Blocking IL-23 Signaling Mitigates Cigarette Smoke-induced Murine Emphysema

Abstract

Inflammatory cell infiltration is a characteristic feature of COPD and correlates directly with the severity of the disease. Interleukin-23 (IL-23) is a proinflammatory cytokine that regulates Th-17 inflammation, which mediates many pathophysiological events in COPD. The primary goal of this study was to determine the role of IL-23 as a mediator of key pathologic processes in cigarette smoke-induced COPD. In this study, we report an increase in IL23 gene expression in the lung biopsies of COPD patients compared to controls and identified a positive correlation between IL23 gene expression and disease severity. In a cigarette smoke-induced murine emphysema model, the suppression of IL-23 with a monoclonal blocking antibody reduced the severity of cigarette smoke-induced murine emphysema. Mechanistically, the suppression of IL-23 was associated with a reduction in immune cell infiltration, oxidative stress injury, and apoptosis, suggesting a role for IL-23 as an essential immune mediator of the inflammatory processes in the pathogenesis of CS-induced emphysema.

Introduction

COPD is a chronic, debilitating, progressive lung disease induced by cigarette smoking (CS) in most cases ¹. Characteristics of COPD are airway and alveolar inflammation, mucus hypersecretion, alveolar destruction, and remodeling/fibrosis of the airway mucosa ². Microscopically, the lungs of COPD patients are persistently infiltrated by inflammatory cells, including macrophages, neutrophils, and B /T lymphocytes ². The extent of infiltration by certain immune cells, such as macrophages and lymphocytes, has been reported to correlate directly with the severity of the disease, suggesting a direct role for persistent immune cell infiltration as “drivers” of disease progression in COPD ^3–5.

Interleukin-23 (IL-23) is a proinflammatory cytokine that regulates Th-17 inflammation. Characterized by the generation of IL-17A secreting T cells, Th-17 inflammation is increasingly appreciated as a mediator of many pathophysiological events in COPD ^1,6,7, including macrophage, neutrophil, and B/T lymphocytes recruitment and activation, mucus hyper-secretion, airway hyper-reactivity, and airway remodeling ^8,9. Although the roles of IL-17A and other Th-17 cytokines in COPD are well recognized, relatively little is known regarding the potential roles of IL-23 in the pathogenesis of CS-induced COPD. Levels of IL-23 were increased in the bronchial biopsies ⁶ and the peripheral blood and sputum ¹⁰ of patients with COPD, while suppression of IL-23 was recently reported to attenuate elastase-induced emphysema in a murine model ¹¹, which was mechanistically linked with a reduction in macrophage activation and diminished neutrophil recruitment in the lungs ^6,12. Despite these observations, the role of IL-23 in the pathogenesis of CS-induced COPD remains poorly understood. The primary goal of this study was to determine the role of IL-23 as a mediator of key pathologic processes in COPD. Using a murine model of COPD, we show that the inhibition of IL-23 in vivo attenuates lung oxidative stress, inflammation, cellular apoptosis, and emphysema severity.

Materials and Methods

Reagents and Miscellaneous Materials

Reagents used in this study include rat IL-23 (R&D, catalog# 3136RL010), human IL-23A (R&D, catalog# 1290IL010), and N-Acetylcysteine (Thomas, catalog# C822F69). Primary antibodies used in this study include: goat anti-IL-23 receptor (GeneTex, catalog# GTX39947), rabbit anti-Nrf2 (Sigma, catalog# SAB4501984), rabbit anti-HO-1(Sigma, catalog# 374090), rabbit anti-caspase 8 (Proteintech, catalog# 13423-1-AP), mouse anti-GAPDH (Invitrogen, catalog# 21068028), rabbit anti-S100/A9 (Proteintech, catalog# 26992-1-AP), rabbit anti-MPO(abcam, catalog# ab208670), mouse anti-CD11c (abcam, catalog# ab52632), mouse anti-IL-23 p19 monoclonal antibody (LS Bio, catalog# LS-C757848), and isotype matched mouse IgG2b kappa monoclonal antibody (abcam, catalog# ab18428).

Human Lung Tissue Biopsies

Formalin-fixed and paraffin-embedded (FFPE) lung tissue sections of patients with COPD were obtained from the archived tissue bank at Mayo Clinic (n=6). The diagnosis of COPD was made according to the standard American Thoracic Society guidelines for COPD diagnosis ¹³. The control tissue sections were taken from lung tissue adjacent to lung benign nodules from patients undergoing surgical resection without COPD (n=6). Control patients were selected based on the review of imaging and clinical evaluation that excluded other pulmonary diseases, such as pneumonia, bronchiectasis, interstitial diseases, and lung cancer. All lung tissue sections were subsequently subjected to an RNAscope assay to detect the expression and sources of IL23A mRNA. The use of human lung tissue in this study was approved by the Institutional Review Board of Mayo Clinic.

