Alveolar macrophage-derived microvesicles mediate acute lung injury

Background Microvesicles (MVs) are important mediators of intercellular communication, packaging a variety of molecular cargo. They have been implicated in the pathophysiology of various inflammatory diseases; yet, their role in acute lung injury (ALI) remains unknown. Objectives We aimed to identify the biological activity and functional role of intra-alveolar MVs in ALI. Methods Lipopolysaccharide (LPS) was instilled intratracheally into C57BL/6 mice, and MV populations in bronchoalveolar lavage fluid (BALF) were evaluated. BALF MVs were isolated 1 hour post LPS, assessed for cytokine content and incubated with murine lung epithelial (MLE-12) cells. In separate experiments, primary alveolar macrophage-derived MVs were incubated with MLE-12 cells or instilled intratracheally into mice. Results Alveolar macrophages and epithelial cells rapidly released MVs into the alveoli following LPS. At 1 hour, the dominant population was alveolar macrophage-derived, and these MVs carried substantive amounts of tumour necrosis factor (TNF) but minimal amounts of IL-1β/IL-6. Incubation of these mixed MVs with MLE-12 cells induced epithelial intercellular adhesion molecule-1 (ICAM-1) expression and keratinocyte-derived cytokine release compared with MVs from untreated mice (p<0.001). MVs released in vitro from LPS-primed alveolar macrophages caused similar increases in MLE-12 ICAM-1 expression, which was mediated by TNF. When instilled intratracheally into mice, these MVs induced increases in BALF neutrophils, protein and epithelial cell ICAM-1 expression (p<0.05). Conclusions We demonstrate, for the first time, the sequential production of MVs from different intra-alveolar precursor cells during the early phase of ALI. Our findings suggest that alveolar macrophage-derived MVs, which carry biologically active TNF, may play an important role in initiating ALI.

microscope and an external light source, a fine catheter was briefly passed 1cm below the cords, and 20µg 'Ultrapure' lipopolysaccharide (LPS) (E. coli O111:B4; InVivoGen, San Diego, CA) in 50µl was instilled intratracheally (i.t.) as previously described 1,2 . The mice were suspended in an upright position for 30 seconds to allow equal bilateral distribution of the LPS. Animals were then placed in a heated box until they had fully recovered from anesthesia. At the end of the experiment, mice were euthanized with anesthetic overdose and exsanguination, an endotracheal tube was inserted via tracheostomy and bronchoalveolar lavage fluid (BALF) was obtained by flushing and gently aspirating 700µl of 0.9% saline in and out of the lungs via the endotracheal tube three times. This was kept on ice and immediately analysed for MV content.

Microvesicle identification
10µL of BALF was incubated with fluorescence-conjugated antibodies against CD11c (clone N418; eBioscience, San Diego CA) and F4/80 (BM8; eBioscience) to identify alveolar macrophage-derived MVs, EpCAM (G8.8; eBioscience) for epithelial cell-derived MVs, or CD11b (M1/70; Biolegend San Diego, CA) and Ly6G (1A8; Biolegend) for neutrophil-derived MVs, for 30 minutes at 4ºC. Samples were resuspended in 1ml of PBS and then analysed with flow cytometry. MVs were identified as events that were under 1µm in size and positive for specific surface staining markers. Forward scatter, correlating with particle size, and side scatter (trigger threshold 0.02), corresponding with particle granularity, were used to elucidate a 1µm gate that was delineated by specific sizing beads. Absolute MV count was assessed by co-treatment of the sample with a known quantity of 6µm AccuCheck counting beads (PCB100; Life Technologies, Paisley UK). The identity of MVs was validated through their sensitivity to 0.1% Triton X-100 detergent, which solubilizes lipid membranes 3 . Counts of MV populations of different cell origins were compared to untreated wild-type controls. Data was acquired with a Cyan flow cytometer (Beckman Coulter, High Wycombe, UK) and analysis of data was performed with Flowjo software (Tree Star, Ashland, OR).

