Influenza A Virus Infection Causes Chronic Lung Disease Linked to Sites of Active Viral-RNA Remnants

Abstract

Clinical and experimental observations suggest that chronic lung disease is linked to respiratory viral infection. However, the long-term aspect of this relationship is not yet defined using a virus that replicates at properly high levels in humans and a corresponding animal model. Here we show that influenza A virus infection achieves 1×106-fold increases in viral load in the lung and dose-dependent severity of acute illness in mice. Moreover, these events are followed by persistence of negative- and positive-strand viral-RNA remnants for 15 weeks and chronic lung disease for at least 26 weeks after infection. The disease is manifested by focal areas of bronchiolization and mucus production that contain increased levels of viral-RNA remnants along with mucin Muc5ac and Il13 mRNA compared to uninvolved areas of the lung. Excess mucus production and associated airway hyper-reactivity (but not fibrosis or emphysema) are partially attenuated with loss of IL-13 production or signaling (using mice with IL-13- or STAT6-deficiency). These deficiencies cause reciprocal increases in l17a mRNA and neutrophils in the lung, however, none of these disease endpoints are changed with IL-13–IL-17a-compared to IL-13-deficiency or STAT6-IL-17a-compared to STAT6-deficiency. The results establish the capacity of a potent human respiratory virus to produce chronic lung disease focally at sites of active viral-RNA remnants, likely reflecting locations of viral replication that reprogram the region. Viral dose-dependency of disease also implicates high-level viral replication and severity of acute infection as determinants of chronic lung diseases such as asthma and COPD with IL-13-dependent and IL-13/IL-17-independent mechanisms.

Introduction

Respiratory viral infections are linked to chronic lung diseases, both in childhood asthma and in adults with severe asthma and chronic obstructive pulmonary disease (COPD) (1). In humans, the link between viral infection and chronic lung disease was initially associated with severe infections due to respiratory syncytial virus (RSV) (2, 3) but later extended to human metapneumovirus and human rhinovirus (HRV) (4–7). Moreover, some reports suggest that the type of virus is not critical to the subsequent development of post-viral disease (8). In experimental mouse models, chronic lung disease can develop after a severe infection with mouse parainfluenza virus also known as Sendai virus (SeV) (9). In this case, the mechanism for chronic lung disease is based on long-term activation of a type 2 immune response that includes IL-13-driven mucus production and airway hyper-reactivity (10–12). The same disease traits are characteristic of patients with chronic lung disease due to asthma and COPD (10, 11, 13–15) and thereby help to validate the proposal that transient viral infections might trigger long-term consequences for initiation, exacerbation, or progression of chronic lung disease. However, the SeV-driven model relies on a virus that infects but does not cause significant disease in humans (16) and therefore might not be representative of more pathogenic human respiratory viruses.

In the present study, we address the need for a model of post-viral lung disease based on a potent human pathogen, focusing on the proposal that the severity of acute infection might determine the subsequent development of chronic lung disease. In a search for other viral pathogens responsible for respiratory illness, we recognized that some of the most common human pathogens linked to chronic lung disease do not replicate efficiently in mice. For example, viral loads in the lung for RSV (e.g., A2 strain) and HRV (e.g., A16 or B1 strains) do not increase from the initial amount of virus administered to the animal, even in mice that are immunocompromised due to age, genetic background, or targeted modifications to delete anti-viral host defense genes or overexpress viral receptors (17–25). Moreover, these immune alterations might have effects on acute infection that may or may not coincide with effects on the development of chronic lung disease, e.g., interferon (IFN) signaling separately influences viral replication and type 2 immune response (26, 27).

To improve a post-viral lung disease model using a human pathogen, we aimed to explore preliminary associations between influenza A virus (IAV) infection and subsequent wheezing illness in infants as well as an established connection of influenza A virus with exacerbations of asthma and COPD in children and adults (28–31). In addition, there were reports of persistent pod-like and cystic structures at 21–200 days after IAV (PR8 strain) infection in mice (32–34). Although these investigators connected these structures to lung regeneration and fibrotic disease, there were also reports of mucous-like cells after IAV infection (with PR8 and HKx31 strains) in mice (35, 36). These initial studies did not examine long-term time points or define mechanism. Our findings show that IAV replicates in mouse lung to levels that are a 1×10⁶-fold higher than initial dosing and causes infection that progresses to chronic lung disease for at least 6 months. The disease is linked to initial viral load and severity of acute infection as well as subsequent IL-13 and mucin (MUC5AC) but not IL-17a induction at sites of viral-RNA remnants. The findings thereby establish the capacity of a virus with high-level replication in humans and mice to cause a severe acute illness and switch the susceptible host to chronic lung disease. This disease is linked at least in part to a localized, long-term type 2 immune response with excess inflammatory mucus production but also to an IL-13/IL-17-independent process for mucus formation as well as fibrosis and emphysema.

Materials and Methods

Mice

Wild-type C56BL/6J mice (000664), Stat6^–/– mice (005977) (37), and Il17a^–/– mice (016879) (38) mice were obtained from The Jackson Laboratory. Breeding pairs of Il13^–/– mice were obtained from T. Wynn (39). Breeding was used to generate Il13^–/––Stat6^–/– and Il13^–/––Il17a^–/– mice. All mice were maintained on a C57BL/6J background and were co-housed in a barrier facility using cages fitted with micro-isolator lids. Animal husbandry and experimental procedures were approved by the Animal Studies Committees of Washington University School of Medicine in accordance with the guidelines from the National Institutes of Health.

