Central Role of the NF-κB Pathway in the Scgb1a1-Expressing Epithelium in Mediating Respiratory Syncytial Virus-Induced Airway Inflammation

ABSTRACT

Lower respiratory tract infection with respiratory syncytial virus (RSV) produces profound inflammation. Despite an understanding of the role of adaptive immunity in RSV infection, the identity of the major sentinel cells initially triggering inflammation is controversial. Here we evaluate the role of nonciliated secretoglobin (Scgb1a1)-expressing bronchiolar epithelial cells in RSV infection. Mice expressing a tamoxifen (TMX)-inducible Cre recombinase-estrogen receptor fusion protein (CreERTM) knocked into the Scgb1a1 locus were crossed with mice that harbor a RelA conditional allele (RelA^fl), with loxP sites flanking exons 5 to 8 of the Rel homology domain. The Scgb1a1^CreERTM/+ × RelA^fl/fl mouse is a RelA conditional knockout (RelA^CKO) of a nonciliated epithelial cell population enriched in the small bronchioles. TMX-treated RelA^CKO mice have reduced pulmonary neutrophilic infiltration and impaired expression and secretion of NF-κB-dependent cytokines in response to RSV. In addition, RelA^CKO mice had reduced expression levels of interferon (IFN) regulatory factor 1/7 (IRF1/7) and retinoic acid-inducible gene I (RIG-I), components of the mucosal IFN positive-feedback loop. We demonstrate that RSV replication induces RelA to complex with bromodomain-containing protein 4 (BRD4), a cofactor required for RNA polymerase II (Pol II) phosphorylation, activating the atypical histone acetyltransferase (HAT) activity of BRD4 required for phospho-Ser2 Pol II formation, histone H3K122 acetylation, and cytokine secretion in vitro and in vivo. TMX-treated RelA^CKO mice have less weight loss and reduced airway obstruction/hyperreactivity yet similar levels of IFN-γ production despite higher levels of virus production. These data indicate that the nonciliated Scgb1a1-expressing epithelium is a major innate sensor for restricting RSV infection by mediating neutrophilic inflammation and chemokine and mucosal IFN production via the RelA-BRD4 pathway.

IMPORTANCE RSV infection is the most common cause of infant hospitalizations in the United States, resulting in 2.1 million children annually requiring medical attention. RSV primarily infects nasal epithelial cells, spreading distally to produce severe lower respiratory tract infections. Our study examines the role of a nonciliated respiratory epithelial cell population in RSV infection. We genetically engineered a mouse that can be selectively depleted of the NF-κB/RelA transcription factor in this subset of epithelial cells. These mice show an impaired activation of the bromodomain-containing protein 4 (BRD4) coactivator, resulting in reduced cytokine expression and neutrophilic inflammation. During the course of RSV infection, epithelial RelA-depleted mice have reduced disease scores and airway hyperreactivity yet increased levels of virus replication. We conclude that RelA-BRD4 signaling in nonciliated bronchiolar epithelial cells mediates neutrophilic airway inflammation and disease severity. This complex is an attractive target to reduce the severity of infection.

INTRODUCTION

Human infections by the paramyxovirus respiratory syncytial virus (RSV) produce significant morbidity in the young and elderly. With no effective vaccine, RSV remains a significant human pathogen worldwide (1–3), responsible for over 64 million cases of acute infections globally (4). In the United States, RSV infects a majority of children, where it is responsible for acute emergency care of an estimated 2.1 million children under 5 years of age annually (3, 5–7). Typically transmitted by self-inoculation in the nose, RSV can also spread into the lower airways, producing lower respiratory tract infection (LRTI), including bronchiolitis and pneumonia in normal children (8–11). Consequently, RSV represents the most common cause of infant hospitalizations in the United States (12); these events are associated with subsequent decreased quality of life, increased health care expenditures, and long-term morbidity. RSV-induced LRTI has been linked with recurrent episodic wheezing, allergic sensitization, and decreased lung function (11, 13). Consequently, much attention has been focused on the acute consequences of RSV infection and its mechanism for evading protective immunity (1, 2, 14).

The normally sterile airway is highly adapted to detect the presence of infectious organisms, eliciting protective innate responses (15, 16). Pulmonary innate responses are triggered by pathogen-associated molecular patterns (PAMPs) expressed by sentinel epithelial cells. Luminal viral PAMPs, including double-stranded RNA (dsRNA) and 5′-phosphorylated RNA, are primarily bound by membrane-associated Toll-like receptor 3 (TLR3) present on airway epithelial cells. In contrast, intracellular viral replication is detected by retinoic acid-inducible gene I (RIG-I), a cytoplasmic germ line-encoded pattern recognition receptor (17, 18). Downstream, activated TLR3 and RIG-I activate the NF-κB arm of the innate immune response (19, 20). NF-κB/RelA plays a major role in innate inflammation, controlling the expression of inflammatory chemokines as well as mucosal interferons (IFNs) through a process of regulated transcriptional elongation (21–23). Here RelA complexes with the CDK9–bromodomain-containing protein 4 (BRD4) complex responsible for activating inactive RNA polymerase II (Pol II) present on immediate early innate response genes, resulting in rapid transcriptional activation. Consequently, RelA directly induces the expression of CC, CXC, and CX₃C chemokine networks that play diverse roles in dendritic cell activation, Th lymphocyte populations, and the activation of innate lymphocytes (24, 25). In addition, RelA indirectly activates mucosal type I and III IFNs via the IFN regulatory factor (IRF)–RIG-I pattern recognition receptor “cross talk pathway” (20, 23, 26). In this cross talk pathway, activated RelA expresses the inducible IRF isoforms (IRF1 and -7); all three proteins synergistically activate the RIG-I pattern recognition receptor at the transcriptional level. RIG-I upregulation further promotes IRF1 and -7 expression, enhancing type I and III IFN production (27) and resulting in a positive-autoamplification loop. After secretion, mucosal IFNs play important roles in the initial restriction of viral replication.

