Human Respiratory Syncytial Virus Subgroup A and B Infections in Nasal, Bronchial, Small-Airway, and Organoid-Derived Respiratory Cultures

Human respiratory syncytial virus (HRSV) is the major cause of bronchiolitis and pneumonia in young infants and causes almost 200,000 deaths per year. Currently, there is no vaccine or treatment available, only a prophylactic monoclonal antibody (palivizumab).

All viruses expressed enhanced green fluorescent protein (EGFP) or dTomato (dTom) as a reporter protein from the 5th position in the genome, facilitating sensitive detection of HRSV-infected cells in the absence of visible cytopathic changes. We found nasal, bronchial, or small-airway cultures (SACs) were all susceptible to HRSV infection, despite induction of a type III interferon (IFN) response. This was recapitulated in AOs grown at ALI that resembled the bronchial cultures. Subgroup A viruses replicated to higher titers of cell-free virus than the subgroup B virus in all our cultures at 3 days postinfection (DPI), but the clinical isolates resulted in more infected cells (measured as percentage of virus-positive area) at 2 DPI. HRSV infection led to visible cytopathic effects, such as rounding of ciliated cells, cilia disruption, shorter cilia, and formation of small syncytia. Viral loads correlated with the production of innate cytokines, dominated by type III IFN.

RESULTS
Replication kinetics in well-differentiated nasal, bronchial, and small-airway cultures. Well-differentiated airway cultures from the nose, bronchus, and SACs (bronchioles) were purchased from Epithelix and infected with a relatively low inoculum of rHRSV A2 (laboratory adapted), rHRSV A11 , and rHRSV B05 (clinical isolate based). After a single round of replication (24 h), we detected slightly more foci of HRSV infection in nasal and bronchial cells than in the SACs (Fig. S1 in the supplemental material). This shows that with a standardized inoculum and single replication cycle, differences in susceptibility and permissiveness between the cells could already be shown. Further dissemination was studied by measuring the production of cell-free virus and the surface area of HRSV-infected cells based on fluorescence over time (Fig. 1). Both rHRSV-A strains replicated to significantly higher cell-free viral titers (610 5 to 10 6 50% tissue culture infective dose [TCID 50 ]/ml) than the rHRSV-B strain (10 4 TCID 50 /ml) over time, especially in nasal and bronchial cultures. However, the percentage of EGFP-positive (EGFP 1 ) surface area at 1 and 2 DPI was significantly higher for both clinical isolate-based strains than for the laboratory-adapted strain. The clinical-based rHRSV A11 virus disseminated to a higher infected surface area than the clinical-based rHRSV B05 .
Cytopathology in well-differentiated nasal, bronchial, and small-airway cultures. To assess the tropism of HRSV and the associated cytopathic effect (CPE) in primary cultures, we performed a histological analysis of airway cultures at 3 DPI, in combination with indirect immunofluorescence staining (IIF). The uninfected nasal and bronchial cultures consisted of a pseudostratified epithelium, and ciliated cells were abundantly present at the apical side ( Fig In the HRSV-infected nasal and bronchial cultures, infected cells were mainly present at the apical side of the epithelium and were predominantly ciliated epithelial cells. We observed hyperplasia of HRSV-infected cells and the formation of small channels in the culture ( Fig. 2A and B; Fig. S2A and B). Microscopically, we observed an increase in mucus production, and cilia actively transported the mucus through these channels (data not shown). These alterations were not observed in the uninfected cultures. Furthermore, we observed cilia degeneration (shortened and damaged cilia, rounding of ciliated cells) and damaged cell layer, the latter visualized by the loss of tight junctions caused by small syncytium formation, in both nasal and bronchial cultures ( Fig. 2C and D; Fig. S2C and D). Differences in CPE between rHRSV-A and rHRSV-B strains were not observed. Compared to nasal and bronchial cultures, uninfected SACs contained a thinner basal cell layer and fewer ciliated epithelial cells ( Fig. S3A and C). In SACs, HRSV also mainly targeted ciliated epithelial cells ( Fig. S3B and D), but the infection was more focal than the nasal and bronchial cultures. We observed infected cell hyperplasia compared to the uninfected control throughout the epithelial layer, but cilia and tight junctions remained intact ( Fig. S3A and B).
