Development and Characterization of an In Vivo Central Venous Catheter Candida albicans Biofilm Model

Animals.

Specific-pathogen-free female Sprague-Dawley rats weighing 400 g (Harlan Sprague-Dawley, Indianapolis, Ind.) were used for all studies. Animals were housed in an environmentally controlled room with 12-h light-dark cycle and were maintained on a standard ad libitum rat diet. Animals were maintained in accordance with the American Association for Accreditation of Laboratory Care criteria (37). Animal studies were approved by the Animal Research Committee of the William S. Middleton Memorial Veterans Administration Hospital.

Catheter.

The catheter diameter was chosen in an attempt to permit blood flow around the extraluminal catheter surface. The catheter tubing material was chosen to mimic a material used in patients. A polyethylene tubing from Intramedic met these criteria (PE 100 [inner diameter, 0.76 mm; outer diameter, 1.52 mm]; Intramedic catalog no. 427430; Becton Dickinson and Company, Franklin Lakes, N.J.).

The length of the catheter was determined based upon the sum of the planned intravascular distance (from the insertion in the internal jugular vein to above the right atrium) (2 cm), the subcutaneous tunneled distance, and the length of catheter to run in an external protective device from the animal to the cage top for ease of catheter access. These 50-cm segment lengths were gas sterilized with ethylene oxide prior to placement in the animals. The volume of this catheter length was determined to be 500 μl.

Catheter placement.

Prior to catheter surgery, the rats were anesthetized by intraperitoneal injection (1 mg/kg) of a mixture of xylazine (AnaSed; Lloyd Laboratories, Shenadonoah, Iowa) (20 mg/ml) and ketamine (Ketaset; Aveco Co., Fort Dodge, Iowa) (100 mg/ml) in a ratio of 1:2 (vol/vol). Hair was shaved from the midscapular space, anterior chest, and neck. The rat was placed in the supine position; the surgical area was prepped and draped in a sterile fashion. A vertical incision was made in skin of the anterior neck just right of midline. The internal jugular vein was identified and exposed with blunt surgical dissection. A longitudinal incision of a few millimeters was then made to the vein wall with vein scissors. The heparinized (100 U/ml) catheter was placed in the opening and advanced to a site above the right atrium (approximately 2 cm). The catheter was secured to the vein with 3-O silk ties. The proximal end of the catheter was then tunneled subcutaneously to the midscapular space and externalized through the skin via a button, which was secured with sutures. The anterior neck site was closed with surgical staples (Ethicon Endo-Surgery, Cincinnati, Ohio), and the midscapular button was secured with 2-0 Ti-cron suture. Recovery of the animal after the catheter surgery was assessed according to a standard protocol approved by the Veterans Administration Animal Committee.

Animal and catheter maintenance.

The animals were examined for signs of distress every 8 h through the study. The anterior neck incision and the catheter exit site were examined twice daily for signs of inflammation or purulence. The catheters were flushed daily with heparinized (100 U/ml) 0.85% NaCl.

Organism and inoculum.

C. albicans K1 was used for all studies. The organism is a clinical isolate from a systemic Candida infection. The organism was maintained, grown, subcultured, and quantified on Sabouraud dextrose agar (SDA) plates (Difco Laboratories, Detroit, Mich.).

The organism was subcultured at 35°C on SDA 24 h prior to infection. The inoculum was prepared by placing three colonies into 5 ml of sterile pyrogen-free 0.85% NaCl warmed to 35°C. The final inoculum was adjusted to 5.60 log₁₀ CFU/ml via hemacytometer count. Viable fungal counts of the inoculum were confirmed by plating on SDA and were 5.54 to 5.74 log₁₀ CFU/ml.

Infection.

A time course schematic of the experiment is shown in Fig. 1. The catheters were placed 24 h prior to infection to allow a conditioning period for deposition of host protein on the catheter surface. After the conditioning time period, 100 μl of blood was obtained from the catheter and cultured to ensure sterility. The inoculum of C. albicans K1 was instilled in the catheter in a 500-μl volume (the entire catheter volume). The inoculum was allowed to dwell for 4 h, after which the catheter volume was withdrawn and the catheter was flushed and locked with heparinized 0.85% NaCl. This infection was undertaken with 10 animals.

