Biomimetic Tri-Culture Models to Study Gut Microbiota- Epithelial-Immune Interactions In Vitro

Size: px
Start display at page:

Download "Biomimetic Tri-Culture Models to Study Gut Microbiota- Epithelial-Immune Interactions In Vitro"

Transcription

1 Biomimetic Tri-Culture Models to Study Gut Microbiota- Epithelial-Immune Interactions In Vitro We developed an in vitro model to study human gut microbiota-epithelial-immune interactions in Transwell culture inserts (Figure 1). For this, we formed co-cultures of epithelial (i.e., CBBE1 and HT29-MTX) and immune (i.e., Dendritic cells, DC) cells to mimic the lining of the intestinal epithelium. We then incorporated four bacterial species (i.e., Escherichia coli, Lactobacillus rhamnosus, Bacteroides fragilis, and Ruminococcus gnavus) that are representative of microbial populations in the human gut. These results demonstrated that all bacterial species could efficiently grow in the apical compartment of the co-cultures. Among the four species tested, only L. rhamnosus did not lead to a significant disruption of the integrity of the epithelial monolayer. Furthermore, exposure of epithelial-immune co-cultures to microbe-free spent media resulted in significant increases in the concentration of reactive oxygen species (ROS). This culture platform could be used to study alterations in the functional relationships between bacterial, immune, and epithelial cells, which are associated with the onset of a variety of diseases, including anxiety and depression. Moreover, these results demonstrated the potential of L. rhamnosus to be used as a psychobiotic, owing to its ability to colonize the human gut without disrupting the barrier function on the intestinal epithelium. Figure 1. Biomimetic model of the human gut. The schematic shows the disposition of the different cell types that comprise the culture system. Epithelial (i.e., C2BBe1 and HT29) and immune (i.e., dendritic) cells were cultured on Transwell inserts to mimic the lining of the gut. Four representative species of bacterial populations in the gut (i.e., E. coli, L. rhamnosus, B. fragilis, and R. gnavus) were then incorporated into the co-cultures to study the influence of microbial communities on the function of epithelial and immune cells in the gut. The apical and basolateral compartments of the Transwell culture systems are shown.

2 Results and Discussion In the physiological context, the epithelial lining of the intestine constitutes a physical barrier that prevents the permeation of noxious agents into the bloodstream. Pathogenic bacteria could trigger inflammatory responses that compromise the barrier function of the epithelium, which has been associated with the development of various diseases including anxiety and depression. Therefore, we aimed to develop an in vitro model of the human gut that could be used to study the interactions between the microbiota, the immune system, and the intestinal epithelium. Figure 2. Bacterial growth kinetics in the Transwell system. (a) Bacteria maintained in apical media in the absence of mammalian cells and in aerobic conditions exhibited comparatively lower growth rates. (b) All species of bacteria grown in the presence of mammalian cells exhibited similar growth kinetics, which were in the range of CFU/ml at 48 hours post-seeding. Epithelial co-cultures were formed using C2BBe1 and HT29-MTX cells. C2BBe1 cells are enterocytes that form a polarized monolayer with an apical brush border that is morphologically similar to that of the human colon. On the other hand, HT29-MTX cells are epithelial cells that form confluent monolayers with discernable tight junctions, which are characterized by their mucus-secreting phenotype. Mature dendritic cells, which were differentiated from human peripheral blood mononuclear cells, are immune cells whose main function is the processing and presentation of antigens to T cells to trigger immune responses. Single-microbe experiments were then carried out using four different species of bacteria that are normally present in the human gut. These four species were chosen to span a variety of phenotypic characteristics, such as motility, oxygen requirements, and Gram status. This strategy enabled the accurate in vitro representation of a mucusproducing epithelial barrier with an immune component, as well as the bacterial species that comprise the normal human microbiota.

3 The characterization of the bacterial growth kinetics revealed that bacteria cultured in the absence of mammalian cells under aerobic conditions exhibited growth rates in the range of 104 CFU/ml (Figure 2a). In contrast, all four species grown in the presence of mammalian cells exhibited final microbial concentrations in the range of CFU/ml. Moreover, there were no statistically significant differences in the ability of the four species to colonize the apical compartment of the Transwell inserts (Figure 2b). Figure 3. In vitro evaluation of epithelial barrier integrity. Fluorescent calcein AM staining of epithelial monolayers growing on Transwell inserts (a) in the absence of bacteria (control), and in the presence of (b) L. rhamnosus and (c) E. coli. Green fluorescence is indicative of viable (live) cells. Scale bars = 1 mm. Although there were no statistically significant differences in the growth rates of the four species tested, their effect on the integrity of the epithelial monolayer was radically different. Fluorescent calcein AM staining was used to visualize viable (live) epithelial cells growing on the surface of the Transwell inserts in the absence of bacteria (control, Figure 3a). These results demonstrated that L. rhamnosus did not exert any detrimental effect on the integrity of the epithelial monolayers (Figure 3b). In contrast, incubation in the presence of E. coli led to significant cell death and delamination of the monolayer, which significantly compromised the barrier function of the epithelium. The increase in membrane permeability triggered by bacterial colonization were further confirmed through a Lucifer Yellow permeability assay. The permeability of the epithelium was assessed before and after 48 hours of incubation in the presence of the four bacterial species tested. These results demonstrated that incubation with E. coli, B. fragilis, and R. gnavus led to 5.5 to 12.2-fold increases in apparent monolayer permeability (Figure 4). In contrast, incubation with L. rhamnosus did not lead to any significant differences in monolayer permeability, when compared to monolayers incubated in the absence of bacteria (Figure 4). Taken together, these results demonstrated the remarkable potential of L. rhamnosus to be used as a vehicle for the delivery of therapeutic agents to the gut. This is mainly because of the ability of the L. rhamnosus to colonize the intestinal epithelium