Dual in situ Hybridization and Immunohistochemistry

To identify the expression level and cell source of IL23A mRNA in lung tissue sections of patients with COPD, a chromogenic RNAscope in situ hybridization (ISH) assay (ACD, RNAscope 2.5 High Definition—BROWN kit, catalog# 322300) was operated with the Hs-IL23A probe (ACD, catalog# 562851) in combination with immunohistochemistry (IHC). Briefly, FFPE lung tissue sections were deparaffinized and rehydrated with ethanol series from 100% to 75% and dH₂O. Then, hydrogen peroxide was applied to deactivate the endogenous peroxidase to control the false positive background staining. Subsequently, tissue sections were submerged into the boiled Target Retrieval solution for 15 min, washed in distilled water, coated with Protease Plus, and incubated at a 40°C humidified oven for 30 minutes. Probe hybridization and signal visualization were achieved by two hours of IL23A probe incubation in an oven, followed by successive cultures of amplifiers (Amp 1-5) for signal amplification and a mixture of equal volumes of BROWN-A and -B for signal coloration. The IL23A mRNA ISH-probed lung tissue sections were then antigen retrieved for subsequent IHC against macrophage and dendritic cell marker CD11c by submerging into preheated boiling 1mM EDTA for 20 minutes, followed by 0.2% Triton/PBS for permeabilization and incubated with anti-CD11c overnight at 4°C. HRP-conjugated goat anti-rabbit secondary antibody was then applied at room temperature for 45 minutes. After rinsing and washing in 0.1% PBST for 3×10 minutes, the slides were incubated with AEC single solution (Abcam, IHC detection kit, catalog# ab236467) for 10 minutes. Following counterstaining with 50% hematoxylin for 2 minutes, the slides were coverslipped with an aqueous mounting medium.

RNAscope Quantification

The images of positively stained IL23A mRNA punctate dots were captured with the microscope (Olympus cellSens Dimension system) under brightfield at ×20 magnification for further semi-quantification using ImageJ (Fiji) following a modified method provided by ACDBio (https://acdbio.com/dataanalysisguide). Color deconvolution was performed using the trainable Weka segmentation plugin to assess the expression of IL23A, followed by the Analyze Particles tool to calculate the number of nuclei ¹⁴. The relative expression of IL23A mRNA in each group was defined as the average integrated density of probe signals normalized by cell numbers in each captured image field.

Gene Expression Profiling Dataset

To investigate the IL23A expression profile in COPD, we traced a gene expression dataset (GSE47460) (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE47460) from the National Center for Biotechnology Information. This dataset was composed of microarray data from total RNA extracted from the lung tissue of patients with COPD (n=220) and control lung tissue biopsies (n=108) obtained from patients who had undergone surgery for benign nodules or tumors and had no chronic lung diseases by computerized tomography scan (CT) or pathology review. IL23A transcriptional expression values were analyzed and visualized using the GEO2R online analyzer.

Cigarette Smoke-Induced Mouse Emphysema Model

A/J mice (17-22g, 8 weeks) were selected to create a CS-induced emphysema model due to its high susceptibility to the development of emphysema ^15,16. This work involving the use of animals was approved by the Institutional Animal Care and Use Committee of Mayo Clinic (IACUC: A00002404-16). Briefly, 29 mice were either exposed to CS (N=23) or room air (negative controls, N=6) for 5 hours a day (20 cigarettes per day), 5 days per week for three months. On day 1 and day 15, mice received intratracheal LPS instillation (0.3mg/kg) in a single dose to induce lung inflammation. Starting in the last four weeks following the CS exposure, mice were randomly assigned to different treatment groups, including a positive control group that received isotype mouse IgG (0.6mg/kg, n=10) and a treatment group that received IL-23 monoclonal antibody (0.6mg/kg, n=13). Mouse anti-IL-23 monoclonal antibody or isotype-matched mIgG were administered intraperitoneally once every two days for 4 weeks. During the study, two mice were lost from the mIgG control group due to emphysema development. At the end of the study, mice were connected to a flexiVent for determination of lung mechanics and then euthanized with ketamine(100mg/kg) and xylazine(10mg/kg) for final data collection.