For two animals the catheter infection was attempted by injection of a 0.2-ml volume of a similar inoculum (5.82 log₁₀ CFU of C. albicans K1 per ml) via the rat tail vein 24 h after initial placement.

Quantitative culture from catheter, peripheral blood, and kidney.

For the directly infected catheters, two animals were sacrificed by CO₂ asphyxiation at four time points (8, 12, 24, and 72 h) after the inoculum of C. albicans was removed. After sacrifice the catheter was aseptically removed from infected animals. One milliliter of sterile 0.85% NaCl was slowly flushed through the catheter to remove blood and nonadherent organisms. The distal 2 cm of catheter was cut from the entire catheter length, and the segment was placed in 1 ml of 0.85% NaCl in a 1.5-ml tube. The tube was sonicated for 10 min (FS 14 water bath sonicator and 40-kHz transducer [Fisher Scientific]) and vigorously vortexed for 30 s (43). Venous blood (100 μl) was removed from the tail veins of animals at the same time points. To ascertain the extent of metastatic disease, the right kidney was also removed from each animal, placed in a volume of 0.85% NaCl to achieve an initial 1:10 dilution by weight, and homogenized. Following serial 1:10 dilution of the catheter fluid and kidney material, the samples were plated on SDA for viable fungal colony counts after incubation for 24 h at 35°C. The lower limit of detection was 100 CFU/ml or 100 CFU/kidney. Results were expressed as the mean CFU per milliliter or kidney from two mice.

The catheters from two distantly infected rats were removed 24 h after infection and processed similarly.

Catheter biofilm imaging. (i) Confocal microscopy.

The distal 2 cm of catheter was further cut perpendicular to the catheter length with an 11-blade scalpel into three 2- to 3-mm-long “doughnut” segments. The catheter segments were stained with the FUN-1 (50 μM) component of the LIVE/DEAD yeast viability kit (Molecular Probes, Eugene, Oreg.) and concanavalin A (ConA) conjugate (Alexa Fluor 488 conjugate) (200 mM) according to the manufacturer's instructions (20, 33). Both live and dead cells are labeled with this dye. All yeast cells are stained with the FUN-1 dye and visualized with a diffusely distributed green fluorescence. Metabolically active cells process this dye, which results in a shift from green to red fluorescence within intravacuolar structures. ConA binds to glucose and mannose residues of cell wall polysaccharides and is visualized as green fluorescence. Due to colocalization of fluorescent dyes (FUN-1 and ConA), the metabolically active cells appear yellow to orange during multichannel image capture. Prior study has demonstrated the utility of this dye pair for imaging of both the extracellular biofilm, which is made up largely of polysaccharides, and the associated cells simultaneously (8).

Stained catheter segments were transferred to a glass-bottom petri dish (coverslip 1.5, 35-mm disk P325G 1.5-14C; MatTek, Ashland, Mass.) in an on-end orientation. Biofilms were observed with a Nipkow disk-based confocal microscope (Axiovert 200; Zeiss, Gottingen, Germany) equipped with a mercury arc lamp. The objectives used included 20× and 40× differential interference contrast oil immersion objectives. Depth measurements were taken at regular intervals across the biofilm. Confocal images of green (ConA) and red (FUN-1) fluorescence were conceived simultaneously using the Z-stack mode. A line drawing of the imaging procedure is shown in Fig. 2.

Imaging of FUN-1-stained cells was accomplished by using a protocol with an excitation wavelength of 488 nm and emission bands at 528 and 617 nm for green and red, respectively. The ConA protocol included excitation and emission wavelengths of 490 and 528 nm, respectively. Images were processed for display by using Axiovision 3.x software (Zeiss).

(ii) Scanning electron microscopy.

