Interferon regulatory factor 5 involves the pathogenesis of emphysema through NLRP3 and Ly6C expressing cells

TL;DR

This study shows that a protein called IRF5 contributes to lung damage in emphysema by affecting immune cells and a type of cell death called pyroptosis.

Contribution

The study reveals a novel role for IRF5 in emphysema pathogenesis via NLRP3-mediated pyroptosis and Ly6C-expressing immune cells.

Findings

Irf5-knockout mice showed reduced alveolar destruction and suppressed NLRP3 expression after cigarette smoke exposure.

Ly6Chigh monocytes and T cells from Irf5-KO mice attenuated lung damage when transferred to emphysema mice.

IRF5 expression was significantly elevated in lung tissues from patients with emphysema.

Abstract

Interferon regulatory factor 5 (IRF5) is a key regulator of inflammatory responses; however, its role in chronic obstructive pulmonary disease remains unknown. A previous study showed increased IRF5 expression in the lungs of cigarette smoke (CS)-induced emphysema. Here we investigated the function of IRF5 in emphysema using Irf5-knockout (KO) mice. Alveolar destruction, inflammatory cell infiltration, cytokine levels and pyroptosis-related gene expression were assessed in CS-induced emphysema. To investigate the role of immune cells, Ly6C++ (Ly6Chigh) monocytes and Ly6Chigh T cells from Irf5-KO mice were introduced into emphysema mice. The correlation between IRF5 levels and emphysema in humans was also evaluated. Irf5-KO mice showed decreased alveolar destruction after CS exposure. NLRP3 expression was suppressed, and gasdermin D cleavage was altered in Irf5-KO mice, suggesting a protective effect against pyroptotic cell death. Moreover, Ly6Chigh monocytes and Ly6Chigh T cells were more abundant in the lungs of Irf5-KO mice after CS exposure, and their transfer attenuated NLRP3 expression and alveolar damage. Furthermore, IRF5 expression was significantly elevated in lung tissues from patients. Our findings highlight IRF5 as a critical regulator of emphysema pathogenesis via NLRP3-mediated pyroptosis and Ly6Chigh-expressing immune cells. Targeting IRF5 may be a potential therapeutic strategy for chronic obstructive pulmonary disease.

Chronic obstructive pulmonary disease is a lung condition often caused by smoking, leading to inflammation and lung damage. This study explores a protein called IRF5, which is involved in inflammation and immune responses, as a potential target for new treatments. Researchers used mice genetically modified to lack IRF5 (Irf5-knockout mice) and exposed them to cigarette smoke to study emphysema. They found that these mice had less lung damage compared with normal mice. The study focused on how IRF5 affects cell death and immune cell behavior in the lungs. They discovered that IRF5 influences the expression of NLRP3, a protein involved in a type of cell death called pyroptosis, and affects immune cells known as Ly6C-expressing macrophages. Results suggest that targeting IRF5 could help reduce lung damage in emphysema by altering immune responses and cell death pathways. This summary was initially drafted using artificial intelligence, then revised and fact-checked by the author.

Figures

IRF5 deficiency reduces the lung damage caused by CS exposure.a Images of immunohistochemistry for IRF5 staining in the lung tissue from WT and *Irf5-*KO mice after exposure on fresh air (CTL) or CS. Scale bars, 50 μm. b A flow cytometry analysis showing the intensity and frequency of CD11c- or EpCAM-positive cells among IRF5-expressing cells isolated from lung tissue in WT mice following CS exposure. CD11c was used for analysis of AM and EpCAM for lung epithelial cells. c Representative images of lung tissue were stained with H&E from WT and *Irf5-*KO mice, with or without CS-induced emphysem

IRF5 deficiency attenuates the extent of NLRP3-mediated cell death caused by exposure to CS.a The CS-induced cell death-related protein levels in the lung tissue of WT and *Irf5-*KO mice were evaluated using western blotting. Left, representative immunoblot images of IRF5, NLRP3, caspase-1, caspase-3 and GAPDH. Right, relative expression levels were normalized to GAPDH. b The mRNA expression of NLRP3 and the related genes, including *Caspase 1*, GSDMD, IL-1β and IL-18, was confirmed in the lung tissue of WT and *Irf5-*KO mice after exposure to fresh air (CTL) or CS. c The confirmation of IRF5

MDSD- and Ly6C^high^-expressing cells increased by IRF5 deficiency in emphysema mice.a MDSC was identified using CD11b and Gr-1 from lung cells of WT and KO mice after exposure on fresh air (CTL) or CS by flow cytometry. b Macrophages subpopulations were gated and analyzed for CD86-positive (M1) or CD206-positive (M2) macrophages by flow cytometry. Antigen-presenting cells, along with M1 macrophages, were detected using MHCII. c Within the CD45^+^ population, myeloid cells were classified into neutrophils/PMN-MDSCs (CD11b^+^ Ly6G^+^), M-MDSCs (CD11b^+^Ly6G^-^Ly6C^+^), eosinophils (Ly6G^-^CD11b

Ly6C^high^ cells separated from *Irf5-*KO mice attenuated lung damage in CS-induced emphysema model.a A schematic diagram of the experimental design to confirm the effect of Ly6c strong-expressing cells. b A representative image obtained from lung tissue of control and CS-exposed groups including PBS-, Ly6c^−^ cell- or Ly6C^high^ cell-infected mice. c The quantification of injury in lung immunohistochemistry images by measuring MLI score. d The number of total cells isolated from BALF. e A confirmation of protein amount of cell death-related factors including NLRP3, caspase-3 and GSDMD in lung

The effect of IRF5 on Ly6C expression.a The efficiency of *IRF5* knockdown was evaluated by qRT–PCR. the mRNA levels of *NLRP3*, *GSDMD*, *IL-1β* and *Ly6C* were measured in pcDNA-IRF5-transfected Raw264.7 cells and compared with those in pcDNA-transfected control cells. Data are presented as means ± s.e.m. P* < 0.001; two-tailed, unpaired *t*-test with Welch’s correction. b Protein levels of IRF5 and Ly6C were assessed by flow cytometry following IRF5 knockdown to determine whether IRF5 expression altered Ly6C expression. c The transcriptional regulation of Ly6C by IRF5 was examined using luc

