P. aeruginosa Infected 3D Co-Culture of Bronchial Epithelial Cells and Macrophages at Air-Liquid Interface for Preclinical Evaluation of Anti-Infectives

Summary

We describe a protocol for a three-dimensional co-culture model of infected airways, using CFBE41o^- cells, THP-1 macrophages, and Pseudomonas aeruginosa, established at the air-liquid interface. This model provides a new platform to simultaneously test antibiotic efficacy, epithelial barrier function, and inflammatory markers.

Abstract

fDrug research for the treatment of lung infections is progressing towards predictive in vitro models of high complexity. The multifaceted presence of bacteria in lung models can re-adapt epithelial arrangement, while immune cells coordinate an inflammatory response against the bacteria in the microenvironment. While in vivo models have been the choice for testing new anti-infectives in the context of cystic fibrosis, they still do not accurately mimic the in vivo conditions of such diseases in humans and the treatment outcomes. Complex in vitro models of the infected airways based on human cells (bronchial epithelial and macrophages) and relevant pathogens could bridge this gap and facilitate the translation of new anti-infectives into the clinic. For such purposes, a co-culture model of the human cystic fibrosis bronchial epithelial cell line CFBE41o^- and THP-1 monocyte-derived macrophages has been established, mimicking an infection of the human bronchial mucosa by P. aeruginosa at air-liquid interface (ALI) conditions. This model is set up in seven days, and the following parameters are simultaneously assessed: epithelial barrier integrity, macrophage transmigration, bacterial survival, and inflammation. The present protocol describes a robust and reproducible system for evaluating drug efficacy and host responses that could be relevant for discovering new anti-infectives and optimizing their aerosol delivery to the lungs.

Introduction

Pseudomonas aeruginosa is a relevant pathogen in cystic fibrosis (CF) that contributes to lung tissue impairment¹. The production of polysaccharides, such as alginate and other mucoid exopolysaccharides, coordinates the progress of the disease, which leads to tenacious bacterial adherence, limits the delivery of antibiotics to bacteria and protects the bacteria against the host immune system². The transition of P. aeruginosa from the planktonic stage to biofilm formation is a critical issue in this context, also facilitating the occurrence of antibiotic tolerance.

In the context of CF, the mouse has primarily been used as a model. Mice, however, do not spontaneously develop this disease with the introduction of CF mutations³. Most of the bacterial biofilm development and drug susceptibility studies have been performed on abiotic surfaces, such as Petri dishes. However, this approach does not represent the in vivo complexity. For instance, important biological barriers are absent, including immune cells as well as the mucosal epithelium. Though P. aeruginosa is quite toxic to epithelial cells, some groups have managed to co-cultivate an earlier P. aeruginosa biofilm with human bronchial cells. These cells originated from cystic fibrosis patients with CFTR mutation (CFBE41o^- cells)⁴ and allowed to assess antibiotic efficacy⁵ or assess the correction of the CFTR protein during infection⁶. Such a model was shown to improve the predictability of drug efficacy, in addition to enabling characterization of issues with drugs that failed in later phases of drug development⁷.

However, in the lung, the mucosal epithelium is exposed to air. Moreover, immune cells present in the airways, like tissue macrophages, play an essential role against inhaled pathogens or particles⁸. Macrophages migrate through the different cell layers to reach the bronchial lumen and fight the infection. Furthermore, inhaled drugs also have to cope with the presence of mucus as an additional non-cellular element of the pulmonary air-blood barrier⁹. Indeed, several complex three-dimensional (3D) in vitro models have been developed, aiming to increase the in vivo relevance. Co-culture systems not only increase the complexity of in vitro systems for drug discovery but also enable to study cell-cell interactions. Such complexity has been addressed in studies about macrophage migration¹⁰, the release of antimicrobial peptides by neutrophils¹¹, the role of mucus in infection⁹, and the epithelial cell reaction to excessive damage¹². However, a reliable CF-infected in vitro model that features the genetic mutation in CF, that is exposed to the air (increased physiological condition), and integrates immune cells is still lacking.

To bridge this gap, we describe a protocol for stable human 3D co-culture of the infected airways. The model is constituted of human CF bronchial epithelial cells and macrophages, infected with P. aeruginosa and capable of representing both a diffusional and immunological barrier. With the goal of testing anti-infectives at reasonably high throughput, this co-culture was established on the permeable filter membrane of well plate inserts, using two human cell lines: CFBE41o^- and THP-1 monocyte-derived macrophages. Moreover, to eventually study the deposition of aerosolized anti-infectives¹³, the model was established at the air-liquid interface (ALI) rather than liquid covered conditions (LCC).

As we report here, this model allows assessing not only bacterial survival upon an antibiotic treatment but also cell cytotoxicity, epithelial barrier integrity, macrophage transmigration, and inflammatory responses, which are essential parameters for drug development.