4 without triggering inflammatory responses from the immune system that could compromise its barrier function. Figure 4. In vitro evaluation of changes in epithelial monolayer permeability. The apparent changes in the permeability of epithelial monolayers growing on Transwell inserts were assessed using a Lucifer Yellow assay before and after 48 hours of incubation with different bacterial species. Asterisks denote statistically significant changes after 48 hours of incubation with bacteria, compared to the permeability before exposure and to the unexposed controls (p < 0.05). The ability of pathogenic microorganisms to trigger cytotoxic responses has been attributed not only to the presence of live bacteria but to an increase in the production of ROS by epithelial and immune cells. Therefore, we aimed to characterize the presence of ROS in the culture media of co-cultures incubated with microbe-free spent media using a chemiluminescent assay. These results demonstrated that incubation in the presence of microbe-free media from all four bacterial species led to a significant increase in the concentration of ROS in the culture media (Figure 5). However, incubation with media from L. rhamnosus cultures was corresponded to the lowest concentration of ROS observed, when compared to the other three species. Moreover, the highest levels of ROS corresponded to incubation of the monolayers with spent media from E. coli cultures. These results suggested that the extensive cell death and monolayer delamination observed after

5 48 hours of incubation with E. coli occurred due to a significant increase in the concentration of ROS in these cultures. These results indicated that soluble agents secreted by live bacteria could trigger inflammatory responses in epithelial and dendritic cells that lead to an increase in oxidative stress and the subsequent cell death. Figure 5. In vitro evaluation of reactive oxygen species in Transwell cultures. The concentration of ROS produced after incubation of epithelial monolayers with different species of bacteria was determined using a chemiluminescent assay. Asterisks denote significant differences from the unexposed controls (p < 0.05). Conclusion Static co-cultures of individual microbial species and mammalian epithelial/immune cells in Transwell inserts are promising alternatives to model complex physiological phenomena in vitro. These results suggested that the growth of bacteria in low nutrientand aerobic environments is heavily dependent on mammalian cells. However, the coculture of certain bacterial species has a significant detrimental effect on the viability of human epithelial cells in vitro. This occurs mainly by increasing the permeability of the monolayer and triggering the production of ROS, which in turn leads to cell death and delamination of the epithelial barrier (Figure 6). Taken together, these results demonstrated that this culture platform could be used to study the immunomodulatory properties of live microorganisms in a biomimetic model of the human gut. Furthermore, these results also highlight the remarkable potential of L. rhamnosus to be used as a therapeutic agent that can target the gut-brain axis. This is mainly because this microorganism can efficiently colonize the human gut without compromising the integrity of the intestinal epithelium. In

6 addition, the intrinsic psychobiotic properties of L. rhamnosus as well as its role in the prevention of a variety of disease make it a highly attractive alternative for the development of Lactobachill. Figure 6. Differences in between intact and compromised epithelial barriers. The schematic shows the increases in non-specific paracellular traffic between intact and compromised epithelia. Pathogenic microorganisms could impair the normal barrier function of the epithelium (left) due to an increase in the production of ROS and the subsequent increase in cell death (right). Experimental procedures Establishment of bacterial/mammalian tri-cultures in Transwell inserts C2BBe1 and HT29-MTX cells were seeded on 24-well Transwell inserts at a 9:1 ratio and maintained in culture for 21 days. Dendritic cells were differentiated from human peripheral blood monocytes for 7 days and seeded on the underside of the Transwell inserts on day 14. Apical media was replaced with PBS containing Ca2+ and Mg2+ supplemented with 10mM HEPES on day 20. Cryopreserved bacterial stocks were thawed on day 20 and grown in MOPS microbe media in an anaerobic chamber. Bacteria were then counted and introduced to the Transwell cultures at a concentration of 6.7 x 104 CFU/mL on day 21. Bacteria were allowed to attach for 2 hours.

