Diacetyl inhalation impairs airway epithelial repair in mice infected with influenza A virus

Abstract

Bronchiolitis obliterans (BO) is a debilitating disease of the small airways that can develop following exposure to toxic chemicals as well as respiratory tract infections. BO development is strongly associated with diacetyl (DA) inhalation exposures at occupationally relevant concentrations or severe influenza A viral (IAV) infections. However, it remains unclear whether lower dose exposures or more mild IAV infections can result in similar pathology. In the current work, we combined these two common environmental exposures, DA and IAV, to test whether shorter DA exposures followed by sublethal IAV infection would result in similar airways disease. Adult mice exposed to DA vapors 1 h/day for 5 consecutive days followed by infection with the airway-tropic IAV H3N2 (HKx31) resulted in increased mortality, increased bronchoalveolar lavage (BAL) neutrophil percentage, mixed obstruction and restriction by lung function, and subsequent airway remodeling. Exposure to DA or IAV alone failed to result in significant pathology, whereas mice exposed to DA + IAV showed increased α-smooth muscle actin (αSMA) and epithelial cells coexpressing the basal cell marker keratin 5 (KRT5) with the club cell marker SCGB1A1. To test whether DA exposure impairs epithelial repair after IAV infection, mice were infected first with IAV and then exposed to DA during airway epithelial repair. Mice exposed to IAV + DA developed similar airway remodeling with increased subepithelial αSMA and epithelial cells coexpressing KRT5 and SCGB1A1. Our findings reveal an underappreciated concept that common environmental insults while seemingly harmless by themselves can have catastrophic implications on lung function and long-term respiratory health when combined.

INTRODUCTION

Diacetyl (DA; 2,3-butanedione) is a flavoring chemical added or naturally occurring in many foods, beverages, and e-cigarette liquids (1–3). When inhaled at occupationally relevant concentrations, DA can cause significant lung disease in the form of bronchiolitis obliterans (BO), a fibrotic lung disease of the small airways. This debilitating airways disease can progress to respiratory failure and even lung transplantation (4, 5). Although prior preclinical models of DA inhalation exposures have shown the capacity of DA to induce airway remodeling (6–8), there remains a continued need to better understand mechanistically how DA exposure contributes to fibrotic airways disease.

Preclinical modeling of DA vapor exposures in both mice and rats has shown significant toxicity associated with acute, subacute, and chronic DA inhalation (8–10). Acute DA vapor exposure at occupationally relevant concentrations damages the airway epithelium (6, 9,10). In subacute exposures, rats exposed to equivalent concentrations of DA vapors develop BO-like lesions with airway epithelial metaplasia, subepithelial collagen deposition, and bronchial lumen narrowing (11). As a potential link between acute toxicity and persistent airway remodeling, Sprague-Dawley rats exposed repetitively to DA vapors for 5 days develop damaged and ubiquitinated proteins that localize to the airway epithelium (12). This airway epithelial damage precedes BO-like lesions and is primarily mediated through nonenzymatic reaction of DA with common airway epithelial proteins, such as the intermediate filaments keratin 5 and 8 (6, 12–16). When monitored for weeks after DA exposure cessation, this damaged airway epithelium fails to fully repair with epithelial metaplasia and ulceration adjacent to basement membrane fibroproliferation (12). These findings suggest that DA inhalation not only causes acute airway injury but also impairs epithelial repair, contributing to the airway pathology.

Respiratory tract infections are a second environmental exposure associated with BO development, often referred to as postinfectious bronchiolitis obliterans (PIBO) (17–19). Adenovirus is the most common respiratory infection associated with PIBO worldwide; however, other common respiratory pathogens including influenza, respiratory syncytial virus (RSV), measles, parainfluenza, and mycoplasma pneumoniae have also been associated with PIBO in both children and adults (19). Yet, thousands of individuals are infected with these common respiratory viruses annually and do not develop PIBO. Similar to DA inhalation, the primary target for most of these respiratory viruses is the small airway epithelium. Hence, damage to the airway epithelium is an important event common to most environmental exposures and associated with BO disease pathogenesis. However, damage to the airway epithelium alone is insufficient to explain BO development. The severity of viral infection and the need for mechanical ventilation are two known risk factors for BO development (20–22). Otherwise, our ability to identify those infected with a respiratory tract infection who will progress to end-stage airways disease (i.e., BO) remains poor. Other yet-to-be identified environmental or genetic risk factors likely contribute substantially to PIBO development, especially in those infected with common viruses like influenza.

The purpose of this work was to evaluate whether exposure to the mild, airway-trophic influenza A virus (IAV) strain H3N2 (HKx31) and lower cumulative doses of DA vapor inhalation could impair airway epithelial repair and promote airway remodeling in C57BL/6J mice. We developed a “two-hit” exposure model of chemical inhalation exposure to diacetyl (DA) vapors followed by respiratory tract infection with influenza A virus (IAV) for testing this hypothesis of impaired airway epithelial repair in airways disease development following combined environmental exposures. Although DA vapor or IAV exposure alone failed to impair airway epithelial repair nor promote airways lesions in mice, we discovered that combined exposures to DA + IAV significantly impair airway epithelial repair and promote airway remodeling.

