Figure 1 : Dual-chamber microreactor fabrication and assembly. A) Components of chip and assembled device. B) PDMS chips fabrication process. C) Silicon gaskets fabrication process. D) Interface establishment - ALI was established when the bottom chamber was filled with liquid and the top chamber was filled with air, while the LLI was created when both chambers were filled with liquid.
Device sterilization
Before inoculation, the PDMS chips, acrylic Plexiglas covers, and silicon gaskets were sonicated with Milli-Q water for 480s, rinsed with an ethanol solution (80% v/v), and exposed to UV light for sterilization for 30 min. The PETE membranes and tubing were sterilized using the autoclave (121°C/ 20min). All the components were assembled aseptically in the biosafety cabinet.
Device inoculation and biofilm formation
P. aeruginosa from frozen stocks was grown on agar plates for 16-18h at 37°C. To prepare liquid pre-culture, one colony was transferred into 1mL of CAMHB media (3g/L beef extract, 17.5g/L casein hydrolysate, 1.5g/L starch, pH 7.0), incubated for 16-18h at 37°C, and shaken at 200 revolutions per minute (RPM). The overnight pre-culture was diluted 1:30 v/v in fresh media, incubated for 2h at 37°C, and shaken at 200 RPM. The inoculum size was optimized by testing two different standardized densities, which are OD600 = 0.04 and 0.4, respectively, to obtain a fully covered membrane within 48 h of growth. Before inoculation, the inoculum size controls were also obtained by viable colony forming units (CFU) counts.
For the device inoculation, 100μl of CAMHB media was injected into the basal chamber through the basal chamber inlet before 100μl of bacterial inoculum was injected into the top chamber through the top channel inlet. After inoculation, the excess media and inoculum were discarded by the tubing removal, and the chip was sealed with a sterile and breathable film (AeraSealTM) and incubated at 37°C for 2h under static conditions to allow bacteria attachment. Two different interface models were tested – ALI and LLI. For the ALI establishment, the culture media was removed from the apical chamber after 2h of bacterial adhesion. For LLI conditions, the apical media was removed after 2h of attachment and replaced by fresh culture media in the apical chamber. Biofilms were grown at 37°C for 48h, and the media on the basal chamber was refreshed after 24h of culture.
Device repeatability test
The repeatability of this microreactor was evaluated by quantifying biofilm growth under ALI and LLI models, using CFU counts. Three biological replicates were performed for biofilm formation on ALI and LLI. Two devices containing six flow cells were inoculated for each test model, and another device infused with media was used as controls. Controls for inoculum size were performed after each inoculation by viable colony counts to ensure that differences in the inoculum size were minimal. After 48h, each membrane was aseptically removed from the device, gently dip-washed in 1mL of sterile PBS, and subsequently transferred to a tube containing 1 mL of sterile PBS. To disrupt the biofilm, the membranes were sonicated for 280s at 47kHz and 1.8W/cm2, previously shown to cause no deleterious effect on the colony-forming ability of P. aeruginosa [18]. Following the disruption, ten-fold serial dilutions of the detached bacterial suspensions were performed in sterile PBS, plated on LB agar plates, and incubated at 37 for 16-18h. Viable colony counts were calculated using equation 1:
\(\frac{\text{CFU}}{\text{mL}=\frac{N\times 10}{10^{-D}}}\) Eq. (1)
where N represents the colony number and D represents the number of 1:10 dilutions. Six randomly selected membranes were analyzed by scanning electron microscopy (SEM) to confirm that bacteria had been removed from the surface.
Scanning Electron Microscopy (SEM)
The morphology of the biofilms grown on the devices was visualized using scanning electron microscopy (SEM, JEOL, JMC-6000 NeoScopeTM Benchtop SEM, USA) at 15kV. The devices were disassembled, and membranes gently dip-washed in sterilized water before fixation in a formaldehyde solution in PBS (4% v/v) for 90min. After fixation, membranes were gently rinsed with water and air-dried at room temperature. Before imaging, the membranes were sputter-coated with gold for 2min using a Smart Coater (JEOL, USA). At least three membranes from each experimental condition were analyzed.
Confocal laser scanning microscopy (CLSM)
The biofilm formed on different interfaces was investigated using FilmTracer™ LIVE/DEAD® Biofilm viability kit according to the manufacturer’s specifications. The samples were gently washed with sterilized water followed by staining with LIVE/DEAD solution for 25mins at 37°C in the dark. Excess dye was removed using sterilized water. The biofilms were then fixed for 40 min by infusing PFA solution (4% v/v in PBS). The sample was rinsed with sterile water and mounted on a glass slide. Fluorescence was observed using a confocal laser scanning microscope (Olympus FluoView, inverted FV 3000RS IX83, 100× magnification oil immersion objective). The image was acquired using 488 nm and 660nm incident light. The z-stack images were taken with a step of 1μm, and the resolution was 1,024×1,024 pixels. All experiments were performed in duplicate. 3D reconstruction view and 3D z-stack projection images were obtained, and fluorescence intensity measured using ImageJ. The ratio of live cells to dead cells was compared using the fluorescent intensity ratio of green to red in 3-D projected z-stack CLSM images to quantify the difference between biofilm formed on different interfaces. For analysis, the images were obtained from two replicates which three images pick on each replicate. Thus, six images in total for each biofilm model were collected for statistical analysis.