Lung Mechanics Measurement

Tracheostomy was performed on deeply anesthetized mice before connecting to a flexiVent (SCIREQ) for lung mechanics measurement. Rocuronium (1mg/kg) was used to inhibit spontaneous breathing. The lung mechanics script was run in a default ventilation setting, and the lung mechanic parameters were captured. The average values from three repeated measurements were adopted as the final values of lung mechanics measurement. The lung static compliance (Cst) and the pressure/volume (PV) curves were analyzed to assess the elasticity or resistance of the lungs.

Lung Histology and H&E Staining

Following the determination of lung mechanics, mice were euthanized, and the left lungs were intratracheally instilled with 4% formaldehyde at 25 cmH₂O pressure and then fixed in 4% formaldehyde overnight, and the right lungs were snap frozen for later analysis. After overnight fixation at 4°C, the left lung lobes were applied for paraffin embedding and generation of formalin-fixed and paraffin-embedded slides. The H&E-stained left lung was digitally scanned for histopathology review and analysis.

Lung Morphometric Assessment

ImageScope eSlide viewer scanning images of the whole left lung section at 20× amplification was used to assess the lung morphometric appearance as previously described ¹⁷. The entire left lung section was reviewed and evaluated for emphysematous change and inflammatory cell infiltration at this magnification. Emphysematous change in the lung was assessed by estimating the air space enlargement using the mean linear intercept measurement (Lm) method ^18,19. A transparent sheet with ten parallel horizontal lines of equal length and width was superimposed onto the scanned H&E-stained histologic images at 20× amplification in the ImageScope eSlide viewer. All the image fields free of bronchioles, vessels, collapsed alveoli, and obvious fibrosis within the entire scanning image were counted. On average, about 30–50 image fields were counted per animal. Only the alveolar septa that crossed the ten horizontal lines were counted as intercepts. The mean linear intercepts were calculated as the total grid length divided by the average value of the intercept count per image field per animal. A reduction in mean linear intercepts was expected to represent a measure of reduced air space enlargement.

The degree of lung inflammation was evaluated using a previously described scoring system ²⁰ to assess the lymphocytic infiltrates in the peribronchial and perivascular regions of the lung in the ImageScope eSlide viewer of the whole left lung section scanning images following the inflammation scale criteria as described by Tournoy et. al. ²⁰ and Curtis et. al. ²¹. A value from 0 to 3 per criterion was applied to each tissue section scored. With this scoring system, a value of 0 represents no inflammatory cells, a value of 1 represents occasional inflammatory cells, a value of 2 is assigned when most vessels or bronchi are surrounded by a thin layer (1 to 5 cells) of inflammatory cells, while a value of 3 is assigned when most vessels or bronchi are surrounded by a thick layer (more than 5 cells) of inflammatory cells. The inflammation scores were presented as a mean value and compared between groups.

Quantitative PCR of Mouse Lung Tissue

The total RNA of fresh lung tissue was isolated and extracted. cDNA was synthesized using the Verso enzyme kit in a thermocycler. The quantitative PCR amplification was performed in triplicates in 7500 Fast Real-Time PCR System (Applied Biosystems, USA) using SYBR GreenER^™ qPCR SuperMix kit (Invitrogen, USA) following cDNA synthesis. The expression levels of the target genes were normalized by the housekeeping gene β-actin. The primers for the mouse-specific qPCR: IL23A: 5’- CCA AGG GCT CGA GAC TTT ATT- 3’(reverse) and 5’ - GAA GGG CAA GGA CAC CAT TA-3’(forward); β-actin: 5’ - AGG AGC CAG AGC AGT AAT CT-3’ (reverse) and 5’- CCG TAA AGA CCT CTA TGC CAA C-3’ (forward).

Immunohistofluorescence Staining of Mouse Lung Tissue Sections

Mouse FFPE lung tissue slides were deparaffinized in CitriSolv, immersed in alcohol series and then distilled water for 5 minutes each. Hydrated slides were steamed for antigen retrieval in preheated 1 mM EDTA for 20 min and then transferred to 0.2% Triton-PBS solution. After a 5-min permeabilization process, the slides were blocked with 10% normal goat serum in 0.2% triton-PBS buffer for 30 min at room temperature and then incubated with primary antibodies against either S100A9 or MPO (myeloperoxidase) overnight at 4°C. After rinsing in 0.1% PBST, the slides were incubated with corresponding secondary antibodies for 1 hour at room temperature. Following sequentially washing in 0.1% PBST and air drying, the slides were mounted with DAPI for nuclei counterstain. Tissue staining for the targeted antigen was scanned, and 8-10 representative images per slide/animal were captured with an Olympus cellSens Dimension system. S100A9 or MPO positively stained cells per image field were counted using the cell counter in Image J fiji.

