Introduction

Several protocols have been developed to understand how Candida spp. biofilms are formed, how they express their virulence potential, and how they can be controlled [1–3]. However, they involve statically grown biofilms, which are not suitable to approach clinical reality [4]. The main problems are the following: (1) failure to permit development of invasive hyphae; (2) culture broths are exchanged after relatively long periods of time with depletion of nutrients and accumulation of toxic metabolites; (3) evaporation causes solute concentration, leading to osmotic imbalances; (4) pHs are not constant, even when using buffered broths; (5) the possibility of cross-contamination among wells (in 12–96 well microtitration/culture plates) must be considered.

Different designs for dynamic biofilms have been developed to avoid such problems. Shaken multi-well plates generate high hydrodynamic flux of nutrients, leading to more cell-containing biofilms [5].

Modified Robbins devices have been used to grow biofilms under continuous flow [6–8]. Despite their robust performance, problems regarding the absence of gas/biofilm interfaces make the extrapolation of mucosa or skin-related infections difficult.

Flow chambers attached to microscopes are very interesting since they permit following the growth of biofilms in real time [9]; however, studies involving high-throughput screenings are difficult to conduct.

The obtaining of a slow laminar flow may be achieved by using drip-flow reactors. At a glance, those systems are very attractive since they allow the growth of biofilms under low shear stress and suitable gas–liquid interface. Nevertheless, such systems have rarely been used to grow C. albicans biofilms [10].

An inexpensive variation in drip-flow reactors was recently proposed [11, 12]. Despite its apparent simplicity, problems regarding repeatability and low throughput limit its use.

Nowadays, there is a tendency to employ microfluidics-based systems [13]. However, the implementation of such technologies demands considerable costs to acquire photolithography and argon plasma equipment, as well as being laborious and time-consuming.

A feasible high-throughput table and low-cost biofilm reactor for C. albicans (and non-albicans Candida spp.) biofilms is proposed here. It is based on the confluent distribution of nutrients through capillary diffusion in ordinary laboratory qualitative filter paper.

Materials and Methods

Paper-Embedded Biofilm Reactor (PEBR)

The PEBR was assembled as presented in Fig. 1. It consists of a stainless steel pipe tube section [100 mm (Ø) × 70 mm (high) × 1 mm (thick)] eccentrically welded onto a metal plate (150 mm × 120 mm × 2 mm). Four holes at the corners were used to adapt adjustable height buttons, and a hole in the bottom permits broth drainage.

Fig. 1
figure 1

Compounding parts of the PEBR (details in the text). 1 Stainless steel needle, 2 two-way controller, 3 rubber stopper, 4 syringe barrel with sterile cotton, 5 glass lid with central hole, 6 stainless steel cylinder, 7 stainless steel base, 8 rubber profile, 9 screw level, 10 sewage exit, 11 glass table, 12 diffusion matrix (filter paper), 13 antibiogram disks with C. albicans, 14 steel brackets, 15 sealing nuts, 16 removable level

Full size image

A glass table [90 mm (Ø) × 5 mm (thick)] was adapted inside the cylindrical jacket, avoiding any contact with its wall. On the table, a 90-mm Ø disk of Whatman™ Qualitative Grade 3 filter paper (General Electric Co., Kent, UK) was laid. This paper disk served as a matrix for nutrient diffusion. Five millimeters from the edge, 6-mm antibiogram paper disks (Cecon Ltd., São Paulo, Brazil) were distributed at a distance of 3–4 mm from one another. A maximum of 30 antibiogram disks could be settled.

To complete the PEBR assembly, a PEXT0640© rubber U-profile segment (PAR Group Co., Lancashire, UK) was tightly attached to the upper jacket’s free edge in order to avoid any gas leakage/external contamination. A flat glass lid (120 mm × 120 mm × 4 mm) with a central hole (20 mm Ø) covered the jacket. A culture broth admission system was built with a 20-mm Ø rubber stopper and a two-way air flow controller (Guandong Boyu Group Co. Ltd, Raoping, China) and a 121-D™ 100 × 20 stainless steel needle (Höppner Veterinary Products Ltd., São Paulo, Brazil).

The glass lid was firmly attached to the jacket using stainless steel straps (120 mm × 20 mm × 2 mm) and bolts/nuts. Once assembled, the PEBR was sterilized at 121 °C, 1 atm. (cm2)−1, and 15 min.

Capillary Diffusion Tests

The PEBR was connected to reservoirs containing either 0.1 or 0.01 % crystal violet (CV). One-liter Kitasato flasks served as reservoirs, to which silicone tubing (1 mm internal Ø) was connected. Between reservoirs and PEBR, a P-1™ peristaltic pump (Pharmacia Biotech Co., Uppsala, Sweden) was placed to control the flow.

