We describe the assembly, operation, and cleaning of a flow apparatus designed to image fungal biofilm formation in real time while under flow. We also provide and discuss quantitative algorithms to be used on the acquired images.
In oropharyngeal candidiasis, members of the genus Candida must adhere to and grow on the oral mucosal surface while under the effects of salivary flow. While models for the growth under flow have been developed, many of these systems are expensive, or do not allow imaging while the cells are under flow. We have developed a novel apparatus that allows us to image the growth and development of Candida albicans cells under flow and in real-time. Here, we detail the protocol for the assembly and use of this flow apparatus, as well as the quantification of data that are generated. We are able to quantify the rates that the cells attach to and detach from the slide, as well as to determine a measure of the biomass on the slide over time. This system is both economical and versatile, working with many types of light microscopes, including inexpensive benchtop microscopes, and is capable of extended imaging times compared to other flow systems. Overall, this is a low-throughput system that can provide highly detailed real-time information on the biofilm growth of fungal species under flow.
Candida albicans (C. albicans) is an opportunistic fungal pathogen of humans that can infect many tissue types, including oral mucosal surfaces, causing oropharyngeal candidiasis and resulting in a lower quality of life for affected individuals1. Biofilm formation is an important characteristic for the pathogenesis of C. albicans, and numerous studies have been done on the formation and function of C. albicans biofilms2,3,4,5, many of which have been conducted using static (no flow) in vitro models. However, C. albicans must adhere and grow in the presence of salivary flow in the oral cavity. Numerous flow systems have been developed to allow for live-cell imaging6,7,8,9,10. These different flow systems have been designed for different purposes, and therefore each system has different strengths and weaknesses. We found that many of the flow systems appropriate for C. albicans were costly, required complex fabricated parts, or could not be imaged during flow and had to be stopped prior to imaging. Therefore, we developed a novel flow apparatus to study C. albicans biofilm formation under flow11. During the design of our flow apparatus, we followed these major considerations. First, we wanted to be able to quantify multiple aspects of the biofilm growth and development in real-time without requiring the use of fluorescent cells (allowing us to study mutant strains and unmodified clinical isolates easily). Second, we wanted all parts to be commercially available with little to no modifications (i.e., no custom fabrication), allowing others to more easily recreate our system, and allowing for easy repairs. Third, we also wanted to allow for extended imaging times at reasonably high flow rates. Lastly, we wanted, following a period of cells attaching to the substrate under flow, to be able to monitor the biofilm growth over an extended time without introducing new cells.
These considerations led us to develop the two-flask recirculating flow system illustrated in Figure 1. The two flasks allow us to split the experiment into two phases, an attachment phase that draws from the cell-seeded attachment flask, and a growth phase that uses cell-free media to continue the biofilm growth without the addition of new cells. This system is designed to work with an incubation chamber for the microscope, with the slide and the tubing preceding it (2 to 5, Figure 1) being placed inside the incubator, and all other components placed in a large secondary container outside the microscope. Additionally, a hotplate stirrer with an attached temperature probe is used to maintain fungal cells in the attachment flask at 37 °C. As it is recirculating, this system is capable of continuous imaging during flow (can be over 36 h depending on conditions), and can be used on most standard microscopes, including upright or inverted benchtop microscopes. Here, we discuss the assembly, operation, and cleaning of the flow apparatus, as well as provide some basic ImageJ quantitative algorithms to analyze the videos after an experiment.
1. Assemble the Flow Apparatus
2. Perform an Experiment
3. Clean the Flow Apparatus
4. Quantifying the Videos
NOTE: All files need to be converted to the tag image file (TIF) format to work. Additionally, to compare between experiments, it is critical that all images are taken with the same microscope and imaging parameters, as discussed above.
Representative images of a normal overnight time-lapse experiment using wild-type C. albicans cells at 37 °C can be seen in Figure 2A and Supplemental Video 1. The images have been contrast enhanced to improve visibility. Quantification of the original data was performed, and representative graphs can be seen in Figure 2B. To generate these graphs, the data were first normalized to the imaging area (i.e., divided by the total imaging area), and the detachment was further normalized to the biomass, as described above. Additionally, the attachment and detachment show the cumulative values over time, rather than the individual frame values generated by the flow biofilm quantification macro. Once the graphs have reached this stage, statistical comparisons can be performed through regression analyses.
