Neutrophil-Airway Epithelial Interactions Result in Increased Epithelial Damage and Viral Clearance during Respiratory Syncytial Virus Infection

This study shows that the RSV-infected human airway drives changes in the behavior of human neutrophils, including increasing activation markers and delaying apoptosis, that result in greater airway damage and viral clearance.

(ARDS), acute lung injury, cystic fibrosis (CF), and chronic obstructive pulmonary disease (COPD) (8)(9)(10). In these and other conditions, it may be that neutrophils recruited to the airways as part of host defense contribute to tissue damage and exacerbate disease. Currently, little is known about how neutrophils behave in the airways during RSV disease. Studies using animal models (11) (e.g., human RSV infection in mice, sheep, and cotton rats) lack sensitivity and do not fully recapitulate the disease in humans. For example, only 18% to 27% of BAL cells recovered from human RSV-infected calves, 3% to 14% from mice (12,13), and Ͻ5% from cotton rats are neutrophils (14). In the human airway, infiltrating neutrophils directly interact with inflamed epithelial surfaces and have adapted mechanisms to respond to RSV-infected airway epithelial cells. However, few studies have investigated neutrophil function during this interaction. These functions include activation, degranulation, apoptosis (programmed cell death), and NETosis (cell death that is characterized by the discharge of decondensed chromatin and granular contents), which can affect the level of inflammation and tissue damage. The primary aims of this study were to evaluate the possible injury caused by neutrophils to RSV-infected ciliated cells in culture, to evaluate the early changes in neutrophil function during RSV respiratory infection, and to assess whether neutrophils had an antiviral effect.

RESULTS
Neutrophils enhance ciliated epithelial layer disruption and cilium loss. We found that neutrophils interacted with motile ciliated nasal airway epithelial cells (nAECs) almost immediately after introduction and gathered in clusters (Fig. 1A). After 1 h of incubation with ciliated nAECs infected with RSV for 24 h and 72 h, neutrophils were shown to decrease the epithelial expression of ZO-1 (a marker of epithelial cell tight junction proteins) compared with mock-infected cultures (P ϭ 0.002). This reduction in ZO-1 expression was also significantly lower than in cultures of RSV-infected cells without neutrophils ( Fig. 1B and C) at 72 h (P ϭ 0.001), but not at 24 h postinfection (P ϭ 0.07). We also found that the addition of neutrophils was associated with greater nAEC loss after 1 h, with significantly fewer RSV-infected nAEC nuclei remaining attached to the membrane insert at 24 h post-RSV infection than in cultures with the respective mock-infected control (P ϭ 0.006) and RSV-infected nAECs without neutrophils (P Ͻ 0.0001) (Fig. 1D). ZO-1 staining and the numbers of epithelial cells were similar whether neutrophils were added to cultures infected for 24 or 72 h.
As RSV has been shown to target ciliated cells for infection, we were especially interested in how ciliary activity is altered during neutrophil interactions with RSVinfected epithelial cells. We found that ciliary beat frequency was unaffected by RSV infection or exposure over the entire study period (Table 1). However, RSV infection led to a higher proportion of dyskinetic cilia ( Table 1). The mean dyskinesia index was significantly higher at 24 h post-RSV infection (25.71% Ϯ 1.14%) and after 72 h of RSV infection (42.38% Ϯ 2.44%) than mock-infected controls (8.44% Ϯ 0.20% and 10.50% Ϯ 0.35%, respectively) (P Ͻ 0.001).
The addition of neutrophils to RSV-infected nAECs did not further enhance ciliary dyskinesia at any time point or infection condition tested (Table 1). Representative slow-motion videos showing ciliary beating are available as supplemental files 1 to 8 (screenshots are shown in Fig. 2A). We found no significant difference in mean fluorescence intensity (MFI) for ␣-tubulin staining (ciliary tubulin marker) under fluorescence microscopy examination at 24 h post-RSV infection compared to that for mock-infected cells or to that following 24 h of RSV infection with and without neutrophil exposure ( Fig. 2B and C). However, we found that at 72 h post-RSV infection, the mean fluorescence intensity for ␣-tubulin was almost half that of the mock-infected group with neutrophils (P ϭ 0.014) (Fig. 2C). This loss of ␣-tubulin staining correlated with a loss in the number of motile cilia observed by light microscopy (referred to as the mean motility index) at 72 h post-RSV infection (57.6% Ϯ 1.89%), which was lower than that of the mock-infected controls   Table 1). As is shown in Fig. 2D, the distribution of ciliary beat frequency of the same field of view before and after the addition of neutrophils produced a similar mean ciliary beat frequency (CBF) of around 10.2 to 10.9 Hz, but fewer areas (region of interest [ROI]) from the RSV-infected ciliated nAECs (bottom right panel of Fig. 2D) show active beating cilia (defined as Ͼ3 Hz) when cocultured with neutrophils.