Cell Isolation and Culture

Human small airway epithelial cells (SAEC) purchased from ATCC (catalog# PCS-301-010) were cultured in epithelial cell growth medium (ATCC, catalog# PCS-300-030) and used for in vitro experiments between passages 5-10. Rat alveolar epithelial Type II cells (ATII) were isolated from pathogen-free, 6- to 8-wk-old female Sprague-Dawley rats as previously described by our group and others ^22–24. Freshly isolated ATII cells were grown in DMEM (Gibco, catalog# 244007) supplemented with 10% fetal bovine serum and streptomycin and penicillin liquid culture for five days to induce transdifferentiation into ATI-like cells as described by others ^25–28. In addition to SAEC, primary rat day 6 ATI-like cells were also used for the in vitro experiments in this report.

Western Blotting of Mouse Lung Tissue and Cultured Cells

To examine the role of IL-23 in CS-induced oxidative stress and apoptosis signaling pathways, the protein expression levels of Nrf2 and caspase 8 in the freshly frozen lung tissue were examined. Briefly, homogenized mouse lung tissue lysate from all groups was analyzed using electrophoresis as previously described ²⁹. Subsequently, blotting membranes were exposed by Odyssey Fc Imaging System (LI-COR Biosciences, USA), and images were quantitated using Image Studio Lite Software (version 5.2, LI-COR).

To test whether IL-23 signaling has a direct role in inducing oxidative stress of alveolar and airway epithelial cells, ATI-like epithelial cells and SAEC cells were stimulated with IL-23 (25 ng/mL) for 24 hours in vitro. Protein expression levels of Nrf2 and HO-1 were detected by western blotting, as described previously. To validate the specificity of IL-23 binding to its homogenous receptor, the cells were pretreated with a blocking IL-23 receptor antibody at the final concentration of 2 μg/mL for 2 hours before the addition of recombinant rat IL-23 for ATI-like cells or human IL-23 for SAEC.

Apoptosis Assay in IncuCyte System

ATI-like cells and SAEC were seeded at 5,000 cells/well density in a 96-well flat bottom tissue culture plate. The cells were either pretreated with rabbit anti-IL-23 receptor antibody (22 μg/mL) alone or N-acetyl cysteine (5mM) alone or anti-IL-23 receptor antibody and N-acetyl cysteine together for 2 hours and followed by mouse IL-23 (25 ng/mL) stimulation for 2 days. IncuCyte caspase 3/7 Green Apoptosis Reagent (SARTORIUS, catalog# 4440) was premixed with IL-23-containing culture medium and added to the culture wells simultaneously with IL-23 stimulation. After the cell culture plate was placed into the IncuCyte system for 30 minutes to equilibrate, nuclear green fluorescence signals were imaged every 3 hours using a ×10 objective and a standard scan type. IncuCyte captured images were then quantified by integrated IncuCyte image analysis software and validated by manually counting caspase 3/7 green fluorescence-stained cell nuclei using Image J fiji cell counter plugins. The apoptotic cells were calculated as the percentage of green-fluorescent positive cells relative to the total cell number at each time point.

Statistical Analysis

Statistical analysis using GraphPad Prism 9.0 was performed to quantitatively analyze study data. The differences between the two groups were determined with an unpaired T-test, and more than two groups’ comparisons were analyzed using one-way ANOVA with Tukey–Kramer posttest multiple comparisons. Statistical significance was accepted for P values <0.05.

Results

IL23A Gene Expression Was Upregulated in COPD Lung Tissues

RNAscope of IL23A in situ hybridization conducted on COPD (n = 6, average age 71.5 yr, 4/6 males, 5/6 smokers) derived lung tissue sections identified a significant increase in IL23A gene expression compared to the lung tissue sections obtained from control lung tissue acquired from patients without COPD ((n = 6, average age 73.7 yr, 3/6 males, 3/6 never smokers, 3/6 unknown smoking history) (P=0.008, Fig. 1A–E). Combined immunohistostaining with cell marker CD11C identified IL23A mRNA-expressing cells as CD11C-positive macrophages and dendritic cells in the perivascular and peritracheal regions. To validate this, a re-analysis of the gene expression microarray data from the publicly available Gene Expression Omnibus (GEO, no. GSE47460) for IL23A gene expression was performed. IL23A mRNA was significantly upregulated in lung biopsies from patients with COPD (n= 220) compared to biopsies from control subjects (n= 108; P<0.0001, Fig. 1F) (patients’ demographic information is summarized in the Supplemental Table). Among COPD patients, IL23A mRNA was significantly upregulated in patients with the more severe COPD, GOLD (Global Initiative for Chronic Obstructive Lung Diseases) stage 4 when compared to GOLD stage 0 (P<0.0001, Fig. 1G). These results suggest that IL23 may play a significant role in COPD pathogenesis.