Virus preparation and infection

IAV strain A/WS/33 was obtained from ATCC (VR-1520) and propagated in chicken embryo allantoic cavities as described previously (40). IAV strain A/PR/8/34 was obtained from Jacco Boon (Washington University). Aliquots of stock virus were stored at −70 °C and were titered using a plaque-forming assay with MDCK cells (ATCC CCL-34), 1 μg/ml of acetylated trypsin (Sigma T-6763), and agarose overlay as described previously (40). Mice were inoculated with IAV-WS/33 (0.04–5.0 pfu) or IAV-PR8 (10 pfu) in 30 μl of PBS given intranasally or with an equivalent amount of UV-inactivated virus or PBS at 5–6 weeks of age under ketamine/xylazine anesthesia. Results from male and female mice were pooled since no significant differences were found between sexes. Mouse lungs were frozen at −70 °C for homogenization in PBS with a cell disrupter (Mini-Beadbeater-96, Biospec Products) followed by viral plaque-forming assay to track pfu level normalized to gm of lung tissue (recognizing that one mouse lung weighs approximately 0.1 gm) (41). For viral titer and other assays in involved versus uninvolved lung tissue, the regions were determined by visual inspection and then separated by gross dissection.

RNA and protein analysis

Total RNA was isolated from lung. tracheobronchial lymph node, spleen, and thymus tissue using the Qiagen RNeasy kit and converted to cDNA using the High-Capacity cDNA Archive kit and random hexamer primers (Applied Biosystems). Levels of viral RNA were monitored using a real-time qPCR assay for the IAV segment 3 containing the polymerase acid (PA) gene or IAV segment 7 containing the matrix (M) gene with forward and reverse primers and MGB probes described in Supplemental Table I. To quantify the level of viral RNA, segments of the PA gene (nt 259–944) or M gene (nt 25–124) were cloned using the TA cloning kit (ThermoFisher Scientific) and used as standards. Lung levels of Muc5ac mRNA were determined using a real-time quantitative PCR (qPCR) assay as described previously (10). Lung levels of Clca1, IL13, IL17a and IL17f were determined using qPCR assay probes described in Supplemental Table II. All target mRNA and viral RNA levels were normalized to Gapdh or Mrlp19 mRNA level using the TaqMan Rodent GAPDH Control Kit or a Mrlp19-specific qPCR assay described in Supplemental Table II. All mRNA values were expressed as copy number based on comparison to the corresponding plasmid cDNA standard. Total lung protein was prepared using bead homogenization in T-per buffer with protease inhibitor (ThermoFisher Scientific). Homogenized samples were cleared with centrifugation, and the supernatant was analyzed using a quantitative ELISA for IL-13 (R&D DuoSet) according to the manufacturer’s instructions.

Strand-specific real-time qPCR assay for IAV RNA encoding the PA and NP genes was performed using tagged PCR primers and high-temperature reverse transcription based on an approach described previously (42) with primers and probes described in Supplemental Table I. Reverse transcription with the tagged primers was performed using SuperScript™ III reverse transcriptase (Invitrogen) and trehalose (Sigma) at 60 °C as described previously (43). For the PA gene, assay specificity was validated using positive and negative RNA strands generated from a Bluescript SK(+) plasmid encoding the PA gene segment (nt 259–944) under separate gene promoters with opposite orientations. For the NP gene, assay conditions were the same as those described previously for this method (43). In both cases, the DNA template was eliminated with TRIzol (Invitrogen) followed by treatment with DNase and the RNeasy Mini kit (QIAGEN).

Tissue staining and microscopy

Tissues were fixed with 10% neutral-buffered formalin, embedded in paraffin, cut into 5-μm thick sections, dewaxed, rehydrated, and incubated with mouse anti-MUC5AC mAb (clone 45M1, ThermoFisher Scientific) or control mAb as described previously (10). Immunofluorescence microscopy for virus was performed using goat anti-IAV H1N1 antibody (20-IG23, Fitzgerald Industries International) and tyramide-based signal amplification with Alexa Fluor 594 fluorochrome (Invitrogen). Slides were counterstained with DAPI-containing mounting media (Vector Laboratories). Sections were also subjected to PAS-hematoxylin and Gomori trichrome staining with quantification using a NanoZoomer S60 slide scanner (Hamamatsu) and Visiomorph (Visiopharm) or ImageJ (44) software with settings specific for the purple-magenta-tone of mucins or light-blue tone of collagen staining. Mean linear intercept values for alveolar size were derived as described previously (45).

Airway reactivity

Airway reactivity to nebulized methacholine (Sigma, St. Louis, MO) was determined by measurements of respiratory system resistance (R_RS) using the FinePointe Resistance and Compliance system (DSI Buxco Research Systems, Wilmington, NC) as described previously (10, 11). Mice were anesthetized, ventilated via tracheostomy, and sequentially challenged with aerosolized PBS (baseline) followed by doubling doses of methacholine ranging from 0.6125 to 20 mg/ml. Methacholine was delivered at 3-min intervals using an in-line nebulizer (Aerogen Laboratory; 2.4–4 μm particle size). Resistance values were recorded during a 3-min period following each challenge. Data were manually verified, and spurious epochs removed from analysis.

Statistical analysis

Unless stated otherwise, all data are presented as mean ± SEM and two-tailed unpaired Student’s t-test was used to assess statistical significance between means. In all cases, the threshold for statistical significance was set at a P value less than 0.05. Airway reactivity was assessed using either two-way repeated measures analysis of variance or restricted maximum likelihood linear mixed model with genotype, infection, and methacholine dose as fixed effects and subject as a random effect.

