The influenza A/Puerto Rico/8/34 (PR8, H1N1) virus was obtained from the American Type Culture Collection (ATCC) and propagated in the allantoic cavities of 10-day-old embryonated SPF chicken eggs. The virus was tittered in MDCK cells (ATCC) using TCID₅₀ (50% tissue culture infectious dose) assay calculated by Reed and Muench method (Reed and Muench, 1938).

The NTHi ATCC49766, a reference strain, was kindly provided by Professor Zhuo from our laboratory. Due to the infectivity and virulence in mice (Kimura et al., 2020), we selected this strain for studies of dual infection-induced AECOPD. NTHi stocks were recovered in chocolate agar plates containing 0.33 mg/L vancomycin (Guangdong Huankai, China) and grown in brain heart infusion (BHI). 100 μL serial dilutions of NTHi in PBS were inoculated in chocolate agar plates for 24 h in a 37°C, 5% CO₂ incubator, after which the colony-forming units (CFU) were determined. All infections experiments were conducted in a biosafety level 2 Plus (BSL-2 +) laboratory following the protocols approved by Guangzhou Medical University.

Wild-type C57BL/6N male mice (6–8 weeks old), were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All the mice were housed in the specific pathogen-free facilities and all experimental protocols were approved by the Animal Care and Use Committee of Kunming Medical University. All mice were allowed to acclimatize for a week prior to cigarette smoke (CS) exposure.

The mice were randomly divided into 5 groups: CS exposure (Group 1), air exposure (Group 2), CS exposure with PR8 infection (Group 3), CS exposure with NTHi infection (Group 4), and CS exposure with PR8 and NTHi coinfection (Group 5). The experimental flow chart was shown in Figure 1. Briefly, group 1, 3, 4, and 5 were exposed to one cigarette (Hongtashan, China) per day for 14 weeks in an oral and nasal exposure system (Beijing Huironghe Technology, China), as previously described (Xiao et al., 2021). Each cigarette yielded 10 mg tar, 11 mg carbon monoxide, and 1.0 mg nicotine. The parameters of exposure were as followed: concentration of oxygen, (20 ± 0.5) %; dilution flow, 10 L/min; air extraction flow, 13 L/min; air humidity, (60 ± 5)%. The cigarette suction parameters were as followed: suction time, 2 s; time interval, 2 s; suction frequency, 10 s; suction volume, 35 mL. At week 14 post-CS exposure, mice from group 3 and 5 were anesthetized with 2.5% isoflurane and intranasally infected with 0.5 LD₅₀ of PR8. After 3 days, group 5 were sequentially challenged with 10⁵ CFU of NTHi. These infectious doses have been established in our previous study of PR8-NTHi coinfection (Wu et al., 2021). At 24 h after NTHi challenge, the invasive lung function was measured. The mice were then sacrificed and the lungs were collected. The whole lungs were weighted and the lung index (lung weight/body weight ratio × 100) was calculated. Left lungs were then rinsed with HBSS containing 2% FBS before flow cytometric analysis of immune cells. Right lungs were homogenized in 1 mL PBS, and 100 μL of homogenate was used for bacterial titer determination. The remaining homogenates were further centrifugated at 3,000 rpm, 4°C, for 10 min, and the resultant supernatant was subjected to measurements of virus titer and cytokine levels. Also, whole lungs were harvested for a hematoxylin and eosin (H and E) staining. In addition, colon luminal content was collected for gut microbiota analysis. In survival study, the body weight and survival of mice were recorded daily for 15 days after PR8 infection.

Pulmonary function in conscious mice was determined once every 2 weeks from week 6 to 14 after initiation of CS exposure using whole body plethysmography (EMKA Technologies, Canada). The spirometric parameters included frequency (F), tidal volume (TV), expiratory volume (EV), and inspiratory time (Ti).

Invasive lung function of mice was measured by using FlexiVent system (SCIREQ, Canada) according to the manufacturer’s protocol. Briefly, after anesthetization with tribromoethanol (Avertin, 240 mg/kg), mice were tracheotomized, intubated, and placed in the chamber of the FlexiVent system. To maintain muscular relaxation, mice were given intraperitoneal injections of vecuronium bromide (6 mg/kg). Mechanical ventilation was subsequently initiated to measure lung function determined by the forced vital capacity (FVC), the forced expiratory volume in 0.1 s (FEV0.1), respiratory resistance (Rrs), tissue damping (G), and respiratory system compliance (Crs). Each measurement was performed in triplicate. The data were analyzed by the FlexiVent software (SCIREQ, Montréal, QC, Canada).