Incubation with neutrophils reduces the amount of infectious virus in the culture. Four hours after neutrophil exposure, we detected significantly smaller amounts of infectious virus, with an RSV titer of 2.4 ϫ 10 4 PFU/ml recovered from cultures infected with RSV for 72 h without neutrophils, compared with 2.8 ϫ10 3 PFU/ml for cultures infected with RSV for 72 h with neutrophils (P Ͻ 0.05) (Fig. 3A). We assessed the numbers of RSV-infected nAECs by measuring the mean fluorescence intensity of GFP, a marker of active cytoplasmic RSV replication. Using fluorescence microscopy ( Fig. 3B), we found that GFP expression was significantly lower 1 h after exposure to neutrophils than for cells not cocultured with neutrophils at both 24 h and 72 h postinfection (P Ͻ 0.05) ( Fig. 3B and C).
Incubation with RSV-infected ciliated epithelial cells increases neutrophil apoptosis. Using flow cytometry (Fig. 4A), we showed that at 24 h post RSV-infection, there was no difference in the percentage of apoptotic (PI lo AnnexinV hi ) neutrophils recovered from RSV-infected nAECs (11.1% Ϯ 5.1%) compared with 4.8% Ϯ 2.7% in the mock-infected cocultures (Fig. 4B). However, at 72 h postinfection, we detected significantly more apoptotic neutrophils after 4 h of exposure to RSV-infected nAECs with 46.7% Ϯ 11.4%, compared with 6.2% Ϯ 0.9% in the mock-infected cocultures (P Ͻ 0.0001) (Fig. 4B). Apoptosis appeared to be the dominant form of cell death, as we did not detect an increase in dead (PI hi AnnexinV lo ) neutrophils at any time point or test condition. After 4 h of exposure to epithelial cells infected with RSV for 72 h, we detected 6.43% Ϯ 7.2% dead neutrophils compared with 5.5% Ϯ 7.3% in the mock control (P Ͼ 0.99) (Fig. 4B).
Incubation with RSV-infected ciliated epithelial cells increases neutrophil expression of CD11B and MPO and augments degranulation. Incubation with ciliated epithelial cells RSV infected for 24 h led to significantly greater expression of CD11B on neutrophils after 4 h, but not after 1 h, than on neutrophils that were incubated with mock-infected ciliated epithelial cells, with a mean (ϮSEM) fluorescence intensity of 1.3 ϫ 10 3 Ϯ 1.21 ϫ 10 2 compared to 9.7 ϫ 10 2 Ϯ 3.9 ϫ 10 1 after 4 h (P ϭ 0.048) (Fig. 5A). At 72 h postinfection, incubation with ciliated epithelial cells infected with RSV led to significantly greater expression of CD11B on neutrophils after 1 h and 4 h than on neutrophils that were incubated with mock-infected ciliated epithelial cells, with MFIs of 1.7 ϫ 10 3 Ϯ 3.6 ϫ 10 2 and 8.9 ϫ 10 2 Ϯ 8.2 ϫ 10 1 , respectively, after 4 h (P ϭ 0.0002) (Fig. 5A). Incubation with epithelial cells infected for 24 h showed significantly greater expression of myeloperoxidase (MPO) on neutrophils after 1 h and 4 h  whether NE correlated with augmented azurophil granule release or was specific to NE alone, MPO activity was also assessed. Here, we found significantly more MPO activity (P ϭ 0.003) after exposure of neutrophils for 4 h to epithelial cells infected with RSV for 24 h (P ϭ 0.006) or 72 h (P Ͻ 0.0001) (Fig. 5D), with a mean Ϯ SEM of 1.0 Ϯ 0.0 optical density (OD) at 24 h, than 0.6 Ϯ 0.1 OD in the mock-infected cultures. This was also significantly different after 1 h at 72 h (P Ͻ 0.003) but not 24 h (P ϭ 0.065) post-RSV infection.