Results

IAV infection results in dose-dependent illness and leaves active viral-RNA remnants

Initial experiments showed that intranasal delivery of IAV (A/WS/33 strain) caused dose-dependent levels of acute illness as monitored by body weight loss (Fig. 1A) in concert with progressive increases in viral load based on plaque-forming assay for infectious virus (Fig. 1B) or qPCR assay for the viral polymerase acidic (PA) gene product (Fig. 1C). The levels of infectious IAV detected in lung tissue indicated marked viral replication, e.g., mice dosed with 2 pfu resulted in recovery of 2 × 10⁷ pfu/gm of lung tissue (equivalent to 2 × 10⁶ pfu/lung) at 3 days after infection (Fig. 1D). Despite clearance of infectious IAV by 12 days after infection based on plaque assay (Fig. 1D) and viral immunostaining (Fig. 1E), IAV-specific RNA was detectable in the lung even at 105 days after infection (Fig. 1F). IAV RNA also persisted in lung lymph node at 77 days, spleen at 49 days, and thymus at 3 days after infection (Fig. 1G). Since conventional qPCR assay can produce non-strand-specific products, we also applied an approach based on the use of tagged PCR primers (i.e., tag-PCR assay) designed for the IAV PA gene to distinguish between positive and negative strands of viral RNA (Fig. 1H). This approach provided for specific detection of positive versus negative-strand RNA from the PA gene (Fig. 1I) and demonstrated persistence of positive-strand viral RNA for 49 days after infection (Figure 1J). Similarly, we found persistence of IAV NP gene expression based on conventional PCR assay (Fig. 1K), and positive-strand IAV RNA (Fig. 1L) using a strand-specific assay described previously (43). Together, these findings indicated that IAV is capable of producing coding RNA for prolonged periods of time, consistent with reports of chronic antigen and T cell activation (46–49).

FIGURE 1.

Acute illness leaves viral remnants after IAV infection. (A) Time course of body weight change in wild-type C56BL/6J mice after infection with IAV (WS/33 strain, 0–5 pfu). Mice were euthanized at body weight <75% of initial value. (B) Corresponding lung levels of IAV monitored with plaque-forming assay for conditions in (A) at 3 d after infection. (C) Corresponding lung levels of IAV RNA encoding the PA gene monitored with PCR assay for conditions in (B). (D) Time course of lung levels of IAV monitored with plaque-forming assay after infection with IAV (2 pfu). (E) Time course for IAV immunostaining in lung sections after infection with IAV (2 pfu). Bar=400 μm. (F) Time course for lung levels of IAV RNA after infection with IAV (2 pfu). (G) Corresponding IAV-PA RNA levels in tracheobronchial lymph nodes, spleen, and thymus. For (H), * indicates p<0.05 versus no infection control. (H) Scheme for strand-specific qPCR assay in which the negative (–) RNA strand of viral RNA (vRNA) (labeled in magenta) or the positive (+) RNA strand of viral RNA (cRNA, mRNA) (labeled in blue) were reverse transcribed to corresponding cDNA using tagged primers that included a portion complementary to each strand (green) and another portion that was not related to viral sequence (orange). The resulting tagged cDNA was amplified using PCR in which one primer was specific to the tagged portion of cDNA (orange) and another primer (green) and probe (black) were specific to viral sequence. (I) Validation of strand-specific tag-PCR assay in which (+) and (–) strand RNA standards were transcribed from a plasmid encoding the IAV-PA gene and cDNA was synthesized using primers specific for the (–) or (+) strand of viral RNA. (J) Time course of lung levels of IAV-PA negative- and positive-strand RNA detected using tag-PCR assay. (K) Time course of lung levels of IAV RNA encoding the NP gene (IAV-NP) detected using PCR assay. (L) Time course of lung levels of IAV (+) strand RNA encoding IAV-NP detected using tag-PCR assay. For (A)-(G) and (I)-(L), values are representative of 3 separate experiments (n=8 mice per condition in each experiment). Values for 0 d were not significantly different than UV-inactivated IAV (IAV-UV) at 3–182 d (data not shown).

IAV infection also causes dose-dependent chronic lung disease

We next considered whether the persistence of IAV-RNA remnants might be associated with similarly long-lasting lung disease after viral infection. Indeed, we found that airway epithelial cell damage and immune cell accumulation was followed by the development of PAS-positive airway epithelial cells and bronchiolization consistent with the formation of mucous cells at 21 days after infection (Fig. 2A). Quantification of PAS-positive staining was performed using image analysis software (Fig. 2B) and showed increased staining in proportion to initial viral dose at 21 d after infection (Fig. 2C). Similarly, mucin Muc5ac and Il13 mRNA were also induced in proportion to viral dosing up to 2–5 pfu (Fig. 2D) and were maintained at increased levels with a maximum at 21 days and persistence to 182 days after infection (Fig. 2E). Consistent with induction of Il13 mRNA, we also detected a significant increase in IL-13 protein levels in lung tissue at 21 d after IAV infection (Fig. 2F), recognizing that this approach reflects IL-13 unbound to IL-13-receptor and the consequent need for additional experiments to establish pathophysiological significance of IL-13 expression in this model.

FIGURE 2.