Figure 1: Schematic of the two-phase recirculating flow apparatus. Connecting black lines indicate tubing, and arrowheads indicate the direction of flow during normal operation. (A) A general schematic overview of the flow system is illustrated. For convenience, the flow system is divided into a green side (upstream of slide) and an orange side (downstream of slide). Bold numbers correspond to parts listed in the Table of Materials. Labels for valves simply mark the location for tubing clamps or hemostats to be placed during experiments. Filter order is as follows: 8–20 µm inline filter, 9–10 µm inline filter, 10–2 µm filter bottle (FB), and 11–0.22 µm single use disposable filter. Schematic is not to scale. (B) A close-up view of the rubber stopper for the attachment flask, illustrating the four components that pass through the ports: the media outlet, the 0.2 µm air filter that allows gas exchange, the temperature probe (requires drilling an extra hole), and the media return. (C) A close-up view of the pulsation dampener (PD) and the FB, as well as the screw cap assembly used for each port. These bottles need to be air-tight to function. The outlet tubing for the PD should reach deeper into the bottle than the inlet tubing for proper functioning. The gray rectangle in the FB represents the steel filter. Please click here to view a larger version of this figure.
Figure 2: Candida albicans wild-type cells grown under flow at 37 °C. (A) Representative darkfield microscopy images of the microcolonies that form under flow at 37 °C at the indicated time points. Scale bar = 100 µm. (B) Representative image quantification data. The total biomass within the imaging region (determined by densitometry analysis), the cumulative rate of cell attachment, and the percent biomass detachment (detachment rate normalized to the biomass) over time are shown for each strain. Data are means of n ≥ 3 experiments. Please click here to view a larger version of this figure.
Supplemental Video 1. Candida albicans wild-type cells grown under flow at 37 °C. This time-lapse darkfield microscopy video shows the attachment of WT cells to the substrate during the attachment phase (time indicated in the upper left-hand corner; images acquired every 2 min), followed by the subsequent growth and development during the growth phase (starts at 2 h; images acquired every 15 min). Cell-seeded media (1 x 106) were used during the attachment phase, while cell-free media were used during the growth phase. Flow is from the right to left. Scale bar = 50 µm. Please click here to view this video. (Right-click to download.)
Using the flow system as outlined above allows for the generation of quantitative time-lapse videos of fungal biofilm growth and development. To allow for comparisons between experiments it is of critical importance to ensure that the imaging parameters are kept the same. This includes ensuring that the microscope is set up for Köhler illumination for each experiment (many guides are available online for this process). Aside from imaging parameters, there are some important steps to keep in mind when working with the flow apparatus. First, it is important to ensure that the bubble trap is maintained under vacuum during fluid flow, as failure to do so will lead to air being pulled in through the bubble trap. Similarly, when the bubble trap is not under vacuum (i.e., when transporting the flow apparatus) both the inlet and outlet must remain clamped shut; otherwise air will leak in through the PTFE membrane. This clamp does not need to be removed until the bubble trap is once again placed under vacuum. Lastly, it is very important to monitor your flow apparatus for potential clogs or leaks. The most efficient way to check for clogs in your system is to check that there are media dripping from the inlets of the attachment or growth flask, the PD, and the FB. The media from these should be dripping at relatively similar rates if everything is operating smoothly. If a clog is present, you can generally determine the location as the tubing just upstream of the clog will be more rigid.
Once the data have been obtained, we provide numerous ImageJ macros to quantitate the videos. These macros determine multiple parameters of biofilm growth and development, including a measure of the biomass, and the rate that cells attach to and detach from the surface or biofilm. Descriptions of the provided macros are provided below.