RSV infection was also associated with a greater release of active MMP-9 from neutrophils (Fig. 5E) after 1 h and 4 h (P ϭ 0.001) in cultures infected with RSV for 24 h, from 529.4 Ϯ .21.9 g/ml in the mock to 949.3 Ϯ 76.9 g/ml in RSV-infected epithelial cells at 1 h (P ϭ 0.02). There was no significant difference in MMP-9 concentrations

DISCUSSION
We have shown that when RSV infected human primary nasal airway epithelial cells are exposed to neutrophils at physiological concentrations, there is increased epithelial layer disruption, ciliary loss, and less infectious virus in these cultures. This suggests that neutrophils are helping to eliminate virus-infected cells and reduce viral spread. The airway epithelial damage that we observed, consistent with previous studies (15,16), may be a necessary consequence of this antiviral effect. We found that RSV infection without neutrophils did not reduce CBF but did increase the number of cilia that presented with an abnormal beat pattern as early as 24 h postinfection, which is similar FIG 5 Neutrophil activation and release of neutrophil-derived products and neutrophil activation after exposure to RSV-infected ciliated epithelial cells. Activation marker expression was calculated by staining for cell surface expression of CD11b (PE hi ) or MPO (allophyocyanin [APC ϩ ]) and determined by flow cytometry. Neutrophils were identified using a PE-positive gate. Using this population, the geometric mean fluorescence intensity of PE or APC fluorescence was calculated. (A) Neutrophils exposed to RSV (red bars)-infected epithelial cells infected for 24 h (plain bars) or 72 h (checkered bars) show increased cell surface expression of CD11b compared to those exposed to mock (blue bars)-infected epithelial cells. (B) Neutrophils exposed to RSV (red bars)-infected epithelial cells infected for 24 h (plain bars) or 72 h (checkered bars) show increased cell surface expression of MPO compared to those exposed to mock (blue bars)-infected epithelial cells. Bars show means Ϯ SEM (n ϭ 5 to 8 epithelial donors with heterologous neutrophils). Statistical significance is shown. Apical surface medium concentrations of neutrophil elastase (NE) (C), MPO (D), and MMP-9 (E) were measured in the apical supernatant by ELISA collected following neutrophil exposure to mock-or RSV-infected (

h [plain bars] or 72 h [checkered bars]) ciliated AECs after 1 or 4 h.
For all graphs, bars represent the mean Ϯ SEM (n ϭ 5 epithelial donors with heterologous neutrophils for HBSSϩ (white), mock-infected (blue bars), or RSV-infected (red bars) cultures. A statistical comparison between all groups was performed using a paired t test. Statistical significance is shown. to our previous findings (3,5). Interestingly, the addition of neutrophils did not further increase ciliary dyskinesia.
The reduction in the number of RSV-infected epithelial cells following neutrophil exposure may result from neutrophil degranulation. Neutrophils are known to mediate direct antimicrobial effects and neutralize several influenza A, RSV, and vaccinia virus strains through effector mechanisms, including degranulation (17)(18)(19). We have shown that neutrophils exposed to RSV-infected human primary airway epithelial cells have greater expression of the activation markers CD11B and MPO and release greater amounts of NE, MPO, and MMP-9. It is recognized that the inflammatory processes in the airways of infants with RSV bronchiolitis are dominated by an intense neutrophil influx (7,20) and that neutrophil products, such as MPO and NE, are released into the airway lumen (21). Indeed, the degree of neutrophilic inflammation correlates with disease severity in patients with RSV-induced bronchiolitis (22). Our study has shown that RSV-infected epithelial cells increased NE and MPO and gelatinase (MMP-9) granule concentrations, at the same time reducing neutrophil membrane integrity. This appears to be relevant to the pathophysiology of viral respiratory infections (23,24). Likewise, both NE and MMP-9, which may be necessary for clearance of bacteria, are linked to airway damage and the progression of cystic fibrosis (25). Our results are consistent with these clinical findings, suggesting that the cytotoxicity of neutrophil antimicrobial proteases may be important in viral clearance but may also potentiate RSV-induced lung injury.
A surprising finding was that at 72 h, but not at 24 h, post-RSV infection, neutrophil exposure led to an increase in the numbers of apoptotic neutrophils. This finding is consistent with a clinical study that found neutrophil apoptosis was accelerated in nasopharyngeal aspirates and peripheral blood of infants with RSV bronchiolitis (26). Neutrophil apoptosis is thought to be associated with reduced levels of degranulation and other proinflammatory capacities (27). We found that RSV-infected epithelial cells increased neutrophil degranulation in regard to increased neutrophil elastase (NE) and myeloperoxidase (MPO) activity in apical surface media at the same time as we detected the increased neutrophil apoptosis. These data suggest that within this model, there may be at least two subsets of neutrophils that respond differently to RSV-infected airway epithelial cells. It is possible that this balance in neutrophil function could be a predictor of disease severity (see Fig. 6).