A major unresolved question is the identity of the major sentinel cell of the airway responsible for detecting replicating RSV. Although RSV does not productively replicate infectious virions in leukocytes, macrophages are activated upon RSV exposure, respond with a cytokine burst, and express viral proteins (28, 29). Recent work has also implicated the role of group 2 innate epithelial lymphoid cells (IELs) in the host response to RSV, mediating interleukin-13 (IL-13) release (30). Finally, RSV productively replicates in the epithelium throughout the airway, triggering a global reprogramming of the epithelial genome (31). Our recent work using unbiased proteomics of the epithelial cell secretome and exosomes has shown that small airway epithelial cells produce unique secretory products implicated in the pathogenesis of disease, including macrophage inhibitory protein (MIP1), CCL20, thymic stromal lymphoprotein (TSLP), and others (31, 32). To provide insight into the sentinel role of epithelial cells in the response to viral PAMPs, we developed a conditional knockout (CKO) mouse by selectively deleting RelA in secretoglobin (Scgb1a1)/CC10-expressing epithelial cells. Scgb1a1 is expressed by nonciliated epithelial (Clara) and stem cell precursors that are located predominately in the bronchioles of the lower airways (33). Tamoxifen (TMX)-treated RelA^CKO mice were challenged with RSV intranasally, and initial inflammatory responses were measured. We observed that RelA inhibition significantly blocked neutrophil recruitment, the chemokine response, and type I IFN expression despite enhanced RSV transcription. Mechanistically, we observed that RelA is required for RSV induction of mucosal BRD4 histone acetyltransferase (HAT) activity and RNA Pol II carboxy-terminal domain (CTD) kinase activity. Longer-term studies demonstrated that RelA^CKO mice showed reduced disease severity and blunted airway obstruction and yet similar levels of IFN-γ production after RSV infection. We conclude that the Scgb1a1-expressing bronchiolar epithelial cell is an innate sensor of mucosal virus infection, where the RelA-BRD4 axis plays a critical role in the acute neutrophilic response. These findings validate targeting of the RelA-BRD4 pathway to prevent RSV-induced lung inflammation.

RESULTS

Conditional knockout of RelA in Scgb1a1-expressing epithelial cells.

To address the role of the distal epithelium as an innate sensor of RSV infection, we selected a strategy using a tissue-specific and inducible RelA knockout. Knowing that NF-κB plays a critical role in alveolar cell expansion (34) and growth factor-regulated lung branching (35), an inducible knockout enables experimentation after normal development and immunological maturation of the lung. We previously reported the production of a mouse strain carrying a RelA conditional knockout-ready allele (RelA^fl) that has loxP sites flanking exons 5 to 8; upon Cre-mediated recombination, the region encoding the Rel homology domain is deleted, producing an out-of-frame mRNA degraded by the nonsense-mediated RNA decay pathway (36). We crossed this mouse with a “knock-in” mouse in which an inducible Cre recombinase-estrogen receptor fusion protein (CreER) was targeted to the 3′ untranslated region (UTR) of the Scgb1a1 gene (Fig. 1A). The mouse carrying Scg1a1^CreERTM expresses Cre selectively in nonciliated epithelial cells and basal progenitor cells primarily located in the bronchioles (33).

FIG 1.

Disruption of RelA signaling in Scgb1a1-expressing epithelial cells of the bronchioles. (A) Strategy for generating a conditional knockout (CKO) of RelA in Scgb1a1⁺ epithelial cells. A RelA^fl/fl mouse was crossed with an epithelial cell-specific Scgb1a1 promoter-driven inducible Cre recombinase-estrogen receptor fusion protein (CreER) transgenic mouse. CreER does not disrupt Scgb1a1. For RelA depletion, Scgb1a1^CreERTM/+ × RelA^fl/fl mice were injected (intraperitoneally) with tamoxifen (TMX) for 10 days and infected with pRSV 3 weeks later. IRES, internal ribosome entry site; RHD, Rel homology domain. (B) Q-RT-PCR for mouse RelA (mRelA) mRNA in oil- or TMX-treated Scgb1a1^CreERTM/+ × RelA^fl/fl mice. Mice were mock infected or infected with sucrose cushion-purified RSV (1 × 10⁷ PFU, via the intranasal route). Total lung RNA was extracted and subjected to Q-RT-PCR. Shown are fold changes after normalization to cyclophilin A (PPIA) expression. #, P < 0.001 compared to uninfected mice; *, P < 0.001 compared to WT mice. The numbers of mice used in each group were 7 for the WT/PBS group, 8 for the WT/RSV group, 6 for the CKO/PBS group, and 8 for the CKO/RSV group. (C) Validation of epithelial RelA conditional knockout. Costaining of RelA (in green) and the epithelium-specific marker Scgb1a1/CC10 (in red) was performed in paraffin-embedded lung sections. (D) Quantifications of total fluorescence intensities (FI) of Scgb1a1 (CC10) and RelA, shown as fold changes of relative total fluorescence intensities compared to oil-treated mice. *, P < 0.001 (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group). (E) Scgb1a1 expression is localized to the small bronchioles. Shown is a low-magnification image of Scgb1a1/CC10-stained lung sections, including the large airways and alveoli.

Confirmation of RelA depletion in the Scgb1a1-expressing bronchiolar epithelium.

We first evaluated the effectiveness of Cre-loxP-mediated RelA depletion. Scgb1a1^CreERTM/+ × RelA^fl/fl mice were treated with oil (wild type [WT]) or TMX to excise RelA and infected with sucrose cushion-purified RSV (pRSV) (Long strain) (1 × 10⁷ PFU via the intranasal route). Uninfected mice were exposed to a sucrose cushion in phosphate-buffered saline (PBS). To examine the effect of CreER induction on RelA expression, total lung RNA was extracted, and the expression of mouse RelA (mRelA) was quantitated by quantitative real-time PCR (Q-RT-PCR). In WT mice, we observed that RelA expression was increased by 2-fold after RSV infection; both basal and RSV-induced RelA expression levels were reduced by 50% in TMX-treated RelA^CKO mice relative to controls (Fig. 1B). These data indicate that bronchiolar RelA contributes substantially to the entire lung RelA RNA pool; the upregulation of RelA in wild-type mice is consistent with our previous observations that RelA is induced by RSV by an autoregulatory mechanism (23).

To confirm RelA depletion in the small airways, the level of RelA was quantitated in Scgb1a1/CC10-expressing cells by colocalization by immunofluorescence confocal microscopy (IFCM). Scgb1a1^CreERTM/+ × RelA^fl/fl mice were exposed to oil or TMX and sacrificed 14 days later for IFCM. In WT mice, RelA was expressed abundantly throughout the lung, including the small bronchioles. Scgb1a1/CC10 was also highly expressed in epithelial cells lining the small airways of the lung, resulting in significantly enhanced merged images (Fig. 1C). TMX produced a significant depletion of RelA in Scgb1a1/CC10-expressing cells, resulting in the loss of colocalization in the merged images (Fig. 1C) and total fluorescence signals quantified in multiple images (Fig. 1D). In contrast, only very faint CC10 staining was detected in epithelial cells lining large airways and alveoli (Fig. 1E). These data indicate that Cre-loxP-mediated RelA depletion was robust in Scgb1a1/CCL10-expressing cells primarily localized in the small bronchioles.