Cell culture
The murine lung epithelial (MLE-12) cell line was cultured and maintained in 175cm 2 cell culture flasks (Thermo Scientific, Hampshire, UK) in DMEM supplemented with 10% heat inactivated fetal calf serum, penicillin and streptomycin at 37ºC in a humidified 5% CO 2 atmosphere. Cells were seeded at a density of 10 5  and its corresponding isotype control, IgG2b κ isotype control (RTK4530; Biolegend), to ascertain surface ICAM-1 expression by flow cytometry as a marker of epithelial cell activation 4,5 . A variety of other controls were used: PBS; a low dose of LPS (11ng/ml) calculated to be the maximum dose of LPS that may have been retained from the initial instillation within the MV sample despite the washing steps; and a high dose of LPS (100µg/ml), to evaluate responses in the event that MVs carried over LPS.

Biological activity of primary alveolar macrophage-derived MVs
Untreated mice were euthanized (via overdose of anesthetic and exsanguination) and tracheostomy was performed. In order to enhance alveolar macrophage recovery, BALF was harvested with 700µL of calcium-free PBS supplemented with 2mM EDTA, warmed to 37ºC, which was flushed into and gently sucked out of the lungs via the endotracheal tube three times. Each flush was accompanied by gentle massaging of the rib cage to encourage macrophage recovery. Macrophages were isolated by centrifugation (200 g, 5 mins, at 4ºC) and washed 3 times in calciumcontaining PBS to neutralize EDTA. They were then placed in 24-well plates and primed for 1hour with 1µg/ml of LPS to attain a pro-inflammatory phenotype as previously described 6 or incubated with PBS alone. Macrophages were then stimulated with 6mM ATP 7 (Tocris, Bristol, UK), 1mM ecto-ATPase inhibitor (ARL67156, Tocris) and 40µM calcium ionophore (A23187, Tocris) 8 for 2 hours to generate either 'inflammatory' or 'non-inflammatory' alveolar macrophage-derived MVs. Stimulation of the purinergic receptor P2X7 results in MV release and ATP plus calcium ionophore synergistically combine to activate P2X7 receptors with resultant MV formation 9 . Whereas ATP alone and ATP plus calcium ionophore generated appreciable levels of alveolar macrophage-derived MVs, the addition of ARL67156 produced much greater quantities (data not shown) of MVs, very similar to numbers found within the BALF of mice 1 hour after LPS instillation, and hence this combination of stimulants was used to generate MVs from primary alveolar macrophages. Supernatants were subsequently collected and centrifuged to remove cells (200g for 5 mins at 4ºC). Cell-depleted supernatants were then centrifuged at high speed (20,000g for 30 mins at 4ºC) to isolate primary alveolar macrophagederived MVs.
MVs or the equivalent associated supernatant fraction were incubated with MLE-12 cells for 4 hours and surface ICAM-1 expression was measured by flow cytometry. In separate experiments, MVs were resuspended in 10µg/ml polyclonal anti-TNF antibody (R&D Systems, Abingdon, UK) and then added to MLE cells which had also been pre-treated with 10µg/ml anti-TNF antibody for 10 minutes.

In vitro generated alveolar macrophage-derived MVs initiate ALI in vivo
In vitro generated inflammatory primary alveolar macrophage derived MVs were whereas type 2 epithelial cells were classified as CD45 -, CD31 -, EPCAM + and T1alphaas previously described 10 (see additional Figure 1).

Statistical analysis.
Shapiro-Wilk normality tests were carried out and wherever possible, non-parametric data was transformed. Comparisons between two data sets were performed using either paired T-tests or Wilcoxon Rank Sum test. Where three or more datasets were present, ANOVA with Tukey HSD or Kruskal Wallis with Dunn's test were used.
Parametric data are presented as mean with SD (untransformed data) or 95% confidence interval (transformed data), whereas non-parametric data are displayed as median with interquartile range. Statistical significance was defined as p<0.05 and data were analysed and graphed using IBM SPSS and Prism software.