One limitation to our study was the number of neutrophils used. The exact ratio of interacting neutrophils and airway epithelial cells in the lungs of children with RSV bronchiolitis is unknown. We used the equivalent concentration of 5 ϫ 10 6 /ml neutrophils, which is the upper limit of the number of neutrophils recovered from BAL specimens of infants with RSV bronchiolitis (1.78 Ϯ 3.3 ϫ 10 6 /ml) that was reported previously by McNamara and colleagues (7). This suggest that our data may indicate airway-immune cell interactions that occur in the lungs of infants with large neutrophil infiltrate or severe RSV bronchiolitis. Another limitation of our study was that we exposed epithelial cells to naive neutrophils directly isolated from peripheral blood. In the lungs, neutrophils are recruited to migrate from the basal subepithelial space across the vasculature and epithelium to the airways following RSV infection (28). In this context, the changes that we observed are all the more striking.
In conclusion, this study has revealed that neutrophils exposed to ciliated epithelial cell cultures infected with RSV have increased degranulation, deploying harmful proteins and proteases to the apical surface media and increasing the capacity for tissue injury. We have shown that neutrophils contribute to RSV-associated ciliary loss combined with epithelial damage, which is likely to result in reduce mucociliary clearance. These effects may contribute to viral clearance and provide important insights into the role of neutrophils in host responses in the airway. Primary human nasal airway epithelial cells (nAECs) were grown to a ciliated phenotype on 0.4-m pore Transwell inserts (Corning) at air-liquid interface (ALI) as described previously (3). Recombinant GFP-tagged RSV A2 strain was kindly provided by Fix et al. (29) and propagated using HEp-2 cells (multiplicity of infection [MOI], 0.1) for 3 to 5 days in Opti-MEM. Virus was purified as described previously (30), collected in bronchial epithelial basal medium (BEBM; Life Technologies), and frozen at -80°C. Aliquots of RSV were thawed immediately prior to use. For ALI culture infection, the apical surface of the ALI cultures was rinsed with medium (BEBM), and 200-l viral inoculum (MOI, 1) in BEBM was applied to the apical surface for 1 h at 37°C and then removed (see Fig. 7). Mock wells received BEBM alone. All cells were fed basolaterally with fresh ALI medium prior to infection. The infection continued for 24 h and 72 h.

MATERIALS AND METHODS
Neutrophil isolation and purification. Neutrophils were isolated from peripheral venous blood using a Percol density gradient, as described previously (31), and were purified using an EasySep human neutrophil enrichment kit (Stemcell Technologies) per the manufacturer's instructions (32). Neutrophils were resuspended in Hanks balanced salt solution (HBSSϪ) (Life Technologies), and fluorescenceactivated cell sorter (FACS) analysis was performed using anti-human CD49d-APC (BioLegend, 304307) and anti-human CD66a/c/e Alexa Fluor-488 (BioLegend, 342306) antibodies to confirm purity. Neutrophils were counted using a hemocytometer.
Neutrophil airway epithelial cell coculture. Neutrophils (5 ϫ 10 5 in 100 l HBSSϩ) were added to the top (apical) chamber of Transwell inserts containing either RSV-or mock-infected ciliated nAECs and incubated at 37°C ϩ 5% CO 2 . After 1 h or 4 h of incubation, neutrophils and apical surface media from the top chamber of Transwell were collected and each membrane was washed once with 0.1 ml HBSSϩ. The apical surface media and washing medium were pooled, spun, and stored at -80°C. The cell pellet was resuspended in 500-l FACS buffer (phosphate-buffered saline [PBS] [Ca2ϩMg2ϩ free], 0.5% bovine serum albumin [BSA], and 2.5 mM EDTA) for further analysis.