Bronchiolar epithelial RelA^CKO blocks RSV-induced neutrophilic inflammation.

Having confirmed the effective silencing of bronchiolar RelA, we next challenged TMX-treated Scgb1a1^CreERTM/+ × RelA^fl/fl mice with RSV intranasally. In this established mouse model, a peak in cellular inflammation occurs in the lung 1 day after RSV challenge (23, 37–40); we therefore focused our experiments on this time point. To examine the extent and localization of RSV replication, paraffin-embedded lung sections were analyzed for RSV antigen expression by IFCM. We found that RSV was highly expressed in the small airway epithelium in both WT and TMX-treated RelA^CKO mice (Fig. 2A). Also, the intensity of RSV staining in TMX-treated RelA^CKO mice was stronger than that in oil-treated mice, suggesting higher levels of replication (P < 0.001 by a t test) (Fig. 2B). Under a lower-power magnification, RSV replication was evident throughout the airway, including both the epithelium of the large airways as well as that of the alveoli (Fig. 2C). The magnitude of infiltrating leukocytes was quantified by counting DAPI (4′,6-diamidino-2-phenylindole)-stained nuclei. We observed a 2.4-fold increase in cell numbers in the lungs of oil-treated mice versus a 1.6-fold increase in the lungs of TMX-treated RelA^CKO mice (P = 0.007 by a t test) (Fig. 2D).

FIG 2.

RelA^CKO blocks RSV-induced leukocytosis in bronchoalveolar lavage fluid. TMX-treated WT or Scgb1a1^CreERTM/+ × RelA^fl/fl mice (RelA^CKO) were infected with sucrose cushion-purified RSV (1 × 10⁷ PFU, via the intranasal route). Uninfected mice were treated with sucrose in PBS as a control. (A) Distribution of RSV infection. Paraffin-embedded lung sections were stained with goat polyclonal anti-RSV primary antibody (Bio-Rad). Secondary detection was performed by using Alexa 568 (red)-conjugated goat anti-rabbit IgG. Nuclei were counterstained with DAPI (blue). Images were acquired by confocal microscopy at a ×63 magnification. Note the diffuse distribution of RSV antigen throughout large and small airways. (B) Quantifications of total fluorescence intensities of RSV staining, shown as fold changes of relative total fluorescence intensities compared to oil-treated mice without RSV infection. *, P < 0.001 compared with WT mice (n = 5). (C) RSV infection in large airways and alveolar cells of the lung, shown as images acquired by confocal microscopy at a ×20 magnification. (D) Quantification of infiltrating leukocytes. The number of cells in each image was quantified by counting DAPI-stained nuclei using Fiji-ImageJ. *, P = 0.007 compared with WT mice (n = 5). (E) Total cells in the BALF, expressed as the total number of cells (10⁵) per milliliter. *, P < 0.001 compared with WT mice (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group). (F) BALF neutrophils expressed as total numbers of neutrophils (10⁵) per milliliter. (G) BALF lymphocytes expressed as total numbers of lymphocytes (10⁴) per milliliter. (H) BALF macrophages expressed as total numbers of macrophages (10⁵) per milliliter.

In a manner consistent with the quantification of infiltrating leukocytes, the total number of leukocytes in the bronchoalveolar lavage (BAL) fluid (BALF) was also significantly reduced in RelA^CKO mice (32 × 10⁵ versus 14.6 × 10⁵ leukocytes; P < 0.001 by a t test) (Fig. 2E). This reduction was primarily due to reduced neutrophilic infiltration (28 × 10⁵ versus 10.6 × 10⁵ neutrophils; P < 0.001 by a t test) (Fig. 2F), the most populous cell type, and also a reduction in lymphocyte infiltration (1.48 × 10⁴ versus 0.56 × 10⁴ lymphocytes; P < 0.001 by a t test) (Fig. 2G). Although RSV induced a smaller population of macrophages, there was no detectable changes in monocytes in RSV-infected lungs in comparisons of WT mice versus TMX-treated RelA^CKO mice (Fig. 2H).

To further examine the effects on acute tissue inflammation, lung sections from oil- or TMX-treated Scgb1a1^CreERTM/+ × RelA^fl/fl mice in the absence or presence of RSV infection were examined 1 day after infection. In WT mice, RSV infection produced marked polymorphonuclear leukocytic and lymphocytic infiltration around bronchioles and within the air spaces (Fig. 3A). In contrast, numbers of peribronchial and alveolar leukocytes were dramatically reduced in the airways, in a pattern similar to that observed in the BALF analysis. We extended the histological analysis to 3 days of RSV infection; here the polymorphonuclear leukocytic inflammation in the airspaces was reduced, but peribronchial mononuclear inflammation was still seen (Fig. 3B). The histological analysis of RSV-infected RelA^CKO mice showed no residual airspace inflammation and no consistent differences compared with uninfected controls. These data indicate that RelA signaling in the Scgb1a1-expressing epithelial cell population is required for acute RSV-induced neutrophilic inflammation.

FIG 3.

Reduced pulmonary inflammation in TMX-treated RelA^CKO mice. (A) Oil- or TMX-treated Scgb1a1^CreERTM/+ × RelA^fl/fl mice were infected with sucrose cushion-purified RSV (1 × 10⁷ PFU, via the intranasal route). Uninfected mice were treated with sucrose in PBS as a control. Mice were sacrificed after 1 day, and lungs were fixed for histochemistry. Shown are H&E stains of paraffin-embedded pulmonary sections. Image magnifications are shown at the left. (B) H&E stains of pulmonary sections 3 days after RSV infection.

RelA^CKO blocks the pulmonary RSV-induced chemokine network.

RSV infection produces time- and inoculum-dependent increases in the transcriptional levels of CC and CXC chemokine mRNAs that coordinate downstream inflammatory and adaptive responses (41). Total RNA was extracted from WT mice or TMX-treated RelA^CKO mice with or without RSV infection and subjected to analysis by Q-RT-PCR. We first examined the magnitude of RSV transcription. In WT mice, we observed marked increases in RSV G and N mRNA levels within 1 day of RSV infection, consistent with productive RSV replication in the mouse airway (Fig. 4A). In TMX-treated RelA^CKO mice, RSV transcription was further enhanced, indicating that RelA signaling in the tracheobronchiolar epithelium plays a role in innate immune defense against experimental RSV infection.

FIG 4.

Effect of RelA^CKO on RSV-induced cytokine expression. TMX-treated WT or Scgb1a1^CreERTM/+ × RelA^fl/fl mice (RelA^CKO) were infected in the absence or presence of sucrose cushion-purified RSV. Mice were sacrificed after 1 day, and total lung RNA was extracted for Q-RT-PCR. (A) RSV transcription. Shown are fold changes of RSV G and N mRNAs after normalization to mPP1A. *, P < 0.001 compared with WT mice (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group). (B) NF-κB-dependent chemokine expression. (C) Expression of the IRF-IFN pathway. *, P < 0.001. Data are means ± SD (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group).