This protocol combines two relevant cell types for inhalation therapy of the pulmonary airways: macrophages and CF bronchial epithelium. These cells are seeded on opposite sides of permeable support inserts, allowing cell exposure to air (called the air-liquid interface (ALI) conditions). This co-culture of host cells is subsequently infected with P. aeruginosa. Both host cell lines are of human origin: the epithelial cells represent the cystic fibrosis bronchial epithelium, with a mutation on the CF channel (CFBE41o^-), and the THP-1¹⁴ cells are a well-characterized macrophage-like cell line. A confluent epithelial layer is first allowed to form on the upper side of well plate inserts before the macrophage-like cells are added to the opposite compartment. Once the co-culture is established at ALI, the system is inoculated with P. aeruginosa at the apical side. This infected co-culture system is then used to assess the efficacy of an antibiotic, e.g. tobramycin. The following end-points are analyzed: epithelial barrier integrity in terms of transepithelial electrical resistance (TEER), visualization of cell-cell and cell-bacteria interactions by confocal laser scanning microscopy (CLSM), bacterial survival by counting of colony-forming units (CFU), host cell survival (cytotoxicity) and cytokine release.

Protocol

1. Growth and differentiation of cells in permeable support inserts

1. Cultivate CFBE41o^- in a T75 flask with 13 mL of minimum essential medium (MEM) containing 10% fetal calf serum (FCS), 1% non-essential amino acids and 600 mg/L glucose at 37 °C with 5 % CO₂ atmosphere. Add fresh medium to the cells every 2–3 days.
   1. Detach the cells after reaching 70% confluence in the flask with 3 mL of trypsin- Ethylenediaminetetraacetic acid (EDTA) at 37°C for 15 min. Add 7 mL of fresh MEM and centrifuge the cells at 300 x g for 4 min at room temperature (RT). Discard the supernatant and add new 10 mL of MEM while disrupting the clumps by gently pipetting up and down.
   2. Count the cells with an automated cell counter or hemocytometer chamber. Seed cells with a density of 2 x 10⁵ cells/well in 12-well plates with permeable supports (pore size of 3 μm, see Table of Materials).
      NOTE: The automated cell counter determines cell number, size distribution, and viability of the cells (see Table of Materials). Permeable supports with a pore size of 0.4 μm could be used here; however, the macrophages, in this condition, should be added directly to the apical side, and their transcellular migration will not be assessed in this case.
   3. Seed cells at liquid-liquid condition (LLC) by adding 500 µL of the cell suspension on the apical side of the permeable support and 1.5 mL of fresh medium in the basolateral side. Then incubate cells at 37°C under 5% CO₂, for 72 h.
   4. To shift to the air-liquid interface (ALI) culture, on the third day after seeding, remove the medium from the basolateral side first, then from the apical side. To the basolateral side, add 500 µL of fresh MEM and change the medium every second day until cells form a confluent monolayer.
      NOTE: For the conditions used in this protocol, the CFBE41o^- cells usually are confluent after 3-7 days in culture.
   5. Assess the epithelial barrier properties on day 7 by incubating CFBE41o^- cells with 500 μL cell medium in the apical side and 1.5 mL in the basolateral side for 1 h, at 37°C under 5% CO₂.
   6. Measure barrier properties via transepithelial electrical resistance (TEER), with an STX2 chopstick electrode and an epithelial volt-ohmmeter; after 7 days this is higher than 300 Ω×cm².
      NOTE: Eventually, in some membrane inserts, the cells have low TEER. Therefore permeable inserts with TEER < 300 Ω×cm² are not used.
2. To cultivate THP-1 cells, grow them in a T75 flask using 13 mL of Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% FCS, and incubate at 37°C under 5% CO₂. Split cells every second day by seeding 2 x 10⁶ cells/mL cells in a new T75 flask.
   NOTE: Non-differentiated THP-1 cells are grown as monocytes in suspension.
   1. Differentiate the THP-1 cells as follows. Centrifuge contents of a T75 at 300 x g for 4 min. Discard the supernatant, resuspend the pellet in fresh medium and put in a new T75. Add 10 ng/mL Phorbol 12-myristate 13-acetate (PMA) incubatig the cells in RPMI for 48 h at 37 °C and 5% CO₂ atmosphere¹⁵.
      NOTE: After the differentiation with PMA, cells do not proliferate anymore and attach to the flask.
   2. To detach THP-1 macrophage-like cells, wash once with phosphate-buffered saline (PBS) at 37 °C and incubate with 3 mL of cell detachment solution (e.g. accutase) containing 0.5 mM EDTA for 10 min at room temperature.
   3. Inspect the cells under an inverted microscope to look for cell detachment. Add 7 mL of fresh medium and centrifuge at 300 x g for 4 min at RT.
      NOTE: Macrophages can also be detached with trypsin-EDTA, 37 °C for 20 min. However, trypsin is harsher to macrophages than the chosen cell detachment solution (see Table of Materials).
   4. After removing the supernatant, re-suspend macrophage cells in 3 mL of THP-1 medium into a 15 mL conical tube, count the cells as described in 1.1.2. and incubate for a maximum 1 h at 37 °C under 5% CO₂ before setting up the co-culture.
      NOTE: THP-1 cells in suspension can be stained with viability dyes to image the co-culture further. At this step, use the procedure below (step 1.2.5).
   5. Stain macrophages with 10 µM of a cell viability dye (based on the conversion of acetate moieties by intracellular esterases, see Table of Materials) in which 3 µL of the cell viability dye is applied to the cell suspension. Incubate cells for 20 min at 37 °C, 5% CO₂, then wash 1x with PBS 37°C to remove the dye.
      NOTE: Centrifuge the cells to remove the dye at 300 x g for 4 min at room temperature (RT).