Cytokine responses in well-differentiated nasal, bronchial, and small-airway cultures. To evaluate the antiviral response of primary differentiated epithelial cells to HRSV infection, we harvested apical and basolateral washes at 1, 2, and 3 DPI and measured cytokine levels. Cytokines were predominantly detected in the apical washes. Type III interferons (IFN) (interleukin 28 A and B [IL-28A/B] and IL-29) were the main cytokines produced in all airway cultures. IL-29 increased at least 100-fold compared to uninfected cultures. IL-28A/B also increased 10-to 100-fold over time and was most abundant in the nasal cultures and SACs (Fig. 3A). Type I IFNs (IFN-a2 and IFN-b) were produced to a lesser extent than type III IFN. The concentration of IFN-a2 hardly increased after infection compared to uninfected controls, and IFN-b was 2-to 3-fold elevated in all cultures compared to uninfected controls (Fig. 3B). Overall, both rHRSV A11 and rHRSV B05 induced higher levels of IFNs at 2 DPI than rHRSV A2 , which is in accordance with the percentages of EGFP 1 surface area (Fig. 1). Notably, the interferon response clearly lagged behind the viral replication, and viral titers correlated positively with interferon production (Fig. S4). Early-stage inflammatory cytokine IP10 was increased in all the HRSV-infected airway cultures and also correlated positively with viral replication ( Fig. 3C; Fig. S4). We also measured IL-6, IL-1b, tumor necrosis factor alpha (TNF-a), IFN-g, IL-8, IL12p70, granulocyte-monocyte colony-stimulating factor FIG 1 Replication kinetics of rHRSV A2 EGFP(5), rHRSV A11 EGFP(5), and rHRSV B05 EGFP(5) in nasal, bronchial, and small-airway cultures. Nasal cells (A), bronchial cells (B), and small-airway cells (C) were infected with rHRSV A2 EGFP(5), rHRSV A11 EGFP(5), or rHRSV B05 EGFP(5) at a standardized low viral inoculum. The percentage of EGFP 1 surface area was determined by confocal microscopy, and the viral titers were determined by endpoint titrations of apical washes (TCID 50 /ml). One independent experiment is shown, and all experiments were performed in triplicate. Differences between the growth curves were statistically analyzed by two-way ANOVA (*, P = 0.05; **, P = 0.01; ***, P = 0.001; ****, P , 0.0001). Mean and individual replicates are shown.
(GM-CSF), and IL-10. These cytokines were produced in low quantities after HRSV infection, with the exception of IL-6. This cytokine was more upregulated after infection with the two clinical-based isolates compared to the laboratory-adapted strain in nasal cultures and SACs (Fig. S5). Cytokines were also measured in the basolateral compartments, but in general, low levels of cytokines were detected, with the exception of IL-6 (data not shown).
HRSV replication kinetics in well-differentiated airway organoid cultures at ALI. In addition to studying HRSV replication kinetics and innate immune responses in commercially available primary cells, we also studied these in AOs cultured at ALI as described before (25). These AOs were obtained from adult human lung stem cells from the distal airways and cultured at ALI for 4 weeks to allow differentiation of the cells into pseudostratified epithelium, including ciliated epithelial cells. Cells were infected with a low standardized inoculum of rHRSV, and the replication kinetics, dissemination, cytopathology, and innate cytokine responses were assessed. In general, HRSV replication kinetics were comparable with the commercially available bronchial cultures: infection with the clinical isolate-based viruses resulted in a higher percentage virus-infected surface area (especially at 2 DPI), and infection with rHRSV-A strains resulted in higher viral titers over time ( Fig. 4; Fig. S1). rHRSV A11 disseminated better than the other viruses, resulting in significantly more infected cells.
Cytopathology in well-differentiated airway organoid cultures at ALI. To assess the tropism of HRSV and the associated CPE, we used AOs cultured at ALI. First, we created organoids (spheroids in Matrigel) from human bronchial tissue obtained after lung resection surgery. We cultured these undifferentiated organoids in Matrigel for 10 to 14 days and then dissociated them into single cells. These cells were seeded on Transwell membranes and grown at ALI for 4 weeks, allowing differentiation. We characterized the cultures by hematoxylin and eosin (H&E) and IIF staining at 3 DPI. The uninfected cultures had multiple cell layers and an abundance of ciliated cells ( Fig. 5A and C). HRSV infected mainly ciliated epithelial cells. Degenerated cilia, rounding of infected cells, and small syncytia were observed in infected cultures compared to uninfected controls, and the integrity of the tight junctions was affected by HRSV infection . We observed an increase in mucus production and the formation of small channels ( Fig. 5A and B), similar to the commercially available nasal and bronchial cultures ( Fig. 2A and B; Fig. S1A and B). Differences in CPE between rHRSV-A and rHRSV-B strains were not observed.