Viral dose drives the development of chronic lung disease after IAV infection. (A) Time course for PAS-hematoxylin staining in lung sections after infection with IAV (WS/33 strain, 2 pfu). Bars=400 μm. (B) PAS-hematoxylin staining of lung sections for airway and parenchymal tissue at 21 d after IAV infection (2 pfu) with green colorization indicating computer-assigned PAS⁺ areas. Bar = 200 μm. (C) Quantitation of PAS⁺ area based on computer-assignment and quantification for conditions in (B). (D) Lung levels of Muc5ac and Il13 mRNA at 21 d after infection with IAV (0–5 pfu). (E) Time course for lung levels of Muc5ac and Il13 mRNA after infection with IAV (2 pfu). (F) Levels of IL-13 protein in lung tissue at 21 d after infection with IAV or IAV-UV. Dashed lines indicate the lower limit of detection for the assay. For (A)-(F), values are representative of 3 separate experiments (n≥8 mice per condition in each experiment). * indicates p<0.05 versus no infection or IAV-UV control (no difference was found for no infection versus IAV-UV).

Meanwhile, however, a second IAV strain (A/PR/8/34) was dosed to produce a similar degree of weight loss as IAV-WS/33 (Fig. 3A) and was also found to produce similar lung disease (Fig. 3B) along with persistence of viral RNA and induction of Muc5ac and Il13 mRNA expression (Fig. 3C) at 21 d after infection, indicating that these post-viral effects were not specific to the WS/33 strain. Together, the results suggested a link between the severity of initial IAV infection and the subsequent development of the inflammatory mucus formation that is characteristic of chronic lung diseases.

FIGURE 3.

Severe acute illness leads to chronic lung disease after IAV-PR8 infection. (A) Time course of body weight after infection with IAV (PR8 strain, 10 pfu or WS/33 strain, 2 pfu) or corresponding UV-inactivated IAV-PR8 in wild-type C56BL/6J mice. (B) PAS and-hematoxylin staining of lung sections at 21 d after infection with IAV-PR8 (10 pfu) or IAV-PR8-UV. Bar=400 μm. (C) Corresponding lung levels of IAV-PA RNA and Muc5ac and Il13 mRNA for conditions in (B). For (A,C), values are representative of 3 separate experiments (n≥8 mice per condition in each experiment). For (C), * indicates p<0.05.

Detection of disease foci containing viral remnants and mucous cells

We recognized that chronic lung disease in humans is characterized by considerable heterogeneity across the lung tissue, so that some parts of the lung are markedly abnormal whereas other parts show little pathology even at advanced stages of disease (11, 50–52). We observed similar heterogeneity for post-viral lung disease in whole lung specimens of mice that were infected with IAV (Fig. 4A). We took advantage of this finding by dissecting involved from uninvolved sections of lung to assess levels of viral-RNA remnants and mucus production on a regional instead of whole lung level. PAS staining confirmed the gross pathology and showed evidence of mucous cell formation in the involved versus uninvolved sections at 21, 49, and 182 days after IAV infection (Figure 4B). In concert with the abnormalities in histopathology, we found corresponding increases in IAV-PA RNA in involved relative to uninvolved sections of lung at each of the same time points after IAV infection (Fig. 4C), although (as noted in Fig. 1F and Fig. 1K), IAV RNA was no longer detectable by 182 days after infection. We also found increased levels of Muc5ac and Il13 mRNA in involved versus uninvolved sections, with maximal values at 21 days declining to lower values at 182 days after IAV infection (Fig. 4C). The findings imply that the areas of the lung that were initially infected with virus and then leave viral-RNA remnants are most likely to manifest disease, but the continued presence of these remnants is not required for persistence of disease.

FIGURE 4.

Identification of focal areas of chronic lung disease after IAV infection. (A) Representative photographs of lung samples with grossly uninvolved or involved sections at 49 and 182 d (without and with separation by dissection) after infection with IAV (WS33 strain, 2 pfu). Bar=1 mm. (B) Representative PAS-hematoxylin staining of uninvolved or involved lung sections at 21, 49, and 182 d after infection with IAV (2 pfu). Bars=400 μm. (C) Levels of IAV-PA RNA and Muc5ac and Il13 mRNA in lung samples for conditions in (B) and control IAV-UV. For (A)-(C), values are representative of 3 separate experiments (n≥8 mice per condition in each experiment). For (C), * indicates p<0.05 versus uninvolved section.

Post-IAV disease is linked in part to IL-13 induction

Levels of IL-13 production and signaling via the IL-13-receptor–STAT6 complex is a major determinant of airway mucus production after infection with SeV in mice (10, 11, 53). Accordingly, we next studied post-IAV disease in Il13^–/– and Stat6^–/– mice that exhibit body weight loss and viral levels in the lung that were no different than wild-type mice (Fig. 5A,B). In contrast to the similarities for acute illness among these mouse strains, the levels of Muc5ac mRNA induction were significantly attenuated in Il13^–/– and Stat6^–/– mice compared to wild-type control mice at 21 days after IAV infection (Fig. 5C). In addition, the usual induction of Clca1 mRNA was completely blocked in Il13^–/– and Stat6^–/– mice compared to wild-type control mice at 21 days after IAV infection (Fig. 5C). Expression of Clca1 mRNA is highly sensitive to IL-13 stimulation (14, 54), indicating that the IL-13-sensitive component of mucus production is effectively blocked in these mice and that there is an additional IL-13-independent component of inflammatory disease and/or mucus production after IAV infection. To address this issue, the level of PAS-hematoxylin staining (as an index of mucus-containing disease foci) was assessed using whole-lung scans of sections from Il13^–/– and wild-type control mice at 21 d after infection with IAV or IAV-UV (Fig. 5D). With this approach, we found significant attenuation of PAS⁺-staining area in Il13^–/– mice compared to wild-type mice (Fig. 5E). We also found that the persistent levels of IAV RNA and induction of Muc5ac mRNA in involved versus uninvolved lung sections from microdissection were no different in Il13^–/– versus wild-type mice at 21 days after infection (Figure 5F). Together, these findings indicate that IAV drives focal lung disease with bronchiolization and mucus production that is at least partially but not completely dependent on IL-13 production and signaling. In addition, the results indicated that the levels but not the viral-RNA and mucus characteristics of disease foci were sensitive to IL-13 deficiency.