The correlation between IRF5 expression and emphysema.a To evaluate IRF5 expression, protein samples were extracted from lung tissue of 15 control individuals (1–15) and 16 patients with emphysema (16–31). ‘B’ represents the protein sample extracted from IRF5-overexpressing BEAS-2B cells. This sample was used as a reference for normalizing protein levels across different immunoreacted membranes in the Western blot analysis. β-actin was used for normalization of IRF5 expression. b The density of the IRF5 and β-actin bands was measured with ImageJ and analyzed quantitatively. P* < 0.001; two-tai

Full text

Introduction

Chronic obstructive pulmonary disease (COPD) is a progressive inflammatory disease commonly caused by prolonged exposure to noxious gases or particles, such as cigarette smoke (CS)^1^. CS induces both structural and functional damage to airway epithelial cells^2–5^ and promotes inflammation and cell death in the lungs^6–8^, leading to emphysema. Notably, even after smoking cessation, pulmonary and systemic inflammation persists in individuals with COPD^9^. Although bronchodilators—the current mainstay of treatment for COPD—improve lung function, quality of life and survival rate^10,11^, they are ineffective in attenuating the chronic inflammatory status in COPD. Thus, novel therapeutic approaches to effectively manage patients with COPD are urgently required.

The interferon regulatory factor (IRF) family has various functions in the cell cycle, cell death, carcinogenesis and gene regulation in response to pathogens^12^. Within the IRF family, IRF5 has a central role in inflammation, inducing proinflammatory cytokines such as interleukin 6 (IL-6), IL-12, IL-23 and tumor necrosis factor α (TNF-α)^13,14^. IRF5 is also highly expressed in macrophages, monocytes, dendritic cells (DCs) and B cells^13,15^. Notably, all of these cytokines and immune cells are related to COPD pathogenesis^16^. Previous studies have found an association between IRF5 and autoimmune diseases, suggesting that IRF5 may serve as a potential therapeutic target^17,18^. Regarding airway diseases, positive associations between IRF5-expressing macrophages and asthma have been reported^19,20^, although the associations between IRF5 and COPD have not been thoroughly reported. As previously shown, increased mRNA levels of IRF5 were observed in the lungs of CS-induced emphysema mice; this increase was attenuated with improvement in emphysema^21^.

In this study, we investigated the role of IRF5 in the pathogenesis of emphysema using in vivo approach. Through the genetic ablation of Irf5 in mice (Irf5-knockout (KO))^14,22^, we demonstrated its pivotal role in regulating cell survival and Ly6C expression in a CS-induced emphysema model. We also explored the cellular and molecular mechanisms underlying IRF5-mediated lung injury by assessing NLRP3/gasdermin D (GSDMD)-dependent pathways, immune cell phenotypes and downstream cytokine responses. Furthermore, we evaluated the clinical relevance of IRF5 by analyzing its expression in lung tissues of patients with emphysema and its correlation with pulmonary function metrics.

Our findings provide novel insights into the role of IRF5 in the pathogenesis of emphysema, highlighting its regulatory effects on NLRP3 expression, GSDMD activation and immune cell recruitment. These results not only advance our understanding of the molecular mechanisms underlying COPD but also suggest that IRF5 is a potential therapeutic target for alleviating lung damage in emphysema.

Materials and methods

Irf5-KO mice generation

All animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee of Asan Medical Center (approval number 2022-14-170). All mice were maintained in the specific pathogen-free facility of the Laboratory of Animal Research at Asan Medical Center. Mouse genomic DNA sequences were analyzed, and target sequences were selected using the web tool Benchling (https://benchling.com/). CRISPR–Cas9-mediated gene targeting in mice was performed as described previously^23,24^. In brief, C57BL/6N (B6N) and ICR mice (Orient Bio) were used as zygote donors and foster mothers, respectively, and were prepared as previously described^24^. The single guide RNA (sgRNA) and Cas9 messenger RNA (mRNA) were co-microinjected into the cytoplasm of B6N zygotes. After incubation at 37 °C, the surviving embryos were transferred into the oviducts of synchronized pseudo-pregnant foster mothers. Genomic DNA samples were prepared from tail biopsies to screen for founder mice with potential targeted mutations in Irf5, and genotyping PCR was conducted using the primers 5′-caggtgaacagctgccagta-3′ and 5′-ctctttcctagggaggggct-3′ (179 bp). PCR products from Irf5 mutant mice were cloned for sequence analysis using a T-Blunt PCR Cloning Kit (SolGent), and the mutations were identified using Sanger sequencing (Macrogen). The selected mutant allele was a 7-bp deletion in Irf5, which induced a frameshift followed by a premature stop codon. WT and Irf5-homozygous KO mice were generated by intercrossing Irf5-heterozygous KO mice (F_1_), which were subsequently killed at 6 weeks of age to obtain spleen samples. Tissue lysates were prepared and analyzed by immunoblotting using specific primary antibodies against IRF5.

CS-induced emphysema animal model

A mouse model of CS-induced emphysema was established as previously described^21,25^. In brief, 8-week-old WT and *Irf5-*KO mice were exposed to CS of 12 commercial cigarettes per day (four cigarettes per session, three sessions per day, 8.0 mg of tar per cigarette and 0.7 mg of nicotine per cigarette; Camel, R.J. Reynolds Tobacco Company) for 4 weeks via a whole-body apparatus with 50 μg poly (I:C) administration through nasal aspiration twice a week at 3 and 4 weeks. The control group inhaled only clean-room air (filtered air). All mice were anesthetized via isoflurane inhalation and killed after treatment.

Analysis of inflammatory cells in BALF

After 2.0 ml of PBS was slowly injected via a cannula inserted into the horizontally incised trachea of the mouse, the bronchoalveolar lavage fluid (BALF) was collected by gently drawing the injected PBS back into the syringe. The total cells in the collected BALF were counted and placed onto glass slides by cytocentrifugation. The slides were stained with the Diff-Quik Stain set (Sysmax, catalog number 38721) for differential cell counts.

Histomorphological analysis

After ligating the right main bronchus, the left lung was filled with 0.5% low-melting-point agarose (Invitrogen). The left lung tissue was fixed in 4% paraformaldehyde and embedded in paraffin. Thereafter, lung tissue was sectioned (5 μm thickness) and stained with hematoxylin and eosin (H&E). Slides stained with H&E were reviewed to measure the mean linear intercept (MLI), which represents the average distance between the alveolar septal walls^26^, determined by the number of interruptions in the 1-mm lines of the alveolar wall. Five random fields were selected per mouse, and four lines were drawn in each field for evaluation.