MATERIALS AND METHODS

Mice

All experiments followed Institutional Animal Care and Use Committee guidelines and were approved by the University Committee on Animal Resources at the University of Rochester. Wild-type, 8–10 week-old C57BL/6J female mice were purchased from Jackson Laboratories (Bar Harbor, ME). Female mice were used considering their well-established response to influenza A virus, as described previously (23, 24). Prior to any exposures, mice were assigned to an individual group matched between groups for average starting weight (19.2 ± 0.2 g). Four groups included 1) room air (RA) + sham (phosphate-buffered solution; PBS), 2) diacetyl (DA) + Sham, 3) RA + influenza A virus (IAV), or 4) DA + IAV.

In subsequent DA exposures to assess airway epithelial repair following IAV infection, mice were infected with IAV first and then exposed to DA. The chosen time period of DA exposure (days 9–14 after IAV inoculation) is when lungs are no longer actively infected with IAV and the associated acute inflammatory response has subsided (24, 25). Groups for these experiments included 1) Sham + RA, 2) Sham + DA, 3) IAV + Sham, or 4) IAV + DA. All exposures were performed two times independently for both models: DA + IAV or IAV + DA.

Diacetyl Vapor Exposures

Diacetyl (DA; 2,3-butanedione; 98% purity; Sigma-Aldrich, St. Louis, MO) vapor inhalation exposures were performed at the University of Rochester’s Inhalation Exposure Facility (IEF), as described previously (12). Whole body inhalation chambers were used for all exposures (RA or DA). On the day of exposures, mice were placed into compartmentalized inserts (stainless steel-reinforced) Plexiglas boxes located within a vented chemical fume hood with continuous monitoring of DA concentrations (Thermo Electron MIRAN1A CVF; Fisher Scientific Inc., Waltham, MA), temperature and humidity (24°C –28°C and 30%–60%, respectively). Animals were exposed to 200 parts per million (ppm) for 1 h/day for 5 days. Justification for the use of 200 ppm as an exposure concentration is supported by studies in mice using acute, subacute, and chronic DA exposures of DA (8, 26). For acute DA exposures in mice, a 1 h/day exposure limited nasal and laryngeal toxicity (in comparison to a 6 h/day) but also led to peribronchiolar inflammation without significant airways remodeling nor changes in BALF cell counts or composition (8). For extrapolation to human exposures, 200 ppm DA is comparable with 12 ppm in humans when corrected for nasal scrubbing due to larger surface area in rodents and DA concentration seen for occupational exposures (5, 27). Twelve ppm is a DA concentration similar to that seen in occupational exposures in popcorn factories or coffee roasting (5, 28–30). Immediately after exposure, animals were returned to housing monitoring weights and activity every other day.

Influenza a Virus

Mice were anesthetized with an intraperitoneal injection of avertin (2,2,2-tribromoethanol; Sigma-Aldrich, Milwaukee, WI) and inoculated intranasally with 120 hemagglutinating units (HAUs) of influenza A virus (strain HKx31, H3N2; IAV) in 25-µL PBS, as described previously (31, 32). This dose of virus is sublethal in immunocompetent adult mice (33, 34). As a control for infection, other mice received the equivalent volume of 25-µL PBS (Sham group). Following inoculation, mice were housed in microisolator cages in a specified pathogen-free environment with weight monitored every other day.

Bronchoalveolar Lavage Analyses

For experiments performed on bronchoalveolar lavage (BAL), the trachea was cannulated; the right mainstem bronchi tied off, and the left lung was lavaged with three (3×) 0.5-mL washes of phosphate-buffered saline (PBS). BAL samples were centrifuged at 500 g for 10 min at 4°C. Supernatants from the first lavage were analyzed for BAL total protein (BCA Assay; Pierce Biotechnology, Rockford, IL), whereas cell pellets from all 3 washes were combined and resuspended using 1-mL PBS, and cytospin slides were prepared using 50,000 cells per slide. Cell counts and differentials (>200 cells/slide) were performed on all prepared slides stained with Diff-Quik (Siemens, DE) and reported as cells per microliter BAL.

Viral Foci Assay

Viral foci assay was performed on mice exposed to IAV, as reported previously (31, 35). Briefly, frozen right lower lobes from individual mice were resuspended and homogenized in 1 mL of ice-cold PBS. Following centrifugation (400 g for 5 min at 4°C), 15 µL of supernatant was diluted (1:10) in serum-free medium containing penicillin (100 U/mL), streptomycin (100 µg/mL), gentamicin (50 µg/mL), amphotericin B (1.25 µg/mL) PSGA with 4 µg/mL trypsin. Madin-Darby canine kidney (MDCK) cells grown to confluence were inoculated with 100 µL of diluted supernatant for 1 h in 96-well flat bottom tissue culture plates (quadruplicate wells/mouse lung). After 1 h, all wells were aspirated and replaced with fresh serum-free media for overnight incubation at 37°C. The following day, cells were fixed in 4% paraformaldehyde for 15 min at room temperature, rinsed with PBS, and then incubated for 1 h with the monoclonal antibody anti-influenza A nucleoprotein (NP; 1:1,000 dilution, Millipore MAB8258B) at 37°C. Following aspiration and PBS wash (3×), cells were incubated with an immunofluorescent secondary antibody (donkey anti-goat FITC, Invitrogen A16006) for 1 h at 37°C. Following wash (PBS 3×), all wells were visualized for fluorescent foci using a microscope with images obtained at ×4 magnification (Leica DM5500). The number of foci/mL in each well was enumerated correcting for dilution [× 4 for total surface area × 10 for dilution × 20 for volume (50 µL)/well]. The number of foci/condition was averaged over the 4 wells enumerated.