Our previous studies have shown that RelA controls the expression levels of over 300 epithelial chemokine and innate genes in response to RSV infection (24, 25). We therefore sampled the expression of a variety of key chemokines important for neutrophil chemotaxis and/or RSV pathogenesis. We observed that mouse granulocyte colony-stimulating factor (G-CSF), Groβ, IL-6, monocyte chemoattractant protein 1 (MCP1), MIP1α/β, and RANTES were all substantially induced by RSV infection; their expression levels were significantly reduced in RelA^CKO mice despite the enhanced level of RSV transcription (Fig. 4B). These findings suggest that distal tracheobronchiolar RelA signaling contributes substantially to the total lung RNA pools in this model.

RelA signaling mediates the IRF–RIG-I amplification loop in RSV infection.

A somewhat surprising finding from our experiment is that RelA signaling contributes to initial virus restriction. To better understand this innate mechanism, we recently proposed that RelA is a master regulator of the mucosal IFN response through cross talk regulation of IFN production (20, 23, 26). Here IRF1 and IRF7 are induced in a RelA-dependent manner; newly formed IRF1 and IRF7 isoforms bind and activate the RIG-I promoter, an event essential for sustained IFN-β production. For RSV-infected WT mice, we observed that IRF1 was highly induced, 17-fold, after RSV infection; this was substantially reduced to <8-fold in RelA^CKO mice (Fig. 4C). Similar findings were observed for IRF7 mRNA expression; here the 30-fold induction of IRF7 in WT mice was reduced to ∼12-fold in TMX-treated RelA^CKO mice. Accordingly, the 15-fold induction of IFN-β mRNA in WT mice was reduced to 7-fold in TMX-treated RelA^CKO mice (Fig. 4C). Collectively, these data indicate that RelA signaling in Scgb1a1-expressing epithelial cells plays a functional role in reducing RSV transcription and activating the type I IFN response through the IRF–RIG-I amplification pathway.

RelA signaling in Scgb1a1-expressing epithelial cells is a major generator of cytokine secretion in BALF.

To demonstrate that the changes in mRNAs were significant enough to affect cytokine secretion, we analyzed BALF for changes in protein expression by a multiplex enzyme-linked immunosorbent assay (ELISA). Consistent with data from our previous studies, we observed that RSV induced the significant upregulation of G-CSF, granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-1α/β, IL-6, MIP1α/β, eotaxin, IL-12(p40), and IL-13, cytokines important for inflammation and mucin production (30, 32, 40) (Fig. 5). In contrast, for RelA^CKO mice, we observed significant reductions in the expression levels of all these proteins. For example, RSV induced MIP1α from undetectable levels to 3,000 pg/ml in RSV-infected WT mice; this was reduced to 1,200 pg/ml in RelA^CKO mice. MIP1α is significant because this chemokine is predominately expressed by small airway epithelial cells, and it contributes to lung inflammation (32, 40). In contrast, IL-13 is a mucin-producing cytokine expressed by innate lymphoid cells in a manner dependent on epithelial factors, including TSLP (30), a factor that we previously demonstrated is produced primarily by bronchiolar epithelial cells (31). Collectively, these data indicate that bronchiolar RelA contributes substantially to cytokine production in response to RSV infection.

FIG 5.

Effect of RelA^CKO on BALF cytokine concentrations. Cytokine secretion in BALF was measured by using a multiplex ELISA (n = 5 mice per group). Protein concentrations (in picograms per milliliter) are shown. *, P < 0.001. Data are means ± SD (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group).

RSV-activated RelA mediates cytokine expression in basal epithelial cells in a replication-dependent manner.

We previously demonstrated that RSV is a potent activator of RelA through canonical and noncanonical pathways, mediating epithelial cytokine expression (42, 43). Because those studies were conducted primarily with transformed type II-like alveolar cells (42–44), we sought to validate this pathway in Scgb1a1-expressing basal cells from the bronchioles (human small airway epithelial cells [hSAECs]) (31). hSAECs are permissive for RSV infection and maintain basal and inducible genomic and proteomic signatures seen in terminally differentiated primary bronchiolar cells (31). To determine whether RSV-activated RelA mediates cytokine expression in these basal epithelial cells in a replication-dependent manner, hSAECs were adsorbed with WT or UV-inactivated RSV, and the expression of the neutrophilic chemokine program was measured. In hSAECs infected with WT RSV, high levels of RSV G and N mRNAs were observed, whereas G and N expression levels were significantly reduced compared to those of controls in hSAECs adsorbed with UV-inactivated RSV, confirming inactivation (P < 0.001 by a t test) (Fig. 6A). We observed that WT RSV induced high levels of expression of IL-6, IL-8, MCP1, and RANTES mRNAs in hSAECs, but this response was completely eliminated by UV inactivation (P < 0.001 by a t test) (Fig. 6A). Furthermore, the replication-competent RSV-induced secretion of the cytokines IL-6, IL-8, MCP1, MIP1α, MIP1β, and RANTES was completely abolished by UV inactivation (P < 0.001 by a t test) (Fig. 6B).

FIG 6.

RSV replication is required for chemokine expression, RelA translocation, and activation of BRD4 activity. WT hSAECs were mock challenged (sucrose cushion) or challenged with replication-competent RSV or UV-inactivated RSV (multiplicity of infection of 1.0 for 24 h). (A) Gene transcription. RSV G, RSV N, IL-6, IL-8, MCP1, and RANTES mRNA levels were measured by Q-RT-PCR. Note that chemokine expression is completely abolished by UV inactivation. *, P < 0.001 compared with mock treatment; #, P < 0.001 compared with RSV only (n = 3). (B) Cytokine secretion. RSV-induced IL-6, IL-8, MCP1, MIP1α, MIP1β, and RANTES levels in conditioned medium were measured by an ELISA. Inducible chemokine expression was completely abolished by UV inactivation. *, P < 0.001 compared with mock treatment; #, P < 0.001 compared with RSV only (n = 3). (C) Effects of WT and UV-inactivated RSV on RelA and BRD4 activities. RelA, phospho-Ser2 CTD RNA Pol II (pPol II), H3K122 Ac, and RSV antigen levels were measured by IFCM. (D) Quantification of total fluorescence intensities, shown as fold changes of relative total fluorescence intensities (FI) compared to mock treatment. *, P < 0.001 compared with mock treatment; #, P < 0.001 compared with RSV only (n = 5).