Single cell dissociation from lung tissue and spleen

To obtain a single cell from mouse lung tissue, tissues soaked in cold PBS were minced into fine pieces and enzymatically digested using an enzyme mix of dispase (Gibco, catalog number 17105041), collagenase (Gibco, catalog number 17101015) and DNase (Invitrogen, catalog number 18047019) at 37 °C while shaking. The cell suspension was pelleted after passing through a 70-μm strainer and resuspended for storage until use.

For splenocyte isolation, the spleens were excised from the mice and placed in a Petri dish containing cold PBS. The crushed spleen with a syringe plunger was filtered using a 70-μm strainer and washed with cold PBS before storage.

DNA construction and transfection

To construct plasmids bearing the IRF5 gene, a coding DNA sequence region of the IRF5 (NCBI reference sequence: NM_001252382.1, https://www.ncbi.nlm.nih.gov/refseq/) was amplified by PCR from mouse complementary DNA (cDNA). The PCR products were cloned into the BamHI and EcoRI sites of the pcDNA3.1 vector.

A 2-kb region of the Ly6C promoter was produced by PCR amplification from mouse genomic DNA on the basis of the NCBI reference sequence NC_000081.7. The fragment was cloned into the SacI and XhoI sites of the pGL4 reporter vector (Addgene, catalog number 48744). For the promoter assay, the luciferase activity was evaluated using the Bright-Glo luciferase assay system (Promega) following the manufacturer’s instructions.

BEAS-2B cells isolated from normal human bronchial epithelium were cultured with RPMI 1640 medium (Welgene, catalog number LM011-01) containing 10% fetal bovine serum (FBS; Biowest, catalog number S1480) and antibiotic–antimycotic solution (Gibco, Thermo Fisher Scientific, catalog number 15240-062). The mouse monocyte macrophage cell line Raw264.7 cells were cultured in DMEM supplemented with 1 mM sodium pyruvate, 10 mM HEPES, 10% FBS and antibiotic–antimycotic solution. The BEAS-2B and Raw264.7 cells were seeded onto six-well plates, and a total of 1 μg DNA per well was used for transfection with FuGene 6 transfection reagent (Promega, catalog number E2691) following the manufacturer’s instructions. The efficacy of overexpression of IRF5 was assessed by western blotting using IRF5-specific antibody or by quantitative PCR using IRF5-specific primers.

RNA interference of IRF5

For the knockdown of IRF5, ON-TARGETplus Nontargeting small interfering RNA (siRNA) Control Pool (catalog number D-001810-10-20) and ON-TARGETplus Mouse Irf5 siRNA (catalog number L-041093-02-0020) were purchased from Dharmacon. The BEAS-2B cells were transfected with 75 pmol IRF5 siRNA or nonspecific control siRNA using Lipofectamine 3000 transfection reagent (Invitrogen, catalog number L3000-008) following the manufacturer’s instructions. The efficacy of IRF5 knockdown was evaluated by quantitative PCR with reverse transcription (qRT–PCR) using IRF5-specific primers.

RNA extraction and real-time PCR

Total RNA was isolated from human bronchial epithelial cells (BEAS-2B, American Type Culture Collection, catalog number CRL-9609) and mouse lung tissue using TRIzol Reagent (Invitrogen, catalog number 15596018) according to the manufacturer’s instructions. The lung tissue was homogenized using stainless steel bead beating before RNA extraction. Single-stranded cDNA synthesis was performed with 2 μg total RNA, oligo(dT)18 primer and ReverseAid M-MuLV reverse transcriptase (RevertAid First Strand cDNA Synthesis kit, Thermo Scientific, catalog number K1621). Primer3 (https://primer3.ut.ee/), a web-based primer design tool, was used to design gene-specific primers^27^ (Supplementary Table 1). Quantitative real-time PCR was performed using a CFX Connect Real-Time PCR System (Bio-Rad Laboratories) with SYBR Green (Bio-Rad, BR1725272). The results were normalized to internal GAPDH as relative expression using the ΔCT method.

Protein isolation and western blotting

Total protein was extracted from cultured cells and lung tissues. The BEAS-2B cells were washed twice with cold PBS and lysed with radioimmunoprecipitation assay buffer (Abcam, catalog number ab156034) containing a protease and phosphatase inhibitor cocktail (Thermo Scientific, catalog number 78444). For lung tissue, tissues excised from mice and humans were homogenized by bead beating in T-PER tissue protein extraction reagent (Thermo Scientific, catalog number 78510) containing a protease and phosphatase inhibitor cocktail.

The protein concentrations were measured using bicinchoninic acid protein assay kits (Abcam, catalog number ab287853). Protein samples were separated by 8% SDS–polyacrylamide gel electrophoresis (SDS–PAGE) for NLRP3, 10% for IRF5, GSDMD, GAPDH, β-actin and vimentin and 12% for caspase-3 and LC3B. The polyvinylidene fluoride membrane (Merck Millipore, catalog number IPVH00010) used to transfer separated protein after electrophoresis was incubated with anti-IRF5 antibody (Abcam, catalog number ab181553), anti-GSDMD (Abcam, catalog number ab219800), anti-GAPDH (Abcam, catalog number ab181603), anti-β-actin (Abcam, catalog number ab8227), antivimentin (Abcam, catalog number ab137321), anti-caspase-3 (Cell Signaling Technology, catalog number 9662) and anti-LC3B (Cell Signaling Technology, catalog number 2775). ECL Pico Plus (Dynebio, catalog number GBE-P100) was used to detect immunoreactive proteins.

Flow cytometry and single-cell sorting

Single cells were suspended in cold PBS containing 2% FBS and incubated with the fluorescent conjugated antibodies listed below for 1 h at 4 °C. For IRF5, the cells were fixed and permeabilized using an intracellular fixation and permeabilization buffer (Invitrogen, catalog number 88-8824-00), according to the manufacturer’s instructions, before antibody incubation. A FACSCantoTM II Cell Analyzer and BD FACSymphony A3 Cell Analyzer (BC Biosciences) and FlowJo v10.8.1 software (BD Life Sciences) were used for multiparameter and data analyses, respectively. In all analyses, doublets were excluded using FSC-A/FSC-H profiles, and mononuclear populations were gated using FSC-A/SSC-A profiles. Single-stained controls and fluorescence-minus-one samples were used as compensation. The antibodies used were purchased from BioLegend and BD Bioscences. These are listed in Supplementary Table 2.

