ABSTRACT
Indwelling urinary catheters are commonly used in the management of hospitalized patients. Candida can adhere to the device surface and propagate as a biofilm. These Candida biofilm communities differ from free-floating Candida, exhibiting high tolerance to antifungal therapy. The significance of catheter-associated candiduria is often unclear, and treatment may be problematic considering the biofilm drug-resistant phenotype. Here we describe a rodent model for the study of urinary catheter-associated Candida albicans biofilm infection that mimics this common process in patients. In the setting of a functioning, indwelling urinary catheter in a rat, Candida proliferated as a biofilm on the device surface. Characteristic biofilm architecture was observed, including adherent, filamentous cells embedded in an extracellular matrix. Similar to what occurs in human patients, animals with this infection developed candiduria and pyuria. Infection progressed to cystitis, and a biofilmlike covering was observed over the bladder surface. Furthermore, large numbers of C. albicans cells were dispersed into the urine from either the catheter or bladder wall biofilm over the infection period. We successfully utilized the model to test the efficacy of antifungals, analyze transcriptional patterns, and examine the phenotype of a genetic mutant. The model should be useful for future investigations involving the pathogenesis, diagnosis, therapy, prevention, and drug resistance of Candida biofilms in the urinary tract.
INTRODUCTION
Hospitalized patients frequently develop urinary tract infections. Catheter-associated urinary tract infection (CAUTI) is the most prevalent nosocomial infection, with over 1 million patients diagnosed yearly in the United States (1–3). Candida spp. account for the third most common cause of infection (2, 4, 5). Many factors have been linked to candiduria, including diabetes, urological procedures, female sex, and urological devices (6). Urinary catheters, devices necessary for monitoring the output of urine and maintaining urine outflow, are used in up to 20% of all hospitalized patients (7). Catheters provide a substrate for adherence of microorganisms and proliferation of biofilms. When growing as a biofilm, Candida is difficult to eradicate due to inherent drug resistance and immune tolerance (8–12).
The identification of Candida in the urine can indicate one of several clinical processes (13). The question of how to differentiate among these scenarios and optimally manage candiduria remains controversial (6, 14). First, Candida may enter the urinary tract from the mucosal surface, adhere to the urinary catheter, and establish a biofilm. Without further invasion, most patients are asymptomatic. However, Candida may produce cystitis or ascend further, reaching the kidneys, producing pyelonephritis. These infections are often symptomatic and require antifungal treatment. In another scenario, candiduria may be a sign of disseminated candidiasis with shedding of organisms from hematogenously seeded kidneys. Alternatively, candiduria may be the result of a contamination of urine, which may occur in the setting of vaginal candidiasis.
Diagnostic tools to differentiate among these clinical states are inadequate, and as a result, many patients receive unnecessary antifungal therapy (6, 15, 16). Models for study of CAUTI and candiduria would be of great value for investigating the pathogenesis of these various clinical presentations. The discovery of diagnostic markers to predict which patients may benefit most from treatment would help clinicians decipher urinary culture results and optimally utilize antifungal therapies.
Here, we describe a model for C. albicans biofilm infection of a urinary catheter in a rat. This model mimics Candida infection of an indwelling urinary catheter in patients. The model represents the clinical scenario with regard to anatomic location, urine flow, and common silicone device material. Over the course of infection, the animals develop progressive candiduria and urinalysis demonstrates inflammation. Ultimately, pathological findings are consistent with cystitis. On microscopic examination, mature biofilms cover the catheter surface. Our findings suggest that this model will be useful for investigations of biofilm pathogenesis and host response to this common clinical infection.
MATERIALS AND METHODS
Organisms and inoculum.Candida albicans strains K1, DAY185, and als1−/− als3−/− mutant were used for studies (17–19). 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 catheter inoculation, cells were grown at 30°C in YPD plus uridine liquid medium with orbital shaking at 200 rpm overnight. To prepare inoculum, cells were enumerated by hemocytometer counting and resuspended in YDP at 108 cells/ml. The final inoculum concentration was confirmed by microbiologic enumeration.