Figure 1.

IL23 RNAscope and gene expression in human lung tissue biopsies. A-D: representative RNAscope images of the dark brown dots or aggregations of IL23A mRNA expression in control lung tissue (A) and COPD lung tissue (C) in the CD11c positively stained lung macrophages or dendritic cells (red cytoplasmic stain). B&D. Enlarged images from A&C. E. Semiquantification of IL23A mRNA expression (n=6, ** P =0.008). F. Increased IL23A gene expression from patients with COPD (n=220) compared to control patients (n=108) (****P <0.0001) (data are obtained from a publicly available dataset (GSE47460)). G. Increased IL23A gene expression in patients with COPD at the GOLD 4 stage (n=54) compared to the GOLD 0 stage (n=116) (****P<0.0001). Scale bar: 50 μm.

IL-23 Monoclonal Antibody Reduces the Severity of CS-induced Murine Emphysema

To determine the effect of IL-23 on CS-induced emphysema in vivo, we conducted experiments using a murine model of CS-induced emphysema in which IL-23 signaling was attenuated by treatment using a blocking antibody. After 12 weeks of CS exposure and two single doses of LPS intratracheal instillation, emphysema development assessed by lung static compliance (Cst) determination showed a significant increase in the mice that received non-immune isotype IgG treatment compared to the room air-exposed control mice (P=0.002). This increase in lung compliance was significantly attenuated in the group of mice that received IL-23 monoclonal antibody treatment compared to the IgG isotype control group (P=0.04) (Fig. 2A). As observed with lung compliance measurement, the PV curves showed an upward shift in the IgG control group, which was reversed by the IL-23 antibody treatment (Fig. 2B), again implying that the treatment with anti-IL-23 attenuated some of the physiological effects of CS and LPS in the model. Morphometric analysis using mean linear intercepts (Lm) on H&E-stained left lung tissue sections further supports the lung compliance data. As shown in Fig. 2C&E, following three months of CS exposure, the isotype IgG control antibody-treated mice developed histologic evidence of emphysema (E) and distal airspace enlargement as determined by Lm measurement of the mean free distance of the alveolar space (C) in the left lung entire tissue sections. Mice that received control non-immune antibody displayed higher Lm compared to the room air control mice (P=0.002), while mice that received IL-23 monoclonal antibody demonstrated a significant decrease in Lm when compared to the isotype control group (P=0.007) (Fig. 2C&F), suggesting that the IL23 antibody attenuates some of the phenotypic effects of CS-induced emphysema in this murine model.

Figure 2.

Lung morphometric assessment. A&B. Lung compliance measurement demonstrates a significant increase in the CS-exposed mIgG isotype controls (**P=0.002) compared to the room air-exposed sham controls, leading to an upward shift of the PV curves. IL-23 monoclonal antibody treatment significantly reduced this CS-induced increase in lung compliance (*P=0.04). C. The Lm measurement for the mean free distance of the alveolar space in the CS-exposed isotype mIgG control group is significantly higher than the room air control group (**P=0.002), whereas IL-23 antibody treatment reduced such an increase considerably (**P=0.007). D-F. Representative images of H&E-stained lung sections. Scale bar: 300 μm.

IL-23 Monoclonal Antibody Attenuates CS-induced Oxidative Stress and Suppresses Inflammation

To define mechanisms by which the IL-23 monoclonal antibody attenuates CS-induced emphysema, a number of parallel approaches were employed. To determine the degree of inflammatory cell infiltrates in the perivascular and peritracheal regions of lung tissue sections, we used the inflammation scale morphometric assessment tool. Quantitative determination of inflammation using representative H&E-stained entire left lung tissue sections demonstrated that IL-23 antibody treatment significantly decreased the inflammatory cell infiltrates (P= 0.01), while the non-immune IgG-treated control group showed a significant increase (P<0.0001) (Fig. 3A). The increase in inflammation in isotype mIgG-treated mice was accompanied by a corresponding increase in lung tissue IL23A mRNA expression (P=0.01) and decreased IL23A mRNA expression in IL-23 antibody-treated mice (P=0.04). This reduced IL23A mRNA expression was possibly a result of the negative feedback of the IL-23 signaling pathway blockade by anti-IL-23 monoclonal antibody treatment. These data confirmed and validated the effect of IL-23 antibody treatment (Fig. 3B).