FIGURE 5.

IL-13 induction and STAT6 activation are linked to chronic lung disease after IAV infection. (A) Time course of body weight change in wild-type (WT), Il13^–/–, and Stat6^–/– mice on the indicated days after infection with IAV (2 pfu) or IAV-UV. (B) Time course for lung levels of IAV-PA RNA in the same mouse strains after infection with IAV (2 pfu). (C) Lung levels of Il13, Muc5ac, and Clca1 mRNA in the same mouse strains at 21 d after infection with IAV (2 pfu) or IA-UV. (D) PAS-hematoxylin staining of lung sections from WT and Il13^–/–mice at 21 d after infection with IAV (2 pfu) or IAV-UV. Bar=1 mm. (E) Quantitation of PAS-positive areas in images in (D). (F) Quantitation of PAS⁺ area for conditions in (D). (F) Levels of IAV-PA RNA and Muc5ac mRNA in uninvolved or involved sections of lung from WT and Il13^–/– mice at 21 d after infection with IAV or IAV-UV. For (A-C) and (E-F), values are representative of 3 separate experiments (n≥8 mice per condition in each experiment). * indicates p<0.05; ns=nonsignificant.

Post-IAV disease is not linked to reciprocal IL-17 production

Other investigators have reported that IL-17 might emerge as a candidate for driving airway inflammation and mucus production, particularly in the setting of IL-13 blockade that appears to enhance IL-17a expression level (55–57). In fact, we also observed an increase in Il17a mRNA in Il13^–/– mice but not in wild-type mice at 21 d after IAV infection (Fig. 6A). We found no significant change in Il17f mRNA in Il13^–/– or combined Il13^–/––Il17a^–/– mice. Accordingly, we next studied post-IAV lung disease Il17a^–/– and combined Il13^–/––Il17a^–/– mice. Initial experiments showed that body weight loss and viral levels in the lung that were no different in these strains compared to Il13^–/– or wild-type mice (Fig. 6B,6C). In addition, the levels of Muc5ac or Clca1 RNA induction were not significantly different in Il17a^–/– mice compared to wild-type mice or in combined Il13^–/––Il17a^–/– mice compared to Il13^–/–mice at 21 days after IAV infection (Fig. 6D), ruling against any effect of either Il17a or Il17f on Muc5ac expression. We also observed selective increases in neutrophil (but not eosinophil, macrophage, or lymphocyte) levels in BAL fluid in Il13^–/– mice compared to wild-type mice, and this increase was blocked in both Il17a^–/– and combined Il13^–/––Il17a^–/– mice (Fig. 6E), ruling against any effect of IL-17-dependent neutrophil influx on mucus production.

FIGURE 6.

Reciprocal IL-17a induction after IL-13 blockade is not linked to chronic lung disease after IAV infection. (A) Lung levels of Il17a and Il17f mRNA in WT, Il13^–/–, Il17a^–/–, and Il13^–/––Il17a^–/– mice at 21 d after infection with IAV (2 pfu) or IA-UV. (B) Time course of body weight change in the same mouse strains as (A) after infection with IAV (2 pfu) or IAV-UV. (C) Time course for lung levels of IAV-PA RNA for conditions in (B). (D) Lung levels of Il13, Muc5ac, and Clca1 mRNA in same mouse strains as (A) at 21 d after infection with IAV (2 pfu) or IAV-UV. (E) Levels of indicated immune cells in BAL fluid for conditions in (D). (F) Levels of airway reactivity using response of respiratory system resistance (R_RS) to inhaled methacholine (MCh) and for baseline R_RS for conditions in (D). For (A-F), values are representative of 3 separate experiments (n=5–8 mice per condition in each experiment). For (A-E), * indicates p<0.05; for (F), * indicates p<0.05 versus IAV-UV and ** indicates p<0.05 versus corresponding IAV-UV and IAV conditions.

In addition to excess mucus production, we also found significant increases in airway reactivity to inhaled methacholine in wild-type mice at 21 d after infection with IAV compared to IAV-UV (Fig. 6F). Similar to the pattern observed for regulation of Muc5ac expression, the development of airway hyper-reactivity was attenuated in Il13^–/– mice compared to wild-type mice but were not significantly different in Il17a^–/– mice compared to wild-type mice or in combined Il13^–/––Il17a^–/– mice compared to Il13^–/–mice at 21 days after IAV infection (Fig. 6F). Comparable to our previous studies of SeV infection, we detected no significant increase in baseline respiratory system resistance (R_RS) at 21 d after viral infection in wild-type or gene-knockout mice (Fig. 6F). Comparisons to pulmonary function testing in asthma and COPD are possible, and perhaps the combination of normal baseline but increased reactivity is more typical of asthma. However, it is prudent to recognize that the measurements in mice are made under general anesthesia that likely attenuates any increase in baseline or post-challenge respiratory resistance (58). Thus, IL-17a-dependent increases in neutrophil influx did not significantly influence key features of Muc5ac induction or airway hyper-reactivity that develop during chronic lung disease due to IAV infection.