Cytokine production in well-differentiated airway organoid cultures at ALI. Next, we harvested culture supernatant from the apical side at 0, 1, 2, and 3 DPI and measured the same cytokines as measured in commercially obtained nasal, bronchial, and small-airway cells ( Fig. 6; Fig. S5). We observed an increase in type III IFNs; IL-29 increased the most upon HRSV infection, but IL-28A/B was also elevated (Fig. 6A). ALI-differentiated AOs produced IFN-a, but IFN-b production was more pronounced (Fig. 6B). General inflammatory marker IP10 was increased after HRSV infection (Fig. 6C). Type III IFN, type I IFN, and IP10 were all induced by HRSV replication, indicated by a positive correlation FIG 4 Replication kinetics of rHRSV A2 EGFP(5), rHRSV A11 EGFP(5), and rHRSV B05 EGFP(5) in primary welldifferentiated airway organoid cultures at ALI. Primary well-differentiated airway organoid cultures at ALI were infected with rHRSV A2 EGFP(5), rHRSV A11 EGFP(5), or rHRSV B05 EGFP(5) at a standardized low viral inoculum. The percentage EGFP 1 surface area was determined by confocal laser scanning microscopy, and the viral titers were determined by endpoint titrations of the apical washes (TCID 50 / ml). One representative experiment is shown that was performed in triplicate. Differences between the growth curves were statistically analyzed with a two-way ANOVA (*, P = 0.05). Mean and individual replicates are shown.
FIG 5 Immunohistochemistry and indirect immunofluorescence on primary well-differentiated airway organoid cultures at ALI infected with HRSV. Well-differentiated airway organoid cultures at ALI were infected with either rHRSV A2 EGFP(5), rHRSV A11 EGFP(5), or rHRSV B05 EGFP(5) at a standardized low viral inoculum. Samples were fixed in formalin at 3 days postinfection (DPI) and embedded in paraffin. (A and B) Paraffin-embedded slides were used for H&E staining and indirect immunofluorescence using antibodies against green fluorescent protein (HRSV, green), acetylated a-tubulin (cilia, red), and Hoechst (nuclei, blue). Arrows indicate putative mucus channels in the cultures. (C and D) Transwell filters were additionally stained with antibodies against zona-occludens 1 (tight junctions, magenta) and acetylated a-tubulin (cilia, orange). Representative images are shown of rRHRSV A11 . between cytokine levels and viral loads (Fig. S6). IL-6, IL-1b, TNF-a, IFN-g, IL-8, IL12p70, GM-CSF, and IL-10 were produced in small amounts (or not at all) in AOs after rHRSV infection (Fig. S7). Overall, cytokine profiles were similar to the cytokine profiles in the commercially purchased bronchial epithelial cells; only IL-28A/B was less produced in these cultures.
HRSV-A and HRSV-B competition experiments in airway organoids cultures at ALI. In infection experiments in nasal, bronchial, SACs, and AOs, rHRSV A11 has an infectivity advantage over rHRSV B05 when comparing both production of cell-free virus and percentage EGFP 1 . To confirm this infectivity advantage of rHRSV A11 over rHRSV B05 , we performed a direct competition experiment. We infected AOs cultured at ALI with rHRSV A11 expressing EGFP and rHRSV B05 expressing dTom. Both viruses were added simultaneously at the same low input, and infections were followed over time. Two different experimental setups were used, daily confocal microscopy accompanied by apical washes after infection (Fig. 7A) or daily confocal microscopy, but only performing an apical wash at 4 DPI (Fig. 7B). In both setups, we found that rHRSV A11 replicated to a significantly higher cell-free viral titer and disseminated to a larger virus-positive area than rHRSV B05 . We observed both single-infected (rHRSV A11 or rHRSV B05 positive) and double-infected cells (rHRSV A11 and rHRSV B05 positive). We quantified the percentage of double-positive cells, which was about 5%. In conclusion, we observed that both viruses can spread well in the culture and were able to coinfect cells (Fig. 7A to D).