Figure 3.

Mouse lung tissue analysis. A. Inflammation scoring of the lymphocytic infiltrates in the peribronchial and perivascular regions of the lung shows a marked increase in the CS-exposed mIgG isotype control lungs (****P<0.0001); in contrast, IL-23 antibody treatment significantly reduced the number and extent of this CS-induced lymphocytic aggregates (*P=0.01). B. IL23A mRNA expression in the lung tissue determined by qPCR shows a marked increase in CS-exposed isotype mIgG control mice (*P=0.01), while this was reduced significantly by IL-23 monoclonal antibody treatment (*P =0.04). C-E. Western blot of fresh frozen lung tissue shows increased expression of Nrf2 (*P=0.01) and caspase 8 (**P=0.001) in CS-exposed mIgG control mice (n=10), and this was significantly reduced by IL-23 monoclonal antibody treatment (n=13) (*P =0.02 for Nrf2 and *P =0.03 for caspase 8). C. A western blot of lung tissue lysate shows the relative expression of Nrf2 and caspase 8 of each individual mouse from all three groups. Although within-group variations exist, the overall densitometry analysis of the western blot bands in each group differs significantly, as shown in D&E. D&E. Semiquantitative analysis of Nrf2 (D) and caspase 8 (E) expression in mIgG control group versus IL-23 antibody-treated group over room air-exposed sham control group (n=6).

To determine whether blocking IL-23 in vivo attenuates CS-induced oxidative stress and apoptosis, the key oxidative stress marker nuclear factor erythroid 2 (Nrf2) and apoptosis signaling pathway marker caspase 8 were evaluated by immunoblotting analysis of the fresh-frozen whole lung tissue lysate. As shown in Fig. 3C–E, IL-23 antibody treatment reduced the Nrf2 (P=0.02) and caspase 8 (P=0.03) expressions in the lung tissue lysate when compared to the isotype antibody controls, whereas the CS exposure samples showed significantly increased expression of both pathway markers when compared to the room air-exposed controls (P=0.01 for Nrf2 and P=0.001 for caspase 8).

Lastly, to define the cellular component of the inflammatory cell infiltrates, we performed immunohistofluorescence staining of the lung tissue sections using selected cell-specific markers, including CD3/CD20 for T/B lymphocytes, S100A9 for lung macrophages and MPO for neutrophils. Immunostaining revealed that the increased inflammatory cell aggregates in the lungs of this CS and LPS exposure model are CD3+/CD20+ positive T/B lymphocytes (Fig. 4A–C). Further immunostaining using antibodies directed to the macrophage marker S100A9 and neutrophil marker MPO revealed markedly increased alveolar and interstitial macrophages (P=0.04) and neutrophils (P=0.01) in the CS exposed mIgG isotype control mice compared to the room air negative control mice (Fig. 4D&E, G&H, and J&M). IL-23 monoclonal antibody treatment significantly reduced the number of S100A9 expressing macrophages (P=0.04) and MPO expressing neutrophils (P=0.001) infiltrating to the CS exposed lungs (Fig. 4F&I, J&M). Together, these results show that blocking IL-23 in vivo suppresses lung oxidative stress and inflammation in this murine COPD model.

Figure 4.

Mouse lung tissue immunohistofluorescence staining. A-C. Immunostaining against CD3 and CD20 identified CD3+ T (A) and CD20+ B (B) lymphocytes of the lymphocytic aggregates (C) infiltrated in the CS-exposed lung tissue. D-F&J. S100A9 positively stained lung macrophages were increased in CS-exposed mIgG control lungs compared to the room air control mice (*P =0.04) and decreased in the IL-23 antibody-treated mice compared to the mIgG controls (*P =0.04). G-I&M. MPO positively stained neutrophils in the lung were increased in CS-exposed mIgG control lungs compared to the room air control mice (*P =0.01) and decreased in the IL-23 antibody-treated mice compared to the mIgG controls (**P=0.001). K&L. Enlarged representative images show S100A9 and MPO staining patterns. Scale bar in A–C: 20 μm, scale bar in D-I, K&L: 50μm.