Cell sorting based on Ly6C expression in splenocytes from the *Irf5-*KO mice was performed using a BD FACSAria III sorter (BD Biosciences). Cells that were CD3 and/or CD11 positive were gated on live DAPI-negative regions and separated on the basis of the intensity of Ly6C expression. Ly6C-negative or strongly positive populations were each isolated in cold PBS containing 20% FBS. Before injection into the mice, the cell number was counted, and viability was determined by trypan blue staining.

Correlation of IRF5 and emphysema in humans

We included 31 patients with lung cancer who underwent lung resection surgery at the Asan Medical Center in Seoul, Korea. Patient clinicopathological information was obtained from medical records after de-identification. This study was approved by the Institutional Review Board of Asan Medical Center (institutional review board no. 2024-0267). All the participants provided written informed consent. For the human data, we performed a BEAS hoc analysis of the correlation between IRF5 levels and emphysema.

Statistical analysis

All statistical analyses were performed using GraphPad Prism version 5.0 (GraphPad Software). We used an unpaired t-test with Welch’s correction to identify any significant differences between two groups and one-way ANOVA followed by the Bonferroni post hoc test for three or more groups. Statistical differences were determined using significance levels of P < 0.05, P < 0.01 and P < 0.001.

Results

Reduced emphysema in Irf5-KO mice

Irf5-KO mice were generated using the CRISPR–Cas9 system for targeted deletion to study the in vivo function of IRF5 in CS-mediated emphysema. As an initial step, protein expression levels of IRF5 were evaluated in lung tissues of wildtype (WT) and *Irf5-*KO mice. IRF5 protein expression was confirmed in the lung tissue of WT mice, but it was not detected in any tissue of the *Irf5-*KO mice (Fig. 1a). The IRF5 expression increased in the lung tissue of CS-exposed mice, especially in alveolar macrophages (AMs), but remained very low in *Irf5-*KO mice, regardless of CS exposure. Further analysis was performed to determine the cell type that expressed IRF5 using flow cytometry (Fig. 1b). The IRF5-positive population increased by exposing CS to WT mice (9.1% in control, 17.1% in CS group). The population was then gated for analysis using CD11c and epithelial cell adhesion molecule (EpCAM) and classified into CD11c^+^ AMs or EpCAM^+^ alveolar epithelial cells. Among IRF5-positive cells, the proportion of AMs was 48.6% in the fresh air-exposed group and 74.9% in CS-induced emphysema group, indicating that AMs were a major population among CS-induced IRF5-expressing cells. In addition, the epithelial cells did not show differences between the fresh air-exposed and CS-induced emphysema groups (16.5% versus 14.4%).Fig. 1IRF5 deficiency reduces the lung damage caused by CS exposure.a Images of immunohistochemistry for IRF5 staining in the lung tissue from WT and *Irf5-*KO mice after exposure on fresh air (CTL) or CS. Scale bars, 50 μm. b A flow cytometry analysis showing the intensity and frequency of CD11c- or EpCAM-positive cells among IRF5-expressing cells isolated from lung tissue in WT mice following CS exposure. CD11c was used for analysis of AM and EpCAM for lung epithelial cells. c Representative images of lung tissue were stained with H&E from WT and *Irf5-*KO mice, with or without CS-induced emphysema. Scale bars, 200 μm. d An evaluation of airspace enlargement performed using MLI score from the images stained with H&E as shown in c. Data represent mean values ± s.e.m. e Left: the number of total cells was collected from BALF isolated from fresh air-exposed WT (CTL, N = 4), CS-exposed WT (CS, N = 5), fresh air-exposed *Irf5-*KO (CTL, N = 5) and CS-exposed *Irf5-*KO mice (CS, N = 6). Right: the immune cells including neutrophils, lymphocytes and monocytes in BALF were counted on the basis of their shape and size following Diff-quick staining. f BALF was analyzed by multiplex cytokine array profiling of eight inflammatory proteins including GM-CSF, IFN-γ, TNF-α, IL-1β, IL-6, IL-18, IL-22 and IL-27. g The quantification of cytokines mRNA expression in the lung tissue separated from WT and *Irf5-*KO mice. The relative expression of CSF2, IL-6, TNF-α and IFNs was normalized with 18s rRNA expression levels. *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA followed by Bonferroni post hoc test.

CS exposure led to the destruction of the alveolar walls, which were imaged (Fig. 1c) and evaluated by quantitative histological analysis and MLI (Fig. 1d). The MLI values were significantly higher in the CS group than in the control group in WT mice. However, no such difference was observed in *Irf5-*KO mice (Fig. 1d). The total number of cells collected from the BALF increased in the CS-exposed groups compared with that in the control groups, regardless of IRF5 expression (Fig. 1e). Neutrophil and monocyte counts showed similar patterns, although the increased number of lymphocytes observed in CS-exposed WT mice was reduced in CS-exposed *Irf5-*KO mice (Fig. 1e). In addition, the secretion levels of cytokines into BALF were analyzed by multiplex cytokine assay (Fig. 1f). The levels of IFN-γ, TNF-α, IL-1β, IL-6 and IL-27 increased by CS exposure in WT mice. The amount of GM-CSF, TNF-α, IL-1β, IL-6 and IL-18 in BALF was highest in CS-exposed *Irf5-*KO group compared with all other groups. As IRF5 is one of the IFN regulatory factors and functions as a transcription factor, IFN expression, along with CSF2, IL-6 and TNF-α expression, in lung tissues was also analyzed. The IFN-γ expression significantly increased in WT mice following CS exposure but did not change in *Irf5-*KO mice, whereas the expression of IFN-α and IFN-β was higher in CS groups than in control groups (Fig. 1g). In addition, IL-6 expression had a similar pattern to IFN-γ expression, with its expression being repressed in *Irf5-*KO groups. CSF2 and TNF-α expression was significantly higher in CS-exposed groups than in control groups in both WT and *Irf5-*KO mice. The MLI values and IFN-γ expression, which were increased by CS exposure, were remarkably reduced in *Irf5-*KO mice compared with WT mice, suggesting that IRF5 may regulate CS-induced emphysema by transcriptionally controlling IFN-γ expression.