DISCUSSION
In this study, we aimed to compare the replicative fitness and HRSV-induced innate cytokine responses of rHRSV-A and rHRSV-B strains in disease-relevant cell culture models. We used two recombinant clinical isolate-based HRSV strains, rHRSV A11 and rHRSV B05 , and one recombinant laboratory-adapted HRSV strain, rHRSV A2 , to infect commercially available nasal, bronchial, and small-airway cultures. Epithelial cells from all anatomical locations were susceptible to HRSV infection despite the induction of a robust type III interferon response. Subgroup A viruses disseminated and replicated faster than the subgroup B virus. Additionally, we studied HRSV infection and innate responses in AO cultured at ALI. The results in this model were similar to the results in the commercially obtained bronchial cells. In summary, we show that HRSV replicates FIG 7 rHRSV A11 EGFP(5) and rHRSV B05 dTom(5) competition experiment in primary well-differentiated airway organoid cultures at ALI. Primary well-differentiated airway organoid cultures at ALI were infected with rHRSV A11 EGFP(5) and rHRSV B05 dTom(5) at a standardized low viral inoculum per virus and either washed apically daily (A) or not (B). The percentage of EGFP 1 or dTom 1 surface area was determined by confocal microscopy, and the viral titers were determined by endpoint titrations of the supernatant from the apical side (TCID 50 /ml) at 1, 2, or 3 days postinfection (DPI). The dissemination throughout the culture was imaged by using laser scanning confocal microscopy (C), and Transwell filters were excised and analyzed at high resolution to identify double-infected cells (indicated by arrows) (D). The percentages of EGFP 1 and dTom 1 cells were quantified by colocalization analysis in Fiji. Experiments were performed in triplicate. Differences between the growth curves were statistically analyzed with a two-way ANOVA or a Student's t test (*, P , 0.05; **, P , 0.01; ***, P , 0.005). Mean and standard deviation are shown. well in cells from both the upper and the lower airways, with a slight infectivity advantage for subgroup A viruses. That finding fits with clinical observations that HRSV-A causes more severe disease than HRSV-B. Lastly, we showed that AOs cultured at ALI are a valuable model for studying HRSV in vitro. This model can be used for future experiments studying intrinsic and extrinsic factors that influence HRSV infection.
HRSV susceptibility studies with cells from different anatomical sites of the respiratory tract have rarely been performed. We found that, with a relatively low standardized inoculum, nasal and bronchial cells seemed more susceptible than SACs. However, HRSV replicated and disseminated efficiently in all the cultures over time. HRSV replication kinetics in nasal and bronchial cultures have also been shown in a previous study in which they used well-differentiated nasal and bronchial epithelial cells (29). These results match with what is observed in HRSV-infected children, where viral loads in nasal aspirates and deep tracheal aspirates correlate (30). We show that HRSV replicates in SACs. Infection of smaller bronchioles in infants can lead to mechanical obstruction due to inflammatory response, one of the hallmarks of the pathogenesis of HRSV bronchiolitis (14). The combined results of our study and the previously mentioned studies highlight that HRSV efficiently replicates and disseminates in airway cells from different anatomical locations.
In human HRSV disease, ciliated epithelial cells are the main target cell, and the infection is mostly restricted to the luminal side of the airways (14). Several studies have shown this in vitro, and our results in primary cells and with clinical isolate-based viruses confirm these observations (12,17,(31)(32)(33). Additionally, we observed clear CPE at 3 DPI, such as cilia degeneration and remodeling of the epithelium. Results in literature are conflicting: cilia degeneration, such as shorter and damaged cilia, was previously described in other studies using either (pediatric) well-differentiated primary bronchial or nasal epithelial cells in combination with both clinical isolates and laboratory-adapted viruses (17,31,34). However, in other studies with primary well-differentiated bronchial or nasal epithelial cells, syncytium formation or cell damage were not observed (12,34). In the latter studies, the major difference was that another recombinant HRSV strain was used (rgRSV, an HRSV-A2-derived recombinant virus), a highly laboratory-adapted strain with green fluorescent protein in the first position of the genome.