Animals and catheter maintenance.Specific-pathogen-free female Sprague-Dawley rats weighing 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. On the day of catheter placement, animals received a single dose of cortisone acetate, (250 mg/kg of body weight) subcutaneously. Animals also received gentamicin (80 mg/kg) subcutaneously twice daily and drinking water containing penicillin G sodium (0.9 mg/ml). Dosing regimens were chosen based upon those previously shown to be effective in treatment of rodent systemic bacterial infections (Andes lab, unpublished data). During the period of catheter placement, animals were maintained in metabolic cages. The animals were examined for signs of distress every 6 h throughout the study. The catheter sites were examined twice daily for signs of inflammation or purulence.
Urinary catheter placement and infection.Rats were anesthetized by intraperitoneal injection (1 mg/kg) of a mixture of xylazine (AnaSed; Lloyd Laboratories, Shenadonoah, IA) (20 mg/ml) and ketamine (Ketaset; Aveco Co., Fort Dodge, IA) (100 mg/ml) in a ratio of 1:2 (vol/vol). Animals were surgically prepped from midline to tail using surgical scrub (4% parachlorometaxylenol). A silicone catheter (Instech Solomon, 3.5Fr, female luer, round tip, 60 cm, gas sterilized) was inserted in the urethra and advanced to the first marking and secured with surgical glue (VetClose Surgical Glue; Butler Shein Animal Health) (Fig. 1). Following device placement, a protective covering and button (Polysolfone Button Tether; Instech Solomon) was advanced over the catheter and secured to the subcutaneous tissue by nylon suture (4-0) using 3 interrupted surgical knots. In addition, the animal was placed in a rodent jacket and Elizabethan collar (Braintree Scientific) to prevent animal manipulation of the urinary device. Using a syringe, urine was drained from the bladder though the catheter. Next, 700 μl of culture (the entire catheter volume) was instilled into the catheter lumen for 2 h. During this time, the animal was placed on a heated blanket and monitored for signs of distress. After 2 h, the inoculum was removed and the animal was placed in a metabolic cage. The catheter in the protective covering was threaded through the wire floor of the cage. The distal catheter was inserted into a 15-ml plastic conical tube through a hole in the lid and secured with a bolt and washer. Recovery of the animal after the catheter surgery was assessed according to a standard protocol approved by the Veterans Administration Animal Committee. After 24 to 72 h, the animals were euthanized and catheters and/or bladders were collected for analysis, as described below.
Model of a rat urinary catheter C. albicans biofilm infection. (A) A silicone catheter (3.5 French [Fr]) was inserted in the urethra of an anesthetized female rat. (B) A catheter covering and cone harness protect the urinary catheter.
Fungal cultures and urinalysis.To determine the viable burden of C. albicans, microbiologic counts were performed for urine, urinary catheters, and bladders. Urinary catheters were placed in 2 ml 0.15 M NaCl, sonicated for 10 min (FS 14 water bath sonicator and 40 kHz transducer [Fisher Scientific]), and vortexed for 30 s. Dilutions (1:10) were plated in duplicate or triplicate on Sabouraud dextrose agar (SAB). Urine analysis for leukocyte esterase and red blood cells was performed using a commercial urine dipstick (Rapid Response Urinalysis reagent strips; BTNX Inc.) after 24, 48, and 72 h of infection.
SEM.Urinary catheters were processed for scanning electron microscopy (SEM) as previously described for venous catheters (17). Urinary catheters were harvested at 48 h. Following removal, the distal segment (bladder and urethral portion) was placed in fixative (1% [vol/vol] glutaraldehyde and 4% [vol/vol] formaldehyde in phosphate-buffered saline [PBS]) overnight. The samples were washed with PBS, placed in 1% osmium tetroxide buffered with PBS for 30 min, and rinsed with PBS. The samples were subsequently dehydrated with 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), and desiccation was performed by critical-point drying (Tousimis, Rockville, MD). Specimens were mounted on aluminum stubs and sputter coated with gold. Samples were imaged in a scanning electron microscope (SEM LEO 1530) at 3 kV. The images were processed for display using Adobe Photoshop.