The lung histopathology was performed as previously described (Shen et al., 2014). Briefly, the whole lungs of mice were removed and fixed in 4% paraformaldehyde for 24 h. After dehydration and embeddedness, lungs were sectioned to 3 μm thickness and stained with H and E or Masson’s trichrome. The histopathological lesions were observed and recorded under a light microscope. Trichrome staining intensity of 4 randomly selected areas per mouse lung was analyzed using ImageJ software.

Haemophilus influenzae were quantified by incubating 100 μL of 10-fold serial dilutions of lung homogenates in chocolate agar plates containing vancomycin, as described above. After centrifugation of the remaining lung homogenates, supernatants were used to determine TCID₅₀ in MDCK cells.

The supernatants of lung homogenates were also used for the measurement of inflammatory cytokines, chemokines, and adhesion molecules. The cytokines and chemokines were determined by using the Luminex ProcartaPlex kit (Invitrogen, Waltham, MA, USA) and the Mouse magnetic Luminex assays kit (R&D, Santa Clara, CA, USA) in a Bio-Plex 200 Multiplex Testing System (Bio-Rad, Hercules, CA, USA) following manufacturer’s instructions. Mouse carcinoembryonic adhesion molecule-1 (CEACAM-1) (MyBioSource, San Diego, CA, Canada), intracellular adhesion molecules-1 (ICAM-1) (R&D), and fibronectin (Fn) (Abnova, Walnut, CA, USA), were determined by ELISA assays.

The collected left lung was digested by the enzyme mixture in the gentleMACS Dissociator (Miltenyi Biotec, Gaithersburg, MD, USA). The resultant digested fluid was allowed to run through a 70-micron cell strainer to obtain a single-cell suspension. The erythrocytes were further lysed in a red blood cell lysis buffer (Biolegend, San Diego, CA, USA). Cells were resuspended with 1 mL FACS buffer (PBS contained 2% FBS and 5 mM EDTA), and the Fc receptors were blocked by anti-mouse CD16/32 (Biolegend, San Diego, CA, USA) at 4°C for 10 min. Cells were then stained with special anti-mouse antibodies, including GhostDye510 (TONBO Biosciences, USA), FITC anti-mouse CD11b (Biolegend, San Diego, CA, USA, Clone M1/70), PerCP-Cy5.5 anti-mouse Ly-6C (Biolegend, San Diego, CA, USA, Clone HK1.4), PE anti-mouse F4/80 (Biolegend, San Diego, CA, USA, Clone BM8), PE-Cy7 anti-mouse CD86 (Biolegend, San Diego, CA, USA, Clone GL-1), APC anti-mouse Ly-6G (Biolegend, San Diego, CA, USA, Clone 1A8), APC-Cy7 anti-mouse CD45 (Biolegend, San Diego, CA, USA, Clone 30-F11), BV421 anti-mouse CD11c (Biolegend, San Diego, CA, USA, Clone N418), BV605 anti-mouse CD206 (Biolegend, San Diego, CA, USA, Clone C068C2), FITC anti-mouse CD4 (Biolegend, San Diego, CA, USA, Clone RM4-5), PerCP-Cy5.5 anti-mouse CD3 (Biolegend, San Diego, CA, USA, Clone 17A2), PE anti-mouse CD19 (Biolegend, San Diego, CA, USA, Clone 6D5), APC anti-mouse CD8 (Biolegend, San Diego, CA, USA, Clone 53-6.7), and BV421 anti-mouse NK1.1 (Biolegend, San Diego, CA, USA, Clone PK136). Flow cytometric analysis of stained cells were performed using a flow cytometer (PARTEC CyFlow Space, Germany). Acquired data were analyzed using FlowJo software.

Colon luminal contents in mice were collected immediately after euthanasia and were rapidly flash-frozen in liquid nitrogen for 30 min before storing at −80°C. Total bacterial DNA was extracted by using a Magnetic soil and stool DNA kit (TIANGEN, China). The V3-V4 region of 16S rRNA gene was amplified using PCR assay with. The amplicon primer sequences were as follows: forward primer 5′-(GTGCCAGCMGCCGCGGTAA)-3′; reverse primer 5′-(GGACTACHVGGGTWTCTAAT)-3′. The libraries were constructed using NEB next ultra-library prep kit (Illumina, USA) and paired-end sequenced on Illumina Novaseq6000 platform (Illumina, USA) of Novogene Co., Ltd. (Beijing, China). The filtering of raw sequence and taxonomic classification were performed as described previously (Li et al., 2020). The diagrams of alpha and beta diversity and relative abundance of microbiota were drawn using the R packages ggplot2 and vegan. The significance of beta diversity based on Weighted UniFrac dissimilarity was tested by analysis of similarities (Anosim) within the R package vegan.