Based on the reported sensitivity of Stat6^–/– mice to IL-17a stimulation of mucus production (56), we also studied combined Stat6^{–/––/–}–Il17a^–/– mice to check for any effect of IL-17a on chronic lung disease after IAV infection. Initial experiments again demonstrated that body weight loss and viral levels in the lung that were no different in combined Stat6^{–/––/–}–Il17a^–/– mice compared to Stat6^–/–mice (Fig. 7A,7B). In addition, we found induction of Il17a mRNA and attenuated induction of Il13 mRNA in Stat6^–/– mice that were similar to Il13^–/– mice (Fig. 7C). Moreover, the levels of Muc5ac RNA induction were not significantly different in combined Stat6^–/––Il17a^–/– mice compared to Stat6^–/–mice at 21 days after IAV infection (Fig. 7C). Thus, effects of IL-13 and IL-13R-signaling deficiencies were quite similar in the setting of IAV-induced lung disease.

FIGURE 7.

Reciprocal IL-17a induction due to STAT6-deficiency is not linked to chronic lung disease after IAV infection. (A) Time course of body weight change on the indicated days in Stat6^–/– and Stat6^–/––Il17a^–/– mice after infection with IAV (2 pfu) or IAV-UV. (B) Time course for lung levels of IAV-PA RNA in same mouse strains as (A) after infection with IAV (2 pfu). (C) Lung levels of Il17a, Il13, and Muc5ac mRNA in same mouse strains as (A) at 21 d after infection with IAV (2 pfu) or IA-UV. For (A), (B), (C), values are representative of 3 separate experiments (n≥5 mice per condition in each experiment). * indicates p<0.05.

Stable characteristics of lung disease at later times after IAV infection

To determine whether longer-term post-viral disease was still due to a persistent type 2 immune mechanism, we also characterized chronic lung disease in mice at 49 days after IAV infection. We found that the pattern of Il17a, Il13, and Muc5ac mRNA expression in Il13^–/–, Stat6^–/–, Il17a^–/–, and combined Il13^–/––Il17a^–/– mice were remarkably similar at 49 days compared to 21 days after IAV infection (Fig. 8A). Thus, Il13 and Muc5ac RNA induction were similarly attenuated in Il13^–/–, Stat6^–/– and combined Il13^–/––Il17a^–/– mice compared to wild-type mice at 49 days after IAV infection (Fig. 8A). In addition, PAS-hematoxylin and MUC5AC immunostaining of lung sections showed attenuation of mucus production in Il13^–/– mice compared to wild-type mice at 49 days after IAV infection (Fig. 8B). Furthermore, PAS⁺-staining area was significantly attenuated in Il13^–/– mice compared to wild-type mice at 49 days after IAV infection (Fig. 8C,D). The development of airway hyper-reactivity at 49 d was also similar to 21 d after IAV infection, exhibiting partial dependence on Il13 and Stat6 gene function (Fig. 8E). Together, these findings indicate that IAV drives chronic lung disease with excess mucus production that depends at least in part on a persistent IL-13-signaling mechanism that is similar at 21 and 49 d after infection.

FIGURE 8.

Persistence of IL-13–STAT6-linked chronic lung disease after IAV infection. (A) Lung levels of Il17a, Il13, and Muc5ac mRNA in WT, Il13^–/–, Stat6^–/–, Il17a^–/–, and Il13^–/––Il17a^–/– mice at 49 d after infection with IAV (2 pfu) or IA-UV. (B) Representative PAS-hematoxylin staining and Muc5ac-hematoxylin immunostaining of lung sections from WT and Il13^–/–mice at 49 d after infection with IAV (2 pfu) or IAV-UV. Bars=400 μm. (C) PAS-hematoxylin staining of lung sections from WT and Il13^–/–mice at 21 d after infection with IAV (2 pfu) or IAV-UV. Bar=1 mm. (D) Quantitation of PAS-positive areas in images in (D). (E) Levels of airway reactivity using response R_RS to inhaled MCh and for baseline R_RS for conditions in (A). For (A)-(E), values are representative of 3 separate experiments (n≥8 mice per condition in each experiment). For (A), (C), and (D), * indicates p<0.05.

To further address the comparison between the post-IAV mouse model with chronic lung disease in humans, we also assessed the development of pulmonary fibrosis and emphysema during post-viral disease. We detected peribronchiolar and parenchymal fibrosis at 21 d after IAV infection based on Gomori trichrome staining for the blue-tone of collagen expression (Fig. 9A). The level of collagen staining was significantly increased in wild-type mice at 21 and 49 d after IAV infection, however, the same increase was also observed in Il13^–/–, Il17a^–/–, Il13^–/––Il17a^–/–, and Stat6^–/– mice based on precision image-analysis of the blue-collagen signal (Fig. 9B-D). Significant increases in the same fibrotic signal were also detected at 21 d after IAV infection in wild-type and Il13^–/– mice using the additional strain of IAV (A/PR/8/34) (Fig. 9E). In addition, we detected pulmonary emphysema in wild-type mice at 21 and 49 d after IAV infection based on mean linear intercept values for alveolar size (Fig. 9F,G). Similar to the data for fibrosis, we observed no difference in the degree of emphysema between wild-type and the four gene-knockout strains after IAV infection (Fig. 9F,G), and we confirmed similarly significant increases in alveolar size after IAV infection with a second strain of IAV (Fig. 9H). The values for alveolar enlargement were similar to those determined by hyperpolarized ³He MRI-based imaging in the post-SeV model (59). Together, the post-IAV mouse model exhibits key features of chronic lung disease (airway inflammation, mucus production, hyper-reactivity, fibrosis, and emphysema) as found in humans with asthma and COPD, but also provides additional insight into the variable dependence of these disease endpoints on IL-13-versus non-IL-13/IL-17-dependent signals.

FIGURE 9.