Histopathology.To evaluate the host response to C. albicans infection of the urinary catheter, we examined bladder wall histopathology. Animals were sacrificed at 48 h. The urinary catheters were removed, and bladders were dissected, fixed in 10% buffered formalin, and embedded in paraffin (20). Sections were stained with hematoxylin and eosin (H&E) and Gomori's methenamine silver (GMS) for imaging of Candida. Images were obtained at magnifications of ×10 and ×40. For SEM, bladders were fixed in a phosphate-buffered 1.5% glutaraldehyde solution and otherwise processed and imaged for SEM as described above.
Antifungal treatments.The effect of antifungal therapy on viable C. albicans biofilms was assessed by systemic treatment of mature biofilm infections (24 h of incubation) for 2 days. Animals were treated with either fluconazole (25 mg/kg subcutaneously once daily) or amphotericin B deoxycholate (1 mg/kg intraperitoneally once daily) and compared to untreated controls. At the completion of therapy, animals were sacrificed. The catheters and bladders were removed, and the viable Candida burden was determined as described above. Microbiologic assays were performed in triplicate, and significant differences were measured by analysis of variance (ANOVA) with pairwise comparisons using the Holm-Sidak method.
RNA collection and quantitative RT-PCR.Urinary catheters were collected for real-time PCR (RT-PCR) analysis after 24 h of growth and placed in RNAlater (Qiagen). Biofilm cells were dislodged from the catheter by vortexing and sonication. RNA was purified using the RNeasy minikit (Qiagen) and quantified using a NanoDrop spectrophotometer. TaqMan primer and probe sets designed using Primer Express (Applied Biosystems, Foster City, CA) for ACT1, FKS1, BGL2, XOG1, and PHR1 were used as previously described (see Table S1 in the supplemental material) (21). These genes were chosen based upon differential expression in the vascular catheter and denture biofilm models (21, 22). The QuantiTect probe RT-PCR kit (Qiagen) was used in a CFX96 real-time PCR detection system (Bio-Rad) with the following program: 50°C for 30 min, initial denaturation at 95°C for 15 min, and then 40 cycles of 94°C for 15 s and 60°C for 1 min. Reactions were performed in triplicate. The expression of each gene relative to that of ACT1 is presented. The quantitative data analysis was completed using the ΔΔCT method (where CT is the threshold cycle) (23). The comparative expression method generated data as transcript fold changes normalized to a constitutive reference gene transcript (ACT1) and relative to the reference strain (C. albicans K1). The comparative expression method generated data as transcript fold changes normalized to a constitutive reference gene transcript (ACT1) and relative to planktonic C. albicans cells, which were grown for 24 h in YPD at 37°C with orbital shaking at 200 rpm.
RESULTS
Urinary catheter placement and animal well-being.Rats tolerated placement and infection of a urinary catheter well and did not show signs of illness throughout the course of the experiments, which extended for up to 72 h. The animals continued normal intake of food and water. No erythema or purulence was observed at the urethral exit site.
Time course analysis.To assess biofilm formation over time, the viable burden was determined using microbiological plate counts at various time points following C. albicans K1 infection of a rat urinary catheter. Pilot microbiological analysis showed involvement of numerous bacteria (data not shown). As the goal of the current project was to model a monomicrobial C. albicans biofilm, we elected to include antibiotic treatment to reduce the bacterial burden and produce a consistent fungal biofilm. Upon including the antibiotic regimen, we observed an increasing urinary Candida burden over the 24- to 72-h time period. The burden started at less than 103 CFU/ml at 24 h and ultimately reached 106 CFU/ml (Fig. 2A). At the 48-h time point, the catheter viable burden was approximately 104 CFU/device (Fig. 2B). On urinalysis, pyuria and hematuria were evident throughout the course of infection (Table 1). Hematuria occurred in rats with uninfected catheters, suggesting that this may be related to trauma. Pyuria was greater in rats with C. albicans-infected than noninfected catheters.
C. albicans burden in a rat urinary catheter biofilm. (A) Urine was collected from a rat following C. albicans infection of an implanted urinary catheter after 24, 48, and 72 h of growth, and microbiological counts were used to determine the number of organisms present in the biofilms. (B) Urinary catheters from 2 rats were harvested after 48 h of biofilm growth, and adherent Candida cells were enumerated.
