Budesonide, fluticasone propionate, and azithromycin do not modulate the membrane vesicle release by THP-1 macrophages and respiratory pathogens during macrophage infection

Budesonide (BUD), fluticasone propionate (FLUT), and azithromycin (AZI) were from Sigma (Sigma Aldrich, St. Louis, MO, USA). The anti-Haemophilus influenzae type b (α-Hib; clone 1079/457) monoclonal antibody was obtained from Acris (Acris GmbH, Herford, Germany). The rabbit serum against Moraxella catarrhalis (Mrc, strain A 1.39 N, isolated from children in a primary school in Nieuwegein, the Netherlands, 1989) was kindly provided by Dr. J. Hays (Erasmus University, Rotterdam, the Netherlands). The polyclonal antibody against Pseudomonas aeruginosa (Psa) (OAMA02609) was from Antibodies online (Aviva Systems Biology, San Diego, CA, USA). α-CD63 (unconjugated, mouse-anti-human clone H5C6) and α-CD81 (PE-conjugated, mouse-anti-human clone JS-81) were from BD (BD Bioscience, Franklin Lakes, NJ, USA). Purification of antibodies from serum was performed using the antibody serum purification kit based on protein A (Abcam, Cambridge, MA, USA). Antibodies for detection for flow cytometric analyses were PE-conjugated using a PE-labeling kit from Abcam according to the manufacturers’ instructions (Cambridge, MA, USA).

The following bacterial strains were selected: Haemophilus influenzae (NTHi, ATCC-49247), Pseudomonas aeruginosa (Psa, ATCC-27853), Streptococcus pneumoniae (Spn, ATCC-49619), and a clinical Moraxella catarrhalis (Mrc) isolate (University Medical Centre Maastricht (MUMC+), the Netherlands). The ATCC strains are well characterized and recommended by ATCC for quality control and antimicrobial susceptibility testing. All bacteria were cultured overnight on blood plates except for NTHi which was cultured on vitalex-supplemented chocolate agar plates (Oxoid, Wesel, Germany) in 5% CO₂ at 37 °C. After overnight pre-culture, bacteria were resuspended at 0.5 McFarland (1.5 × 10⁸ colony forming units (cfu)/ml) in RPMI1640 and used for infection or culture experiments. For bacterial culture, bacteria were used at 5 × 10⁷ cfu/ml and culture without or with BUD, FLUT, or AZI for 6 h in 10 ml RPMI1640. Next, the conditioned media were processed by centrifugation at 1200×g for 10 min, at room temperature. The pelleted bacteria were washed, diluted in PBS, and the optical density was determined at 600 nm using optical methacrylate disposable cuvettes (Sarstedt, Newton, NC, USA). The supernatants were centrifuged again at 1200×g for 10 min, and the supernatants were filtered through 0.22 µM filters. Hereafter, the supernatants were further concentrated 20 times to a total of 500 µl by centrifugation at 4000×g for 15 min using Amicon Ultra-15 10-kDa centrifugal filter units (Millipore, Billerica, MA, USA).

MVs used for the stimulation of THP1 macrophages were obtained from bacterial cultures (at a density of 1 × 10⁸ cfu/ml) following culturing for 4 h in 30 ml complete vesicle-depleted medium containing 5% FCS, that was obtained as described in the “Cells and media” section. Upon culture, supernatants were depleted from bacteria by two centrifugation steps at 1200×g for 10 min and 0.22 µm filtration. The supernatants that were cleared from bacteria were then further processed by ultrafiltration and size-exclusion chromatography (SEC), as described below.

