and sputter coated with gold. Dentures were imaged on a Jeol 6100 at 10 kV. The images were processed for display using Adobe Photoshop.
For confocal microscopy, denture biofilms were removed, gently rinsed with PBS, and visualized using one of two stain sets. For imaging of live fungal elements, biofilm architecture, and the extracellular matrix, dentures were stained with the FUN-1 (50 μM) component of the Live/Dead yeast viability kit (Molecular Probes, Eugene, OR) and concanavalin A (ConA) conjugate (Alexa Fluor 488 conjugate, 200 mM), according to the manufacturer's instructions. Briefly, the device was incubated with each dye in PBS with 2% glucose for 30 min at 30°C. Both live and dead yeast cells were stained with the FUN-1 dye and visualized with a diffusely distributed yellow-green fluorescence. Metabolically active cells process this dye, which results in a shift to red fluorescence within intravacuolar structures. ConA binds to glucose and mannose residues of cell wall polysaccharides and is visualized as green fluorescence. Colocalization of fluorescent dyes (FUN-1 and ConA) in metabolically active cells results in a yellow to orange fluorescence during multichannel image capture. The biofilms were imaged on an Axiovert 200 with excitation and emission filters at 490/617 nm and 490/528 nm for FUN-1 and ConA, respectively, as previously described for venous catheters (3).
The second dye set was selected for simultaneous imaging of yeast, hyphae, and bacteria within the mixed-species denture biofilm. This set included SYTO 9 dye (Molecular Probes, Eugene, OR) to visualize bacteria in the mixed biofilms, as previously described (23). Calcofluor white was selected for its ability to bind to Candida hyphal cell walls, and FUN-1 was chosen for imaging of yeast. Dentures were stained with calcofluor white (22.5 μg/ml), SYTO 9 (5 μM), and FUN-1 (50 μM) in PBS with 2% glucose for 30 min at 30°C and imaged on Axiovert 200 with excitation and emission filters at 360/457 nm, 490/528, and 490/617 nm, respectively. Images were obtained at 63× or 100× magnification and compiled using Adobe Photoshop.
To discern if the model mimicked denture stomatitis, we performed mucosal biopsy and histopathology at the end of the study period (48 h). After animals were sacrificed and the dentures were removed, the tongues and palates were dissected. The tissues were fixed in 10% buffered formalin and embedded in paraffin (26). Sections were stained with either hematoxylin and eosin (H&E) or Gomori's methenamine silver (GMS) for C. albicans. Images were obtained at 200× magnification.
Previous studies have demonstrated biofilm recalcitrance to antifungal therapy. We included both topical and systemic antifungal treatment with a triazole, fluconazole, and an echinocandin, micafungin, to determine the susceptibility of the denture biofilm to these therapies. For topical therapy, regimens included a 0.15 M NaCl control, topical administration of micafungin at 5 mg/kg daily, or topical administration of fluconazole at 16 mg/kg twice daily. Therapy was applied directly to mature (48-h-old) biofilms. Systemic drug treatments included micafungin at 5 mg/kg intraperitoneally once daily or fluconazole at 16 mg/kg subcutaneously twice daily. The dosing regimens were chosen based on maximal efficacy in treatment of oral candidiasis in animal models (27, 51). All therapies were continued for 48 h, after which devices were removed and processed for viable burden determination as described above.
Ex vivo antifungal efficacy was also assessed on biofilm cells that were detached from the biofilm and resuspended, as described previously for biofilm cells from a central venous catheter (3). Devices were removed at 48 h and placed in PBS. Organisms were dislodged by scraping the surface with a plastic loop, vortexing, and water bath sonicating as described above. Following biofilm removal from the biofilm, cells were resuspended for antifungal susceptibility testing to amphotericin B, micafungin, and fluconazole using NCCLS methods (37). Comparisons were made to susceptibility testing with planktonic cells using the standard NCCLS protocol and identical inocula (103 CFU/ml).
Biofilm cells were dislodged from dentures by scraping, vortexing, and sonicating. Cells were flash frozen with liquid nitrogen in AE buffer (50 mM sodium acetate at pH 5.2, 10 mM EDTA), and RNA was collected using the hot method as previously described (3, 40). The RNA integrity was assessed using an Agilent Bioanalyzer 2100 with an RNA Nano chip before use.