Regulation of cell survival by IRF5 in emphysema

To explore whether IRF5 regulates cell survival in lung tissue during emphysema development, the expression and activation of several factors related to cell death were examined (Fig. 2a). NLRP3 expression increased in WT mice after CS exposure but was repressed in CS-exposed *Irf5-*KO mice. Unexpectedly, the cleaved form of caspase-1 and caspase-3 was detected in the *Irf5-*KO mice after CS exposure, whereas its full-length form remained unchanged. The expression of NLRP3-related genes was measured to investigate whether IRF5 induces pyroptosis, inflammatory response or both in emphysema. Consistent with the protein expression, Nlrp3 gene expression in WT mice was higher than in *Irf5-*KO mice after CS exposure. The expression of GSDMD, IL-1β and IL-18 did not show significant changes among groups, whereas caspase-1 was more highly expressed in CS-induced groups, independent of IRF5 expression (Fig. 2b). The transcriptional regulation of NLRP3 by IRF5 was investigated using an Irf5-expressing plasmid generated by inserting the coding DNA sequence region of Irf5 into pcDNA3.1. IRF5 expression was efficiently increased by transfection with an Irf5-expressing plasmid, leading to a proportional change in NLRP3 expression (Fig. 2c). These results suggest that IRF5 may play a role in NLRP3-induced pyroptosis by regulation of NLRP3 expression.Fig. 2IRF5 deficiency attenuates the extent of NLRP3-mediated cell death caused by exposure to CS.a The CS-induced cell death-related protein levels in the lung tissue of WT and *Irf5-*KO mice were evaluated using western blotting. Left, representative immunoblot images of IRF5, NLRP3, caspase-1, caspase-3 and GAPDH. Right, relative expression levels were normalized to GAPDH. b The mRNA expression of NLRP3 and the related genes, including Caspase 1, GSDMD, IL-1β and IL-18, was confirmed in the lung tissue of WT and *Irf5-*KO mice after exposure to fresh air (CTL) or CS. c The confirmation of IRF5 mRNA and protein expression following transfection of pcDNA or pcDNA-IRF5 into BEAS-2B cells. The mRNA expression of IRF5 and NLRP3 assessed by normalization to GAPDH. d The respective immunoblot images of full length, p40, p30 and p23 of GSDMD (left) and quantitative analysis (right). *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA followed by Bonferroni post hoc test.

Regarding pyroptosis, cleaved caspase-3 may generate p40 or p23 isoforms of GSDMD, which contribute ‘no pyroptosis’^28–30^. As cleaved caspase-3 was only increased in the CS-exposed *Irf5-*KO group, it was necessary to confirm whether cleaved caspase-3 contributed to the suppression of pyroptosis along with reduced NLRP3 expression. Interestingly, p40 and p23 GSDMD were markedly expressed in the CS-exposed *Irf5-*KO group, whereas p30 GSDMD, which is involved in pyroptosis, was highly expressed in the CS groups (Fig. 2d).

Increased Ly6C-expressing cells in CS-exposed Irf5-KO mice

Considering that IRF5 functions as a mediator of various immune responses, the characteristics of immune cells were analyzed using flow cytometry. Myeloid-derived suppressor cell (MDSC) was identified by using CD11b and Gr-1, and the macrophage polarization was confirmed using CD86 and CD206 as an M1 and M2 marker, respectively^31,32^. MHC class II was used for both M1 macrophages and antigen-presenting cells. The cell was dissociated from lung tissue and analyzed by flow cytometry. Interestingly, CD11b^+^Gr-1^+^ cells (MDSC) were significantly higher in the CS-exposed *Irf5-*KO group compared with all other groups (Fig. 3a). In addition, CD11c^+^MHCII^+^ cells were also significantly more abundant in *Irf5-*KO mice following CS exposure than any other groups, whereas CD11b^+^CD86^+^ and CD11b^+^CD206^+^ populations increased by CS exposure in WT and KO groups (Fig. 3b). As the CD11b^+^Gr-1^+^ population encompasses both neutrophil and MDSC, further analysis was performed to characterize these cell subsets. First, various myeloid populations were assessed using cell type-specific surface markers and a refined gating strategy^33^ (Supplementary Fig. 1). CD11b^+^Ly6G^+^ cells, which include neutrophils and polymorphonuclear MDSCs (PMN-MDSC) were significantly increased in the CS-exposed groups compared with fresh air-exposed controls, whereas AMs (CD11b^─^Ly6G^─^CD11c^+^Siglec-F^+^ cells) were reduced in the CS-exposed groups (Fig. 3c). By contrast, the frequencies of monocytic-MDSCs (M-MDSD; Ly6G^─^CD11b^+^Ly6C^+^) and eosinophils (Ly6G^─^CD11b^+^Siglec-F^−^) did not differ significantly among the groups. In CS-exposed *Irf5-*KO mice, DCs were significantly increased, whereas interstitial macrophages (IMs) were significantly decreased compared with all other groups. Because neutrophils and PMN-MDSCs originate from the same source and exhibit numerous morphological and phenotypic similarities^34^, further analysis was performed to characterize the CD11b^+^Ly6G^+^ population using functional surface markers, including Siglec-E, Siglec-H, CD33, CD84, CD115 and CD244.2^35–37^. Siglec-E, a neutrophil-associated marker, was most highly expressed in CS-exposed WT mice (8.8% ± 1.9%) compared with all other groups (1.5% ± 0.8% in WT mice, 5.1% ± 1.6% in KO mice and 5.6% ± 1.1% in *Irf5-*KO mice after CS exposure). By contrast, CD244.2, a PMN-MDSC-associated marker, was significantly elevated in *Irf5-*KO mice following CS exposure (55.1% ± 5.1%) than any other groups (31.2% ± 6.0% in WT mice, 35.9% ± 9.0% in WT mice after CS exposure and 26.2% ± 7.2% in KO mice) (Fig. 3d). Interestingly, CD115 was detectable, although at a low frequency (1.6% ± 0.6%), in CS-exposed *Irf5-*KO mice. Siglec-H expression was generally higher in KO mice, albeit with considerable variability, whereas CD84 expression remained similar across all groups. Conversely, CD33-positive cells were reduced in CS-exposed WT mice compared with all other groups (Fig. 3d). These results suggest that IRF5 may function in CS-induced emphysema by inhibition of immunosuppressive cells.Fig. 3MDSD- and Ly6C^high^-expressing cells increased by IRF5 deficiency in emphysema mice.a MDSC was identified using CD11b and Gr-1 from lung cells of WT and KO mice after exposure on fresh air (CTL) or CS by flow cytometry. b Macrophages subpopulations were gated and analyzed for CD86-positive (M1) or CD206-positive (M2) macrophages by flow cytometry. Antigen-presenting cells, along with M1 macrophages, were detected using MHCII. c Within the CD45^+^ population, myeloid cells were classified into neutrophils/PMN-MDSCs (CD11b^+^ Ly6G^+^), M-MDSCs (CD11b^+^Ly6G^-^Ly6C^+^), eosinophils (Ly6G^-^CD11b^+^Siglec-F^+^), DC (Ly6G^-^Siglec-F^-^F4/80^-^CD11c^+^MHCII^+^), IMs (IMs; Ly6G^-^Siglec-F^-^F4/80^+^Ly6C^int^) and AMs (CD11b^-^Ly6G^-^Siglec-F^+^CD11c^+^). d Within the neutrophil/PMN-MDSC gate (CD11b^+^ Ly6G^+^), subpopulations were further characterized using six surface markers–including Siglec-E, Siglec-H, CD33, CD84, CD115 and CD244.2–across all experimental groups. e CD45, CD3, CD11b and Ly6C-expressing cells were collected from BALF, and their proportions were estimated by flow cytometry. The Ly6C expression depending on its intensity was analyzed as percentage among CD45^+^/CD3^+^ or CD45^+^/CD11b^+^ cells. *P < 0.05; ***P < 0.001; two-tailed, unpaired t-test with Welch’s correction.