2. Establishment of an epithelial-macrophage co-culture on permeable supports

1. Use CFBE41o^- monolayers at ALI with TEER ≥ 300 Ω×cm² (step 1.1.6.). Remove the medium from the lower chamber, carefully invert the support inside a sterile glass Petri dish (50 mm x 200 mm), and remove the cells overgrown through the membrane pores on the bottom side of the membrane using a cell scraper.
   NOTE: Due to the pore size of 3 µm, epithelial cells tend to grow through the pores toward the basolateral side. Therefore, one needs to remove them before adding the macrophages on this side. CFBE41o^- lung epithelial cells can be stained at this step. The procedure in step 1.2.5 can be used; however, instead of a cell suspension, the dye solution in MEM is applied (500 µL apical side only) on the adhered cells on the permeable support.
2. Use 2 x 10⁵ cells/well (in 200 µL of RPMI) from the cell suspension of PMA-differentiated THP-1 macrophages and place the cells on the basolateral side of the inverted inserts.
3. Close the Petri dishes carefully and incubate for 2 h at 37 °C under 5% CO₂.
4. Place the inserts back into the 12-well microplates and add 500 µL of MEM medium in the basolateral side of the permeable insert to maintain ALI conditions. The cells are now ready for infection.

3. Infection by P. aeruginosa

NOTE: All following steps from here must be done in a biosafety level 2 (BSL2) laboratory.

1. Inoculate 15 mL of lysogeny broth (LB) supplemented with 300 μg/mL ampicillin in an Erlenmeyer flask (50 mL) with a single colony of P. aeruginosa PAO1-GFP.
   NOTE: Other strains of P. aeruginosa could also be used here, for instance, PAO1 wild type, PA14, or clinical strains, following their own cultivation protocols.
2. Incubate the bacteria for 18 h at 37°C, shaking at 180 rpm.
3. Transfer the contents after the 18 h to a 50 mL conical tube and centrifuge at 3850 x g for 5 min. Discard the supernatant and add 10 mL of sterile PBS at 37°C.
4. Measure optical density on a spectrophotometer at wavelength 600 nm and adjust the concentration of bacteria using the cell culture medium to a final concentration of 2 x 10⁵ CFU/mL. This corresponds to a multiplicity of infection (MOI) of one bacterium per epithelial cell.
5. Add 100 μL of bacterial suspension to the apical side of the permeable support (step 2.4.) and incubate at 37 °C under 5% CO₂ for 1 h, to allow bacteria attachment to the cells. Then, remove apical liquid carefully with a pipette to restore ALI conditions. Keep some samples uninfected as a control.
   NOTE: At this stage, the bacteria attached should be plated in LB agar (see steps 5.4/5.5) to determine the initial bacteria inoculum.
6. Incubate the drug of interest after bacterial adhesion in the cells. For treatment experiments, add 500 μL of a drug solution diluted in cell medium (in this protocol tobramycin 6 μg/mL was used) to the apical side. Add 1,500 µL of cell medium without drug on the basolateral side.
   NOTE: Instead of instilling the drugs as a solution, the model can also be adapted to aerosol deposition. For such purposes, the cells at ALI are fed from the basolateral side with 500 µL of cell medium. The drug is then first nebulized and allowed to deposit in the apical compartment by an appropriate device (not described here). The infected and treated sample can be checked for the endpoints outlined in the sections 4–7. From this step on, permeable supports can be used to create either images (section 4) or to get results of bacteria growth and mammalian cell viability, amongst others (sections 5–7).

4. Sample preparation for confocal laser-scanning microscopy (CLSM)

1. After the establishment of the co-culture, infection and drug treatment, remove all medium from the apical and basolateral side. Wash 1x with PBS at 37°C and then fix the cells with 3% paraformaldehyde (PFA) for 1 h at RT (300 µL on apical/600 µL on basolateral). Cell nuclei are stained with 5 µg/mL of DAPI-PBS for 30 min at room temperature.
   CAUTION: PFA is hazardous.
2. Cut the membranes using a scalpel and place them between two 12 mm microscopy cover slides using a mounting medium (see Table of Materials). Let it dry inside the flow bench for 30 min before storage at 4 °C. Visualize by confocal scanning microscopy.
   NOTE: After the co-culture and before the mounting, tight junctions immunostaining can be performed. For that, cells are fixed with paraformaldehyde 3% for 30 min, washed again with PBS, and permeabilized with saponin 0.05%/BSA 1% in PBS. In this protocol, the zonula occludens protein (ZO-1) was detected via mouse anti-human ZO-1 antibody (1:400, incubation at 4 °C overnight). The samples were then incubated for 2 h at RT with goat anti-mouse IgG antibody Alexa Fluor 633 (1:2000 in red). Nuclei were stained with DAPI (1 µg/mL) and mounted with mounting medium on coverslips.
3. Use a confocal microscope for imaging the stored membranes. Choose 25x or 63x water-immersion objectives and lasers at 405, 488, 505 or 633 nm for detection. Images should have a 1024 x 1024 pixel resolution.
   NOTE: The lasers are chosen according to the stain used.
4. Acquire apical and cross-section views, and use zeta-stack mode (10–15 stacks) for the construction of a three-dimensional model using imaging software.