Catheter disk segments were washed with phosphate-buffered saline and placed in fixative (1% [vol/vol] glutaraldehyde and 4% [vol/vol] formaldehyde) overnight. The samples were rinsed in 0.1 M phosphate buffer two times for 3 min each and then placed in 1% osmium tetroxide for 30 min, followed by immersion in hexamethyldislilazane (Polysciences, Inc., Warrington, Pa.). 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. Samples were observed in a scanning electron microscope (Hitachi S-5700) in the high-vacuum mode at 10 kV. The images were processed for display by using Adobe Photoshop version 5.0.

Candida biofilm drug susceptibility in vivo.

Three assays were used to determine the fluconazole susceptibility phenotype of in vivo biofilm-associated cells compared to that of planktonically grown cells.

The first experiment compared in vivo growth of C. albicans in the presence and absence of the triazole fluconazole. Two catheters were placed in rats, and the catheters were directly infected as described above. After a period of biofilm growth (18 h), fluconazole was instilled and allowed to dwell in the catheter at a concentration of 8 μg/ml, which is 16 times the planktonic MIC, for 5 h. After the 5 h, the 500-μl catheter volume was removed and the same volume of sterile 0.85% NaCl was instilled. Another animal and indwelling catheter were similarly processed, except sterile 0.85% NaCl was used in place of fluconazole. The catheters were removed from the animals 1 h after drug removal. The catheters were flushed with sterile 0.85% NaCl to remove blood and nonadherent C. albicans. The distal 2 cm of the catheters, representing the intravascular length, was placed in sterile 0.85% sterile saline, sonicated, and quantitatively cultured as described above.

We similarly compared the viability of biofilm-associated cells after drug exposure and that of a no-drug control by using confocal imaging and FUN-1 vital staining of small catheter doughnuts as described above. We counted the number of cells with red fluorescence (metabolically active) per field (magnification, ×40) for both fluconazole-treated cells and NaCl-treated control cells. We examined two catheter doughnuts per catheter and five microscopic fields per segment.

In a second experiment, we performed in vitro susceptibility testing with fluconazole on cells removed from an in vivo biofilm. The catheterized rat was infected via the catheter as described above. After a 24-h period of biofilm development, the catheter was removed. The entire catheter was flushed with 0.85% NaCl, and the distal 2 cm was placed in 1 ml of 0.85% NaCl. The tube was then sonicated and vortexed to dislodge cells. The cells obtained after vortexing of the biofilm-involved catheter were used as the inoculum in a standard broth microdilution in vitro susceptibility test. The cells were quantified by hemacytometer count and transferred to a 96-well microtiter plate for susceptibility testing by the NCCLS M27-A method (36). The cell counts from hemacytometer measurements were confirmed by direct plating. The MIC for these cells was compared to that for planktonically grown cells. MIC determinations were performed in duplicate. Final results are expressed as the mean of these results.

Transcriptional profile of biofilm-associated cells.

The mRNA abundances of four genes that have been shown to be associated with fluconazole drug resistance were measured both in the in vivo biofilm model and in planktonically grown C. albicans K1 (59). For collection of the in vivo biofilm-associated cells, the distal 2 cm of catheter after 24 h of biofilm development was utilized. The catheter segment was cut into multiple 2- to 3-mm-thick doughnuts and placed in 400 μl of Tris-EDTA buffer. The tube containing the segments was sonicated and vortexed as described above. We next proceeded with total RNA isolation by the hot-phenol method of Schmitt et al. (53). For the comparative in vitro reference, cells were grown in yeast nitrogen base with shaking at 35°C to an optical density at 600 nm of 0.6. The quantity of the RNA was determined via spectrophotometry with an ND-1000 spectrophotometer (NanoDrop Technologies, Montchanin, Del.).

Quantitative RT-PCR.

Quantitative real-time reverse transcription-PCR (RT-PCR) was used to compare mRNA abundances of the genes of interest (7). The TaqMan MGB probe and primer sets were designed for the _target genes CDR1, CDR2, MDR1, and ERG11 and for the normalizing gene ACT1 of C. albicans (Table 1), using Primer Express 1.5 software (Applied Biosystems, Foster City, Calif.). All probes have 6-carboxyfluorescein as a 5′ fluorophor and 6-carboxytetramethylrhodamine as a 3′ quencher (28). Primers annealed at 60°C and defined the endpoints of the amplicon. Probes annealed at 70°C and hybridized between, but not overlapping, the two endpoint primers. Oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, Iowa).