Clinical manifestations in HRSV bronchiolitis patients include epithelial cell sloughing, cell death, increased mucus production leading to mucus plugs, and occasional syncytium formation (14). We did not measure cell death or epithelial cell sloughing in our model, but we observed an increase in mucus production upon HRSV infection. We also observed alterations of the epithelium, including loss of tight junction integrity, formation of small syncytia, and formation of channels through which mucus was transported (observational data). These observations have been described in other studies using (pediatric) well-differentiated primary epithelial cells (17,31,35,36). There is one study that did not describe any cytopathology in well-differentiated primary nasal epithelial cells; in this study, laboratory-adapted strains (HRSV-A2 and HRSV-A long) were used at a low multiplicity of infection and measured up to 36 h postinfection (HPI) (37).
Although infection of well-differentiated primary airway cultures with HRSV has been described, studies directly comparing laboratory-adapted strains and clinical isolates or HRSV-A versus HRSV-B are rare. Since we used recombinant HRSV strains, we were able to show that clinical isolates disseminated faster at early time points (1 and 2 DPI) than the laboratory-adapted virus, measured by surface area of infected cells. We did not observe these differences in cell-free virus titers. A limitation when comparing different virus strains in different cell types is that the susceptibility, permissiveness, and cell receptor usage can be different for each cell type (38). However, we believe that dissemination (fluorescence surface area) in combination with CPE is a better proxy for human disease severity than cell-free virus titers. Other studies have shown that HRSV-A2 replicates to higher or similar viral loads than clinical isolates, but this is based on TCID 50 /ml and not infected cells or infected surface area (17,31,32).
We also found that, in general, subgroup A viruses resulted in higher cell-free titers and percentage of virus-positive area than the subgroup B virus, which is in agreement with clinical observations that HRSV-A causes more severe disease (26). We also compared the infectivity of these viruses by both performing stand-alone infections and a direct competition experiment. rHRSV A11 always replicated and disseminated better than rHRSV B05 , but both strains were able to disseminate and infect cells in the culture. We also found dual-infected cells in our cultures. In another study, rHRSV A11 expressing EGFP and rHRSV B05 expressing dTom were used in cotton rats and cotton rat cell lines. Occasional double-infected cells were observed, both in vitro and in vivo (27). Our data confirm these observations, with as added value that we use primary human epithelial cells. Future studies elucidating coinfections and the factors underlying the replicative advantage of HRSV-A subgroup viruses are required.
Early dissemination in the airways can be influenced by host innate immune responses. The postinfection cytokine response in our cultures was dominated by type III IFNs. Type I IFNs and IP10 were additionally produced as described previously (17,31,36,37). We also measured background cytokine levels in our cultures, which is probably an inherent property of the cultures in combination with daily washing of the cells, which can cause immune activation. Nevertheless, infected cultures showed a clear increase in cytokine production. Surprisingly, the abundance of type III IFNs did not hinder HRSV replication; the innate immune response might either be too late or insufficient. Another option is that immune cells are required to clear the virus effectively. Experiments with IFNs and cocultures with innate immune cells would be better suited to study the interactions between innate cytokines and HRSV infections.
As experiments in commercial primary airway cultures are expensive and dependent on suppliers, we validated a robust model system to study HRSV infections. It has been reported that AOs are susceptible to HRSV infection and reproduce several characteristics of HRSV disease (epithelial cell shedding, mucus production) (22)(23)(24). However, these studies used AOs in a basement matrix. We decided to evaluate this model further by culturing in-house-developed AOs at ALI to create a well-differentiated stem cell-based epithelial cell model, reflecting the natural epithelial barrier in the human respiratory tract with a direct interface between air and submucosal fluids. We assessed replication kinetics of the three rHRSV strains and found that replication in AOs cultured at ALI was comparable to replication kinetics in bronchial cultures from Epithelix. Similar to observations in commercially available cells, we found that in AOs cultures at ALI, mainly ciliated cells were infected, with a loss of tight junction integrity and an increase in mucus production. Finally, we showed the cytokine response upon infection was dominated by type III IFNs. Also, IP10 was increased, which was shown previously in AO cultures in Matrigel (22). Taken together, we concluded that an AObased well-differentiated model system accurately resembles commercially obtained bronchial cells.