5. Measurement of bacterial proliferation via colony-forming units (CFU)

1. Collect the apical and basolateral medium (containing bacteria) to assess CFU of non-attached bacteria. Collect 500 µL from the apical and basolateral sides and pool them.
   NOTE: Use this suspension directly to count bacteria (step 5.4) or centrifuge at 21,250 x g for 10 min to evaluate lactate dehydrogenase (LDH) from the supernatant (section 6) and/or re-suspended in PBS to count extracellular bacteria (step 5.4).
2. Assess survival of bacteria attached and/or internalized in the cells by adding 500 µL of sterile deionized cold water in each compartment of the permeable support. Incubate cells for 30 min at room temperature.
   NOTE: The samples can either be plated on LB agar (see step 5.4) or frozen (as whole insert plate) at -20 °C for plating later on.
3. For assessing CFU of adherent/internalized bacteria, thaw samples at 37°C for 10 min (if frozen). Using pipette tips for each well, scrape the membrane surface and pipette up and down to remove all adhered content.
   NOTE: At this step, all the epithelial cells are lysed and adherent/internalized bacteria are available as a suspension to be plated.
4. With the bacterial suspension from both fractions, perform a 1/10 serial dilution using Tween-80 0.05% in PBS and plate the bacteria on LB agar plates.
   NOTE: Dilutions in between 1 to 10 are recommended. The bacteria should be counted in the highest dilution, where single colonies are first identified.
5. Incubate agar plates at 30°C for 16–72 h to count colonies, and calculate CFU accordingly.
   NOTE: A temperature of 30°C at the time of plate incubation is essential for treated-samples and to observe delayed-growth of colonies.

6. Evaluation of cell cytotoxicity via lactate dehydrogenase assay

1. Use the supernatant of infected cells containing bacteria (from step 5.1) for cell viability assessment for LDH assay¹⁶. Centrifuge the supernatant at 21,250 x g for 10 min to pellet the bacteria and eventually rest of the cells. Use the bacteria-free supernatant to measure LDH release.
   NOTE: The supernatant should not be frozen before measuring LDH by this assay.
2. Transfer 100 µL of the supernatant to a 96-well plate, and add 100 µL of the LDH assay solution (see the Table of Materials). Incubate at room temperature for 5 min in the dark, then read absorbance at 492 nm.

7. Assessing the release of human cytokines

1. For cytokine quantification, use either ELISA or cytometric bead array immunoassay¹⁷(see the Table of Materials). For this, centrifuge supernatant from step 5.1 at 21,250 x g for 10 min and measure either immediately or store -80 °C for up to 15 days until analysis.
2. Evaluate supernatants with a commercially available ELISA kit.
   NOTE: The procedure follows the manufacture instructions, which include the coating of plates with the capture antibody, addition of the samples (100 µL) and cytokine standards, incubation, washing, and addition of detection antibody to provide a colorimetric measurement of cytokine presence. Alternatively, flow cytometry can be used to measure cytokines via commercially available kits (see Table of Materials).

Representative Results

Figure 1A shows the morphology of the resulting co-culture of human bronchial epithelial cells and macrophages after growing both for 24 h on the apical and basolateral side of permeable supports, respectively. The epithelial barrier integrity is shown by higher TEER (834 Ω×cm²) and CLSM by immunostaining for the tight junction protein ZO-1 (Figure 1B). The same results observed in terms of barrier integrity of uninfected CFBE41o^- monoculture could be seen in the uninfected epithelial-macrophage co-cultures.

To model a bacterial infection, P. aeruginosa was inoculated at a multiplicity of infection (MOI) of 1:1 on CFBE41o^- cells. Six hours after infection (Figure 2A), macrophages were observed on the apical side of the co-culture. After the infection, the TEER dropped from 834 to 250 Ω×cm², indicating a compromised epithelial barrier, as also visualized by ZO-1 staining (Figure 2B).

Figure 3 shows macrophage transmigration through the permeable membrane pores and bacteria uptake by THP-1 cells on the apical side. The samples were fixed at 1, 3, and 6 h post-incubation from independent experiments. In the THP-1 monocultures (Figure 3A–C), macrophages migration was observed as early as 1 h, while in the co-culture (Figure 3D–F), this was seen after 3 h infection. Bacteria uptake in THP-1 was observed after 3 h of infection, in both monoculture and co-culture. No bacterial uptake by CFBE41o- could be seen. Cross-sectional views were placed such that the permeable membrane support was in the middle as a separation of the apical and basolateral compartment.

Figure 4 shows confocal scanning laser microscopy pictures of infected co-cultures (CFBE41o^- + THP-1) treated with or without tobramycin for 6 h (Figure 4A, B) or 20 h (Figure 4C,D). Without treatment, both the epithelial cells and the macrophages died after 20 h of infection (Figure 4C). However, upon tobramycin treatment, the host cells are preserved after 20 h; still, some bacteria can be observed in the culture. Despite being seen after 6 h of treatment in the microscopy pictures (Figure 4B), the bacteria did not proliferate as observed in CFU assays in Figure 4E. Nevertheless, after 20 h treatment, the bacteria recovered the proliferation capability, as seen by the colonies in the CFU assay (Figure 4F). The cell lysis protocol with cold water and scraping can release bacteria attached and possibly internalized in cells. At the same time, the cells are destroyed (Supplementary Figure S1A, B). The centrifugation steps used in this paper for epithelial cells (300 x g) or bacteria (21,250 x g) did not hamper the viability of both (Supplementary Figures S1C, D). All CFU assays were performed by freezing the samples at -20°C, followed by thawing and plating. This procedure reduced the number of bacteria by 2-logs, compared to fresh samples (Supplementary Figure S1E). As this procedure is done simultaneously for all experimental groups (treated and untreated) at different time points, this reduction will be incorporated in the final results (Supplementary Figure S1E). Moreover, the concentration of tobramycin used here showed no toxicity for the uninfected cells (Supplementary Figure S2A) and also no further inflammatory response (Supplementary Figure S2B). However, it was within the range of the minimum inhibitory concentration to kill P. aeruginosa.