AnnexinV/PI apoptosis assay. The annexin V apoptosis detection kit (Miltenyl Biotec) was used to carry out this assay. Neutrophils collected from the cell pellet (described above) were resuspended in 50-l FACS buffer plus 1:250 dilution APC anti-human CD11B conjugate (Insight Biotechnology) and incubated at 4°C for 20 minutes in the dark. Cells were washed once in FACS buffer, resuspended in 50-l annexin binding buffer with 0.3 l of annexin V for 15 minutes in the dark at room temperature. They were then flash stained with propidium iodide (PI), 10 l in 900 l of annexin binding buffer, and  Transwell inserts containing ciliated differentiated airway epithelial cells were infected with GFP RSV or mock infected for 1 h before the inoculum was removed. Transwell inserts were maintained in media and infection was allowed to progress for 24 or 72 h. Ultrapure (Ͼ99.9%) neutrophils (5 ϫ 10 5 ) were then added to the apical side of the Transwell inserts and cocultured for 1 or 4 h. After coculture, the apical neutrophils were collected and downstream analysis was performed. controls confirmed no cross-reactivity. Cells were washed once in FACS buffer, resuspended in 1% (wt/vol) paraformaldehyde (PFA), and stored at 4°C. Directly prior to being run, samples were centrifuged (1,400 rpm) and resuspended in FACS buffer.
Samples were analyzed using a Beckton Dickenson LSR II flow cytometer and FlowJo v10.0 FACS analysis software. Neutrophils were first identified as being CD11B positive (PEϩ). Using this population, the mean fluorescence intensity for PE and APC was calculated.
Determination of neutrophil degranulation. Degranulation was assessed by measuring neutrophil elastase (NE), myeloperoxidase (MPO), and matrix metalloproteinase-9 (MMP-9) in the apical surface media. The amount of NE or MPO in apical surface media was measured using commercial activity assay kits (Cayman, USA). MMP-9 release was measured using a commercial ELISA kit (Biolegend, USA). All protocols were conducted according to the manufacturers' instructions.
CBF and beat pattern. Beating cilia were observed via an inverted microscope system (Nikon TiE; Nikon, UK) equipped with an incubation chamber (37°C and 5% CO 2 ) as previously described (3). To determine ciliary beat frequency (CBF), videos were recorded using a 20ϫ objective using a CMOS digital video camera (Hamamatsu) at a rate of 198 frames per second and image size of 1,024 by 1,024 pixels. CBF (Hz) was calculated using ciliaFA software (33). The number of motile ciliated cells in each sample area was counted (motility index). The dyskinesia index was calculated as the percentage of dyskinetic ciliated cells (those that displayed uncoordinated motile cilia or those that beat with a stiff, flickering, or twitching motion) relative to the total number of motile ciliated cells.
Immunofluorescence microscopy. Following fixation, cells were washed three times with PBS and treated with PBS containing 0.1% Triton X-100 for 10 min to permeabilize the cells. Cells were incubated with 5% fetal calf serum (FCS) in PBS for 0.5 h at room temperature to block nonspecific interactions and washed again three times with PBS. All subsequent antibody incubations were carried out in 5% FCS in PBS ϩ 0.1% Triton X-100. Reagents used in this study were rabbit anti-ZO-1 polyclonal antibody (1:200; catalog number sc-5562; Santa Cruz Biotechnology) and mouse anti-acetylated ␣-tubulin monoclonal antibody (6-11B-1; 1 g/ml; Sigma). Primary antibody incubations were carried out in a humidified chamber overnight at 4°C, followed by three washes with PBS. The detection of primary antibodies was carried out for 1 h using the following reagents: fluorescein fluorescein isothiocyanate (FITC; catalog number F2012; Sigma) conjugated rabbit anti-mouse (1:64) or Alexa Fluor 594-conjugated rabbit anti-donkey antibody (1:250; Invitrogen, Paisley, UK). All secondary antibodies were tested and found to be negative for cross-reactivity against human epithelial cells. Following three washes in PBS, DNA was stained with Hoechst 33258. After a final wash in distilled water, the insert was cut from the support and mounted under coverslips in 80% (vol/vol) glycerol and 3% (wt/vol) n-propylgallate (in PBS) mounting medium. Images were captured with a confocal laser microscope (Zeiss Observer Z.1) using a 40ϫ water immersion objective. The pinhole was set at 1 airy unit (AU). For z-stack images, the slice thickness was 1 m.
Statistical analysis. Statistical analysis was performed using GraphPad Prism 5 (GraphPad, San Diego, CA, USA). Differences between the mock-and RSV-infected group were analyzed using paired t tests or a paired two-way analysis of variance (ANOVA) for multiple comparisons with a Bonferroni correction (GraphPad Prism v5.0).  The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health.