In conclusion, we have shown that the combination of primary airway cultures with recombinant clinical isolate-based HRSV strains, expressing reporter proteins, is a powerful tool to study HRSV-host interactions in vitro. Additionally, we have refined stem cell-based cultures developed in-house for HRSV infection studies. Using these models, we demonstrated how HRSV rapidly disseminates throughout anatomically different airway cultures in the face of innate immune response. In follow-up experiments, we will use these models to investigate the influence of host factors, such as interactions with innate immune responses and bacterial and viral coinfections, on HRSV dissemination.

MATERIALS AND METHODS
Viruses. Virus stocks were grown on HEp-2 cells by inoculating cell pellets at a multiplicity of infection of 0.01 for 1 to 2 h at 37°C, followed by seeding into T175 cm 2 tissue culture flasks (for growth kinetics in HEp-2 cells, see Fig. S8 in the supplemental material). When 90 to 100% of CPE was observed, cells were scraped from the flasks and sonicated three times for 30 s, and subsequently, cell debris was removed by centrifugation for 15 min at 600 Â g. The supernatant was mixed with sucrose to a final concentration of 25% (vol/vol) for stability, snap-frozen, and stored at 280°C. rHRSV A2 [rHRSV A2 EGFP(5)] was commercially obtained from ViraTree (product no. R125), and previously rescued rHRSV A11 [rHRSV A11 EGFP(5), passage 5]; rHRSV B05 rHRSV B05 EGFP(5), passage 6; and rHRSV B05 dTom(5), passage 7 have been described before (the 5 between brackets refers to the position of the additional transcriptional unit in the genome) (27,28).
Human primary airway culture and differentiation. MucilAir (primary nasal or bronchial human airway epithelial cells) and SmallAir (primary small-airway human airway epithelial cells) were obtained from Epithelix (Epithelix Sàrl, Geneva, Switzerland) and cultured according to the manufacturer's instructions. The culturing of human airway organoids was developed in-house, and differentiation on ALI was performed as described before (22,25). In short, nontumor lung tissue, obtained from patients undergoing lung resection surgery for lung cancer, was used to derive adult human lung stem cells from the distal airways (Medical Ethical Committee of the Erasmus MC Rotterdam, METC 2012-512) (22). Undifferentiated AOs were cultured in Matrigel (Corning) droplets (;30 ml) with 250 ml AO medium. To obtain differentiated organoid-derived cultures at ALI, organoids were made into single cells with TrypLE Express and consequently seeded on Transwell membranes (Corning) coated with rat tail collagen type I (Fisher Scientific). The single AO cells were seeded in AO medium and complete base medium (CBM; Stemcell Technologies; PneumaCult-ALI) at a 1:1 ratio. When a confluent monolayer was formed (2 to 4 days), the cultures were put on ALI in CBM. Cultures were differentiated for 4 weeks, and the medium was changed every 5 days.
Virus quantification. Growth kinetics in the primary airway cells cultured at ALI were assessed by washing the cultures twice with Dulbecco's phosphate-buffered saline (DPBS) with Ca and Mg (0.9 mM MgCl 2 and 0.49 mM CaCl 2 ), followed by HRSV inoculation from the apical side at a standardized low viral inoculum (multiplicity of infection, 0.5), based on titrations on HEp-2 cells, for 1 to 2 h at 37°C and 5% (vol/vol) CO 2 (experiment performed once in triplicate). After inoculation, cultures were washed three times with DPBS plus Ca and Mg. At the indicated time points, supernatants from the apical side were collected by adding 200 ml Dulbecco's modified Eagle medium (DMEM) (catalog no. LO BE12-733F; Lonza), incubating for 10 min at 37°C and 5% (vol/vol) CO 2 , and harvesting for direct titration on HEp-2 cells (39). Leftover supernatant was kept for cytokine detection. Additionally, the percentage EGFP 1 or dTom 1 surface area was determined by imaging complete wells by performing tile scans on an LSM700 confocal microscope, analyzed with ZEN software (Zeiss) and Fiji (40). For the analysis of fluorescencepositive area, the tile scan was selected, and Fiji was used to automatically set the optimal brightness, contrast, and threshold. Then, the software measured the percentage fluorescence area, independent of particle size (pixels).