Figure 5 shows the transepithelial electrical resistance (TEER) and cell viability. Figure 5A–B illustrates the TEER of monocultures and co-cultures. The co-culture of CFBE41o^- cells with THP-1 did not induce any change in the epithelial barrier integrity compared to the monoculture (red bars). Upon the infection, the TEER value dropped (green bar). After 1 h of infection, some samples were treated with the antibiotic tobramycin (blue bar), for 6 or 20 h. The treatment preserved the epithelial barrier integrity, as observed by the higher TEER. Figure 5C shows the percentage of LDH release as an indication of cell toxicity upon infection and tobramycin treatment after 6 h. The co-culture itself induced a release of LDH, which was the same for the infected cells (around 20%). After 20 h of infection, no signal of LDH could be detected. To prove LDH reliability for long-term infection, PAO1-GFP was incubated in medium with and without LDH 1 U/mL compared to respective uninfected controls (Supplementary Figure S2C). The LDH signal was lost after prolonged incubation (20 h) with P. aeruginosa, indicating that LDH is only stable in shorter incubation times in infected cultures.

Figure 6 shows the kinetics of pro-inflammatory cytokines detected via ELISA. The advantage of an infected co-culture of CFBE41o- and THP-1 cells was observed with the higher secretions of pro-inflammatory cytokines. The secretion of some pro-inflammatory cytokines was either similar (IL-6) or higher (IL-8, TNF-α, IL-1β) in the infected co-culture (Figure 6C) than in the corresponding monocultures (Figures 6A-B). Unexpectedly, some cytokines in THP-1 monocultures (Figure 6B) are downregulated in infected samples (Il-8, TNFα, IL-1β).

Figure 7 demonstrates the release of cytokines in mono- and co-cultures upon infection and treatment with tobramycin measured via fluorescence activated cell sorting (FACS). The secretion of the pro-inflammatory cytokine IL-8 (Figure 7A) and the anti-inflammatory cytokine IL-10 (Figure 7F) was higher in the co-cultures of epithelial cells and macrophages, compared to the monocultures. However, for all other cytokines (IL-1 α, IL-12p40, IL-23 and GM-CSF) (Figures 7B–E), the levels of cytokine secretion were not higher in the co-culture that in the respective monocultures.

Figure 1: Cross-sections and apical views of the uninfected epithelial-macrophage co-culture. (A) Cross views of uninfected 24 h epithelial-macrophage co-culture. CFBE41o^- stained red, THP-1 macrophages in yellow and nuclei in blue (DAPI). (B) Apical views of the uninfected CFBE41o- monolayer immuno-stained for ZO-1 (red). DAPI: nuclei. Scale bars: 50 µm. Please click here to view a larger version of this figure.

Figure 2: Cross-sections and apical views of the infected epithelial-macrophage co-culture. (A) Cross views and (B) apical view of epithelial-macrophage co-culture at 6 h post-infection (hpi) with P. aeruginosa PAO1-GFP. CFBE41o^- stained in red, THP-1 macrophages in yellow, nuclei in blue (DAPI) and P. aeruginosa PAO1-GFP in green. (B) Apical views of the 6 h infected CFBE41o- monolayer. Scale bars: 50 µm. Please click here to view a larger version of this figure.

Figure 3: Kinetics of macrophage transmigration and bacteria uptake visualized by cross-section of the 3D model. PAO1-GFP infection kinetics in monocultures of THP-1 macrophages (A–C) or co-culture (D–F). THP-1 macrophages are stained in red, nuclei of epithelial cells in blue (DAPI), and P. aeruginosa in green (GFP). Each figure is divided into apical and basolateral sides, the space in between is considered to be the membrane, which is empty or occupied by the CFBE41o^- confluent layer (D–F). Inserts in the figures show bacteria uptake by macrophages at different times (A–F). Scale bars: 50 µm. Please click here to view a larger version of this figure.

Figure 4: Characterization of PAO1-GFP survival in tobramycin-treated co-culture. (A–D) Confocal micrographs co-cultures with and without treatment. (A) Untreated co-culture after 6 h of infection. (B) Infected co-culture treated with tobramycin 6 µg/mL (Tob) for 6 h. (C) Untreated co-culture 20 h post-infection, (D) infected co-culture treated with tobramycin 6 µg/mL for 20 h. Nuclei stained with DAPI (blue), macrophages (red) and P. aeruginosa GFP (green). (E) Colony-forming units (CFU) of adherent/internalized bacteria after 6 and (F) 20 h with tobramycin 6 µg/mL treatment. Empty membrane insert was used as an abiotic substrate to grow PAO1-GFP. Two-way ANOVA with Tukey’s multiple comparisons test (# no colonies) was used, *p < 0.05; ***p < 0.001; ****p < 0.0001; ns: not significant. Error bars indicate standard deviation, n = 9–27 replicates of 3–9 independent experiments. Please click here to view a larger version of this figure.