Previously, we observed that RSV induced complex formation between activated RelA and its cofactor BRD4, a complex required for the transcriptional elongation of NF-κB-activated target genes (21, 23, 45). BRD4 plays a major role in the process of transcriptional elongation through intrinsic atypical HAT activity coupled with its serine kinase activity. BRD4 HAT activity produces destabilizing H3K122 acetylation (Ac), evicting nucleosomes from chromatin (23, 46), and its serine kinase activity produces activating phosphorylation of the RNA Pol II CTD, licensing it to become processive (21, 23, 47). Infection with WT RSV produced strongly positive RelA nuclear translocation, an ∼16-fold increase in phospho-Ser2 RNA Pol II formation, and a 12-fold increase in H3K122 Ac (Fig. 6C and D). In contrast, UV inactivation completely abolished RSV-induced RelA nuclear translocation and phospho-Ser2 RNA Pol II and H3K122 Ac formation. Consistent with RT-PCR results, cells infected with WT RSV expressed high levels of RSV antigen and formed multinucleated cells; both of these responses were not observed for UV-inactivated RSV-treated cells (Fig. 6C and D). Collectively, all of these data clearly demonstrated that RSV-activated RelA mediates cytokine expression, RelA activation, and BRD4 enzymatic activity in basal epithelial cells in a replication-dependent manner.

RSV-activated RelA mediates BRD4 activation and transcriptional elongation.

We next asked if RSV-induced BRD4 HAT activity is regulated by RelA. For this purpose, control (WT) or clustered regularly interspaced short palindromic repeat (CRISPR)-Cas9 genome-edited hSAECs depleted of RelA (RelA^−/− hSAECs) were infected by WT RSV and analyzed by IFCM. RSV infection produced uniformly strongly positive RelA nuclear translocation in WT hSAECs, but no staining was detectable in RelA^−/− hSAECs (Fig. 7A). Turning to the analysis of BRD4 activation, we observed that the 15-fold increase in H3K122 Ac formation was significantly reduced to <3-fold in RelA^−/− hSAECs (Fig. 7A, middle row). Finally, RSV infection produced a 22-fold increase in phospho-Ser2 CTD RNA Pol II formation in WT hSAECs, but this induction was reduced to <4-fold in RelA^−/− hSAECs (Fig. 7A, bottom row). To confirm that RelA was functionally silenced in these experiments, we analyzed RelA-dependent chemokine expression. The highly inducible expression of G-CSF, Groβ, IL-8, MCP1, and RANTES was significantly reduced in RelA^−/− hSAECs (Fig. 7C). Collectively, these data indicate that RSV activates BRD4 in a RelA-dependent manner.

FIG 7.

RelA is required for activation of mucosal BRD4 HAT and RNA Pol II kinase activities. (A) RelA is required for activation of BRD4 HAT and RNA Pol II kinase activities. Wild-type (RelA^+/+) and CRISPR-Cas9 genome-edited (RelA^−/−) hSAECs were mock challenged or infected with RSV (multiplicity of infection of 1.0 for 24 h). Cells were fixed and stained with primary Abs to RelA, acetylated H3K122 (H3K122ac), or phospho-Ser2 CTD RNA Pol II, as shown. Secondary detection was done with Alexa Fluor 488 (green, for RelA)-, Alexa Fluor 568 (red, for pPol II)-, and Alexa Fluor 647 (deep red, for H3K122ac)-conjugated goat anti-rabbit IgG, respectively. Nuclei were counterstained with DAPI (blue). Images were acquired by confocal microscopy at a ×63 magnification. (B) Quantification of total fluorescence intensities upon immunofluorescence staining, shown as fold changes compared to control hSAECs. *, P < 0.001 (n = 5). (C) Gene expression. Total RNA from the same experiment was extracted, and expression levels of CSF3/G-CSF, CXCL2/Groβ, CXCL8/IL-8, CCL2/MCP1, and CCL5/RANTES were quantified by Q-RT-PCR. Shown are the fold changes in mRNA abundances normalized to the value for the cyclophilin A gene (PPIA). *, P < 0.001 compared to control hSAECs. Data are the means ± SD from 3 experiments.

Scgb1a1⁺ bronchiolar RelA mediates BRD4 activation and transcriptional elongation in vivo.

To establish the relevance of the RelA-BRD4 pathway in the mouse model, TMX-treated C57BL/6 wild-type or Scgb1a1^CreERTM/+ × RelA^fl/fl mice were challenged in the absence or presence of RSV, and the activation of the phospho-RelA-BRD4-HAT pathway was assessed by IFCM. In control mice challenged with PBS, only faint phospho-Ser276 RelA signals were produced (Fig. 8A, top row). In contrast, in response to RSV infection, a 9.5-fold increase in phospho-Ser276 RelA was observed, primarily in the bronchiolar epithelium; the phospho-Ser276 RelA signal was significantly reduced in TMX-treated RelA^CKO mice (Fig. 8A, top row, and B). Because these signals were located in the epithelium, RelA activation is probably mediated by Scgb1a1-negative epithelial cells; we note that <40% of bronchiolar cells are Scgb1a1 positive (Scgb1a1⁺) (33). Additional phospho-Ser276 RelA signals are seen in cells in the interstitium, probably representing fibroblasts or innate cells whose identities will require further study. An examination of H3K122 Ac staining showed a similar pattern of inducible expression. Faint H3K122 Ac staining was observed in PBS-challenged oil-treated RelA^CKO mice that increased 10-fold in response to RSV infection (Fig. 8A, bottom row). Strikingly, this RSV-induced activation of H3K122 Ac was effectively blocked in TMX-treated RelA^CKO mice (Fig. 8A, bottom row, and B).

FIG 8.

Scgb1a1⁺ bronchiolar RelA mediates BRD4 activation and transcriptional elongation in vivo. (A) Scgb1a1^CreERTM × RelA^fl/fl mice were treated with oil or TMX prior to RSV or mock infection. Tissues were fixed, paraffin embedded, and subjected to immunofluorescence staining for phospho-Ser276 RelA (top) (in red) and H3K122 Ac (bottom) (in green) on paraffin-embedded lung sections. (B) Quantifications of total fluorescence intensities of phospho-Ser276 RelA and H3K122 Ac staining, shown as fold changes of relative total fluorescence intensities compared to WT mice. *, P < 0.001 (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 6 for the CKO/RSV group). (C) RSV induces RelA-BRD4 interactions in airway mucosa in vivo. PLA assays of RelA-BRD4 molecular interactions were performed with lung sections from oil- or TMX-treated Scgb1a1^CreERTM/+ × RelA^fl/fl mice in the absence or presence of RSV. Foci of interactions are amplified as red foci; sections are counterstained with DAPI (blue). Magnification, ×63. At the right are quantifications of data from the PLA assay. *, P < 0.001 compared to WT mice (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group). (D) Quantifications of total fluorescence intensities from a PLA assay of RelA-BRD4 molecular interactions, shown as fold changes of relative total fluorescence intensities compared to WT mice. *, P < 0.001 (n = 7 for the WT/PBS group, n = 8 for the WT/RSV group, n = 6 for the CKO/PBS group, and n = 8 for the CKO/RSV group).