In addition, total cells were isolated from the BALF of WT and *Irf5-*KO mice after CS exposure, and CD45, CD3, CD11b and Ly6C were stained to distinguish between leukocytes (CD45^+^), monocytes (CD11b^+^), T cells (CD3^+^) and inflammatory cells (Ly6C^+^) (Fig. 3e). CD3^+^ T cells were slightly decreased in *Irf5-*KO mice compared with WT mice, whereas the proportions of CD45^+^ and CD11b^+^ cells were similar between the groups. Interestingly, Ly6C expression in both CD11b^+^ and CD3^+^ populations differed between the WT and *Irf5-*KO groups. Ly6C-positive cells were divided into Ly6C^+^ (positive) and Ly6C^++^ (strongly positive, Ly6C^high^). In the CD11b^+^ population, Ly6C^high^ cells were significantly more abundant in *Irf5-*KO mice than in WT mice following CS exposure (%, 20.6 ± 5.7 in *Irf5-*KO mice versus 4 ± 2.6 in WT mice). This trend was also observed in CD3^+^ T cells (%, 25.9 ± 4.5 in *Irf5-*KO mice versus 14.4 ± 5.4 in WT mice, Fig. 3e). In further analysis, Ly6G-CD11b^+^ Ly6C^high^ cells mainly consist of 52.7% ± 8.6% F4/80^−^ and 30.2% ± 8.1% F4/80^+^ (Supplementary Fig. 2).

Reduced CS-exposed emphysema by Ly6-expressing cells

To investigate whether Ly6C^high^ cells mediate alveolar cell survival and contribute to emphysema recovery, Ly6C^high^ cells were isolated from the splenocytes of CS-exposed *Irf5-*KO mice and injected into the tail veins of C57BL/6 mice during CS exposure (Fig. 4a). In CS-exposed groups, mice injected with Ly6C^high^ cells exhibited less alveolar enlargement (Fig. 4b) and had a significantly lower MLI than the mice injected with PBS- or Ly6C^−^ cell-injected groups (Fig. 4c). In addition, the total cell number in the BALF increased by CS was reduced in both Ly6C^-^ and Ly6C^high^ cell-injected groups to a similar extent as that in the control group (Fig. 4d). The protein expression of NLRP3 was lower in Ly6C^high^ cell-injected mice than in PBS or Ly6C cell-injected mice, whereas the expression levels of caspase-3 and GSDMD were not significantly different between the groups (Fig. 4e,f). Ly6C^−^ cell administration increased p23 and p30 GSDMD levels, whereas Ly6C^high^ cell injection did not alter GSDMD isoform expression. These results suggest that the increased number of Ly6C^high^ cells in CS-exposed *Irf5-*KO mice may enhance alveolar cell survival by regulating NLRP3 expression.Fig. 4. Ly6C^high^ cells separated from *Irf5-*KO mice attenuated lung damage in CS-induced emphysema model.a A schematic diagram of the experimental design to confirm the effect of Ly6c strong-expressing cells. b A representative image obtained from lung tissue of control and CS-exposed groups including PBS-, Ly6c^−^ cell- or Ly6C^high^ cell-infected mice. c The quantification of injury in lung immunohistochemistry images by measuring MLI score. d The number of total cells isolated from BALF. e A confirmation of protein amount of cell death-related factors including NLRP3, caspase-3 and GSDMD in lung tissue from mice injected Ly6C negative- or strong positive-expressing cell after exposure on fresh air (CTL) or CS. f The relative band intensity of NLRP3, caspase-3 and GSDMD against GAPDH. *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA followed by Bonferroni post hoc test.

Regulation of Ly6C expression by IRF5

To determine whether IRF5 regulates Ly6C expression in immune cells, both Ly6C expression and promoter activity were examined under different IRF5 expression conditions. As IRF5 was previously shown to regulate NLRP3-induced pyroptosis (Fig. 2), the expression of NLRP3-related genes was confirmed in a macrophage cell line, Raw264.7 cells following IRF5 overexpression. In IRF5-overexpressing cells, NLRP3 expression remained unchanged, whereas GSDMD and IL-1β expression levels were reduced. Notably, IRF5 overexpression led to a marked decrease in Ly6C expression (Fig. 5a). Conversely, IRF5 knockdown using IRF5-specific siRNA increased Ly6C expression, as demonstrated by flow cytometry: Ly6^+^ cells increased from 4.0% in control to 7.4% following IRF5 silencing (64.9% in control siRNA-transfected cells versus 55.7% in siIRF5-transfected cells) (Fig. 5b).Fig. 5. The effect of IRF5 on Ly6C expression.a The efficiency of IRF5 knockdown was evaluated by qRT–PCR. the mRNA levels of NLRP3, GSDMD, IL-1β and Ly6C were measured in pcDNA-IRF5-transfected Raw264.7 cells and compared with those in pcDNA-transfected control cells. Data are presented as means ± s.e.m. ***P < 0.001; two-tailed, unpaired t-test with Welch’s correction. b Protein levels of IRF5 and Ly6C were assessed by flow cytometry following IRF5 knockdown to determine whether IRF5 expression altered Ly6C expression. c The transcriptional regulation of Ly6C by IRF5 was examined using luciferase reporter assays under varying levels of IRF5 expression, achieved by transfection with pcDNA-IRF5 (pIRF5) and/or IRF5-specific siRNA. *P < 0.05; **P < 0.01; one-way ANOVA followed by Bonferroni post hoc test.