Pulmonary Function Testing

Pulmonary function tests were performed on mice, as described previously (36). Briefly, mice were anesthetized with a ketamine/xylazine mixture [100 mg/kg (Par Pharmaceutical, Chestnut Ridge, NY) and 20 mg/kg (Acorn, Inc., Lake Forest, IL), respectively] and muscle-relaxed with pancuronium bromide (10 mg/kg, Sigma-Aldrich, St. Louis, MO) to ensure passive ventilation. Mice were then ventilated with a tidal volume of 10 mL/kg, 150 breaths/min, PEEP of 3 cmH₂O, and $\text{F}{\text{I}}_{{\text{O}}_2}$ of 21% (SCIREQ Inc., Montreal, Canada). Baseline data and step-wise pressure-volume (PV) curves were obtained after ∼5 min of equilibration without evidence of spontaneous respiratory effort. Respiratory function data were obtained using the forced oscillation technique (37–39) and analyzed using the constant phase model (40, 41) to calculate the following parameters in triplicate: Respiratory system resistance (R_rs), Newtonian airway resistance (R_N), respiratory system compliance (C_rs), and tissue damping (G). Pressure volume curves for each group were analyzed using a customized computer script in an unbiased fashion.

Lung Harvest and Histology

A subset of mice, who did not undergo lung function testing, received a terminal dose (100 mg/g ip) pentobarbital/phenytoin (Virbac; Nice, FR) followed by exsanguination. Lungs were removed 14 days after the last exposure (either day 19 in DA + IAV or day 28 in IAV + DA). The right lobes were fixed and processed for histology. The left lobe was flash frozen in liquid nitrogen for future protein analysis. For histology, lung tissues were inflation fixed overnight at 4°C in 10% neutral buffered formalin, embedded in paraffin, sectioned, and stained as previously described (42). Sections were stained with hematoxylin and eosin (H+E) and antibodies against KRT5 (rabbit, 1:500; Thermo Fisher Scientific), SCGB1A1 (goat, 1:1,000; gift from B. Stripp, Cedars-Sinai Medical Center, Los Angeles, CA), cytochrome p450 enzyme 2f2 (CYP2F2; mouse, Santa Cruz Biotechnology, Dallas, TX), and α -smooth muscle actin (αSMA; mouse, 1:200; Sigma-Aldrich A2547). Sections were incubated with fluorescently labeled secondary antibody (1:200; Jackson immune Research) and stained with 4', 6-diamidino-2-phenylindole (DAPI; Life Technologies). Collagen was detected using Gomori’s trichrome stain (Richard-Allan Scientific Co., Kalamazoo, MI). Sections were examined and photographed using a Nikon E800 Fluorescence microscope (Nikon Instruments, Microvideo Instruments, Avon, MA) and a SPOT-RT digital camera (Diagnostic Instruments, Sterling Heights, MI).

Lung Histopathology Scoring

Semiquantification of histopathology was performed on all fixed and embedded lung tissue using a grading scheme, as published previously, with the reviewer blinded to the underlying exposure condition (8, 43). Briefly, terminal bronchioles were defined by airways immediately proximal to the bronchoalveolar duct junction (BADJ) with airways identified in each mouse lung section identified at ×20 magnification. Four terminal bronchioles were assessed per mouse. For each terminal bronchiole, a severity grade was assigned for both inflammation and airway epithelial remodeling. Inflammation severity was graded as: <1/4 field of view obstructed by inflammatory cells = 1, <1/2 obstructed = 2, >1/2 but <3/4 obstructed = 3, and >3/4 obstructed = 4. Airway epithelial remodeling was graded as: minimal = 1, mild = 2, moderate = 3, and marked = 4. The average grade was calculated per condition by taking the average of the 4 airways assessed/mouse and averaging for the 5 exposed mice for each exposure condition (20 total airways assessed/exposure condition).

Western Blot Analysis

For Western blot analysis, lungs were homogenized in ice-cold lysis buffer containing protease and phosphatase inhibitors (16, 42). Soluble material was recovered by centrifugation, and protein was quantified using a BCA protein assay (Sigma, St. Louis, MO). Equivalent amounts of protein were resolved on Tris-HCl SDS–PAGE gel, transferred to a pure nitrocellulose membrane (PALL Corp, New York, NY) and blocked in 5% nonfat dried milk. Membranes were incubated with antibodies against secretoglobin 1 A family 1 (SCGB1A1; rabbit, 1:1,000; Millipore 07623), keratin 5 (KRT5; rabbit, 1:2,000; Biolegend 905504), and α-smooth muscle actin (αSMA; mouse, 1:600; Sigma-Aldrich A2547) overnight at 4°C. Immune complexes were detected with horseradish peroxidase-conjugated secondary antibody (rabbit, 1:10,000; Abcam ab97051 and mouse, 1:5,000; Bio-Rad 1706516) and visualized with enhanced Supersignal West Pico chemiluminescence (Thermo Scientific, Waltham, MA). Total protein served as loading controls semiquantitated from each corresponding stain-free gel (Mini-PROTEAN; Bio-Rad, Hercules, CA). Semiquantification and protein normalization were performed using Image Lab software (Bio-Rad, Hercules, CA).