Reservoirs were filled with 800 mL of crystal violet solutions. The systems were assembled, and solutions were dropped (200 μL min−1) in the central area of the diffusion filter disks. These tests were carried out at 37 °C and normal atmosphere.

After 48 h, the systems were disassembled and the antibiogram paper disks were carefully collected with thin-tip forceps. They were placed into wells of U-bottomed microtitration plates. Such wells were filled with 200 μL of absolute ethanol, and the absorbed dye was eluted for 5 min. No signals of non-partitioned dye remains were observed on the paper disks, which lead us to infer that all CV was extracted from the disks. Eluates were combined and transferred to flat-bottomed plates. The OD540nm were measured in a TP-Reader™ (ThermoPlate Co., Shenzhen, China).

Capillary diffusion tests were carried out in four different occasions, performing 120 repetitions (4 days × 30 disks).

Growing C. albicans Biofilms

Candida albicans SC5314 was grown in yeast extract–peptone–glucose (YEPD) at 37 °C, normoxia, and 120 rpm, to achieve OD540nm equal to 1.00. Aliquots of 10 μL of cell suspension were dropped onto 30 sterile 6-mm pro-antibiogram paper disks. Prior to cell application, the disks had already been set on the sterile 90-mm Ø disk of Whatman™ Qualitative Grade 3 filter paper within the PEBR, as stated before. The bioreactor was assembled and connected to the YEPD broth reservoir. The broth was admitted at a 200 μL min−1 flow rate, and biofilms were aerobically grown at 37 °C (Fig. 2).

Fig. 2
figure 2

Putative events happening inside assembled PEBR

Full size image

After 72 h of continuous broth admission, the PEBR was disassembled and the antibiogram paper disks were carefully collected with sterile thin-tip forceps. They were disposed into wells of U-bottomed microtitration plates. Such wells were filled with 200 μL of 1 mg mL−1 MTT (in 150 mM PBS) [14]. After incubation for 5 h at 37 °C, the MTT solution was removed and the paper disks were carefully washed by immersion three times with 0.15 M PBS (2 mL each) to remove excessive MTT. Absolute isopropanol (1 mL) was then added to solubilize the MTT formazan product. MTT formazan formation was measured at 540 nm.

Thirty non-inoculated paper disks soaked in YEPD (72 h, 37 °C) served as controls. MTT processing was carried out as stated above.

Independent parallel experiments were carried out with 6-mm disks of 0.45-µm-pore-sized Immobilon-P™ membrane (Merck KGaA, Darmstadt, Germany) and 6-mm disks of ordinary cellophane paper (Emblema Ltd., São Paulo, Brazil). The results of such tests were unsatisfactory, since there was a great variation among repetitions (≥312 %). Thus, they were not followed through, and results concerning them are not presented here.

Experiments with fungal biofilms were carried out in five different occasions, performing 150 repetitions (5 days × 30 disks). Controls with non-inoculated paper disks were carried out in four different occasions, performing 120 repetitions (4 days × 30 disks).

Scanning Electron Microscopy

Biofilms were fixed with a fresh solution containing 4 % paraformaldehyde and 100 mM cacodylate buffer (pH 6.2), for 4 h at room temperature. They were dehydrated in a graded series of ethanol and critical point dried in CO2 (Mod. CPD-030 Critical Point Dryer, Bal-Tec AG, Balzers, Liechtenstein) and were mounted on aluminum stubs. In this step, six out of twelve disks were inverted-mounted, i.e., with the paper–disk interfaces exposed upwards. Disks were sputter-coated with gold (Mod. SCD030, Balzers AG, Balzers, Liechtenstein), and images were achieved in a Phenom™ Tabletop Scanning Electron Microscope (Phenom-World BV, Eindhoven, Netherlands). The entire surface of each sample was examined, and images that were representative of the sample were taken.

Statistics

For each test, descriptive statistics data (average, median, and standard deviation) were calculated. To test the data homogeneity, the Pearson’s coefficient of variation for each individual experiment was determined. Finally, to test whether the data had normal distribution, the Kolmogorov–Smirnov normality test was applied. The Tukey HSD test was used to access differences between MTT reductive activities in the biofilm-covered disks. A p value of 0.05 was considered as threshold.

Results

Descriptive statistical data for capillary diffusion of crystal violet revealed mean OD540nm values of 2.828 ± 0.070 for 0.1 % CV and 0.597 ± 0.072 for 0.01 % CV (Table 1).

Table 1 Descriptive statistics for capillary diffusion tests with crystal violet
Full size table

Pearson’s coefficients of variation were 2.467 and 12.055 %, respectively, revealing that in both situations, it had achieved good homogeneity of distribution. In both cases, the Kolmogorov–Smirnov (K–S) test revealed normality of data distribution (p > 0.05).