As two putative IRF5 binding sites were identified within the 2-kb region of the Ly6C promoter, it was investigated whether IRF5 acts as a transcriptional repressor in immune cells. The overexpression of IRF5 markedly reduced Ly6C promoter activity, whereas IRF5 knockdown restored promoter activity to levels exceeding those observed in control cells (pcDNA plus scrambled siRNA (siScr)-transfected cells) (Fig. 5c). These findings suggest that IRF5 may modulate CS-induced immune response through the regulation of Ly6C expression.

Analysis of IRF5 expression in human lung tissue

To validate the findings in mice experiments in humans, human lung samples were obtained from 31 patients with lung cancer and categorized as control (diffusing capacity of the lungs for carbon monoxide (DL_CO_) ≥95%) or emphysema (DL_CO_ ≤75%). Of them, 15 control individuals (mean age 63.5 ± 8.4 years of age) had a predicted forced expiratory volume in 1 s (FEV_1_%pred) of 101.5% ± 11.4%, whereas 16 patients with emphysema (mean age 66.3 ± 5.2 years) had an FEV_1_%pred of 69.7% ± 8.8%—significantly lower than that of the control group (Table 1). Crude proteins were extracted from the 31 human lung tissue samples, separated and analyzed for IRF5 expressions (Fig. 6a). The relative amount of immunoreactive IRF5-specific bands was quantified against β-actin, with normalization performed using common reference samples included in every membrane (Fig. 6b). Notably, IRF5 expression was significantly elevated in lung tissues from patients with emphysema than in the tissues from controls. These results suggest that IRF5 is a potential therapeutic target for patients with COPD with emphysema.Fig. 6. The correlation between IRF5 expression and emphysema.a To evaluate IRF5 expression, protein samples were extracted from lung tissue of 15 control individuals (1–15) and 16 patients with emphysema (16–31). ‘B’ represents the protein sample extracted from IRF5-overexpressing BEAS-2B cells. This sample was used as a reference for normalizing protein levels across different immunoreacted membranes in the Western blot analysis. β-actin was used for normalization of IRF5 expression. b The density of the IRF5 and β-actin bands was measured with ImageJ and analyzed quantitatively. P < 0.001; two-tailed, unpaired t-test with Welch’s correction.Table 1. Clinical characteristics of patient samples used in Fig. 6.Control (n = 15)Emphysema (n = 16)Age, years63.5 ± 8.466.3 ± 5.2Sex, male5 (33.3%)15 (93.6%)Lung function Post-bronchodilator FEV_1_/FVC79.4 ± 4.462.0 ± 8.9 Post-bronchodilator FEV_1_ (%pred)101.5 ± 11.469.7 ± 8.8 Post-bronchodilator FVC (%pred)96.1 ± 7.582.4 ± 11.9 DL_CO_ (%pred.)109.1 ± 13.666.4 ± 5.0*Smoking history, ex-smoker3 (20.0%)14 (87.5%)Data are presented as numbers (%) or means ± s.d. FVC, forced vital capacity; %pred, percent of the predicted value. *P < 0.001.

Discussion

In this study, we investigated the role of IRF5 in emphysema pathogenesis by focusing on its interaction with the NLRP3 inflammasome and Ly6C^high^ macrophages. *Irf5-*KO mice showed decreased alveolar damage after exposure to CS. The expression of NLRP3, a major mediator of pyroptosis was attenuated in *Irf5-*KO mice following CS exposure, and transfection of pcDNA-IRF5 into BEAS-2B cells confirmed the role of IRF5 as a transcriptional regulator in pyroptosis. The proportion of Ly6C^high^ macrophages increased in the lungs of *Irf5-*KO mice after CS exposure, and the introduction of Ly6C^high^ immune cells isolated from *Irf5-*KO mice attenuated NLRP3 expression and CS-induced alveolar damage. In addition, it showed that IRF5 suppresses Ly6C expression in immune cells by directly repressing Ly6C promoter activity, as shown by decreased Ly6C levels with IRF5 overexpression and increased expression with IRF5 knockdown. We also observed correlations between IRF5 and emphysema in human lungs. Taken together, these results indicate that IRF5 plays an important role in COPD development via NLRP3-mediated pyroptosis and Ly6C^high^ macrophages. This is a report on the role of IRF5 in emphysema development thorough the NLRP3-induced pyroptosis pathway.

The role of IRF5 in immune regulation has been extensively studied; however, its potential contribution to COPD pathogenesis remains largely unexplored. Recent evidence suggests that IRF5 acts as a master regulator of myeloid cell differentiation, influencing macrophage polarization and monocyte recruitment in chronic inflammatory diseases^15,38^. COPD is characterized by an imbalance between tissue destruction and repair, partly driven by dysregulated immune responses^39^. Our study provides new insights into the contribution of IRF5 to this imbalance, particularly through its effects on the recruitment and function of Ly6C-expressing immune cells. In the context of lung injury, Ly6C^+^ monocytes can differentiate into inflammatory macrophages that exacerbate tissue damage or reparative macrophages that promote inflammation resolution^40^. The increased proportion of Ly6C^high^ macrophages in IRF5-deficient mice suggests that IRF5 acts as a key regulator of monocyte differentiation, possibly promoting a proinflammatory phenotype that worsens emphysema. This observation is consistent with previous reports demonstrating that IRF5-deficient macrophages exhibit impaired inflammatory cytokine production and enhanced tissue repair functions^41^. Thus, IRF5 inhibition may be a potential strategy to modulate macrophage plasticity and shift the immune response toward tissue preservation in COPD.