Statistical Analysis

Prism 8.0 software (GraphPad Software) was used for statistical analysis with ANOVA followed by Tukey’s post hoc analysis for multiple comparisons for normally distributed results, whereas Kruskal–Wallis test was used for nonparametrically distributed results. A log-rank (Mantel-Cox) test was performed for survival analysis. Mean values were reported with standard deviation for normally distributed results and medians with interquartile ranges for nonparametrically distributed results. A P-value < 0.05 was considered significant with correction for multiple comparisons for most results. A 30% difference in the primary variable was assumed to achieve 80% power based on historical controls. Using these parameters, the minimal number of animals per group was calculated at 4 (8, 12, 32).

RESULTS

Diacetyl Inhalation Enhances Influenza-Associated Mortality

Adult mice were exposed to diacetyl (DA) vapors 1 h per day for 5 consecutive days. Immediately following DA exposures, mice were inoculated with influenza A virus (IAV) and then monitored for 14 days (exposure schematic; Fig. 1A). Mice exposed to DA failed to gain weight during exposure, but did not differ statistically from RA control mice (Fig. 1B). Mice exposed to IAV lost weight for 8 days after exposure differing significantly from those mice exposed to DA alone or room air (RA). Percent weight loss did not differ significantly between the two groups exposed to IAV (IAV alone vs. DA + IAV). However, mortality was significantly increased in those mice exposed to both DA + IAV (survival: 44%) compared with those mice exposed to a single exposure alone (survival: 100% for all other groups; Fig. 1C) with most animal deaths occurring 5–10 days after IAV infection in those mice exposed to both DA + IAV.

Figure 1.

A: schematic outline of exposures to diacetyl (DA) vapors and influenza A (IAV) infection (H3N2; x31). C57BL/6J female mice exposed to DA vapors for 1 h/day for 5 consecutive days from days 0 to 5. Following DA exposure, mice were inoculated with IAV x31, and monitored for 14 days until harvesting at day 19. B: percent weight change relative to starting weight for 4 groups: room air (RA) + Sham (n = 10; green inverted diamonds), RA + IAV (n = 10; black diamonds), DA + Sham (n = 10; blue squares), and DA + IAV (n = 18; red circles). Percent weight change relative to baseline differed significantly at day 9 in groups exposed to IAV compared with non-IAV-exposed groups remaining significantly lower until day 19 (****P < 0.0001; two-way ANOVA with correction for multiple comparisons). C: Kaplan–Meier percent (%) survival for each of the four exposed groups. Survival at day 19 differed significantly for mice exposed to DA + IAV compared with RA + Sham, DA + Sham, and RA + IAV (44% vs. 100%, ****P < 0.0001; log-rank test).

Diacetyl Inhalation Prior to Influenza a Infections Enhances Airway Neutrophils but Does Not Change Lung Permeability, Viral Load, or Viral Clearance

To further characterize the impact of DA inhalation on IAV infection, we used bronchoalveolar lavage (BAL) to measure changes in cell count and cell differential as a marker of pulmonary inflammation. The total number of cells in BAL did not differ significantly between groups exposed to IAV (DA + IAV and RA + IAV) at any of the time point assessed after IAV inoculation (Fig. 2A). However, there were statistically significant differences between DA + IAV and RA + IAV groups in the BAL cell composition. In particular, the percentage of neutrophils was significantly higher (Fig. 2B), whereas the percentage of macrophages (Fig. 2C) was significantly diminished in mice exposed to DA + IAV compared with RA + IAV. The percentage of BAL lymphocytes did not differ significantly between groups.

Figure 2.

Bronchoalveolar lavage (BAL) total cell count, BAL neutrophils, BAL macrophages, BAL total protein, and viral titers from whole lung homogenates in mice exposed to RA + IAV (black) and DA + IAV (red) at days 3, 5, and 9 after IAV inoculation (n = 4 or 5/group). A: BAL cell count (cells ×10⁵/mL) did not differ statistically between groups at days 3, 5, and 9 after IAV. B: BAL neutrophil percent (%) increased significantly at day 3 in mice exposed to DA + IAV compared with RA + IAV mice, but did not differ between groups at days 5 and 9 (*P < 0.05, two-way ANOVA). C: BAL macrophage percent (%) decreased significantly at day 3 in mice exposed to DA + IAV compared RA + IAV mice but did not differ between groups at days 5 and 9 (*P < 0.05, two-way ANOVA). D: BAL total protein (mg/mL) did not differ statistically between groups at days 3, 5, and 9 after IAV. E: the number of viral fluorescent units (log₁₀ × VFU/mL) did not statistically differ between groups at days 3, 5, and 9 after IAV. DA, diacetyl; IAV, influenza A; RA, room air.

Considering changes in BAL cell composition can be due to changes in lung permeability, BAL total protein was assessed at similar time points after IAV (44). There were no statistically significant differences in BAL total protein between DA + IAV and RA + IAV groups (Fig. 2D). Next, viral load was also measured using whole lung homogenates after IAV. IAV was detected in the lung on days 3, 5, and 9 after IAV infection; however, viral load nor viral clearance did not differ significantly between groups for any of the time points assessed after IAV exposure (Fig. 2E). These results suggest that DA exposure before IAV promotes an exaggerated neutrophilic response in the mouse lung, and the increased mortality associated with combined DA + IAV exposures is less likely due to increased lung permeability, enhanced viral load, or impaired viral clearance.