Quantitative real-time reverse transcription-PCR (RT-PCR) was used to quantify the mRNA abundances of a subset of transcripts (CDR1, CDR2, ERG25, ALS1, SAP5, GDB1, and BGL2). Transcripts for study were chosen based on differential expression in the central venous catheter biofilm model and included those involved in drug resistance, biofilm adhesion, and glucan modification (40, 45). The TaqMan probe and primer sets were designed using Primer Express (Applied Biosystems, Foster City, CA) (Table (Table1).1). The QuantiTect probe RT-PCR kit (Qiagen, Valencia, CA) was used in an ABI Prism 7700 version 1.7 sequence detection system (Applied Biosystems) as previously described (3, 40). Reactions were performed in triplicate and analyzed using the threshold cycle (2−ΔΔCT) method (34). The comparative expression method generated data as transcript fold changes normalized to a constitutive reference gene transcript (ACT1) and relative to log-phase planktonic cells grown in RPMI at 37°C on an orbital shaker set to 200 RPM.
Following oral device placement, the animals did not appear ill throughout the course of the experiments (72 h) and continued normal intake of the food and water. The oral cavity did not appear visibly inflamed, and no purulence was noted.
Time course experiments showing viable counts from the dentures demonstrated an increasing viable burden of both Candida spp. and bacteria from 6 to 48 h after denture placement and inoculation (Fig. (Fig.22 A). At the 6-h time point, <2 log10 CFU/device of viable Candida spp. and bacteria was adherent to the device surface. By the final time point (48 h), there were approximately 6 log10 CFU/device of viable Candida spp. and 8 log10 CFU/device of viable bacteria present on each denture, based on microbiologic plate counts. Classification of bacterial organisms suggested the presence of a polymicrobial biofilm consisting of both aerobic and anaerobic organisms (Fig. (Fig.2B).2B). The aerobic Gram-positive coccus species included viridans group Streptococcus spp., Enterococcus spp., and Streptococcus agalactiae. In early studies, Escherichia coli was a prominent species identified. We felt this was likely related to the rodent practice of coprophagy. We attempted to address enteric oral colonization in two ways. First, we administered a broad-spectrum antibiotic (ampicillin-sulbactam) prior to denture placement. Antibiotic treatment resulted in a 10-fold decrease in the number of biofilm-associated bacteria and a 100-fold increase in C. albicans bacteria (Fig. (Fig.2C).2C). For the next experiment, we housed the animals so that coprophagy was not possible and administered an antibacterial agent with a narrow aerobic Gram-negative bacterial spectrum of activity, aztreonam. Both conditions were undertaken 2 days prior to denture placement, and the housing continued throughout the denture biofilm study. These conditions resulted in a >1,000-fold reduction in aerobic Gram-negative rods (3.5 log10 CFU/device compared to 7 log10 CFU/device) (data not shown).
We used SEM to assess biofilm formation and architecture over the 48-h period following denture placement and inoculation. The choice of the imaging time course experiment was based upon the identification of architecturally early and mature biofilms in previous studies (3, 11, 24, 33, 55, 56). Imaging of the denture demonstrated adherent yeast cells after 6 h (Fig. (Fig.3).3). After 24 h, a confluent layer of yeast cells had covered the denture surface, and an extracellular matrix became apparent. Over the course of 48 h, a mature biofilm composed of two fungal morphologies (yeast and hyphae) and host cells encased in extracellular matrix was visualized. In addition, bacteria were embedded in the biofilm. These bacilli and cocci were found growing in groups associated with Candida biofilm cells and matrix material.
Two sets of fluorescent dyes were used to examine biofilm formation on the denture surfaces. First, imaging with the Live/Dead FUN-1 dye (red) confirmed the presence of viable yeast in a mature biofilm attached to the denture surface after 48 h (Fig. (Fig.44 A). The extracellular matrix surrounding the cells was visualized using ConA carbohydrate staining and appears green in the image capture.
The second set of dyes attempted to image both the Candida spp. and associated bacteria simultaneously. We utilized SYTO 9 staining to image the bacteria attached to the denture and embedded in the biofilm. The utility of the combination of SYTO 9 and calcofluor white in the imaging of mixed-species biofilms has previously been demonstrated for Candida and Staphylococcus aureus biofilms (23). In addition to these dyes, we included FUN-1 to visualize live yeast cells. In the multichannel image, bacteria appear yellow, and the predominant form was coccus (Fig. (Fig.4B).4B). The clusters were found attached to calcofluor white-stained hyphae and yeast cells, which appear blue.