Statistical analysis was performed by using GraphPad Prism 8.0 and SPSS 16.0 software. The differences among groups were compared by using the unpaired two-tailed student’s t-test, welch’s t-test or non-parametric test. Correlations between cytokines and gut microbiota were assessed by using Spearman’s rank correlation. The value of p < 0.05 was considered statistically significant.

To replicate COPD in humans, we first developed a mouse model by CS exposure for 14 weeks. Mice exposed to CS had impaired lung function, as indicated by an increase in F and a decrease in Ti, TV, and EV (Figures 2A–D). Besides, CS-exposed mice gained weight significantly slower than control mice (Figure 2E). These results suggested that COPD mice were successfully established. We further developed the AECOPD model by sequentially infecting COPD mice with PR8 and NTHi. Coinfected mice exhibited 100% lethality (p < 0.001), but PR8 or NTHi alone had no effect on mortality, indicating that the combination of two pathogens caused synergistic mortality (Figure 3).

As expected, CS exposure significantly impaired lung function, as shown by a decline in FEV0.1/FVC and Crs, and a increase in Rrs and G compared to control mice (Figures 4A–D), but PR8 or NTHi alone did not result in further impairment (Figures 4A–D). Interestingly, coinfection deteriorated the impairment of lung function in COPD mice (Figures 4A–D).

The combination of PR8 and NTHi resulted in significant higher levels of NTHi load in mice lungs at 24 h post-NTHi challenge (Figure 5A), but the PR8 titer in coinfected mice did not differ from that in PR8-infected mice (Figure 5B).

Mice exposed to CS had emphysema-like pathological changes, including bronchial wall thickening, alveolar wall thinning and destruction, and alveolar enlargement below the pleura, as compared with control group (Figures 6A–D). Mild inflammation was observed in COPD mice infected with NTHi, as shown by the recruitment of neutrophils into bronchial lumen (Figures 6E, F). In PR8-infected COPD mice, lymphocytic inflammation was present in peribronchial alveolar (Figures 6G, H). In addition, coinfected mice had more severe lung injury, as evidenced by infiltration of neutrophils into bronchiolar cavity, profuse hemorrhage and edema, and destruction of ciliated columnar epithelium (Figures 6I, J). In accordance with histopathological findings, a single pathogen resulted in a higher lung index than CS exposure alone, and coinfection resulted in the highest lung index (Figure 6K). In addition, mild collagen deposition was observed in smooth muscle layer of trachea and lung blood vessels of normal control mice (Figures 6L, Q). Compared to control group, mice exposed to CS exhibited bronchial wall thickening and significantly increased collagen deposition (Figures 6M, Q). Collagen expression was not significantly different between the CS, CS + NTHi, and CS + PR8 groups (Figures 6M, N, O, Q), but it was more extensive and significant in coinfected mice (Figures 6P, Q). These findings indicate that coinfection significantly increases the collagen deposition in lungs of COPD mice.

Since pathological inflammation was pronounced in AECOPD mice, we further determined the expression levels of inflammatory cytokines and immune cells in the lungs. NTHi alone induced increased levels of cytokines, including IL-1β, IL-6, TNF-α, IL-22, IL-17, KC, and MIG, as compared with non-infected COPD mice (Figures 7B–H). Besides, PR8 infection alone also triggered significantly higher expression of inflammatory cytokines and chemokines, including CRP, IL-1β, IL-6, TNF-α IL-22, IL-17, KC, and MIG (Figures 7A–H). Furthermore, coinfected mice had significant higher levels of CRP, IL-1β, IL-6, TNF-α, IL-22 and KC as compared with PR8 or NTHi alone (Figures 7A–E, G). We also found that the levels of the adhesion molecules, including ICAM-1 and CEACAM-1, as well as the extracellular matrix Fn, were significantly higher in coinfected mice than the CS group (Figures 7I–K).