Exposure to DA ± IAV Promotes Mixed Obstructive and Restrictive Lung Disease in Mice

Next, we evaluated pulmonary function among mice surviving to the end of the study (14 days after IAV). Lung function testing did not differ significantly in mice exposed to single exposures (DA + Sham or RA + IAV) compared with RA + Sham (Fig. 3). Pressure-volume curves were slightly reduced in total volume from mice exposed to DA or IAV alone (Fig. 3E); however, constant phase lung function parameters did not differ significantly compared with RA controls (Fig. 3, A–D). Conversely, pulmonary function was markedly worse in DA + IAV compared with controls. More specifically, inspiratory capacity decreased; total lung resistance increased; total lung compliance decreased; and tissue damping increased in mice exposed to DA + IAV. Collectively, these parameters shifted the pressure-volume curves downward and outward in coexposed mice consistent with a mixed obstructive and restrictive phenotype.

Figure 3.

Pulmonary function testing at day 19 in mice exposed to RA + Sham (green), DA + Sham (blue), RA + IAV (black), and DA + IAV (red; n = 4 or 5/group). Inspiratory capacity (IC; A), total lung resistance (R_rs; B), total lung compliance (C_rs; C), tissue damping (D), and pressure-volume loops (E) differed significantly in those mice exposed to DA + IAV compared with RA + Sham, RA + IAV, or DA + Sham (****P < 0.0001, **P < 0.01 or *P < 0.05, ANOVA with Tukey’s). DA, diacetyl; IAV, influenza A; RA, room air.

Abnormal Airway Remodeling in Mice Exposed to DA ± IAV

Similar to lung function testing, lungs of mice exposed to single exposures alone (DA + Sham or RA + IAV) appeared structurally similar to controls (RA + Sham), whereas lungs from mice exposed to DA + IAV appeared markedly dysplastic with airway-centric remodeling (Fig. 4, A and B). Airways from mice exposed to DA or IAV alone showed regions of airway epithelial dysplasia. In contrast, airways of those mice exposed to DA + IAV demonstrated a thickened and hypertrophied airway epithelium with adjacent subbasement membrane thickening. To provide further semiquantitative assessment of the lung pathology identified in exposed mice, a scoring system was adopted from prior publications for histopathologic assessment of DA exposures in rodents (8, 43). Both the relative amount of airway epithelial remodeling and airways inflammation were assessed at 2 wk post-exposure in four airways per mouse and in 5 mice per exposure condition (total of >20 airways assessed/exposure condition). Consistent with the differences observed in pulmonary function testing, the amount of distal airway remodeling and inflammation increased significantly in mice exposed to DA + IAV compared with RA + Sham controls and did not differ significantly in mice exposed to RA + IAV or DA + Sham alone compared with RA + Sham controls (Fig. 4, C and D).

Figure 4.

A: representative trichrome-stained lung sections at day 19 from mice exposed to RA + Sham (top left), RA + IAV (top right), DA + Sham (bottom left), or DA + IAV (bottom right). Lower magnification image with scale bar of 2.0 mm and higher magnification image (insets) with scale bar of 200 µm. Airway remodeling with epithelial cell hypertrophy and adjacent matrix deposition seen in those mice exposed to DA + IAV (black arrow; insert bottom right). B: representative hematoxylin and eosin (H&E)-stained lung sections from those mice exposed to RA + Sham (top left), RA + IAV (top right), DA + Sham (bottom left), or DA + IAV (bottom right) at day 19. Lower magnification image with scale bar of 300 µm, and higher magnification insight image with scale bar of 100 µm. Semiquantitation of histopathology for airway epithelial remodeling (C) and airways inflammation (D) in mice exposed to RA + Sham (green), DA + Sham (blue), RA + IAV (black), and DA + IAV (red; n = 5/group). Airway epithelial remodeling and inflammation increased in mice exposed to DA + IAV compared with RA + IAV mice (**P < 0.01 and *P < 0.05, respectively, ANOVA). DA, diacetyl; IAV, influenza A; RA, room air.

Abnormal Club Cell Expression with Increased α-Smooth Muscle Actin in DA ± IAV Exposed Distal Airways

To further characterize the airway phenotype following DA + IAV, we examined common club cell markers as well as α-smooth muscle actin (αSMA) in mouse lung sections and in lung homogenates (Fig. 5). Club cells are the predominant cell type of the distal airway epithelium and express the protein secretoglobin family 1 A member 1 (SCGB1A1). The majority of cells in the distal airways of mice exposed to RA + Sham expressed SCGB1A1 (Fig. 5A). Airway expression of SCGB1A1 was patchy after exposure to DA or IAV. The relative abundance of SCGB1A1 in whole lung homogenates was mildly reduced by Western blot in single exposures (DA or IAV alone) but did not differ statistically from RA controls (Fig. 5, C and D). Conversely, SCGB1A1 expression was absent in large patches of distal airways exposed to DA + IAV with a significant reduction in total abundance of SCGB1A1 in lung homogenates, supportive of impaired epithelial repair with combined exposures. In contrast, the expression of α-SMA in distal airways as well as relative abundance in total lung homogenates increased significantly in mice exposed to DA + IAV compared with all other groups (Fig. 5, B–D).