IL-23A Induces Oxidative Stress and Apoptosis of Transdifferentiated ATI-like Cells and Small Airway Epithelial Cells

In addition to regulating the inflammatory response, we sought to determine whether IL-23 directly impacts epithelial cells. We first determined whether primary alveolar epithelial type I (ATI) and type II (ATII) cells, cultured transdifferentiated ATI-like cells, as well as SAEC express the IL-23 receptor and identified that it is universally present in all epithelial cells examined (Fig. 5). We then determined the cell response to recombinant IL-23A stimulation (24 hours) in the cell culture environment for oxidative stress and apoptosis signaling pathways. Direct stimulation of primary ATI and ATII cells by IL-23 did not alter the expression of oxidative stress signaling or cell apoptosis markers (data not shown). However, following a 24-hour stimulation with IL-23A, we identified increased Nrf2 (P=0.02 for ATI-like cells and P=0.009 for SAEC) and heme oxygenase-1 (HO-1) (P=0.02 for ATI-like cells and P=0.003 for SAEC) expression in the cultured transdifferentiated ATI-like cells and SAEC, as a compensatory mechanism of the cell-protective response ³⁰(Fig. 5A–F). Preincubation with 2 μg/ml IL-23 receptor antibody (2 hours) protected ATI-like cells from IL-23-induced oxidative stress demonstrated by the reduced expression of HO-1 (P=0.002) and Nrf2 (P=0.02). Preincubation of SAEC with 2 μg/ml IL-23 receptor antibody also reduced Nrf2 (P=0.008) and HO-1(P=0.03) expression.

Figure 5.

Examination of IL-23 induced primary epithelial cell responses. A-F. Western blot of ATI-like cells and SAEC in response to IL-23 stimulation with or without IL-23 receptor blocking antibody pretreatment. The result shows both cells have IL-23 receptor expression and response to IL-23 stimulation. 24 hours of IL-23 stimulation induced upregulation of oxidative stress marker Nrf2 (*P=0.02 for ATI-like cells and **P=0.009 for SAEC) and compensatory upregulation of HO-1 (*P=0.02 for ATI-like cells and **P=0.003 for SAEC) in both cells. Blocking the IL-23 receptor by a specific antibody significantly reversed this IL-23-stimulated increase in the expression of Nrf2 (*P=0.02 for ATI-like cells and **P=0.008 for SAEC) and HO-1 (**P=0.002 for ATI-like cells and *P=0.03 for SAEC). G-J. Dynamic detection of apoptotic cells labeled with caspase3/7 green fluorescence IncuCyte apoptosis reagent using IncuCyte cell culture and imaging system. The result shows increased apoptotic ATI-like cells (**P=0.001) and SAEC (**P=0.002) after IL-23 stimulation and decreased apoptotic cells after the IL-23 receptor was blocked by IL-23 receptor antibody pretreatment before IL-23 simulation (**P=0.003 for ATI-like cells and *P=0.02 for SAEC). Cells pretreated with antioxidant N-acetyl cysteine did not affect IL-23-induced cell apoptosis.

Next, we determined whether IL-23 has a direct role in inducing cell apoptosis. Using the IncuCyte cell culture and imaging system, we identified a significant increase in the fluorescence caspase3/7 positively labeled apoptotic cells over 48 hours of IL-23 stimulation in both ATI-like cells (P=0.001) and SAEC (P=0.002). Preincubation with IL-23 receptor antibody protected both ATI-like cells (P=0.003) and SAEC (P=0.02) from IL-23-induced apoptotic cell death (Fig. 5G&H). To determine whether IL-23-induced cell apoptosis is oxidative stress injury-dependent, cells were preincubated with the antioxidant N-acetyl cysteine (NAC) before IL-23 stimulation. The lack of suppression of IL-23-induced cell apoptosis by NAC suggests that this apoptotic response may be due to an oxidative stress-independent mechanism (Fig. 5I&J).

Discussion

The increased IL23 gene expression in lung tissue biopsies of patients with COPD and the effect of an IL-23 blocking antibody on immune cell infiltration in a model of CS-induced emphysema suggest important roles for IL-23 in the pathogenesis of COPD. In this study, we identified an increased IL23 gene expression in the lung biopsies of COPD patients and showed that the levels of IL23 expression correlate with the severity of COPD. Using a murine model of COPD, we also report that suppressing IL-23 in vivo through the administration of a monoclonal antibody reduced emphysema severity. Suppression of IL-23 in this model led to reduced inflammatory cell infiltration, reduced oxidative stress injury, and apoptosis, suggesting a role for IL-23 as a key mediator of the immunoinflammatory process and oxidative stress/apoptosis signaling pathways in the pathogenesis of COPD.