The QuantiTect probe RT-PCR kit (Qiagen, Valencia, Calif.) was used in an ABI PRISM 7700 v1.7 sequence detection system (Applied Biosystems). Reactions were performed in duplicate according to the kit manufacturer's instructions. Final concentrations or amounts in the 50-μl reaction included 1× QuantiTect probe RT-PCR master mix, 1 μM forward primer, 1 μM reverse primer, 0.2 μM probe, 1 U of heat-labile uracil-N-glycosylase, 0.5 μl of QuantiTect probe RT mix, and 50 ng of total RNA. The ABI PRISM 7700 sequence detection system was programmed to reverse transcribe RNA at 50°C for 30 min, followed by 15 min of initial HotStart Taq DNA polymerase activation at 95°C. Two-step cycling, at 94°C for 15 s and 60°C for 1 min, was used for cDNA amplification. Fluorescence data were collected during the 60°C annealing and extension step.

The quantitative data analysis was completed using the ΔC_t (2^−ΔΔC_t) method of Livak and Schmittgen (27). For each amplification run, the threshold cycle, C_t, of the _target was normalized to the C_t of the ACT1 gene and related to the C_t of the control or baseline condition. The comparative ΔC_t, ΔΔC_t, was defined as the difference between the ΔC_t values: ΔΔC_t = (C_{t, _target} − C_{t, normalizer}) − (C_{t, baseline} − C_{t, normalizer}). The comparative expression value was transformed to an absolute value by using the following formula: fold change = 2^−ΔΔC_t.

The comparative expression method generated data as fold change in gene expression normalized to a constitutive reference gene and relative to a control or baseline condition. The statistical significance of differences in expression between biofilm and planktonic cells was analyzed by analysis of variance. The baseline condition in this case was mRNA isolated from planktonically grown C. albicans K1 cells. Real-time quantitative RT-PCR was performed on four 10-fold cDNA dilutions to ensure similar primer efficiencies among the various amplified _targets (data not shown).

General animal well-being and catheter site observations.

The catheterized animals appeared well over the 96-h period of study (24 h preinfection and up to 72 h postinfection). Both the surgical site and catheter exit site remained free of signs of inflammation or purulence in all animals studied.

Quantitative culture from catheter, peripheral blood, and kidney.

Time course viable counts from the catheter, rat kidney, and peripheral blood are shown in Fig. 3. At the first time point (8 h) there were more than 3 log₁₀ CFU of viable cells per ml present on the 2-cm catheter segment. The burden of organisms associated with the biofilm and in the distant rat kidney continued to increase over the 72-h period of study. However, the majority of the maximal burden of viable cells was present in the first 24 h. The burden of organisms on the 2-cm length of catheter increased nearly 1 log₁₀ unit from our initial measurement at 8 h to the 24-h time point. Over the same period of time the organisms grew 1.5 log₁₀ units in the rat kidney, suggesting a somewhat slower rate of growth within the biofilm than at a distant organ site of metastasis. Culture of 100 μl of peripheral blood was not able to detect growth of Candida at any of the time points. However, growth of organisms in the kidney demonstrates the disseminated nature of this infection model. The distantly (tail vein) infected catheter contained a similar burden of viable cells at the 24-h time point (4.50 log₁₀ CFU/ml).

Catheter biofilm imaging. (i) Confocal microscopy.