Complete analysis performs all the analyses listed below, and automatically outputs the data. This macro can be executed without an open image file while all others require an open video. When executed, it will prompt the user to open an attachment stack file, then a growth stack file. Following this, it will automatically analyze the images and output a data folder containing all the data tables, as described below, to the same folder as the attachment image file. The attachment counter macro is performed only on the attachment file; all other analyses are performed on a concatenated stack of the attachment and growth files. The output data files generated are text files, but should be imported into excel for ease of use.
The sum intensity analysis will analyze each frame of the active window. It adds up all gray values for each pixel that is above the lower threshold designated in step 4.3.7, and outputs one cumulative value per frame. The values generated are proportional to the biomass present within the frame, up until any camera saturation that occurs. The data should then be normalized to the area of the imaging region; this is not performed by the macro.
The coverage area analysis will analyze each frame of the active window for the area of the frame that is covered by cells (above the lower threshold value) as a percentage.
The attachment counter will use frame subtraction to determine the sum intensity of all the cells that attach between each frame. Thus, the first data point is the biomass of the cells that attach between frames 1 and 2; the second data point is the biomass of the attaching cells between frames 2 and 3, etc. These data should be normalized to the area of the imaging region. For easier readability, it is also helpful to integrate this value prior to graphing.
The detachment counter works the same as the attachment counter, but reverses the frame subtraction, such that it determines the sum intensity of the cells that detach between each frame. These data should also be normalized to the area of the imaging region, and integrated prior to graphing. Prior to integration, these data can be further normalized to the total sum intensity of the preceding frame calculated in the sum intensity analysis. This new value represents the proportion of cells that detach from the biofilm at that time point, which is often more valuable data, since the biomass of the detaching cells will increase with increasing biofilm biomass.
While the flow system presented here is more complicated to build and operate than other flow systems, it does offer several advantages. Many of these advantages result from our use of a commercially available channel slide. The tissue culture treatment available for these slides is sufficient to allow Candida cells to adhere to the surface. Additionally, the profile of this channel slide being similar to a traditional slide allows it to easily be used on a wide variety of microscope systems, including upright microscopes using transmitted light at low magnification. Using this type of microscope allowed us to use darkfield microscopy, which made the quantification of data much easier, especially compared to fluorescent microscopy (as there was no photobleaching and low phototoxicity). Traditional microscopy (without optical sectioning), is power conservative, meaning out of focus cells contribute similar numbers of photons to an image as in focus cells of similar size12. This means that, despite our single plane of imaging, the full 3D growing biofilm is still being quantified throughout the experiment, even though the higher regions are out of focus. This single-plane imaging has the advantage of dramatically lowering the phototoxicity damage to cells, but does not provide any information on the 3D architecture of the biofilm. However, this flow system can also be used with fluorescent cells and confocal microscopes to obtain this information13.
The unique two flask recirculating setup of our flow apparatus also has many advantages. First, many flow systems require that the slides be pre-seeded with cells, however our use of a separate cell-seeded attachment flask allows us to image and quantify cells as they adhere to the slide while under flow, and we feel that this is more similar to what occurs in vivo. Additionally, we have previously been able to adjust our microscope for high-speed imaging and image adhesion events as they occurred in real-time, as opposed to quantifying them after the fact11. Second, having a cell-free growth flask that recirculates and can be maintained cell-free over an extended duration allowed us to understand how biofilms grow under flow for over 24 h, a duration that cannot typically be accomplished with non-recirculating systems. We have not yet determined the upper time limit of what can be achieved with our flow system, but we have successfully completed 36 h experiments; however, the longer the experiment, the greater the chance of a leak or clogged filter. Numerous factors can affect the potential duration of an experiment, including the growth rate of the cells, how adhesive they are, and the degree of hyphae formation, making it difficult to define an upper limit on the duration of an experiment. However, if much longer durations are desired than can be achieved with the flow apparatus as presented, the filters can be replaced with an in-line ultraviolet (UV) sterilization box as has been previously described8. This sterilization box may also allow this flow apparatus to be used to image bacteria; our previous attempts to image bacterial strains resulted in rapid clogging of the 0.2 µm filter. Ultimately, we opted not to adopt UV sterilization, as the box is custom fabricated, and as this would result in recirculating dead cells.