Previously, we demonstrated that RSV induces the formation of BRD4-RelA complexes in situ in RSV-infected hSAECs (23). To determine changes in molecular interactions between RelA and BRD4 in the airway mucosa in vivo, we conducted proximity ligation assays (PLAs). PLAs detect atomic-distance interactions that are detected by the enzymatic ligation of antibody (Ab)-selective oligonucleotides. Rabbit anti-RelA and mouse anti-BRD4 were used for PLAs of RSV-infected lung sections. After ligation and PCR amplification, RelA-BRD4 interactions appear as fluorescent foci by confocal immunofluorescence (23). We observed that RSV induced molecular interactions between RelA and BRD4 and that this binding was disrupted in TMX-treated RelA^CKO mice (Fig. 8C and D). Collectively, these data provide compelling evidence that RelA activation in Scgb1a1-expressing bronchiolar epithelial cells is a major sensor of RSV infection, mediating BRD4-RelA complex formation, BRD4 HAT activity, and the global activation of RNA Pol II for neutrophilic inflammation and the expression of mucosal IFNs.

Scgb1a1⁺ bronchiolar RelA mediates disease and airway obstruction and participates in virus restriction in vivo.

After the rapid induction of chemokines and neutrophils (40), a time-dependent induction of both weight loss and airway obstruction are observed in the mouse model of RSV infection. These complex responses are mediated by the effects of ongoing viral replication, oxidative damage, and the adaptive immune response (48, 49). To understand the contribution of the RelA signaling pathway in the Scgb1a1⁺ epithelium on the evolution of RSV disease, we conducted a longer-term time course experiment. WT or TMX-treated RelA^CKO mice were exposed to RSV and monitored over 5 days (n = 4 males). In WT mice, RSV infection produced a characteristic, transient, 15% decline in body weight after 2 days, followed by recovery after 5 days, a pattern consistent with data from previous studies (23, 48). In contrast, TMX-treated RelA^CKO mice showed a significantly reduced body weight loss of 5% (Fig. 9A). Functionally, it is well established that RSV infection produces acute airway obstruction. For RSV-infected WT mice, we observed a significant induction of baseline enhanced pause (Penh); this obstructive physiology was significantly attenuated in TMX-treated RelA^CKO mice (Fig. 9B). As a measure of the adaptive immune response, IFN-γ concentrations in the BALF were measured. These values were indistinguishable between WT and RelA^CKO mice (Fig. 9C), despite the finding that viral replication was slightly enhanced (Fig. 9D). Collectively, these data demonstrate that RelA signaling in Scgb1a1⁺ bronchiolar epithelial cells mediates disease and airway obstruction and participates in virus restriction in vivo.

FIG 9.

Effect of RelA signaling in Scgb1a1-expressing epithelial cells on RSV disease. WT or TMX-treated RelA^CKO mice were exposed to RSV and monitored over 5 days. (A) Changes in body weights during the course of the experiment. The RSV infection-induced decline in body weights of WT mice was significantly attenuated in TMX-treated RelA^CKO mice. **, P < 0.01; ***, P < 0.001 (n = 4 [males]). (B) Effects on airway obstruction and methacholine sensitivity. Enhanced pause (Penh) was measured by whole-body plethysmography at baseline and in response to methacholine. Penh values are presented as means ± standard errors of the means (n = 4 mice/treatment group). RSV infection-induced airway obstruction was attenuated in TMX-treated RelA^CKO mice. *, P < 0.05; **, P < 0.01 (compared with WT mice infected with RSV). (C) Effects on adaptive immune responses. IFN-γ concentrations in BALF were analyzed by an ELISA and are presented in picograms per milliliter. n.s., not significant. (D) RSV replication. RSV titers in lung tissue were determined by a plaque assay. *, P < 0.01 compared with WT mice infected with RSV.

DISCUSSION

RSV is a major human pathogen that triggers host inflammatory responses, cellular infiltration, mucus production, and long-term airway remodeling. Despite an extensive understanding of intracellular pattern recognition receptors (50), the structure of the virus-induced innate signaling pathway (22, 23, 51), and mechanisms of activation of lymphocyte subsets (30, 52, 53), the initial innate cellular sensors of RSV infection have not been definitely established. Here we establish and characterize an inducible RelA knockout in the tracheobronchiolar epithelial cell lineage and demonstrate that the progeny of this cell type plays a key role in triggering the initial host neutrophilic response, chemokine expression, and type I IFN production in experimental RSV infection. We demonstrate that RelA is coupled to transcriptional elongation in vivo, validating data from our previous studies of differentiated epithelial cells.

The pulmonary mucosal innate defense is coordinated by a complex interaction between resident epithelial cells, alveolar macrophages, innate lymphoid cells, and other innate leukocytes (15, 16). Epithelial cells are a dynamically responsive and structurally diverse population of cells lining the respiratory airway, forming a semi-impermeable barrier that plays a role in clearing inhaled microparticles by the ciliary escalator, secreting antiviral mucins, and expressing innate cytokines that play a major role in antiviral host defense (15). Despite this understanding, the relative contributions of epithelial versus other sentinel cells of the airways (macrophages, innate epithelial lymphocytes, NK cells, and others) have not been definitely established. Here we demonstrate that an Scgb1a1/CC10-expressing lineage of epithelial cells represents the major sensor of viral patterns in the airway lumen and contributes substantially to the production of inflammatory mediators. Scgb1a1/CC10 is selectively expressed in nonciliated epithelial cells by differentiated Clara cells and basal progenitor cells but not ciliated or neuroepithelial cells that are enriched in the bronchioles of the airway (33). The Scgb1a1-expressing population of epithelial cells plays important roles in xenobiotic metabolism and as stem cell precursors to repopulate the postnatal airway after injury.