The human monocytic cell line THP-1 (ATCC-TIB202) was maintained in RPMI1640 (Sigma, St. Louis, MO, USA) supplemented with 100 mM sodium pyruvate, 22.5% glucose, 25 mM β-mercaptoethanol, and 10% fetal calf serum (FCS) (Lonza, Verviers, Belgium) and cultured in 5% CO₂ at 37 °C. For monocyte differentiation, cells were seeded in a 24-well plate at 0.5 × 10⁶ cells/well or in a 96-well plate at 1 × 10⁴ cells/well and stimulated for 72 h with 200 nM phorbol 12-myristate 13-acetate (PMA; Sigma, St. Louis, MO, USA). THP-1 macrophage stimulations were performed in vesicle-depleted medium containing 5% FCS (complete vesicle-depleted medium). This medium was obtained by combining vesicle-depleted RPMI1640 medium with 30% FCS with FCS-free medium (both supplemented with sodium pyruvate and glucose). Vesicle-depleted medium was generated by overnight centrifugation at 100,000×g using a 70Ti-rotor, κ-factor 44 in an Optima L-90 K ultracentrifuge (both Beckman Coulter, Fullerton, CA, USA).

THP-1 macrophages seeded in 24-well plates were washed three times with PBS, and medium was replaced with complete vesicle-depleted medium. Hereafter, the cells were pre-treated with BUD (0.1 µM), FLUT (0.1 µM), or AZI (3 µg/ml) for 1 h, and the concentrations were previously calculated to represent a concentration as can be obtained locally upon administration by inhalation (Wagner et al. 2015; Olsen et al. 1996; Ek et al. 1999). After pretreatment, macrophages were infected with one of the bacteria at a multiplicity of infection of ten for 6 h. After infection, the medium was harvested, processed by centrifugation at 300×g and 1200×g, filtrated with a 0.22-µM micropore filter, and subsequently analyzed by flow cytometry.

The bacteria-cleared conditioned media obtained from bacterial cultures were subjected to ultrafiltration (Lobb et al. 2015) and SEC (Böing et al. 2014). First, the cleared conditioned media were concentrated to 500 µl in 2 runs at 4000×g for 15 min at room temperature using Amicon Ultra-15 10-kDa centrifugal filter units (Millipore, Billerica, MA, USA) (Lobb et al. 2015). Then, the concentrates were purified by SEC using Sepharose columns as described by Boïng et al. with minor modifications (Böing et al. 2014). For this, a 15-ml TELOS filtration column (Kinesis Scientific Experts, St. Neots, Cambridgeshire, UK) was stacked with a total of 10 ml Sepharose CL-2B (GE Healthcare, Uppsala, Sweden). The concentrates were loaded onto the column and fractions of 0.5 ml were eluted using PBS. The fractions that were found to be highly enriched for MVs and negative for free protein (fractions 7–11 as were determined by flow cytometry and Micro BCA, Pierce, Rockford, IL, USA) were pooled and stored at −80 °C until further use.

A bead-based flow cytometric assay based on a method that was previously described by Volgers et al. (2017) was used for the semi-quantitative analysis of bacterial MVs and CD63⁺/CD81⁺ host cell-derived MVs. Four-micrometer-sized aldehyde–sulfate beads were washed in MES buffer, coated with α-CD63 antibody, or antibodies directed against bacterial vesicles (with the exception of Spn as we could not establish this assay for Spn). Antibody-coated beads were incubated overnight with processed supernatants (200 µl) obtained after infection under constant agitation at 1000 rpm at room temperature. Next, the beads were washed twice with PBS filtered through a 0.22-µM filter containing 2% (w/v) bovine serum albumin (BSA) and incubated with PE-conjugated detection antibodies (α-CD81-PE for host cell vesicles or an PE-conjugated α-bacterial antibody) for 90 min under continuous agitation at room temperature. Bacterial antibodies were PE-conjugated using a PE-conjugation kit (Abcam, Cambridge, MA, USA) according to the manufacturer’s instructions. Then, the vesicle-bead complexes were washed, suspended in PBS (300 µl), and used for flow cytometric analysis using a FACSCanto™ (BD Bioscience, Franklin Lakes, NJ, USA). The lower threshold for detection was set at 2% based on the percentage of PE-positive control beads (based on culture medium). Analyses were performed using FACSDiva Software. The relative amount of MVs was calculated by the multiplication of the percentage of positive beads with the median fluorescence intensity and expressed relative to the control (as % of control).