We further performed flow cytometric analysis of inflammatory cells, including macrophages (CD45 + /CD11b + /F4/80 + /Ly6G-), M1 macrophages (CD45 + /CD11b + /F4/80 + /Ly6G-/CD86 +), M2 macrophages (CD45 + /CD11b + /F4/80 + /Ly6G-/CD206 +), alveolar macrophages (CD45 + F4/80 + CD11c + CD11b-), neutrophils (CD45 + /Ly6C + /Ly6G + /CD11b +), monocytes (CD45 + Ly6C + Ly6G-CD11b +), NK cells (NK1.1 + CD3-), T cells (CD45 + /CD3 +), CD4 + T cells (CD45 + /CD3 + /CD4 + /CD8-), CD8 + T cells (CD45 + /CD3 + /CD4-/CD8 +), and B cells (CD45 + /CD3-/CD19 +). COPD mice infected with NTHi alone had significantly increased M1 and M2 macrophages but decreased alveolar macrophages as compared with non-infected COPD mice (Figures 8B, C, E). PR8 infection alone resulted in an increase in the number of M1 and M2 macrophages, neutrophils, monocytes, and NK cells in the infected COPD mice (Figures 8B, C, F–H). In coinfected mice, total macrophages, M1 macrophages, neutrophils, monocytes, and NK cells were significantly increased (Figure 8A, B, F–H), whereas the alveolar macrophages were significantly decreased (Figure 8E). We also showed that coinfection resulted in a significant increase in the ratio of M1/M2 (Figure 8D). Regarding adaptive immune cells, coinfection significantly reduced the levels of CD4 + T cells and B cells (Figures 8J, L), but had no effect on total number of T cells or CD8 + T cells (Figures 8I, K).

It has been reported that exposure to CS or biofuels disrupts the gut microbiota of COPD mice, which may influence disease progression (Li et al., 2021). Infection with IAV can also alter the gut microbiota, which increases susceptibility to viral infection and bacterial superinfection, exacerbates lung injury, and even increases mortality (Yildiz et al., 2018). However, the alteration of gut microbiota in AECOPD individuals remained unknown. Hence, we sequenced the 16S rRNA gene of colon contents to further investigate the effect of coinfection on the gut microbiota in COPD mice. After removing low-quantity data, 1,810,199 effective sequence reads from 30 samples were utilized for analysis and 9,403 operational taxonomic units (OTUs) were identified. The number of OTUs was significantly higher in mice exposed to CS than in the control group, but COPD mice infected with NTHi or PR8 alone had a slightly lower number of OTUs, and the decrease was even more pronounced in coinfection (Figure 9A). Chao 1 indices also showed a similar pattern of decline in alpha diversity (Figure 9B). In addition, PCoA results revealed varying degrees of shift across all groups, with the most significant shifts occurring between CS and CPN (p = 0.0149) and between CN and CPN (p = 0.0049) (Figure 9C).

We then analyzed the taxonomic level of microbial communities that may have contributed to the changes in alpha and beta diversity. At the phylum level, all colon contents contained four major bacterial phyla: Bacteroidetes, Firmicutes, Proteobacteria, and Actinobacteria. Mice exposed to CS for 14 weeks had a greater relative abundance of Actinobacteria, which was decreased when COPD mice were infected with NTHi (p < 0.05) or PR8 (p > 0.05) (Figure 9D). COPD mice with coinfection exhibited more pronounced decreases in Actinobacteria and Firmicutes, and an increase in Bacteroidetes (Figure 9D). In comparison to COPD mice, coinfected mice had lower abundance of Muribaculaceae, Lactobacillus, and Lachnospiraceae, and higher abundance of Bacteroides (Figures 9E, F).

Inflammatory cytokine production induced by acute viral respiratory infections is an important cause of gut microbiota changes (Sencio et al., 2021). We further analyzed the correlation between gut microbiota and the expression of lung cytokines. The Mantel test revealed a positive correlation between the diversity of gut microbiota and multiple inflammatory cytokines, including KC, IL-1β, IL-6, and TNF-α, which were significantly elevated in coinfected COPD mice (Figure 9G). We then investigated the association between lung cytokines and individual microbes. Interestingly. the abundance of commensal bacteria, including Muribaculaceae, Lactobacillus, Akkermansia, Lachnospiraceae, and Rikenella, were negatively correlated with most of cytokines. On the contrary, the higher levels of lung cytokines were correlated with increased abundance of Bacteroides, Alistipes, Parabacteroides, and Ruminococcus (Figure 9H), but the last three genus did not significantly differ among all groups.