Extracellular polymeric substances (EPS) staining
Alcian blue staining was used as a colourimetric-based approach to detect the polysaccharides in the extracellular polymeric substances [19]. The membrane was aseptically removed from the device, gently washed with sterilized water, and then fixed using 4% (v/v) formaldehyde solution in PBS for 40min. The membrane was rinsed with water again and then immersed in Alcian blue (1% w/v in acetic acid, 3% v/v pH 2.5) for 15mins. Excess dye was removed by multiple rinses with water. Images were obtained using NanoZoomer-SQ Digital slide scanner C13140-01 (Hamamatsu). Eight images were collected at 40x magnification from each sample, and the mean red, green and blue (RGB) ratio of Alcian blue and standard deviation were obtained using ImageJ.
Antibiotic Treatment Test
Ciprofloxacin (CIP) is a broad-spectrum antibiotic that is effective against PAO1[20] and hence has been used as the model antibiotic in this study. To assess the effects of the different interface on P. aeruginosa susceptibility to CIP, the 48h-old biofilms were exposed to a range of concentrations (50-1600μg/mL) for 6h. For the antibiotic treatments, a stock solution of CIP was prepared in water, and working solutions were freshly prepared by dilution of the stock in culture media. The antibiotic solution was injected into the basal chamber, and the biofilm interacts with antibiotics from the substrate side for 6h at 37°C. The susceptibility was assessed by CFU counts according to the procedure described in section 2.6.
Statistical Analysis
The repeatability of culturing biofilms at different interfaces was examined by calculating the standard deviations across all experiments. All data are expressed as Log (CFU/mL). Unpaired t-tests were performed to determine the significant differences between the biofilm formed at ALI and LLI. The antibiotic susceptibility of biofilm formed on ALI and LLI was investigated using one-way ANOVA. Graphs show means and error bars indicate the standard error of the mean. All the statistical analysis was performed using Graphpad Prism 7.0.
Results and Discussion
An adaptable dual-chamber microreactor has been developed incorporating different interfaces and surfaces to investigate the effects on biofilm development in terms of morphological properties, colony counts, and responses to antibiotics. Compared to conventional mono-interface reactors, the dual-chamber device offers significant advantages in mimicking the relevant physiologic and environmental settings in which bacteria thrive, offering the potential for more targeted experiments to deliver precise outcomes.
Device fabrication and assembly
The fabrication process of PDMS chips using 3D printing illustrated in Fig. 1-C, shows the versatility of the microfluidic system with affordable equipment and straightforward protocols [21]. PDMS is the ideal material for the fabrication of microfluidic devices having great biocompatibility, chemical stability, gas permeability, transparency, low cost and ease of molding [22]. The intermediary gaskets (Fig. 1-D) fabricated with silicone (a non-toxic, inert, and soft material) were introduced to cushion the mechanical force created by the screw and nuts and to prevent leakage from the flow cell. The assembled dual-chamber device comprises six layers (Fig. 1-A), with two acrylic Plexiglas cover layers, two PDMS layers containing the flow chamber, and middle layers of gaskets. The acrylic Plexiglas covers help distribute the bonding force evenly on the PDMS chips to seal the system. Several bonding techniques have been used in microfluidic devices such as surface plasma and corona discharge that permanently bond PDMS layers with glass or other PDMS layers [23]. However, these techniques require specialized equipment. Moreover, several end-point sample analysis methods require the device to be disassembled, which could potentially disrupt the biofilm matrix, and consequently, the use of an irreversibly bonded device is limited. For analysis that requires many replicates, using permanently bonded devices could also be both costly and time inefficient. Therefore, the reversible mechanical bonding developed in our device will offer more flexibility and convenience for endpoint sample manipulation. The ability to fit different membranes in the same device enables the testing of different surfaces, and the device can be reused after sterilisation to reduce the cost and time of production. The establishment of the ALI and LLI methods is illustrated in Fig.1-D using red and blue dyes. The ALI method was established when the bottom chamber was filled with liquid and the top chamber filled with air, while the LLI was created when both chambers were filled with liquid.
Optimization of inoculum concentration and biofilm attachment time
Inoculum size optimization was performed to establish the formation of a mature biofilm throughout the membrane in 48h. The investigated inoculum density was standardized to OD600=0.04 and OD600=0.4 with inoculation times of 2h and 24h. Fig. 2 presents the microphotographs of the biofilms grown from different inoculum densities and with different attachment periods. Biofilms from low-density inoculum (OD600=0.04) with 2h attachment time were scattered throughout the membrane, while biofilms from the same density with attachment time of 24h, completely covered the membranes. Biofilms grown from high-density inoculum (OD600=0.4) formed a homogeneous layer on the membranes after 2h of attachment. This condition was chosen as the optimal parameter to carry out the experiments throughout this study.