Figure 5: Barrier integrity and evaluation of the viability of mono- and co-culture. The following co-culture conditions were assessed: uninfected (gray bars), infected (green bars), or infected and treated with tobramycin (blue bars). (A) Transepithelial electrical resistance after 6 h and (B) 20 h of infection in mono-cultures (CFBE41o^- and THP-1) and co-culture. (C) Cytotoxicity of mono- and co-culture measured via LDH release 6 h post-infection. Two-way ANOVA with Tukey’s multiple comparisons test was used; *p < 0.05; ****p < 0.0001; ns: not significant. Error bars indicate standard deviation; n = 9 replicates of three independent experiments. Please click here to view a larger version of this figure.

Figure 6: Kinetics of cytokine release of uninfected and infected mono- and co-culture supernatants assessed via ELISA. ELISA was done according to the kit manufacturer's protocol. (A) CFBE41o-, (B) THP-1, and (C) co-culture releasing IL-8, TNF-α, IL-1β, and IL-6. Error bars indicate standard deviation. n = 6 replicates of 2 independent experiments. Please click here to view a larger version of this figure.

Figure 7: Supernatant results of cytokine panel measured via FACS with and without tobramycin 6 µg/mL for 6 h post-infection. Supernatants of mono- and co-culture after 6 h post-infection used to analyze the respective cytokines IL-8 (A), IL-1α (B), IL12p40 (C), IL-23 (D), GM-CSF (E) and IL-10 (F). Error bar indicates standard deviation, n = 9 replicates of 3 independent experiments. Please click here to view a larger version of this figure.

Supplementary Figure S1: Control experiments for critical steps of the protocol. (A, B) Micrographs of CFBE41o- cells in 24-well plates grown for 2 days at density of 2 x 10⁵ cells/well. (A) CFBE41o- cells in PBS for 30 min, and (B) water-treated cells after 30 min after scraping with a pipette. (C) Viability of mammalian cells after centrifugation. CFBE41o- cells were removed from T75 cell culture flask as described in step 1.1.1 and 1.1.2. 100 µL of resulting cell suspension was analyzed in 10 mL isotonic solution. An automated cell counter was used to assess the viability of single cells. Then, respective cell suspensions were centrifuged at 300 x g for 4 min, re-suspended and counted again. Error bars indicate standard deviation, n = 6 different flasks of 2 individual experiments. (D) Viability of PAO1-GFP after centrifugation. PAO1-GFP bacteria were diluted to OD = 0.01 in cell medium. CFU was assessed via a 10-fold dilution row and LB plates incubated overnight at 30 °C. Respective plastic tubes were centrifuged at 21,250 x g for 10 min and re-suspended in medium. CFU was assessed accordingly again. Two-tailed student's t-test, * p < 0.033. Error bar indicates standard deviation, n = 6 of 2 experiments. (E) Viability of bacteria after freezing. PAO1-GFP bacteria were prepared as in (D) and CFU was analyzed, then plastic tubes were frozen for one day at -20 °C and thawed to assess CFU again. Two-tailed student´s t-test, *** p < 0.001. Error bars indicate standard deviation, n = 6 of two experiments. Please click here to download this file.

Supplementary Figure S2: Control experiments to assess LDH behavior and influence of tobramycin. (A) Control experiment to assess cytotoxicity after 20 h of incubation with 6 µg/mL tobramycin. Mono- and co-culture was done as described in the protocol, but cells were grown for 2 days on 24-well plates and THP-1 cells were seeded apically. Cells with 6 µg/mL tobramycin or controls were incubated for 20 h. One-Way ANOVA, Tukey's multiple comparisons test, *** p < 0.001. Error bar indicates standard deviation, n = 6 of 2 experiments (CFBE41o-), n = 3 of one experiment (THP-1 and co-culture). (B) Control-ELISA of supernatants of mono- and co-culture with/without tobramycin. Cell culture was done according to (A) to show no cytokine release compared to controls for all conditions. ELISA was done in step 7.1 and 7.2, 10 µg/mL (LPS) was added as control, IL-8 release for LPS-treated controls containing THP-1 was higher than detectable. Two-Way ANOVA, Tukey's multiple comparisons test, ns p > 0.12; * p < 0.033; *** p < 0.001. Error bars indicate standard deviation, n = 6 of two experiments, n = 3 of one experiment (LPS control). (C) LDH degradation due to excessive PAO1-GFP proliferation. LDH was added at concentration of 1 U/mL to the MEM Medium. Either LDH medium or control medium was used to dilute cells to OD = 0.01 (corresponds to 1 x 10⁸ CFU/mL) and then incubated for 20 h. LDH assay was done as described in section 6. One-Way ANOVA, Tukey´s multiple comparisons test, ns p > 0.12; ***p < 0.001. Error bar indicates standard deviation, n = 8–9 of three individual experiments. Please click here to download this file.