Figure 5.

A: immunofluorescent images of airways stained for club cell marker SCGB1A1 (green) and DAPI (blue) from mice exposed to RA + Sham (top left), RA + IAV (top right), DA + Sham (bottom left), or DA + IAV (bottom right) harvested at day 19 (14 days post-IAV exposure; scale bar: 100 µm). Patchy airway staining for SCGB1A1 in mice exposed to both DA + IAV (white arrows; bottom right). B: immunofluorescent images of airways stained for α smooth muscle actin (α-SMA; red) and DAPI (blue) from mice exposed to RA + Sham (top left), RA + IAV (top right), DA + Sham (bottom left), or DA + IAV (bottom right) harvested at day 19 (14 days post-IAV exposure; scale bar: 50 µm). C: representative Western blots for α-SMA and SCGB1A1 from mouse lung homogenates harvested at day 19 from RA + Sham, DA + Sham, RA + IAV, or DA + IAV. Each lane contains lung homogenate from a separate mouse, and at least three samples per group were run on the same gel. Total protein from stain-free gel images for semiquantification as loading control (lower gel). D: semiquantification of and SCGB1A1 (upper graph) and α-SMA (lower graph) with relative abundance compared with total loaded protein in mouse lung homogenates from RA + Sham (green), DA + Sham (blue), RA + IAV (black), or DA + IAV (red; **P < 0.01; Kruskal–Wallis test with Dunn’s correction; n = 4 or 5 lungs/group). DA, diacetyl; IAV, influenza A; RA, room air.

Emergence of Keratin 5-Expressing Airway Epithelial Cells with DA ± IAV

Previous investigators have shown the emergence of lineage-negative epithelial progenitor cells (LNEPs) that express the basal cell marker keratin 5 (KRT5) in the distal lung parenchyma following inoculation with a different flu strain (H1N1, PR8) (45–48). Similar to these prior publications, a small number of KRT5⁺ cells emerged in airways of those mice exposed to IAV alone (Fig. 6). KRT5⁺ cells were not seen in RA- or DA-exposed airways. In airways exposed to RA + IAV, KRT5⁺ cells were adjacent to the basement membrane and did not coexpress the club cell marker SCGB1A1 (Fig. 6C). In mice exposed to DA + IAV, KRT5⁺ expression was seen in the distal airways and coexpressing the club cell markers SCGB1A1 and CYP2F2 compared with mice exposed to DA or IAV alone (Fig. 6, A and C). The relative abundance of KRT5 also increased significantly in total lung homogenates exposed to DA + IAV compared with RA controls (Fig. 6B). These findings of KRT5⁺ cells that coexpress club cell markers additionally support abnormal airway epithelial repair after combined exposures.

Figure 6.

A: coimmunofluorescent images of airways stained for club cell marker CYP2F2 (red), basal cell marker keratin 5 (KRT5; green) and DAPI (blue) from those mice exposed to RA + Sham (top left), RA + IAV (top right), DA + Sham (bottom left), or DA + IAV (bottom right) harvested at day 19 (scale bar: 50 µm). B representative Western blots for KRT5 detected from mouse lung homogenates at day 19 in RA + Sham, DA + Sham, RA + IAV, or DA + IAV (n = 4 or 5 lungs/group). Total protein from stain-free gel images for loading control and semiquantification (lower gel). Semiquantification of KRT5 (right) relative abundance compared with total loaded protein in mouse lung homogenates from RA + Sham (green), DA + Sham (blue), RA + IAV (black), or DA + IAV (red; **P < 0.01; Kruskal–Wallis test with Dunn’s correction; n = 4 or 5 lungs/group). C: coimmunofluorescent images of airways stained for club cell marker SCGB1A1 (red), KRT5 (green), and DAPI (blue) from those mice exposed to RA + Sham (top left), RA + IAV (top right), DA + Sham (bottom left), or DA + IAV (bottom right) harvested at day 19 (scale bar: 50 µm). DA, diacetyl; IAV, influenza A; RA, room air.

DA Inhalation Impairs Influenza-Induced Airway Epithelial Repair

Changes seen in lung function and histology in mice exposed to DA first and then IAV infection could be due to 1) increased IAV viral pathogenesis; 2) a heightened immune response to IAV, and/or 3) impaired airway epithelial repair. To test the hypothesis that DA exposure impairs airway epithelial repair, we reversed the order of DA exposure, inoculating first with IAV (day 0) and then exposing to DA vapors on days 9–14 after IAV inoculation (Fig. 7). The percent weight loss in mice exposed to IAV did not differ significantly between those mice exposed to Sham + DA or Sham + RA. Survival also did not differ significantly between groups. However, lungs from mice exposed to IAV + DA demonstrated abnormal airway remodeling similar to that seen with the initial order (DA, then IAV) with KRT5⁺ cells adjacent to SMA⁺ staining in terminal bronchioles. Collectively, these results support abnormal airway epithelial repair in mice following combined exposures to IAV + DA and not seen with single exposures to DA or IAV alone.

Figure 7.