Histology. Primary airway cultures were fixed in 4% (wt/vol) paraformaldehyde (PFA) for at least 30 min, after which the filters from the Transwell were excised, stored in formalin, and subsequently embedded in paraffin. Thin sections (3mm) were prepared from the formalin-fixed, paraffin-embedded tissues and stained using hematoxylin and eosin. For IIF staining, slides were incubated at 37°C overnight and then deparaffinized (twice for 4 min in xylene, twice for 2 min in 100% ethanol, once for 2 min in 96% ethanol, once for 2 min in 90% ethanol, once for 2 min in 70% ethanol, and, lastly, 5 min of PBS). Antigens were retrieved with citrate buffer (10 mM, pH 6.0 [Sigma]) (4.2 g C 6 H 8 O 7 · H 2 O in 2 liters Milli-Q, set pH, with 10 N NaOH). Slides were blocked in 10% normal goat serum (NGS) in phosphate-buffered saline (PBS) for 30 min at room temperature (RT). Primary antibodies (acetylated a-tubulin [Santa Cruz Biotechnology, catalog no. sc-23950 AF488] and rabbit anti-GFP [Invitrogen, catalog no. A11122]) were added for 1 h of incubation at RT and after washing secondary antibodies (acetylated a-tubulin and goat-anti-rabbit [AF594; Invitrogen, catalog no. A11012) were added for 1 h of incubation at RT. After two washings, Hoechst (Thermo Scientific, catalog no. 62249) was added and incubated for 15 min at RT. Samples were washed twice and mounted with Prolong Diamond antifade mounting medium and analyzed using an LSM700 confocal microscope, ZEN software (Zeiss), and Fiji (40).
Immunofluorescence microscopy. Transwell inserts were fixed in 4% (wt/vol) PFA for 30 min and then stored in PBS for further analysis. For indirect immunofluorescence staining, one-quarter of the filter was excised, washed twice, permeabilized in 0.2% Triton-X, and blocked in 10% NGS in staining buffer (DPBS with 150 mM MgCl 2 , 150 mM CaCl 2 , and 10 mM HEPES) for 30 min. The cells were incubated with conjugated mouse monoclonal antibodies for 60 min in staining buffer containing 10% NGS and 2% (wt/vol) bovine serum albumin (BSA). Tight junctions were stained using anti-zona-occludens 1 (clone 1A12; Alexa Fluor 550; Santa Cruz Biotechnologies), and cilia were stained using anti-acetylated tubulin (Alexa Fluor 647; Santa Cruz Biotechnologies, clone 6-11B-1). In the last 10 min of the incubation with antibodies, Hoechst 33342 was added (Life Technologies/Invitrogen, catalog no. 10150888). After staining, the cells were washed three times with staining buffer and mounted in Prolong antifade mounting medium (Life Technologies/Invitrogen, catalog no. 9P36961). Samples were imaged on an LSM700 confocal microscope using ZEN software (Zeiss) and Fiji (40).
Cytokine detection. Cytokines were measured in apical and basolateral cell culture supernatant from the primary airway cells using a human antivirus response panel (13-plex) kit (LEGENDplex; BioLegend). In short, cell culture supernatant was mixed with beads coated with capture antibodies specific for IFN-a2, IFN-g, IFN-b, IL-28A/B, IL-29, IL-6, IL-8, IL-10, TNF-a, IL-12p70, GM-CSF, IP-10, and IL-1b and incubated for 2 h. Beads were washed and incubated with biotin-labeled detection antibodies for 1 h, followed by a final incubation with streptavidin PE . Beads were analyzed by flow cytometry, and final analysis was performed using the LEGENDplex analysis software v8.0. The quantity of each respective cytokine is calculated on basis of the intensity of the streptavidin PE signal and a freshly prepared standard curve (as described in Weiskopf et al.) (41).
Statistics. The mean and standard deviation are depicted in graphs. Experiments were performed once in triplicate. Growth curves were compared by one-way analysis of variance (ANOVA) test or Student's t test to determine if there was statistically significant variation between groups. A P value of ,0.05 was considered statistically significant. Pearson's correlation coefficient was calculated to test correlation between viral loads and cytokine levels. All analyses were performed using GraphPad Prism version 8.0 for Windows.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.