Persistent peribronchial and parenchymal fibrosis and pulmonary emphysema after IAV infection. (A) Gomori trichrome staining of lung sections from WT mice at 21 d after infection with IAV (2 pfu) or IAV-UV (WS/33 strain). Bar=1 mm. (B) Quantitation of trichrome-blue-positive areas for lung sections from the indicated mouse strains after infection with IAV or IAV-UV as described in (A). (C) Quantitation of trichrome-blue-positive areas for lung sections from the indicated mouse strains at 21 d after infection with IAV (10 pfu) or IAV-UV (PR8 strain). (D) Mean linear intercept values for indicated mouse strains after infection with IAV or IAV-UV (WS/33 strain) as described in (A). (E) Mean linear intercept values for indicated mouse strains after infection with IAV or IAV-UV (PR8 strain) as described in (C). For (B)-(E), values are representative of 3 separate experiments (n≥8 mice per condition in each experiment), and * indicates p<0.05.

Discussion

This study demonstrates previously unreported capacities of IAV to persist as active viral-RNA remnants in the lung and to cause a switch to chronic lung disease. The results provide for new host-pathogen principles as well as implications for the development of chronic inflammatory disease based on the following key findings: First, we show that IAV persists in a transcriptionally active form that is incapable of replication or generation of infectious virus but survives in the lung cell compartment for up to 3–6 months in wild-type mice. Second, in this mouse model, we demonstrate that IAV infection triggers long-term lung disease that continues after IAV-viral remnants are no longer detectable even in foci of inflammatory disease that are enriched for viral RNA, mucus production, and IL-13 expression. We further show that these disease foci of immune cell accumulation and mucus production are linked to induction of IL-13 expression and STAT6 activation but not reciprocal induction of IL-17a expression. Here we discuss the novelty and significance of these findings for the virus and the host.

In regards to the first issue related to IAV persistence, a combination of previous studies showed that viral antigen could be detected for prolonged periods. For example, viral antigen can be found for at least 70 days in draining lymph node based on activation of virus-specific CD4⁺ and CD8⁺ T cells (46–49, 60). These observations indicated the presence of a persistent depot of viral antigen long after clearance of infectious virus but left open the mechanism for the source of this antigen. Viral RNA was detected in the lymph node for up to 10 days and in the lung for up to 30 days after IAV infection in mice, leading to a proposal that respiratory dendritic cells serve to transport viral antigen from the lung to the draining lymph nodes as an explanation for long lasting stimulation of T cells at that site (49). However, the nature of viral RNA production remained uncertain. Here we show that viral RNA can be detected for at least 77 days in lung lymph nodes, suggesting this site as a possible source of viral antigen for the reported T cell stimulation. In support of this possibility, we also demonstrate that IAV remains transcriptionally active in this setting. We also find that IAV RNA can be present in the lung for at least 105 days after infection., and we localize the viral RNA to foci of disease. These findings provide a basis for even longer-term stimulation of the immune system that may have escaped detection in previous studies. The results support the possibility that the host may retain virus to keep the system armed for the next infection or that the virus may use the protected site in the host as an opportunity to improve infectious capability. For both strategies, it appears that virus and lung cells are able to co-exist without immune-mediated destruction for prolonged periods, in contrast to the cytopathic process found during acute infection. Further studies must determine the host cell for viral RNA remnants and the possible function of these cellular and molecular components in causing chronic inflammatory disease.

Co-persistence of viral RNA and post-viral lung disease is not specific to IAV infection. In particular, there is also persistent viral RNA in the lung for at least 49 d after infection after SeV infection in mice (10). In this case, we did not yet establish whether viral-RNA remnants are transcriptionally active or whether they become undetectable despite persistent lung disease. Similarly, we do not yet know whether SeV-RNA remnants are localized to a specific host tissue compartment or cell type that might be associated with lung disease in this model. This determination should also prove quite useful given the extensive definition for the immune mechanism of chronic lung disease after SeV infection (10–12, 53) and the possible overlap with at least the type 2 immune process after IAV infection that was found in the present study.

In regards to the second issue concerning post-viral lung disease, to our knowledge, our results are the first to fully model this type of disease for a potent human respiratory viral pathogen with highly efficient replication in a mouse model. In particular, IAV achieves peak viral titers that are increased 1 × 10⁶-fold from the initial inoculum and maintains persistence of infectious virus for 10 days after inoculation. In contrast, RSV peak titer barely reaches the level of the initial inoculum, and virus is fully cleared by 8 days after infection (17–21). Even less impressive, HRV peak titer is generally 1×10³-fold lower than initial inoculum and virus is cleared by 4 days after infection (22–25), thereby often requiring high-viral dosing or genetically-modified mice that still show relatively little viral replication (22, 24, 25). Moreover, the previous studies of IAV infection in mice suggest a degree of mucous cell metaplasia at 3 weeks after infection but did not study long-term outcome or viral persistence (35). Similarly, other studies suggest that IAV infection in mice might enhance or inhibit allergen-induced lung disease but do not assess the direct effects of viral infection on long-term lung disease (36, 61–63). None of these previous studies analyzed the critical effect of viral inoculum size and consequent viral titer and severity of acute illness on the subsequent development of chronic inflammatory disease as we show in the present study. This finding is consistent with the epidemiology of RSV-induced asthma, where children with the most severe cases of acute bronchiolitis are the most likely to develop chronic wheezing illnesses (2, 64). We would further expect that individuals with the most severe forms of acute influenza illness will also be at greatest risk to develop chronic lung disease, perhaps with the same chronic inflammatory foci as we observed in the mouse model of this process. The recent outbreak of H1N1 swine-origin influenza virus further suggests that the development of post-viral asthma could be compounded by an increased susceptibility to severe influenza in the asthmatic or asthma-prone population (65).