Results of urinalysis in rats with Candida albicans urinary catheter-associated biofilm infectionsa
SEM of urinary catheter biofilms.We used SEM to assess biofilm extent and architecture of the in vivo C. albicans biofilms. SEM has been a valuable tool for examining biofilm cell morphology, the extracellular matrix, and the relative extent of biofilm formation (17, 24). After 48 h, a confluent layer of biofilm had covered most intraluminal surfaces. However, compared to the vascular catheter model, which employed a polyethylene catheter, the silicone urinary catheter biofilm was less resilient in the face of the processing required prior to SEM imaging (17). The biofilm was frequently observed to be peeling from the surface, a finding that we suspect may have been related to the swelling of silicone during the dehydration process (Fig. 3, top panel [×100]). At higher magnification, the mature biofilm was composed primarily of hyphae with an extracellular matrix material covering sections of the biofilm (Fig. 3, bottom panel [×1,500]). Due to the biofilm disruption observed with SEM processing, viable-burden counts were utilized for comparisons of mutant strains and antifungal treatments in subsequent experiments.
Scanning electron microscopy (SEM) images of a C. albicans urinary catheter biofilm. Intact urinary catheter C. albicans biofilms were harvested after 48 h of growth, processed for SEM, and imaged. Scale bars for images with magnifications of ×100 and ×1,500 represent 250 μm and 10 μm, respectively. Arrows point to areas of extracellular matrix. The arrowhead denotes an area with hyphae and yeast.
Bladder microscopy.To discern the impact of the C. albicans urinary catheter infection on the host, we harvested the bladders for histology and SEM. H&E staining of the bladder revealed inflammation, marked by infiltration of polymorphonuclear cells (Fig. 4). GMS staining for fungal elements confirmed tissue invasion by Candida. Both yeast and hyphal forms were observed on the uroepithelial surface. The finding of fungal invasion and neutrophilic inflammation is consistent with acute cystitis. On SEM imaging of the bladder, the urothelial surface was covered by a heterogenous, fibrinous material (Fig. 5). Given the density of this material, identification of the underlying cells was somewhat limited. Host blood cells could be identified. In several areas, there was the appearance of yeast beneath the extracellular material, suggesting the presence of a surface-associated biofilm infection (25).
Bladder histopathology for C. albicans urinary catheter biofilm infection. Rat urinary catheters were infected with C. albicans. After 48 h, animals were sacrificed and dissected samples were fixed. Sections were stained with hematoxylin and eosin (H&E) and for C. albicans with Gomori's methenamine silver (GMS). Images were obtained at magnifications of ×10 and ×40. The outline box in each of the ×10 images marks the approximate location where the ×40 image was obtained.
Bladder SEM for C. albicans urinary catheter biofilm infection. Rat urinary catheters were infected with C. albicans. After 48 h, animals were sacrificed and dissected samples were processed for SEM and imaged. Scale bars represent 400 μm and 20 μm for the images with magnifications of ×50 and ×1,000, respectively. Arrows point to yeastlike structures within the biofilm.
Impact of antifungal drug treatments.Both in vitro and in vivo Candida biofilms exhibit tolerance to antifungal drugs (8, 17, 24, 26–32). We tested the impact of systemic administration of antifungal therapy on the in vivo urine catheter biofilm cell viability and bladder Candida viable burden. An azole (fluconazole) and amphotericin B were selected based on their achievable urinary concentrations and their clinical use for treatment of urinary candidiasis. An echinocandin was not included, as minimal amounts of active drugs accumulate in the urine for this drug class. At the completion of the experiment, the urinary catheters of untreated animals contained nearly 5 log10 CFU/device (Fig. 6). Treatment with either fluconazole (25 mg/kg/day) or amphotericin B deoxycholate (1 mg/kg/day) minimally impacted the catheter viable burden of C. albicans. However, these antifungal treatments decreased the viable fungal burden in the bladder by approximately 2 and 3 log10 CFU/bladder for fluconazole and amphotericin B, respectively. The doses selected for study are typically effective in nonbiofilm rodent infection models (19).