Discussion

This paper describes a protocol for a 3D co-culture of the infected airways, constituted by the human cystic fibrosis bronchial epithelial cell line CFBE41o- and the human monocyte-derived macrophage cell line THP-1. The protocol allows the assessment of epithelial barrier integrity, macrophage transmigration, bacteria survival, and inflammation, which are important parameters when testing drug efficacy and host-responses simultaneously. The novelty in the model lies within the incorporation of epithelial cells (i.e. human CF cell line and macrophages) with acute bacterial infection (i.e. P. aeruginosa). The acute infection in the epithelial cells is demonstrated to be controlled by an antibiotic (i.e. tobramycin). Besides the use of a human CF cell line, the entire model is set up at ALI conditions, which is considerably closer to the physiological conditions in CF. The use of a CF cell line implements some of the characteristics of the disease in the model. The mutation of the cystic fibrosis transmembrane conductance regulator (CFTR) is directly related to the dysregulation of epithelial fluid transport in the lungs. Furthermore, mutations in the CF gene, such as the ΔF508, result in thick mucus, with inflammation and severe lung damage upon infection with P. aeruginosa⁴. These pathological manifestations caused by dysfunctional CFTR potentially involve autophagy impairment as an important cellular mechanism associated with the pathogenesis of CF lung disease¹⁸. However, the CFBE41o^- fail to secret mucus, which is a limitation of this cell line. If it is intended to study the role of mucus more specifically, the protocol can be adapted by using other bronchial cell lines (e.g., Calu-3).

One critical step to set up this protocol is the combination of epithelial and immune cells and the subsequent infection with P. aeruginosa at ALI. The infection of CFBE41o^- by P. aeruginosa in vitro has already been described, mainly using a flow-cell chamber, supplemented with arginine in the culture media, to improve epithelial cell survival and support biofilm formation¹⁹. The present protocol aimed for a new model using only human cells, which moreover, could be grown at ALI on permeable well plate inserts for higher sample throughput. The inclusion of THP-1 differentiated macrophages as a human immortalized cell line, instead of being dependent on obtaining reproducible primary cells from donors, is another advantage of our model. By adding these macrophages to the basolateral side of permeable membrane support, it was observed that macrophages protruded and eventually transmigrated to the apical side of the filter-grown epithelial barrier. A variation of this protocol could be the addition of macrophages directly on the apical side on top of the epithelial cells, as described by Kletting et al.¹⁴. The co-culture of non-human immune and lung cells has already been described before. Ding et al.¹⁰ used mouse Lewis lung carcinoma cells on permeable insert supports in combination with macrophages on the basolateral side and infected with S. aureus, another critical pathogen of chronic infection in CF patients. However, in this study, there was no focus on cystic fibrosis or to use the co-culture as a platform for the evaluation of drug efficacy. Our protocol can be adapted for other bacterial infections, such as Staphylococcus aureus, Mycobacterium abscessus, or Burkholderia cepacia— important pathogens in CF lung.

Another critical step is the addition of THP-1 macrophages to the cells by flipping the permeable inserts upside-down (section 2). This is crucial to assess macrophage transmigration through the well to the side of infection. The later imaging process from the 3D models with z-stacks, and cross-section view, can be performed to observe the inside of macrophages and detect bacteria uptake (Figure 3). At 1 hpi, bacteria applied on the apical side migrate through the membrane, while macrophage migration and uptake only take place at 3 hpi. Therefore after 1 h of infection, it was appropriate to start the treatment with tobramycin and have the possibility to address both host cell and bacteria survival for a long period (20 h). In the course of the protocol, maintaining sterility is a critical issue due to the multitude of steps that each carry the risk of contamination. Nevertheless, experienced cell-culture personnel will be able to follow this protocol after appropriate preparation and training. Cell medium should be regularly checked for contamination, preferably after all critical steps.

As with any model, the infected co-culture also has some limitations; for instance, the integration of the macrophage-like cells. Here, it was important to have macrophages on the basolateral side; however, the manipulation of the insert with a previously grown epithelial layer may have provided early damage and disturbances to the co-culture. Although, the permeable support model provided high-throughput characteristics, which has not been observed in previous co-cultures of the CF infected lung^20,21. With that, further experiments need to assess the limitations of using THP-1 as macrophage substitute. While this cell line is widely used, it is less responsive to LPS²² and it lacks full activation and the entire population is not differentiated from monocytes to macrophage-like cells²³. Another limitation is the lack of other key components in CF infection and drug delivery. The CFBE41o^- cell line does not possess cilia nor does it produce mucus, which usually happens 20-30 days of cell culture at ALI. As this was not the case for CFBE41o^- cell line, we used the cells after seven days when a tight epithelial barrier was formed. Mucociliary clearance alters the residence conditions for either microbes²⁴ or drug particles^9,25 and in vitro models assessing lung deposition should take this into account. Differently from what is observed by other cells, the tissue culture inserts coating with an extracellular matrix material (like fibronectin or collagen I) do not show a significant difference for CFBE41o^-, for instance in TEER²⁶. Therefore, the permeable filters were not coated with an extracellular matrix material in this protocol.