Another advantage of this flow system is that it is reasonably inexpensive relative to commercial systems, especially if you need to purchase a microscope with it. In our lab, we were able to purchase a basic transmitted light benchtop microscope and place the entire microscope inside a large standard convection incubator. The only major requirement is that the microscope should have a shutter function (either mechanical or electrical) in order to perform time-lapse microscopy.
While this system is versatile and offers many advantages, it is a low throughput method. Our flow apparatus is unable to grow multiple strains in parallel, unlike other available flow systems. Due to the extensive preparation and cleaning time, we are only able to perform two experiments a week. However, many other flow systems are rather costly, and may clog when Candida cells are grown under hyphae forming conditions.
Additionally, this flow system is quite complex compared to others, and can be difficult to keep in operation. After many experiments, filters begin to clog, tubing begins to wear thin, and parts start to rust or become loose; thus requiring these components to be replaced. The use of filters makes this system incompatible with growth conditions of some fungal strains; in particular, anything that induces flocculation will rapidly clog the 20 µm in-line filter. However, with sufficient experience using the flow system, it becomes easier to detect potential issues before they result in a failed experiment. One thing that can be done to make the everyday operation of the flow apparatus a little simpler is to have a machinist make a replica of the bubble trap housing out of an autoclavable material (such as aluminum or stainless steel), allowing you to autoclave the bubble trap with the rest of the flow apparatus, as the PTFE membrane and adapter components of the bubble trap are autoclavable.
In conclusion, the two-phase recirculating flow apparatus presented here represents a unique model to image and quantify in vitro biofilm formation of fungi under flow and in real-time. While the system has its limitations, it is highly adaptable and works well with most microscopes.
The authors have nothing to disclose.
The authors would like to acknowledge Dr. Wade Sigurdson for providing valuable input in the design of the flow apparatus.
Pump | Cole Parmer | 07522-20 | 6 |
Pump head | Cole Parmer | 77200-60 | 6 |
Tubing | Cole Parmer | 96410-14 | N/A |
Bubble trap adapter | Cole Parmer | 30704-84 | 3 |
Bubble trap vacuum adapter for 1/4” ID vacuum line | Cole Parmer | 31500-55 | 3 |
In-line filter adapter (4 needed) | Cole Parmer | 31209-40 | 8,9 |
Orange-side Y | Cole Parmer | 31209-55 | 7 |
Green-side Y | ibidi | 10827 | 2 |
* Slides | ibidi | 80196 | 4 |
* Slide luers | ibidi | 10802 | 4 |
Vacuum assisted Bubble trap | Elveflow/Darwin microfluidics | KBTLarge – Microfluidic Bubble Trap Kit | 3 |
Media flasks | Corning | 4980-500 | 1 |
0.2 µm air filter | Corning | 431229 | 1 |
Threaded glass bottle for PD and filter flask (2 needed) | Corning | 1395-100 | 5,10 |
Ported Screw cap for PD and filter flask (2 needed) | Wheaton | 1129750 | 5,10 |
Screwcap tubing connector | Wheaton | 1129814 | 5,10 |
Tubing connector beveled washer | Danco | 88579 | 5,10 |
Tubing connector flat washer | Danco | 88569 | 5,10 |
Clamps for in-line filters and downstream Y (7 needed) | Oetiker/MSC Industrial Supply Company | 15100002-100 | 7,8,9 |
Clamp tool | Oetiker/MSC Industrial Supply Company | 14100386 | N/A |
20 micron in-line media filter | Analytical Scientific Instruments | 850-1331 | 8 |
10 micron in-line media filter | Analytical Scientific Instruments | 850-1333 | 9 |
2 micron inlet media filter | Supelco/Sigma-Aldrich | 58267 | 10 |
* 0.22 µm media filter | Millipore | SVGV010RS | 11 |
* 0.22 µm media filter “adapter” | BD Biosciences | 329654 | 11 |
Rubber stopper | Fisher Scientific | 14-131E | 1 |
Hotplate stirrer with external probe port | ThermoFisher Scientific | 88880006 | N/A |
Temperature probe | ThermoFisher Scientific | 88880147 | N/A |