We next sought to test the ability of the model to detect the phenotype of a Candida strain previously shown to produce poor biofilms. We chose the bcr1Δ/bcr1Δ strain, which lacks a zinc finger transcription factor and produces minimal biofilms in vitro and in a central venous catheter in vivo biofilm model (43, 44). We hypothesized that the mutant would also exhibit a biofilm defect in the rat denture model. Compared to an otherwise isogenic reference strain, the bcr1Δ/bcr1Δ infection of the rat denture resulted in 4-fold-fewer adherent cells (bcr1Δ/bcr1Δ, 2 × 105 Candida CFU/device; reference strain, 8.7 × 105 Candida CFU/device). Interestingly, a 50%-higher burden of bacteria was found within the bcr1Δ/bcr1Δ biofilm (1.5 × 103) compared to that associated with the reference strain biofilm (1 × 103).
Similar to reports from in vitro biofilm studies and the central venous catheter in vivo biofilm model, the triazole, fluconazole, did not effectively reduce the burden of viable Candida in this denture model (3, 6, 11, 28, 30, 32, 35, 53). The devices of control animals contained approximately 5.3 log10 CFU/device, while fluconazole treatment decreased the burden by less than 0.5 log10 CFU/device (to 4.5 and 4.8 log10 CFU/ml for topical and systemic therapies, respectively). Although previous in vitro and in vivo catheter biofilm studies have shown echinocandins to successfully treat these infections, both systemic and topical echinocandin therapies were fairly ineffective in the denture model. Over the 5.3 log10 CFU/device growth in control animals, treatment with 48 h of topical or systemic micafungin therapies reduced the viable Candida to 4.8 and 4.7 log10 CFU/device, respectively. The lack of effect of the systemic micafungin may be a result of reduced drug accumulation in the oral mucosa (69).
The phenotype of the biofilm-dispersed Candida observed was similar to that described using other biofilm models (3, 53). The suspension of biofilm cells remained viable in the presence of concentrations of amphotericin B, fluconazole, and micafungin that far exceeded the planktonic MIC (amphotericin B, 0.03 μg/ml; fluconazole, 0.5 μg/ml; micafungin, 0.015 μg/ml). The biofilm MIC, which was defined as the concentration necessary to produce a 50% reduction in the number of viable dislodged cells, was >256 μg/ml for each compound.
To test the ability to utilize the model for examination of Candida biofilm-associated gene expression, we compared the transcriptional profile of the denture-associated biofilm cells to that of free-floating planktonic cells. Over 1.5 μg of total RNA was isolated from a single denture. We chose a subset of genes which we had previously found to have increased expression during biofilm formation in an in vivo central venous catheter biofilm model relative to planktonic growth (40). For each of the selected genes, transcripts were more abundant for denture-associated biofilm cells than for planktonic cells (Table (Table2).2). While the venous catheter and denture expression studies were not undertaken at the same time, it is interesting that the fold change in expression of this subset of genes during biofilm growth varied by as much as 10-fold between the two in vivo biofilm models.
H&E staining of the rat oral mucosal surface demonstrated inflammation, marked by infiltration of polymorphonuclear cells in the keratinized and deeper layers of the epithelium (Fig. (Fig.5).5). GMS staining confirmed tissue invasion by Candida hyphae, consistent with the findings of pseudomembranous candidiasis. Hyphae were also observed in the superficial keratinized layer of the palate by GMS staining. The constellation of fungal invasion and neutrophilic inflammation is consistent with acute denture stomatitis (47, 58). Unlike models with prolonged infection, epithelial atrophy was not found (47).
Candida, one of the most common human fungal pathogens, adopts a biofilm lifestyle, often growing attached to a prosthesis or medical device (19, 22, 29). Many devices have been demonstrated to serve as substrates for biofilm growth and infection, including oral prostheses, catheters (venous or urinary), vascular stents, pacemakers, and joint implants (29). Advancing age and a number of pathological conditions of the head and neck often result in a decline in oral health and the need for prosthetic devices (70). Candida spp. and oral bacterial flora can grow as a biofilm on the surface of an oral device, resulting in denture stomatitis and often device malfunction. The device biofilm and associated mucosal inflammation negatively impact a patient's ability to eat and speak. In addition, these mixed-species biofilms harbor pathogenic bacterial organisms which commonly cause dental caries and less frequently cause more severe disease, such as pneumonia or endocarditis (14, 60). Biofilm infections in general, including those associated with dentures, are notoriously difficult to eradicate due to associated resistance to both host defenses and anti-infective therapy (12, 25, 55).