Representative images of the intraluminal in vivo biofilms at magnifications of ×20 and ×40 in catheters after 24 h of development are shown in Fig. 4. The catheter and biofilm orientation is end on, with the catheter wall at the top left and the catheter lumen at the bottom right of the image. The thickness of the biofilm at maturity varied from 25 to nearly 200 μm on the intraluminal surface. The density of the biofilm matrix adjacent to the catheter wall makes it difficult to discern individual yeast cells. The hazy green appearance represents the dense extracellular material adjacent to the catheter wall. At the 8- and 12-h time points it is possible to visualize the outlines of individual cells adjacent to the catheter wall. In the mature (24- and 72-h) biofilm, distinct, metabolically active cells (stained yellow or red) and the surrounding extracellular matrix strands are visualized adjacent to the catheter lumen. Figure 4B clearly demonstrates only metabolically active cells utilizing only single-filter capture.

We did not capture the initial attachment of cells to the catheter with our earliest examination time point. By the 8-h time point (not shown), the biofilm was multiple cell layers thick, consisting of both yeast and hyphae. Fully mature biofilms were produced after 24 h and consisted of a dense network of yeasts, hyphae, and pseduohyphae. While the organisms continued to proliferate through the 72-h time point, the dimensions of the biofilm appeared similar to those at 24 h. At maturation the cells are completely encased within the matrix material. At this stage, focus on individual cells is difficult due to the matrix density.

Confocal imaging of the biofilm from the hematogenously infected catheter appeared similar to that for those infected intraluminally, except there was more biofilm on the extraluminal surface (not shown) than with catheters infected via direct catheter dwelling of inoculum.

(ii) Scanning electron microscopy.

Imaging of the biofilm on the catheter segments by scanning electron microscopy more clearly demonstrates the heterogeneous architecture of the biofilm structure. In Fig. 5A, the biofilm is imaged from above the intraluminal biofilm surface. The mature biofilm surface adjacent to the catheter lumen is characterized by both yeast and hyphal cell forms and a network of extracellular matrix strands. In Fig. 5B, the catheter segment is on end, and one can observe the differences between the biofilm adjacent to the wall of the catheter and that on the luminal surface. The basal region consists of densely packed yeasts encased in and often obscured by a surrounding extracellular material. Hyphal structures are also apparent in the basal region; however, they are fewer in number. Adjacent to the lumen, it is easier to discern yeast and hyphal forms, and the matrix is more strand-like and less compacted. In addition, scanning electron microscopy identified cell structures too large to represent C. albicans. We theorize, based on size, that these cells represent host neutrophils. Future study utilizing specific histochemical staining will be necessary to further identify these cell types. The proximity of the host cells within the biofilm points to the relevance of examination of this process in an in vivo system to capture the host-microbe-biofilm interaction.

Candida biofilm drug susceptibility in vivo.

In vivo exposure to fluconazole concentrations many times greater than the in vitro MIC did not impact Candida viability in mature biofilms. The drug-treated catheter contained a similar number of viable cells as determined by both quantitative culture (fluconazole, 5.05 log₁₀ CFU/ml; NaCl, 4.98 log₁₀ CFU/ml) and vital staining (fluconazole, 25 ± 5 cells/40× field; NaCl, 27 ± 8 cells/40× field).

We observed a remarkable rise in the MIC for cells associated with the in vivo biofilm compared to those grown planktonically. The traditional (planktonic) MIC of fluconazole in C. albicans K1 is 0.5 μg/ml. The MIC for the cells recovered from the catheter was more than 128 times higher than the planktonic MIC (>64 μg/ml).

Transcriptional profile of biofilm-associated cell.

RNA isolation from a single catheter yielded enough material for numerous quantitative real-time RT-PCR experiments. More than 1,000 ng of total C. albicans RNA was isolated from a single catheter (260/280 ratio, 2.11). The fold change in abundance of genes examined from the biofilm relative to those from planktonic cells is shown in Fig. 6. The mRNA abundance of the gene encoding the drug _target enzyme, ERG11, was not different between the planktonic and the in vivo biofilm-associated cells. Similarly, expression of the major facilitator pump, MDR1, did not appear to be remarkably affected by biofilm growth in vivo. However, mRNAs from both of the ATP-binding cassette pumps, CDR1 and CDR2, were increased in the biofilm state (P < 0.01). CDR2 was the most remarkably affected, with a nearly 10-fold increase in expression. The results from RT-PCR were extremely similar in assay replicates.