Impact of antifungal treatment on C. albicans urinary catheter biofilms. Rat urinary catheter biofilms were treated with either fluconazole (25 μg/ml subcutaneously once daily) or amphotericin B deoxycholate (1 mg/kg intraperitoneally) for 48 h. The viable burden was determined by microbiological plate counts following disruption of the biofilm from the urinary catheter (A) or following bladder homogenization (B). Two rats were included for each condition in panel A, and 1 rat was included for each condition in panel B. Microbiological replicates were performed in triplicate. ANOVA with pairwise comparisons using the Holm-Sidak method was used to compare viable burdens in treated samples and untreated controls, *, P < 0.05. FLU, fluconazole; AMB, amphotericin B deoxycholate.
Transcriptional analysis of biofilm-associated cells.Approximately 0.3 μg of total RNA was isolated from a single urinary catheter, an amount sufficient to perform RNA analysis using many methods. To test the utility of the model for examination of C. albicans biofilm-associated gene expression, we measured the transcript abundance of glucan-associated genes. These gene products have previously been shown to impact both biofilm matrix production and biofilm drug resistance in vitro and in a rat venous catheter model (21). Urinary catheter-associated biofilm cells were compared to free-floating, planktonic cells by RT-PCR (Fig. 7A). Transcriptional analysis revealed that the glucan-associated genes were upregulated in urinary catheter-associated biofilm cells relative to planktonic cells, consistent with findings from the prior investigations of C. albicans biofilms. Of the transcripts measured, those of BGL2 and PHR1 were the most abundant, with 3-fold-higher levels in the urinary catheter biofilm cells than in planktonically grown C. albicans. Expression of XOG1 and FKS1 was greater in the biofilm but less than 2-fold different.
Role of select gene products in urinary C. albicans biofilm formation. (A) Transcriptional abundance of glucan-associated genes in C. albicans urinary catheter biofilms. The transcript abundance of glucan-modifying enzymes in urinary catheter biofilms was compared to that in planktonic C. albicans. Analysis of two rat catheters was performed in triplicate by RT-PCR with ACT1 normalization. (B) Impact of adhesin disruption on urinary C. albicans biofilm formation. The biofilm-forming capacities of C. albicans als1−/− als3−/− mutant and parent strains were compared with viable burden endpoint. One rat was used for each condition. Microbiological replicates were performed in triplicate. Student's t test was used to compare viable burdens, *, P < 0.05.
Comparison of the reference strain and the adhesion-defective (als1−/− als3−/−) mutant strain.We next sought to test the ability of the model to detect the phenotype of a Candida strain with a biofilm-deficient phenotype. We chose the als1−/− als3−/− mutant, which lacks two adhesins important for C. albicans adherence and biofilm formation in vitro and in an in vivo vascular catheter model (33). We hypothesized that the mutant would also exhibit a biofilm defect in the rat urinary catheter niche. Compared to an otherwise isogenic reference strain, the als1−/− als3−/− urinary catheter biofilm was composed of nearly 100-fold fewer cells on viable-burden testing (Fig. 7B). As theorized based on prior biofilm studies, these adhesins appear to play a critical role in biofilm formation in the urine environment, and this is detectable in the rat urinary catheter model.
DISCUSSION
In the presence of an artificial substrate, Candida transitions to a biofilm lifestyle, engaging with the surface and proliferating as an adherent community (34–37). Numerous medical devices have been associated with biofilm growth and infection, including catheters (venous or urinary), vascular stents, cerebrospinal fluid shunts, pacemakers, and joint implants (36). Among these, CAUTIs represent 70% of all hospital-acquired infections and Candida is the third most common CAUTI pathogen (38–40). Here, we characterize a rat urinary catheter biofilm infection model, which is a close mimic of Candida CAUTI. The model recapitulates the clinical infection in terms of formation of a surface-associated biofilm, anatomic position of the catheter, conditions of the surrounding milieu, incorporation of host immune factors, material of the artificial device, and the flow conditions through the functioning catheter. With this model, we were able to quantify biofilm growth, assess biofilm architecture, study the impact of drug therapy, analyze the biofilm transcriptome, compare the biofilm-forming capacity of mutant strains, and assess the host response to biofilm infection.