7 The apical and basolateral culture media were replaced with sterile media, which was deoxygenated overnight in an anaerobic chamber. The media was replaced every 3 hours with 9-hour overnight periods for 48 hours. Bacterial populations were monitored via standard spot plating in agar at 0, 15, 24, 39, and 48 hours. The viability of the epithelial cells was evaluated using a commercial Lucifer Yellow permeability assay and a LIVE/DEAD assay, before and after the introduction of bacteria to the cultures. The evaluation of the concentration of ROS in the culture media was carried out using a chemiluminescent Acridan assay. Statistical analysis was conducted using ANOVA and paired t-tests. Spot plating Prepare media plates, supplemented with and without antibiotic(s) of interest at desired concentrations. Inoculate an overnight culture and incubate overnight. The next day, determine optical density of diluted cultures and convert into CFU/mL using the conversion factor of the strain if known (example: 1.00 OD600 to 8x108 cells/ml). Serially dilute cultures (e.g., 100 μl of culture into 900 μl of fresh growth media) to obtain a 1x103 CFU/mL culture. Mix or vortex well between dilutions. Aseptically pipette 8 separated spots onto the plate using 20 μl of 1x103 CFU/mL culture. For the control, prepare a 1x10 2 CFU/mL culture from the 1x10 3 CFU/mL culture. Spot plate 2 spots to a new plate with no antibiotics by pipetting 200 μl of the 1x10 2 CFU/mL culture. Let the spots dry completely before incubating the plates overnight at 37ºC. After incubation, count the isolated colonies per spot. Calcein AM fluorescent staining The calcein AM staining provides a simple, rapid, and accurate method to measure cell viability and/or cytotoxicity. Calcein AM is a non-fluorescent, hydrophobic compound that easily permeates intact, live cells. The hydrolysis of Calcein AM by intracellular esterases produces calcein, a hydrophilic, strongly fluorescent compound that is well-retained in the cell cytoplasm.

8 Protocol Tri-cultures are established as described above. Discard the media supplement and add 100 ml of 1X Calcein AM DW Buffer. Remove the 100 ml of 1X Calcein AM DW Buffer and replace with 50 ml of fresh 1X Calcein AM DW Buffer. It is important to remove any carry-over media, as phenol red and serum will interfere with the sensitivity of the assay. Add 50 ml per well of freshly prepared 2X Calcein AM Working Solution. Incubate for 30 minutes at 37 C under CO 2. Record fluorescence using a 490 nm excitation filter and a 520 nm emission filter. The fluorescence intensity is proportional to the number of viable cells. Measurement of Cell Monolayer Integrity using Lucifer Yellow The permeability of C2BBE1 cells can be evaluated by measuring the passive passage of different molecules across the monolayer. Small hydrophilic compounds such as Lucifer Yellow are able to cross the monolayer through the paracellular space (e.g., tight junctions). The passage of this marker can be used to determine whether a given experimental condition can disturb the integrity of the monolayer. Protocol After removing samples for sample analysis, aspirate the remaining liquid from the apical and basal wells. Add 500 μl of 0.1 mg/ml Lucifer Yellow Solution to the apical wells and 1,000 μl of Buffer B to the basal wells. Incubate at 37 C for 60 minutes. Transfer 150 μl from the basal wells to a 96 well plate and read in a spectrofluorometer (excitation: 485 nm, emission: 535 nm). Fluorescence of the Buffer B (blank) and the 0.1 mg/ml Lucifer Yellow Solution should be read and used to normalize the readings. Calculate the percent permeability from the fluorescence values as follows: % permeability = (sample blank) / (Lucifer Yellow blank) 100 Acridan Lumigen PS-3 assay The Acridan Lumigen assay can be used to measure the concentration of ROS in culture media using a 96-well plate format. For this, we used the Acridan Lumigen PS-3 reagent from the Amersham ECL Plus kit (GE Healthcare) as a chemiluminescent substrate for the enzyme horseradish peroxidase (HRP).

9 Protocol Culture media samples were collected and stored at -20 C. Allow solutions A (H 2O 2 in Tris buffer) and B (acridan solution in dioxane and ethanol) to come to room temperature (Approximately 1 hour for 100 ml) Gently invert (4-5 times) solutions A and B in their packaged containers to assure homogeneity prior to dispensing. Avoid vigorous agitation of reagent. Dispense the amount needed of solutions A and B into separate containers. Containers should be opaque or covered with aluminum foil to protect from direct light (daylight or artificial). To prepare the working solution (ALPS-3 substrate), mix solutions A and B in the ratio of 40:1 in a new container. Each culture media sample is placed in an individual well of a 96-well plate. The final volume of each sample is adjusted to 100 µl with PBS (ph 7.4), followed by 50 µl of the ALPS-3 substrate. The plate is incubated in the dark for 5 min at room temperature. Light intensity is measured using a BioTek Synergy 2 plate reader and Gen5 software. Readings recorded by the BioTek plate reader are used for analysis.