Here, we adapted a rat denture model to mimic human Candida denture stomatitis, the most common form of oral candidiasis (67). The model was designed to study Candida biofilm formation on the surface of a prosthetic oral device. The model mimics human denture stomatitis in terms of Candida biofilm formation, host factors, and anatomic location. The ecologic niche for denture stomatitis, the oral cavity, is similar between rats and humans in terms of salivary flow, host immune response, and host salivary proteins. The role of saliva in Candida adhesion is complex (18, 41, 42). Saliva has a physical cleansing effect and innate immune defense components, including lysozyme, histatin, lactoferrin, and IgA, which can act to decrease colonization and adherence (18, 50). However, saliva proteins, such as mucins, may also facilitate adherence by absorbing to Candida or coating the oral device. Full incorporation of the impact of the numerous and complex component makeup of saliva on biofilm formation is difficult to accomplish without use of an in vivo model.
Compared to other commonly used laboratory animals, we found several advantages to employing a rat for the denture stomatitis model. Unlike smaller animals, such as mice, rats have a larger oral cavity, easing denture placement. However, rats are much less expensive and easier to maintain than larger animals, such as primates (58). The current model has the advantage of not requiring a mold for manufacture of a preformed device (47, 58). With this technique, the device can be constructed directly in the oral cavity and has the benefit of a well-fitted apparatus. One limitation of this method of placement is a small amount of variability (less than 10%) in size among the devices.
Using this model, we found histopathologic changes to be consistent with candidiasis and acute denture stomatitis. Similar to previously described mucosal models, we observed hyphal invasion of the epithelial layer and inflammation marked by polymorphonuclear cells (47, 58). Unlike models with prolonged mucosal or device infection, epithelial atrophy was not identified, perhaps due to the relatively short duration of the current studies (47). Future investigation, including a prolonged time course experiment, may be of interest for examination of the histopathology of chronic denture stomatitis. Given the animal tolerance to the denture model, longer studies should be feasible. In addition, this model should be useful for examination of the host response to biofilm infection through histopathological comparisons of animals with mixed-species biofilm infections and those with single-species biofilm infections or by study of the response to genetically manipulated strains.
Microscopic examination of the denture surface demonstrated a C. albicans biofilm. The biofilm consisted of adherent yeast and hyphae embedded in an extracellular matrix. In addition, host cells and bacteria were identified in the biofilm. This biofilm architecture was similar to that described for a rat venous catheter biofilm (3). The main difference was the presence of a mixed microbial biofilm infection of the denture model. Confocal microscopy confirmed the presence of adherent bacteria on the surfaces of hyphae. These biofilm cells were resistant to several antifungal drug classes, a characteristic typical of C. albicans biofilm cells, described in the rat venous catheter model, and mimicking the therapeutic recalcitrance of denture stomatitis observed in patients (1, 3, 7, 16, 35, 39, 46, 55).
The current denture model provides an additional tool for examining the impact of specific gene products on biofilm formation, architecture, and drug resistance in vivo. Using this model, we were able to examine the role of Bcr1p, a transcription factor implicated in biofilm formation, and corroborated its importance in this niche as well (43, 44). A difference in the requirement of an adhesin, Als3p, for biofilm formation in vitro and in vivo has previously been demonstrated using a venous catheter model (43). Similarly, we anticipate that biofilm formation in the oral cavity, subjected to saliva and specific oral host defenses, may require a unique set of gene products. Potential features may include the requirement of gene products involved in quorum sensing, bacterial interactions, or adherence to epithelial cells. The current studies indicate the utility of the denture model for investigation of these questions.
One method of determining the gene products required for growth and proliferation in an environmental niche is through gene expression profiling. This model provides a means to examine the transcriptome of the C. albicans denture-associated biofilm cells. In current experiments, transcript abundances of a subset of biofilm-associated genes were compared directly between in vivo denture biofilm conditions and planktonic conditions (Table (Table2)2) (35, 40, 53, 59, 68). Biofilm cells from both the denture and catheter in vivo models exhibited a direction of expression similar to that of planktonic growth profiles with transcriptional abundance for each of the chosen genes. However, the magnitude of expression varied between the two models. For example, the transcript abundance of SAP5, a member of the extracellular aspartyl proteinase gene family known to function in pathogenicity by degrading tissue barriers, destroying host defense molecules, or digesting proteins for nutrient supply, was higher for the denture biofilm model (>250-fold increase versus 30-fold increase) (36, 57, 64). This finding is consistent with prior studies implicating the gene product in mucosal candidiasis pathogenicity (36, 57, 64). Variability in gene expression between the sites likely reflects the differences in environmental factors, including temperature, nutrient availability, oxygen level, and surface adherence properties. Further exploration for potential biofilm niche similarities and differences in global gene expression may provide important pathogenesis and therapeutic insights into these difficult-to-manage infections.