Our studies extend the function of the Scgb1a1-expressing epithelial cell population as a major mediator of the antiviral response. It is interesting that ciliated bronchial epithelial cells are infected by RSV, where they respond by activating innate pathways yet apparently do not mediate cellular inflammation (54, 55). Why, then, would the depletion of NF-κB/RelA selectively in the tracheobronchial cell population have such a dramatic effect on airway inflammation? One explanation may be provided by our systems-level proteomics studies that have shown cell type differences in innate cytokine production in response to RSV infection. Here we observed that airway epithelial cells dynamically respond to viral patterns in a cell type-specific pattern by synthesizing and secreting over 400 proteins as free and membrane-bound nanoparticles (exosomes) in addition to well-known cytokines, CXC and CC-type chemokines, type I and III IFNs, and IFN-stimulated genes (ISGs) (25, 31, 56). Although significant numbers of proteins were shared between tracheal and Scgb1a1-expressing bronchiolar cells, bronchiolar cells selectively secreted 103 unique proteins and exosomal contents, including MIP1, TSLP, and CCL20, proteins that promote lymphocyte inflammation (40), IEL activation (30), Th2 polarization (57), and mucin expression (31). These mediators coordinate the complex processes important for the pathogenesis of RSV-induced lower respiratory tract infections. These regional differences in chemokine expression patterns are therefore immunologically significant.

Our approach using a TMX-inducible RelA knockout in Scgb1a1-expressing cells is distinct from several studies that have used the same promoter to inhibit the canonical NF-κB pathway using the inducible dominant negative expression of the IκBα inhibitor. Those studies implicated the canonical pathway in airway inflammation and fibrosis in dust mite allergen exposure (58). The expression of the nondegradable IκBα “superrepressor” selectively inhibits NF-κB activation by the canonical pathway (59). It is important to recognize that NF-κB activation by viral patterns involves both the canonical and cross talk pathways (42, 43); our novel RelA^CKO mouse interferes with both the canonical (IκBα) and the cross talk (RelA–NF-κB2) pathways; the latter is not affected by the Scgb1a1-IκBα_SR system.

Although the IRF pathway is thought to be a major mediator of mucosal cytokine production, our studies indicate that NF-κB signaling plays an important role in virus restriction and triggering the mucosal IFN response. Not only is the IFN-β enhanceosome recognized by NF-κB, but also the robust IFN autoregulatory pathway is dependent on the inducible expression of the IRF1 and -7 subunits mediated by direct NF-κB transactivation (20, 23, 26). NF-κB mediates rapid genomic responses by recruiting transcriptional elongation complexes to immediate early genes. RSV replication induces reactive oxygen species (ROS) production that interfaces with DNA damage response pathways to produce Ser phosphorylation of RelA (22), a posttranslational modification coupled with lysine acetylation that promotes complex formation with BRD4 (21). BRD4 is a chromatin-organizing protein containing intrinsic kinase activity and atypical histone acetyltransferase activity that bridges RelA with cyclin-dependent kinase 9 (CDK9) of the positive transcriptional elongation complex. BRD4 is an essential coactivator in the transcriptional elongation complex, playing a pleiotropic role in transcriptional elongation. In addition to serving as a bridging molecule linking NF-κB/RelA to chromatin-bound acetylated histones, BRD4 has intrinsic serine kinase activity, activating both CDK9 and RNA Pol II (46, 47). Phosphorylation on Ser2 of the heptad repeats in the CTD of RNA Pol II is a key regulatory step in the activation of NF-κB-dependent genes. Our findings that Ser2 CTD RNA Pol II activation is dependent on RelA as well as BRD4 suggest that the NF-κB/RelA-BRD4 complex is a global regulator of total nuclear RNA Pol II activity. In addition, BRD4 is an atypical HAT producing H3K122 Ac, a modification that reduces the histone side charge and enables RNA Pol II to transcribe through GC-rich regions of innate immune response genes. Like BRD4 Pol II kinase activity, BRD4 HAT activity is highly upregulated by activated RelA downstream of diverse innate stimuli.

Apart from its acute inflammatory effects, RSV-induced LRTI has been linked with recurrent episodic wheezing and decreased lung function (11, 13). In the mouse model, enhanced obstruction is observed by Penh. Our studies indicate that airway obstruction and methacholine hypersensitivity are downstream of the bronchiolar epithelial RelA pathway. Because our study was not powered to detect sex effects, further studies will be required to examine this issue. Intriguingly, although acute leukocytosis is reduced and viral replication is slightly increased in the RelA^CKO model, the adaptive immune response appears not to be affected. These slightly enhanced levels of viral replication are probably explained by the reduced levels of mucosal IFNs. Our studies here extend this work to indicate that the Scgb1a1-expressing population is primarily responsible for the acute neutrophilic response, linked to disease manifestations and airway obstruction. Further work will be needed to determine whether a blockade of bronchiolar RelA is effective in reducing airway remodeling as a consequence of viral infections.

MATERIALS AND METHODS

Preparation of sucrose cushion-purified RSV.

The human RSV Long strain was grown in Hep-2 cells and prepared by sucrose cushion centrifugation as described previously (44, 60). The viral titer was determined by a methylcellulose plaque assay. pRSV aliquots were quick-frozen in dry ice-ethanol and stored at −70°C until use.

To inactivate replicating virus, pRSV was diluted in 1 ml of minimal essential medium (MEM) containing 2% fetal bovine serum (FBS) and exposed for 3 min to a 254-nm UV light source at a 10-cm distance on ice.

Construction of the RelA conditional knockout.

Animal experiments were performed according to the NIH Guide for the Care and Use of Laboratory Animals (61) and approved by the University of Texas Medical Branch (UTMB) Animal Care and Use Committee (approval no. 1312058A and 9001002). The construction of the RelA^fl/fl mouse, containing loxP sites inserted into introns 4 and 8 of RelA, was described previously (36). RelA^fl/fl mice were crossed with B6N.129S6(Cg)-Scgb1a1^{tm1(cre/ERT)Blh}/J mice (stock no. 016225; Jackson Laboratories). The presence of Scgb1a1^{tm1(cre/ERT)Blh} and RelA^fl alleles was determined by PCR.

For RelA depletion, both adult male and female mice, 3 to 4 weeks old, were injected with tamoxifen (catalog no. T5648; Sigma) at 1 mg/day intraperitoneally for 10 days. Tamoxifen was dissolved in 10% ethanol and 90% corn oil for a 10-mg/ml working solution. After 3 weeks, mice were infected with sucrose cushion-purified RSV. One day later, the mice were euthanized. In one experiment, 4 mice did not show evidence of recombination and were excluded from subsequent analyses. We used both TMX-treated wild-type (C57BL/6 mice) and oil-treated Scgb1a1^CreERTM/+ × RelA^fl/fl mice as controls; both types of mice produced similar levels of RelA expression, cytokine-inducible gene expression, and cellular inflammation.