In vitro models of biofilm infection have been instrumental in many Candida biofilm investigations, including the identification of factors governing biofilm behaviors and their ability to tolerate antifungal therapy (28, 41–46). The models can also be useful for characterizing the impact of surface modifications and treatments. In vitro models have attempted to account for many in vivo infection conditions suspected to be important in clinical infection. Examples include the addition of urine to media to replicate the milieu of urinary biofilms, the incorporation of substrate materials similar to medical devices, and the inclusion of flow conditions (47). Uppuluri et al. examined C. albicans biofilm growth in the presence of synthetic urine media that included defined electrolyte concentrations, a relatively low pH, creatinine, and urea (47, 48). Similar to the current investigation, biofilms formed under these conditions exhibited resistance to antifungals commonly used to treat urinary tract infections, including amphotericin B and fluconazole. However, compared to control biofilms growing in RPMI media, biofilms produced in synthetic urine medium were less dense and fewer cells had transitioned to the hyphal state. This is in contrast to the current investigation, in which hyphae were prominent in the C. albicans biofilms on the luminal urinary catheter surface (Fig. 3). Interestingly, the antifungal therapy was effective against the tissue-associated Candida in the model, suggesting the presence of both biofilm and nonbiofilm cells in the model.
Differences between in vitro and in vivo models are not unexpected. It is difficult for in vitro models to account for all the factors that may be influencing biofilm infection in the host (49). For example, cells in the in vitro systems are not exposed to many immune components and proteins that may condition or coat the surface and promote adherence. Hundreds of proteins have been identified as adsorbing to urinary devices in patients. The protein set is diverse and includes cytokeratins, albumin, and inflammatory proteins (50). These conditioning factors likely arise from surrounding cells under inflammatory conditions, as the protein content of urine is generally low. Mimicking this process in vitro would be very complex. In vitro conditions are also limited in the ability to reproduce the influence of the immune system, a dynamic process over the infection course.
While examining the utility of the model for gene expression analysis, we identified upregulation of several transcripts in the glucan synthesis and modification pathways in C. albicans urinary catheter biofilms (Fig. 7). This was not surprising, given the role of these pathways in extracellular matrix production and biofilm drug resistance (21, 43). We found similarities between our current study and our prior microarray analysis, which compared rat vascular catheter biofilms to planktonic controls (51). For example, transcripts of β-1,3 glucan synthase FKS1 were more abundant in the catheter biofilms (vascular, 1.8-fold more abundant; urinary, 1.3-fold more abundant). Likewise, 1,3-beta-glucosyltransferase BGL2 was upregulated in both catheter models (vascular, 1.5- to 3.1-fold; urinary, 3.7-fold), as were glucanosyltransferase PHR1 (vascular, 2.4- to 24.2-fold; urinary, 3.3-fold) and beta glucosidase XOG1 (vascular, 1.7-fold; urinary, 1.9-fold). In a rat denture model of C. albicans biofilm formation, BGL2 was similarly upregulated (1.6-fold) compared to planktonic controls (24). These findings suggest that there are conserved pathways among the various clinical biofilm niches.
Prior in vivo investigations of urinary catheter biofilms and CAUTI have utilized rodent models. To examine Pseudomonas aeruginosa urinary catheter biofilms, Kurosaka et al. developed a rat model of CAUTI. In that study, a stylet was inserted through the urethra of a rat and a catheter segment was threaded over the stylet and released into the bladder (52). Following transurethral inoculation, bacterial biofilms were established on the catheter surfaces and animals developed the histopathologic findings of acute pyelonephritis. This model was subsequently adapted for use in a mouse for study of Pseudomonas aeruginosa, Proteus mirabilis, Enterococcus faecalis, and Candida albicans (53–55). In an investigation of C. albicans biofilms by Wang and Fries, the catheter segments were secured in the bladder for 5 to 7 days prior to infection, allowing for host proteins to adsorb to the device surface (55). In this model, dense biofilms formed on both the luminal and external catheter surfaces and consisted of yeast and hyphae. As a method to predispose to Candida infection, mice deficient in lysozyme M production, an important effector for mucosal innate immunity, were utilized. The advantages of the model include the smaller animal size of the mouse and the ability to include the defined murine genotypes, such as the lysozyme M-deficient mouse.