A: schematic outline of reverse exposure experiment with influenza A (IAV) infection (H3N2; x31) and then diacetyl (DA) vapors. C57BL/6J female mice were infected with IAV, monitored for 9 days, exposed to DA vapors for 1 h/day for 5 consecutive days from days 9 to 14, monitored for an additional 14 days and harvested at day 28. B: percent weight change relative to starting weight for 4 groups: Sham + room air (RA; n = 10; green inverted diamonds), influenza A virus (IAV) + RA (n = 12; black diamonds), Sham + DA (n = 10; blue squares), and IAV + DA (n = 12; red circles). Percent weight change relative to baseline differed significantly at day 5 in groups exposed to IAV compared with sham-exposed controls, remaining significantly lower until day 21 post-IAV inoculation (****P < 0.0001; two-way ANOVA with correction for multiple comparisons). C: Kaplan–Meier percent (%) survival curve for each of the four exposure groups. Survival at day 28 did not differ significantly for mice exposed to Sham + RA, IAV + RA, Sham + DA, and IAV + DA (P = 0.29; log-rank test). D and E: coimmunofluorescent images of airways stained for basal cell marker keratin 5 (KRT5; green), α-smooth muscle actin (αSMA), and DAPI from those mice exposed to DA + IAV at day 19 (D) or IAV + DA at day 28 (E) (scale bar: 200 μm).

DISCUSSION

The current study demonstrates for the first time that exposures to two relatively ubiquitous environmental exposures, the chemical flavoring diacetyl (DA), and respiratory infection with influenza A virus (IAV) result in significantly worse mortality, impaired lung function, and significant airway remodeling. With combined exposures, bronchoalveolar lavage (BAL) neutrophils increased significantly after IAV infection compared with RA + IAV alone and without a significant increase in lung permeability, viral load, or viral clearance. Abnormal airway remodeling was characterized by increased subepithelial smooth muscle actin expression and the unique costaining of airway epithelial cells with both basal cell marker keratin 5 (KRT5) and club cell markers SCGB1A1 and CYP2F2. Exposure to DA or IAV alone failed to result in worsening mortality nor significant airway remodeling, whereas combined exposure to DA + IAV markedly reduced lung function and impaired airway repair with combined exposures. In addition, abnormal airway repair was seen histologically whether DA was administered before or after infection, demonstrating that DA profoundly impaired the resolution of viral-mediated airway epithelial injury. Our findings indicate exposure to these common environmental factors may seem relatively harmless by themselves but can have catastrophic implications on lung function and long-term respiratory health when combined.

Previous studies by our group and others have begun to delineate the underlying mechanisms of lung disease affiliated with DA vapor exposures also known as “flavoring-related lung disease” (6–8, 12, 14, 16, 49–51). Related to the increased percent neutrophils in BAL seen with combined exposure to DA + IAV, interleukin-8 (IL-8) production increased in rat lungs and BAL following exposure to DA (52). IL-8 is a potent neutrophilic chemoattractant, and increased production of IL-8 was primarily dependent upon epidermal growth factor receptor (EGFR) from airway epithelial cells (52). More recently, repetitive DA exposures in primary human airway epithelial cells that were then inoculated with SAR-CoV2 increased interleukin-1 β (IL-1β) RNA expression, yet another proinflammatory and potent neutrophilic activator (50). In contrast to findings here, where prior DA exposure failed to increase lung viral load nor impair viral clearance of IAV, DA exposures predisposed human airway epithelial cell cultures to SAR-CoV2 with increased total viral load and RNA transcript levels. These differences highlight potential differences in viral susceptibility, exposure models, the dose of DA, and/or species. In mice exposed repetitively to DA vapors at comparable concentrations (200 ppm), BAL neutrophil percentage did not increase significantly (8), whereas combined exposure, as seen here, to DA + IAV enhanced airway neutrophil percentage compared with RA + IAV alone. Collectively, these results suggest, even short-term exposures to DA vapors may prime the host to a proinflammatory and enhanced neutrophilic response when exposed before influenza A viral infection.

In addition to our assessments of the lung permeability and the immune response, we also assessed for persistent airway remodeling with combined exposures. Club cells are considered the predominant intrapulmonary airway epithelial progenitor cell (53–55). Following most airway injuries, epithelial repair occurs through migration, proliferation, and differentiation of airway progenitor cells (55, 56). In contrast, when challenged with certain environmental toxicants, including but not limited to DA (51, 57, 58), significant damage not only occurs to the airway epithelium alone but to this rare population of airway epithelial progenitor cells. When resident lung progenitor cells are significantly injured, alternative epithelial progenitor cells are activated (46–48). These alternatively activated progenitor cells express the protein KRT5, reside in the proximal airways, and migrate to the distal alveolar space following severe lung injury. One well-described lung injury model known to activate these alternative progenitor cells is infection with an H1N1 influenza A virus strain PR8 (46, 47). Mice infected with PR8 develop significant parenchymal lung injury. In those mice who survive the acute infection, parenchymal fibrosis develops marked by failed alveolar epithelial repair. In this context, maladaptive epithelial repair occurs with activation of KRT5⁺ epithelial cells that migrate and proliferate at the site of alveolar injury (45–48). Likewise, conditional depletion of alveolar epithelial type II cells, the primary progenitor of the alveolar epithelium, impairs lung repair following IAV x31 infection, with the development of parenchymal fibrosis and recruitment of KRT5⁺ cells into the distal alveolar space (59). These alternatively activated progenitor cells remain a maladaptive response considering the expression of airway markers in the distal alveolar space is abnormal and not functional for gas exchange. KRT5⁺ cells have also been identified in lung autopsies from those who have died from acute respiratory distress syndrome (ARDS) with diffuse alveolar damage (DAD) and parenchymal fibrosis, further emphasizing that these cells mark a maladaptive repair response in the lung parenchyma (60, 61).