Of the biofilm-associated phenotypes, the profound drug resistance characteristic of biofilm cells is of specific interest due to the difficulty in eradication of these organisms (1, 35, 39, 55). Using the current model, we were able to test the susceptibility of denture-associated C. albicans cells both in vivo and ex vivo. Similar to previous investigations, the biofilms cells were extraordinarily resistant to the triazole, fluconazole, when tested on disrupted biofilm cells or following in vivo systemic or topical treatment (2, 35, 53, 62). Unlike previous investigations, micafungin, an echinocandin, did not demonstrate significant anti-Candida biofilm activity using these methods (4, 30). The reason for this apparent discrepancy is unclear but may reflect C. albicans growth in a mixed-species biofilm at this distinct site. The model may prove to be beneficial as an additional environmental condition for testing of potential antifungal or anti-Candida biofilm drugs.
One unique feature of the Candida denture biofilm relative to most systemic biofilms is the mixed-species nature of the process in the former. Although Candida spp., including C. albicans, C. glabrata, and C. tropicalis, are predominant organisms, many types of bacterial organisms may be present. The most common bacterial organisms include Streptococcus spp., Staphylococcus spp., and anaerobes, but molecular analysis of denture biofilms obtained from patients with stomatitis identified over 50 bacterial species involved in the process (10, 52). Interestingly, the bacterial species found on the dentures of patients with stomatitis appear distinct from those adherent to the dentures of healthy patients. The impact of bacterial microflora during development of denture stomatitis has not been fully explored, but investigations have begun to examine the interaction between Candida spp. and bacteria in mixed biofilms. Candida has been reported to form mixed biofilms with both Staphylococcus and Streptococcus during denture and catheter infection (5, 65). These bacterial species have been shown to interact with Candida and, in some instances, impact adhesion and antimicrobial susceptibility (1, 23). The current denture model provides ideal tools for examination of mixed biofilms.
The intricacies of the bacterial-fungal interaction highlight the importance of the oral microflora during denture placement. One potentially important difference in the oral cavities of rodents and humans is the higher burden of enteric flora present in the mouths of rodents due to coprophagy. We found the rat oral cavity to harbor many Gram-negative organisms, presumably linked to this behavior. To inhibit growth of these organisms and more closely mimic human denture stomatitis, we treated the animals with a broad-spectrum antibiotic prior to denture placement and housed animals in metabolic cages with wire cage floors. In a subset of experiments, animals were treated with an antibiotic specific to Gram-negative organisms and were housed in a tube which did not allow the animal to turn in place (15). These conditions greatly decreased coprophagy and the burden of Gram-negative organisms. We considered the possibility of using gnotobiotic, germ-free animals and colonizing a subset of animals with pathogens typical of human denture stomatitis. However, we currently do not have a gnotobiotic facility, and at this time, the cost would be prohibitive.
In conclusion, these studies indicate the utility of the denture model for testing the impact of gene products, studying biofilm architecture, and examining biofilm drug resistance. Furthermore, the model is potentially valuable for the investigation of mixed-species biofilms, bacterial-fungal interactions, and the impact of individual host factors, such as salivary proteins and immune components. Future investigations may include gnotobiotic animals with defined microflora to examine the impact of specific bacteria on C. albicans biofilm formation and drug resistance. To better understand the role of the host response in this unique process, studies may include animals with genetically defined immune deficiencies or with reduced salivation. Comparisons to other biofilm infection models promise to identify fungal and host elements distinct for the oral biofilm environment.
We thank William R. Gengler for helpful suggestions on the placement of oral rodent devices and Thomas Warner for assistance with histopathology. We thank Aaron Mitchell and Clarissa Nobile for providing strain CJN702.
This work was supported by the National Institutes of Health (grant RO1 AI073289-01).
Editor: G. S. Deepe, Jr.
Published ahead of print on 6 July 2010.