With the protocol described here, mono and co-cultures after 6 h infection provide sufficient cytokine release to be used as a measurement in future drug testing. The co-culture brings an advantage of cell cooperation in modeling immune response. The inefficacy of tobramycin in reducing inflammation was expected since not all bacteria were eliminated during the treatment (Figure 4E, F). Nevertheless, modeling the response to tobramycin in a CF model is crucial, as tobramycin (in higher concentrations) can be effective in P. aeruginosa inhibition, even on biofilm^19,27. One possibility for further use of this protocol is to integrate anti-inflammatory drugs in the treatment. The overall recommendation regarding inflammatory responses would be to use the short duration treatment (6 h), which still has the host cell and bacteria present. After this time point, the host cells are destroyed in untreated samples. Both ELISA and FACS could be used to measure the release of cytokines. Finally, if the samples are stored longer than 15 days at -80 °C, it is recommended to check the reliability of the cytokines by using, for instance, positive control of fresh samples (e.g. cells stimulated with LPS).

Some modifications of the protocol are possible. For example, the current protocol can be expanded to the application of nebulized drugs (step 3.6). This is necessary to model pulmonary drug delivery via oral inhalation. Nebulization of water-soluble drugs, like tobramycin, or nano-carriers thereof, such as liposomal colistin, is relatively straight forward by commercially available devices routinely used in the clinic. Also, there are several commercially available devices to deposit aerosols onto cell culture inserts. In addition, as the model described here is based on permeable membrane supports, it could also be adopted to some contemporary microfluidic (e.g. “lung on a chip”) devices, for example, to study the influence of breathing and the related mechanical stretching and changes in the airflow. Moreover, this protocol could be modified by the addition of mucus or replacement by primary cells depending on the scientific question to be addressed. Another interesting next step would be the testing of nanomedicines, especially as nanotechnology is making progress in the development of novel anti-infectives²⁸, CF correctors²⁹ and co-delivery of antibiotics and pathoblockers³⁰. Overall, the current protocol may be perceived as useful in assessing bacterial survival upon antibiotic treatment in a complex system, together with some host-related readouts: cell cytotoxicity, epithelial barrier integrity, macrophage transmigration and inflammatory response. These are essential parameters for drug development.

Materials

Name	Company	Catalog Number	Comments
Accutase	Accutase	AT104
Ampicillin	Carl Roth, Germany	HP62.1
CASY TT Cell Counter and Analyzer	OLS Omni Life Sciences	-
CellTrace Far Red	Thermo Fischer	C34564
Centrifuge Universal 320R	Hettich, Germany	1406
CFBE41o- cells	1. Gruenert Cell Line Distribution Program 2. Sigma-Aldrich	1. gift from Dr. Dieter C. Gruenert 2. SCC151
Chopstick Electrode Set for EVOM2, 4mm	World Precision Instruments, Sarasota, USA	STX2
Confocal Laser-Scanning Microscope CLSM	Leica, Mannheim, Germany	TCS SP 8
Cytokines ELISA Ready-SET-Go kits	Affymetrix eBioscience, USA	15541037
Cytokines Panel I and II	LEGENDplex Immunoassay (Biolegend, USA).	740102
Cytotoxicity Detection Kit (LDH)	Roche	11644793001
D-(+) Glucose	Merck	47829
Dako Fluorescence Mounting Medium	DAKO	S3023
DAPI (4′,6-diamidino-2-phenylindole)	Thermo Fischer	D1306
Epithelial voltohmmeter	World Precision Instruments, Sarasota, USA	EVOM2
Falcon Permeable Support for 12 Well Plate with 3.0μm Transparent PET Membrane, Sterile	Corning, Amsterdam, Netherlands	353181
Fetal calf serum	Lonza, Basel, Switzerland	DE14-801F
Goat anti-mouse (H+L) Cross-adsorbed secondary Antibody, Alexa Fluor 633	Invitrogen	A-21050
L-Lactate Dehydrogenase (LDH), rabbit muscle	Roche, Mannheim, Germany	10127230001
LB broth	Sigma-Aldrich, Germany	L2897-1KG
MEM (Minimum Essential Medium)	Gibco Thermo Fisher Scientific Inc.	11095072
Non-Essential Amino Acids Solution (100X)	Gibco Thermo Fisher Scientific Inc.	11140050
P. aeruginosa strain PAO1	American Type Culture Collection	47085
P. aeruginosa strain PAO1-GFP	American Type Culture Collection	10145GFP
Paraformaldehyde Aqueous Solution -16%	EMS DIASUM	15710-S
Phosphate buffer solution buffer	Thermo Fischer	10010023
Petri dishes	Greiner	664102
Phorbol 12-myristate 13-acetate (PMA)	Sigma, Germany	P8139-1MG
Precision Cover Glasses	ThorLabs	CG15KH
Purified Mouse anti-human ZO-1 IgG antibody	BD Transduction Laboratories	610966
Roswell Park Memorial Institute (RPMI) 1640 medium	Gibco by Lifetechnologies, Paisley, UK	11875093
Soda-lime glass Petri dish, 50 x 200 mm (height x outside diameter)	Normax, Portugal	5058561
Saponin	Sigma-Aldrich, Germany	S4521
T75 culture flasks	Thermo Fischer	156499
THP-1 cells	Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ; Braunschweig, Germany)	No. ACC-16
Tobramycin sulfate salt	Sigma	T1783-500MG
Trypsin-EDTA 0.05%	Thermo Fischer	25300054
Tween80	Sigma-Aldrich, Germany	P1754