One limitation of the previously described rodent CAUTI models is the placement of the catheter segments (53–55). Although the catheter segments are exposed to urine and host components in the bladder, the catheters lack a urethral component and do not function to drain the urinary system, as would be the case for patient catheters. Without a urethral component, the catheters lack flow, one of the key factors influencing biofilm architecture and extracellular matrix production (56–58). To best account for physiologic flow in the current studies, we utilized urethral catheters that functioned to drain urine from the bladder throughout the course of the experiments. Not only did this have the advantage of mimicking flow, but it also permitted repeated collection of urine samples.
To most closely mimic a patient infection, we chose to use a silicone urinary catheter, as this is the most common urinary catheter material (59). Biofilms formed on the luminal surface over the several days following intraluminal inoculation. However, on SEM, the biofilms were observed to often be dislodged or peeling from the catheter surface (Fig. 3). This is in contrast to what has been observed in a prior CAUTI model of C. albicans infection and in a rat venous catheter biofilm infection, both of which had used polyethylene catheters (17, 55). We suspect that the dehydration process required for SEM altered the silicone, weakening the biofilm binding. Another possibility is that urinary biofilms are less adherent to the device due to unique environmental conditions in the urine. We favor the former hypothesis, given clinical descriptions of extensive biofilm in the literature and our demonstration of a large infectious burden by microbiological counts. The viable plate count method was also useful for assessment of antifungal drug effect and the impact of various genetic mutants on urinary biofilm formation (Fig. 7). Using this model, we identified histopathologic changes consistent with acute cystitis (Fig. 4). This is similar to descriptions from other animal models of bacterial CAUTI (52, 55). On specific fungal staining, mucosal invasion by Candida was evident and reminiscent of denture biofilm-associated mucosal changes (Fig. 4). The adherent community of Candida cells suggested the presence of a mucosal biofilm, as has been described for both oral and vaginal candidiasis (25, 60). On close examination of the bladder urothelial surface by SEM, we observed aggregates of cellular material encased in a fibrinous material, suggesting surface-associated biofilm formation (Fig. 5). Similar findings have been described for Klebsiella pneumoniae infection of a rat bladder (61). It has been proposed that epithelial cells are eventually sloughed during acute cystitis as a protective mechanism to rid the bladder of the surface-associated pathogens (61).
The current studies demonstrate the utility of the rat urinary catheter model for numerous research avenues involving Candida biofilm and CAUTI. Biofilm formation and architecture can be assessed by microscopy, and assays can easily be designed to test the impact of antifungal drugs and the influence of gene products. The model allows for comparisons of genetically manipulated strains and transcriptional analysis. Given the physiologic catheter flow, it may be optimal for preclinical testing of catheters with impregnated or surface-adherent anti-infectives (37). Although the focus of this study was C. albicans, the model could likely be adapted for use of non-albicans species of Candida, such as C. parapsilosis, C. glabrata, or C. dubliniensis, as has been described for other animal models (62, 63). Furthermore, comparisons among the various animal models of Candida biofilm infection may be of interest to identify pathways either unique to individual clinical niches or conserved among diverse clinical biofilms (17, 24, 25, 32, 55, 60, 64, 65).
ACKNOWLEDGMENTS
We thank Aaron Mitchell and Clarissa Nobile for the als1−/− als3−/− strain. We thank Heather Taff for help with SEM imaging.
This work was supported by the National Institutes of Health (R01 AI073289-01 and K08 AI108727) and the Burroughs Wellcome Fund. We acknowledge use of instrumentation supported by the UW MRSEC (DMR-1121288) and the UW NSEC (DMR-0832760).
FOOTNOTES
- Received 3 July 2014.
- Returned for modification 19 July 2014.
- Accepted 25 August 2014.
- Accepted manuscript posted online 2 September 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.02284-14.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