Tunable resistive pulse sensing analysis was used to determine the MV-concentration using the qNano Gold, the Izon Control Suite Software v3.2, and the reagent kit (type RK1) for EV analysis from Izon (Izon Science Ltd., Oxford, UK). The measurements were performed using a NP150-pore. The stretch was fixed at 47 mm, the pressure kept at 6 mbar, and a baseline current of ±100 nA was maintained. Solution G was added (10%) to supernatants that were diluted in solution Q (1:1), and each sample was measured for 10 min and measurements were repeated when system instabilities occurred. The samples were calibrated using 114-nm polystyrene calibration beads (CPC100, Izon Science Ltd., Oxford, UK) at a concentration of 1 × 10⁹ particles/ml diluted in culture medium.

Stimulation of THP-1 macrophages with MVs was performed using macrophages seeded in 96-well plates. Prior to stimulation, the cells were washed three times with PBS whereafter the medium was replaced with complete vesicle-depleted medium. Then the cells were pre-treated with BUD (0.1 µM), FLUT (0.1 µM), or AZI (3 µg/ml) for 1 h. Next, the cells were washed again after which they were exposed to 20 µl of the MV-containing SEC fractions, without or in the presence of BUD, FLUT, or AZI, in complete vesicle-depleted medium for 16 h. Hereafter, the culture supernatants were harvested and used for cytokine measurements.

TNF-α levels in the supernatants from the stimulation experiments were determined by enzyme-linked immunosorbent assay (ELISA) using the human Ready-Set-Go TNF-α ELISA kit eBioscience (Affymetrix eBioscience, Santa Clara, CA, USA).

Statistical analysis was performed using GraphPad Prism 5 Software (GraphPad, San Diego, CA, USA). Statistical dispersion was determined by calculating the standard error of the mean. A Mann–Whitney t test was performed for the statistical analysis of the variance between the means of two groups. P values were considered significant when <0.05.

First, the effect of BUD, FLUT, and AZI on bacterial MV release was determined. This was done semi-quantitatively by flow cytometry based on antibody-coated beads. We found that neither FLUT nor BUD significantly affected the vesicle release by NTHi and Mrc (Fig. 1a, b). Treatment with AZI also did not affect the MV release (Fig. 1c).

The release of MVs is highly conserved and apart from the bacteria the host cells are also known to release membrane vesicles. We therefore assessed the effect of BUD, FLUT, and AZI treatment on the release of CD63⁺/CD81⁺ host cell vesicles by macrophages under control conditions and upon infection. Neither treatment with BUD and FLUT, nor that with AZI, was found to affect the release of host cell MVs (Fig. 2).

Next, we aimed to determine how the bacterial MV release is affected by BUD, FLUT, and AZI in the absence of host cells. Vesicle release after 6 h of culture without or with these drugs was determined by tunable resistive pulse sensing analysis. We observed that treatment with neither FLUT, BUD, nor AZI affected the release of bacterial MVs (Fig. 3a).

Next, we determined if FLUT, BUD, or AZI had any effect on the bacterial growth. To assess this, the relative number of bacteria was determined upon treatment with FLUT, BUD, and AZI, which was done by measuring the optical density. As shown in Fig. 3b, only the Mrc counts were significantly reduced after bacterial culture in the presence of AZI.

Since bacterial MVs can possess a strong pro-inflammatory character (Macdonald 2013; Ellis and Kuehn 2010; Schaar et al. 2011a; Ren et al. 2012; Olaya-Abril et al. 2014), our next aim was to establish if BUD, FLUT, and AZI affected the TNF-α release by naïve macrophages in response to bacteria membrane vesicles. MVs from both NTHi and Mrc induced the release of TNF-α by macrophages, while the response to MVs from Psa and Spn was less pronounced. Treatment with BUD and FLUT significantly reduced the release of TNF-α by macrophages in response to bacterial MVs from NTHi and Mrc (Fig. 4a, b). In contrast, the TNF-α release in response to Spn and Psa MVs was not affected by BUD or FLUT. No significant effect of AZI was found on the TNF-α response to either NTHi-, Mrc-, Spn-, or Psa-derived MVs (Fig. 4c).