For bronchoalveolar lavage (BAL), the trachea was cannulated, and 1 ml of PBS was introduced by syringe. The BAL fluid was analyzed for total and differential cell counts and secreted cytokine/chemokine levels by using multiplex immunoassays (BioPlex). Half of each lung was fixed with 10% (vol/vol) neutral buffered formalin for 3 days, processed into paraffin blocks, and cut into 5-μm sections for hematoxylin and eosin (H&E) staining. The other half of each lung was immediately frozen in liquid N₂ and pulverized. One milliliter of Tri reagent was added, followed by the extraction of total RNA according to the manufacturer's directions. The total RNA was reverse transcribed, and gene expression was quantified by using Q-RT-PCR.

Immunofluorescence confocal microscopy.

Formalin-fixed, paraffin-embedded lung sections were rehydrated by using serial concentrations of ethanol. Antigen retrieval was performed with antigen-unmasking solution based on recommendations from Abcam (Tris-EDTA [TE] buffer, pH 9.0). Paraffin-embedded sections were blocked by using 0.1% Triton X-100–5% normal goat serum and incubated with rabbit anti-RelA (Santa Cruz, Dallas, TX), anti-CC10, anti-acetyl H3K122, and anti-p276 RelA (Abcam, Cambridge, MA) Abs (1:200 dilution) overnight at 4°C. Normal anti-rabbit IgGs were used as staining specificity controls. After washing, cells were stained with Alexa Fluor 488- or 568-conjugated goat anti-rabbit IgG (Life Technologies) in incubation buffer for 1 h and then counterstained with the nuclear marker DAPI (Sigma-Aldrich, St. Louis, MO) and visualized with an LSM510 fluorescence confocal microscope at a ×63 magnification (58, 59).

Cultured hSAECs were plated onto rat tail collagen-treated cover glasses and stimulated for the indicated times. The cells were fixed with 4% paraformaldehyde in PBS and incubated with 0.1 M ammonium chloride for 10 min. Cells were permeabilized with 0.5% Triton X-100, followed by incubation in blocking buffer (5% goat serum, 0.1% IGEPAL CA-630, 0.05% NaN₃, and 1% bovine serum albumin [BSA]) and incubation with anti-RelA (Santa Cruz, Dallas, TX), anti-H3K122 Ac, and anti-phospho-Pol II (Abcam, Cambridge, MA) antibodies in incubation buffer (0.1% IGEPAL CA-630, 0.05% NaN₃, and 2% BSA) overnight at 4°C. After washing, cells were stained with Alexa Fluor 488-, 568-, and 647-conjugated goat anti-rabbit IgG (Life Technologies, Carlsbad, CA), respectively, in incubation buffer for 1 h and then counterstained with the nuclear marker DAPI (Sigma-Aldrich, St. Louis, MO) and visualized with a Nikon fluorescence confocal microscope at a ×63 magnification (23).

BALF cellular analysis.

Total cell counts were determined by trypan blue staining of 50 μl of BALF and counting of viable cells by using a hemocytometer. Differential cell counts were performed on cytocentrifuge preparations (Cytospin 3; Thermo Shandon, Pittsburgh, PA) stained with Wright-Giemsa stain. A total of 300 cells per sample were counted by using light microscopy. Formalin-fixed lungs were embedded in paraffin, sectioned at a 4-μm thickness, and stained with hematoxylin and eosin or Masson's trichrome. Microscopy was performed on a Nikon Eclipse Ti system (17, 23).

Cell culture, treatment, and RelA silencing.

Immortalized hSAECs were previously described (23). hSAECs were grown in small airway epithelial cell growth medium (SAGM; Lonza, Walkersville, MD) in a humidified atmosphere of 5% CO₂. RelA deletion in hSAECs was performed by CRISPR-Cas9 genome editing. A 20-nucleotide (nt) seed sequence in exon 4 of the RelA gene was used as a small guide RNA (sgRNA)/guide sequence to introduce a stop codon that induces nonsense-mediated RNA decay.

Quantitative real-time PCR.

For gene expression analyses, 1 μg of RNA was reverse transcribed by using SuperScript III as previously described (21, 24). One microliter of the cDNA product was amplified by using SYBR green supermix (Bio-Rad) and gene-specific primers. The reaction mixtures were subjected to 40 cycles of 15 s at 94°C, 60 s at 60°C, and 1 min at 72°C in an iCycler apparatus (Bio-Rad). Quantification of relative changes in gene expression levels was performed by using the ΔΔC_T method, and expression as the fold change between experimental and control samples was normalized to the cyclophilin A (PPIA) internal control.

In situ proximity ligation assay.

Paraffin-embedded lung section slides of mock-treated or RSV-infected mice were subjected to antigen retrieval, permeabilized with 0.1% Triton X-100, and incubated with IgG or primary rabbit Ab to RelA (Santa Cruz) and monoclonal antibody (MAb) to BRD4 (Sigma-Aldrich). Slides were then subjected to PLA using the Duolink PLA kit from OLink Bioscience (Uppsala, Sweden) according to the manufacturer's instructions. The nuclei were counterstained with DAPI, and the PLA signals were visualized with an LSM510 fluorescence confocal microscope at a ×63 magnification.

Measurements of IFN-γ.

The IFN-γ levels in BAL fluid samples were measured by an ELISA, according to the manufacturer's protocol (Invitrogen, Waltham, MA). The range of sensitivity of the assay is 0.63 to 40 pg/ml.

RSV titration of lung tissue.

Lungs were removed at day 5 after RSV infection. Tissue samples were homogenized in 1 ml of Dulbecco's modified Eagle's medium and centrifuged twice at 14,000 rpm for 1 min at 4°C. Titers of serial 2-fold dilutions of the supernatant were determined by a plaque assay on Hep-2 cells under a methylcellulose overlay. Plaques were visualized 5 days later, and virus titers are expressed as log₁₀ PFU per gram of tissue.

Pulmonary function testing.

Airway hyperresponsiveness (AHR) was assessed in unrestrained mice at day 5 after infection by using whole-body barometric plethysmography (Buxco, Troy, NY) to record enhanced pause (Penh), as previously described (37). Penh is a dimensionless value that represents a function of the ratio of the peak expiratory flow to the peak inspiratory flow and a function of the timing of expiration. Respiratory activity was recorded for 5 min to establish baseline Penh values. Mice were subsequently exposed to increasing doses of nebulized methacholine (6.25, 12.5, 25, and 50 mg/ml) for 30 s, and data were recorded for another 3 min.

Statistical analysis.

One-way analysis of variance (ANOVA) was performed for time differences, followed by Tukey's post hoc test to determine significance. Mann-Whitney tests were used for nonparametric data. All data subjected to statistical analysis are means ± standard deviations (SD). A P value of <0.05 was considered significant.