Novel to our findings, exposure to DA for relatively short periods of time followed by a sublethal infection with the airway-tropic IAV x31 resulted in marked airway remodeling. Similar to other IAV infections, exposure to both DA + IAV resulted in epithelial cells expressing KRT5. These epithelial changes persisted for at least 2 wk after IAV with increased basement membrane remodeling, further supportive of abnormal airway epithelial repair. In other models of airway injury, a single day’s depletion of airway club cells using diphtheria toxin A did not cause significant airway remodeling, whereas repetitive diphtheria exposures to club cells resulted in failed airway epithelial regeneration and peribronchiolar fibrosis (56). This study emphasizes that a single injury to club cells may be insufficient to impair airway epithelial repair; however, repetitive depletion promotes abnormal repair. A similar model of failed airway epithelial repair was developed using continuous exposures to the club cell toxin naphthalene (62). Again, a single exposure to naphthalene resulted in airway epithelial repair, whereas repetitive naphthalene exposures impaired club cell regeneration (63, 64). Specific to DA, previous studies using acute and subchronic DA vapor exposures in mice resulted in airway epithelial dysplasia (8). When combined here with IAV infection, the effect of DA on airway epithelial repair was accentuated, noted by increased expression of αSMA and the emergence of KRT5⁺ cells with common club cell markers.

Although keratin 5 (KRT5) has been used to define a population of epithelial progenitor cells, it is also important to remember that cells express KRT5 for a purpose. KRT5 is an intermediate filament that provides essential structural integrity for the maintenance of epithelial barrier integrity and attachment to the airway basement membrane. Genetic mutations in the KRT5 gene cause significant fragility of the epithelium, manifesting as the disease epidermolysis bullosa (EBS) (65). Similar to EBS (66), acute DA exposure damages KRT5 causing perinuclear aggregation and subsequent loss of cytoplasmic structural integrity (16). When allowed sufficient time to recover after DA exposure, damaged KRT5 proteins are degraded through the ubiquitin-proteasome system and reorganize to reestablish barrier integrity. Conversely, at high DA concentrations or with repetitive DA exposures, KRT5 aggregates persist resulting in hemidesmosome instability with cleavage of integrin β 4 (ITGβ4) (67). In addition to its functional role of providing structural support to hemidesmosomes, KRT5 has also been used to classify major subclasses of nonsmall cell lung cancer (68) as well as discriminate lung adenocarcinoma from squamous cell carcinoma (69). These studies emphasize the clinical importance of KRT5 as a cancer biomarker in addition to its functional role of epithelial barrier integrity and regeneration.

A limitation of the current study is the use of a single concentration of DA for our vapor exposures. Prior dose-response studies of DA inhalation exposures, using both male and female mice, support the use of 200 ppm DA as an environmentally relevant human exposure (8, 26). This administered dose to mice is comparable to 12 ppm of DA in humans when accounting for the increased surface area of the nasal passages of mice. An exposure of 12 ppm DA in humans aligns with multiple assessments of DA exposures in occupational settings (70, 71). Two other limitations of the current study are the use of a single strain of IAV tested and female mice used for the majority of the outcomes assessed. We deliberately used a mild viral challenge because use of a highly pathogenic strain or lethal dose of virus would have impaired our ability to answer our main research question. Given that humans experience a range of respiratory viruses throughout a lifetime, it is possible that combined exposure to DA with different strains of IAV, other common respiratory viruses, or even bacteria that infect the respiratory tract, may yield different outcomes (72). Female mice were used for these studies considering aspects of the immune response to IAV infection, such as the timing of the peak of various cellular responses, differ significantly between males and females although the mechanistic underpinnings of these sex differences are not fully understood (73, 74).

In conclusion, exposure to DA or IAV alone is insufficient to impair airway epithelial repair, whereas combined exposure to DA + IAV resulted in abnormal airway remodeling. These studies have direct implications for those previously exposed to DA as a risk factor for future airway remodeling following IAV infection. In addition, those exposed to IAV combined with a chemical inhalation exposure, such as DA found in e-cigarette exposures or in coffee roasting, may increase the risk for the development of significant airways disease and lung function decline. Future research into the abnormal airway epithelial repair response to chemical inhalation and respiratory tract infections is needed to delineate the synergistic effects of environmental exposures to persistent lung disease.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

M.D.M. reports grant support of National Institutes of Health (NIH) KL2 TR001999 from National Center for Advancing Translational Sciences. M.A.O. reports grant support of NIH R01 HL091968 through the National Heart, Lung, and Blood Institute. B.P.L. reports grant support of NIH P30 ES001247 and R01 ES030300 Lung, from the National Institute of Environmental Health Sciences.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.D.M., B.P.L., and M.A.O. conceived and designed research; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. performed experiments; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. analyzed data; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. interpreted results of experiments; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. prepared figures; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. drafted manuscript; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. edited and revised manuscript; M.D.M., M.Y., S.-Y.K., A.M.D., B.P.L., and M.A.O. approved final version of manuscript.