Given the possible relevance of our results to human disease, we also defined the immune basis for chronic lung disease after IAV infection. We established that the development of post-IAV disease, characterized by immune cell accumulation, bronchiolization, and excess mucus production depends at least in part on IL-13 expression and signaling via IL-13-receptor-associated STAT6 that is characteristic of the type 2 immune response. These findings are consistent with long-term IL-13-dependent disease after SeV infection in mice (10–12, 53). The combined results also define auto-amplification of IL-13 induction, given the decrease in IL-13 levels in Stat6^–/– mice, consistent with IL-13 feed-forward and IL-13R induction in post-SeV disease in mice and COPD in humans (10–12). The present analysis also shows that the lung disease develops and persists as discrete foci within the lung at sites of active viral-RNA remnants that are likely also the locations of initial viral replication during acute infectious illness. In addition, we find that IL-13 drives an increase in foci of disease with no significant effect on the level of viral-RNA remnants or mucus production within these foci. The full significance of this finding still needs to be determined, but it already seems likely to be linked to the initial severity of infection and level of immune response, similar to the connection to viral dosing. In any case, these observations appear consistent with the localized abnormalities found with lung histology and imaging in humans with chronic lung disease due to asthma and COPD (66–68).

Our results also provide new insight into the relationship between the type 2 immune response and the possible role of IL-17 in the development of chronic lung disease. In that regard, previous work implicated increased IL-17A levels in the pathogenesis of this type of disease in humans and mouse models (69–75) under a mechanism that might be linked to neutrophil function (76–78) and might be increased with IL-13 or IL-13-signal blockade (55–57). However, we find that IL-17 does not contribute to foci of disease or excess mucus production after IAV infection, despite increases in IL-17- and IL-17A-dependent increases in neutrophils. Similarly, we found no effect of IL17a blockade on IL-13 expression, in contrast to other experimental models (57, 79) and no synergy for IL-17 and IL-13 despite reports of this effect in airway inflammatory disease (74). We also found no induction of IL-17F expression that might explain an independent effect on IL-17RA signaling despite reports of this mechanism in other models (80). Our findings are consistent with the lack of therapeutic effect in initial clinical trials of anti-IL-17RA mAb in asthma (81), although analysis of patient subsets and clinical trials with anti-IL-17A mAb (ClinicalTrials.gov NCT03299686) remain ongoing.

A key implication of our study is the extent of extrapolation from the post-viral mouse model to chronic lung disease in humans. In that regard, it is critical to recognize that respiratory viral infection and post-viral disease often involves both airway and alveolar sites. Indeed, the post-IAV mouse model shows elements of airway disease (i.e., airway inflammation, excess airway mucus production, airway hyper-reactivity, and peribronchiolar fibrosis) and alveolar/parenchymal disease (i.e., bronchiolization, emphysema, and fibrosis). Therefore, the model has been proposed to represent experimental counterparts of chronic airway disease (as found in chronic bronchitis/COPD and asthma) as well as pulmonary fibrosis (as found in interstitial lung diseases) and acute lung injury (as found in ARDS) (33, 82, 83). Each of these proposals offers support for corresponding mouse and human cell and molecular candidates for pathogenesis. The model will provide utility to the extent that these candidates prove useful as biomarkers and/or targets for therapeutic interventions in human disease. In the case of airway disease, the post-viral mouse models for IAV and SeV already provide a useful substrate to discover new small-molecule drugs for attenuating mucus production (14). In addition, the data from the post-viral mouse model bolsters ongoing development of biologics directed against IL-13. In that regard, a blocking mAb directed against the IL4Ralpha chain that is common to IL-13 and IL-4 receptors (84–86) but not against IL-13 alone (87–89) might be more effective in decreasing exacerbations of eosinophilic asthma. This difference might be due to the need to interrupt both IL-13- and IL-4-triggered signals in allergic asthma, but this possibility still needs to be tested directly, e.g., by direct comparison of mAb against IL-4/IL-13-receptor versus IL-13-receptor alone. This issue and the related issue of patient stratification will be equally challenging in COPD, where there also appears to be a type 2 immune response as reported by our group (10, 11, 13–15) and by others (90–93).

The present study establishes a new experimental model for developing precise therapeutic strategies for chronic lung disease based on viral-remnant RNA, IL-13-dependent, and non-IL-13–IL-17-dependent mechanisms for chronic lung disease in humans as well as insights into the upstream driver for IL-13 production. This will require definition of expression and function of epithelial cell-derived cytokines such as IL-33, IL-25, and TSLP in the post-IAV model as was done for SeV infection and was translated to humans as was done for COPD (11). This information will also serve to guide therapeutic strategies with mAbs against IL-33 and TSLP that are also being developed for use in humans with chronic lung disease due to asthma and COPD (94, 95). In fact, comparison of mechanism for post-viral lung disease after IAV versus SeV will likely prove useful for stratification of patients with viral exacerbations of chronic lung disease. In particular, the present findings expose disease endpoints that are independent of IL-13 and IL-17 expression and must therefore be controlled through new mechanisms that are not typically considered as a consequence of viral infection.

Abbreviations used in this article:

BAL

bronchoalveolar lavage

Clca1

chloride channel accessory 1

COPD

chronic obstructive pulmonary disease

HRV

human rhinovirus

IAV

influenza A virus

MUC5AC

mucin 5AC

NP

nucleoprotein

pfu

plaque-forming unit

R_RS

respiratory system resistance

RSV

respiratory syncytial virus

SeV

Sendai virus