Mean values for MTT reduction were 1.638 ± 0.091 (C. albicans) and 0.113 ± 0.005 (Control). Pearson’s coefficients of variation were 17.157 and 4.315 %, respectively, revealing that in both situations, it had achieved reasonable homogeneity of distribution. In both cases, the Kolmogorov–Smirnov (K–S) test revealed normality of data distribution (p = 5.96E−9). The Tukey HSD test revealed significantly higher reductive activity in the biofilm-covered disks (p < 0.0001) (Fig. 3).

Fig. 3
figure 3

MTT reduction rate induced by C. albicans SC5314 biofilms grown using the PEBR

Full size image

The inspection of the upper surfaces of the paper disks revealed the growing of C. albicans biofilm rich in pseudohyphae with few true hyphal cells (Fig. 4). On the bottom paper disk surfaces, which remained in contact with the diffusive paper matrix, a high number of true hyphae were observed (Fig. 5). It is noteworthy that large amounts of budding yeast cells and pseudohyphae were seen inside the disk paper nets (Fig. 6).

Fig. 4
figure 4

Panoramic view of C. albicans SC5314 biofilm grown over antibiogram paper disk. In detail, amplified view showing an intense growth of pseudohyphae (PH) and few true hyphae (TH)

Full size image
Fig. 5
figure 5

Antibiogram disk/filter paper interface showing an intense growth of pseudohyphae (PH) and true hyphae (TH)

Full size image
Fig. 6
figure 6

Antibiogram disk/filter paper interface showing an intense growth of pseudohyphae within the antibiogram disc matrix

Full size image

Discussion

The necessity of a feasible system that generates biofilms under a constant nutrient flow with gas-exchanging option seems to be very attractive. Before the development of PEBR, we had considered growing biofilms onto Teflon-covered fiberglass screens (anti-mosquito screens) or synthetic fabrics. After this, the possibility of using sintered glass beads was evaluated, which hypothetically could favor hyphal growth. However, none of such substrates were suitable for generating consistent candidal biofilms.

The idea of using thick filter paper disks was promptly tested, but the problem of disk immersion in the broth flow remained. The system presented here supplies a laminar nutrient flow under low shear stress and mimics those situations in which nutrients come from the bottom layers of biofilms.

In the particular case of C. albicans, it is noteworthy to point out that hyphae can grow, invading the paper disks below them. This may become useful when an investigator is interested in evaluating hyphal growth.

Another favorable point of such a system is the possibility to grow biofilms under distinct atmospheric conditions. It is known that Candida spp. can grow in hypoxia [15, 16] and anoxia [17, 18], using parallel respiratory chain [19] and fermentative pathways [20]. As the PEBR has a gas admission port, the headspace atmosphere can be easily changed as desired.

If the inner glass table is properly leveled, the “center-to-border” diffusion of fluids allows a proximate amount of nutrients to reach the antibiogram disks in a homogeneous manner. This is an advantage directly related to final biofilm biomasses. To evaluate the distribution of water-soluble molecules, crystal violet (CV) was used. CV is a cationic basic dye with molar mass of 407.98 g mol−1 [21]. Its diffusion throughout the filter paper bed occurs without any “chromatographic-like arresting” in a way that CV molecules may flow freely through the paper matrix. It is interesting because many molecules used as nutrients for microbial growth, such as amino acids, sugars, and vitamins, are smaller than CV. Larger molecules that could be entrapped in the bed matrix can flow after the paper saturation, a phenomenon noticeable after 2–3 h post-broth admission. In favor of such a thesis, we obtained final biofilm biomasses with variations <5.76 %, even among sets of repetitions carried out in different days.

When handling biofilms, high-throughput capability is among the most required features [22–24]. This system seems to be very robust in offering such an advantage, since it is possible to grow up to 30 individual biofilms, per run, with low variation among their biomasses.

Besides the experimental advantages, it is also noteworthy to mention that its production costs were inferior to US$ 25.00 (without peristaltic pump and Kitasato flasks), with parts bought in local hardware stores. This is an important point of interest because an effective biofilm reactor with reasonable costs becomes very attractive to research groups with low budgets.

The main negative point noticed is that, in handling greater amounts of microbial biomasses, technical personnel must be aware of contamination risks. Thus, it is imperative to avoid underestimating the possibility of environmental or personnel contamination or neglecting the standard rules of laboratory biosafety.

To summarize, we presented here an inexpensive, reliable, and robust system that allows the obtaining of high-throughput biofilms with minor variations in biomasses due to the uniform nutrient diffusion in the filter paper bed.