Another emerging concept in COPD pathogenesis is the role of programmed cell death mechanisms beyond apoptosis, particularly pyroptosis and necroptosis, in driving alveolar destruction^42,43^. *Irf5-*KO mice exhibited increased caspase-3 activation and altered GSDMD cleavage following CS exposure, as well as significantly reduced NLRP3 expression compared with WT controls (Fig. 2). These findings suggested a complex interplay between apoptosis and pyroptosis in emphysema pathogenesis, with IRF5 playing a critical role in shaping the inflammatory response and cellular fate. Pyroptosis, mediated by NLRP3 inflammasome activation and the caspase-1 pathway, has been implicated in CS-induced lung injury and contributes to the chronic inflammation observed in both human and animal models of COPD^25,44^. Notably, caspase-3 cleavage of GSDMD can generate a nonpyroptotic isoform, effectively suppressing pyroptotic cell death while promoting apoptosis^30^ or not triggering apoptosis^28^. In addition, papain treatment generated the p40 N-terminal form of GSDMD and released IL-33 without activating necroptosis or apoptosis^29^. In our model, the increased cleaved caspase-3 that led to the alteration of GSDMD processing in *Irf5-*KO mice (Fig. 2) indicated a shift toward the inhibition of pyroptosis rather than apoptosis—this may contribute to improved cellular survival and reduced emphysema severity. Furthermore, the reduced NLRP3 expression in *Irf5-*KO mice aligns with our observation of attenuated pyroptosis. As NLRP3-mediated pyroptosis exacerbates inflammatory lung injury^25,45^, its suppression may protect against the progression of emphysema. The increased levels of cleaved caspase-3 observed in *Irf5-*KO mice might facilitate controlled cell turnover by promoting the alteration of GSDMD isoforms, preventing excessive lung tissue destruction. Collectively, these results suggest that IRF5 contributes to emphysema pathogenesis by promoting pyroptosis via NLRP3 activation, whereas its deletion shifts toward cell survival, thereby mitigating lung damage.

In our study, the introduction of Ly6C^high^ splenocytes from *Irf5-*KO mice attenuated alveolar damage in the lungs of CS-exposed mice (Fig. 4). This finding highlights the important role of Ly6C^high^ macrophages in modulating inflammatory responses and lung tissue integrity in emphysema, similar to Irf5 deletion. However, studies on the role of Ly6C^high^ macrophages in airway diseases are limited. IMs in the lungs are thought to be derived from Ly6C^+^ monocytes, which play diverse roles depending on their polarization and environmental stimuli^46–48^. Generally, Ly6C^high^ macrophages are involved in acute inflammation and have a proinflammatory ability, whereas Ly6c^low^ macrophages have a protective role in anti-inflammatory and antifibrotic processes^49^. Interestingly, our findings are consistent with reports suggesting that monocyte-derived macrophages can shift phenotypes depending on local cytokine and chemokine environments^41^. For example, Ly6C^+^ macrophages may switch to a more reparative phenotype in the absence of key inflammatory drivers such as IRF5, which has been implicated in macrophage polarization and inflammasome activation^13^. The dynamic role of Ly6C^+^ macrophages in airway diseases may extend beyond inflammation. Emerging evidence suggests that Ly6C^+^ monocytes are critical for mediating tissue remodeling and adaptive immunity in response to chronic insults such as CS or persistent allergen exposure^50^. This dual role underscores the complexity of macrophage biology and highlights the need for further studies to elucidate the specific contributions of Ly6C^+^ macrophages to the different phases of airway disease progression.

This study showed the increased cytokines in the BALF of *Irf5-*KO mice, although alveolar destruction was decreased, and pyroptosis-related NLRP3 expression was downregulated in *Irf5-*KO mice after CS exposure. Elevated levels of cytokines such as IL-1β, IL-6, IL-12, IL-18, GM-CSF and TNF-α can lead to severe inflammation and tissue damage^51^. However, these effects may be counteracted by various regulatory mechanisms, including the activation of immunosuppressive MDSC^52^. We observed increased CD11b^+^Gr-1^+^ MDSC in lung tissue of CS-induced *Irf5-*KO mice (Fig. 3a). These findings suggest that IRF5 may contribute to balance immune activation and resolution. In addition, macrophage polarization into either M1 (CD86^+^) or M2 (CD206^+^) phenotype remained unchanged in terms of their relative expression (Fig. 3b). By contrast, CD11b^+^MHCII^+^ cells were significantly increased in the lung tissue of *Irf5-*KO mice following CS exposure compared with all other groups (Fig. 3b). These results implied that IRF5 may play a more prominent role in regulating antigen presentation rather than influencing macrophage polarization.

GM-CSF can activate specific signaling pathways depending on its concentration, allowing it to influence cell survival, proliferation and differentiation in a dose-dependent manner^53–56^. GM-CSF influences the differentiation and function of MDSC, particularly the Ly6C-expressing subset, which exhibits immunosuppressive activity^57–59^. In addition, GM-CSF drives the generation of CD11c^+^MHCII^+^ cells, a heterogeneous population that includes both conventional DCs and monocyte-derived macrophages^60^. Taken together, our findings suggest that GM-CSF might act as a key modulator in IRF5-mediated mechanism of CS-induced emphysema.

In the clinical data, a significant difference in IRF5 levels in lung tissue was observed between individuals with COPD and controls; however, correlations between IRF5 levels and clinical parameters, such as lung function or disease exacerbation, could not be drawn because of the small sample size and cross-sectional data. The correlation between IRF5 levels and emphysema in mouse models and humans warrants further investigation into the role of IRF5 in the pathophysiological mechanisms of COPD.

In conclusion, this study identified IRF5 as a key regulator of emphysema pathogenesis and demonstrated its involvement in NLRP3-mediated pyroptosis and immune cell dynamics. Using an *Irf5-*KO mouse model, we demonstrated that Irf5 deletion reduced alveolar destruction and suppressed NLRP3 expression, highlighting its role in modulating cell death pathways in response to CS exposure. Furthermore, the increase in Ly6C^high^ macrophages in *Irf5-*KO mice suggests a potential link between IRF5 and monocyte differentiation, with implications for immune-mediated tissue remodeling in COPD. Importantly, our findings were supported by human lung tissue analysis, which revealed elevated IRF5 expression in patients with emphysema. These results suggest that IRF5 may serve as a novel therapeutic target, offering a potential strategy for mitigating the progression of emphysema and preserving lung function in patients with COPD.

Supplementary information

Supplementary Information