An increased serum level of IL-23 has been recently reported to positively correlate with the severity of clinical COPD as determined by the GOLD stage ³¹. This evidence that IL-23 is increased in human COPD and correlates with clinical disease severity is consistent with our data generated using COPD-derived lung tissue biopsies and the data analyzed from the publicly available microarray data set, showing a similar positive correlation of IL23 gene expression with the disease GOLD stage. The current study sheds additional insight into the potential mechanisms by which IL-23 can mediate disease progression in COPD through the use of an LPS and CS inhalation model. Neutrophil and macrophage infiltration and lymphoid follicle development in the perivascular and peritracheal regions of the lung are characteristic features of the chronic inflammation that occurs in COPD ^{3–5,12,32–36}. Macrophages play a central role in orchestrating the inflammatory response in both murine models and human COPD, and markedly increased macrophage numbers are found in the airway, lung parenchyma, bronchoalveolar lavage (BAL) fluid, and sputum of COPD patients ^37,38. In keeping with prior reports, we have demonstrated increased S100A9-expressing macrophages in the lung parenchyma of CS-exposed IgG-treated placebo control mice by immunostaining lung tissue sections. Following exposure to CS or biomass smoke, macrophages and epithelial cells in the lung are activated, releasing chemotactic factors that attract neutrophils and monocytes to the lung, with the latter differentiating into macrophages ^39–41. As illustrated in the schematic working model (Figure 6), macrophages and dendritic cells release IL-23 that polarizes immunity by promoting downstream IL-17-mediated neutrophilic inflammation, which is initiated by neutrophil infiltration, resulting in the release of proteases, such as MMP9, reactive oxygen species, and elastolytic enzymes, causing elastin degradation and emphysema ^42,43. In the current study, we provide evidence that IL-23 blocking antibody reduces CS exposure-induced macrophage and neutrophil recruitment and lymphocytic infiltrates in the murine lung parenchyma, similar to what was reported by others in the nervous system ⁴⁴, providing additional support that IL-23 regulates key inflammatory processes relevant to COPD pathogenesis. In addition, the elevated IL23 gene expression in the CS-exposed IgG placebo control lungs and the significantly reduced IL23 expression after monoclonal antibody treatment strongly supports a mechanism of IL-23-mediated inflammatory signaling pathway involved pathogenesis of the disease in our model.

Figure 6.

Schematic illustrating how IL-23 may mediate lung injury in the CS&LPS-induced COPD model. CS&LPS activate lung macrophages and dendritic cells, leading to the release of IL-23 and other pro-inflammatory cytokines and factors, such as proteases. IL-23 induces epithelial cell oxidative stress and apoptosis. Through direct and secondary effects by activating epithelial cells, downstream proinflammatory cytokines and chemokines are produced that facilitate the recruitment of CD3+ or CD20+ lymphocytes, S100A9+ macrophages, and MPO+ neutrophils to the lungs. These inflammatory cell infiltrates in the lung further release pro-inflammatory cytokines and proteolytic enzymes, leading to a vicious cycle of inflammatory injury, apoptosis, and proteolysis of the alveolar cells and extracellular matrix proteins. Therapy with the blocking IL-23 monoclonal antibody administration diminishes this inflammatory cascade, reducing inflammation and emphysematous tissue injury.

Oxidative stress injury and apoptosis are other characteristic features of COPD pathogenesis ^1,2. Excessive generation of reactive oxygen species occurs following CS exposure and results in the dysregulation of redox homeostasis, leading to oxidative stress, oxidative cellular injury, and apoptosis of structural cells, especially alveolar epithelial cells ^45–47. Whether IL-23 can directly induce cellular oxidative stress and apoptosis has not been well defined. Using a monoclonal antibody in a murine model of COPD, we demonstrate that suppressing IL-23 reduced oxidative stress and cell apoptosis-related signaling pathways, suggesting the role of IL-23 as a mediator of oxidative stress-induced cell injury in the murine COPD model. Furthermore, our in vitro studies identified the presence of IL-23 receptors in alveolar and bronchial epithelial cells. Using recombinant IL-23A and monoclonal antibodies, the current study supports a role for IL-23 as a direct inducer of epithelial oxidative stress and apoptosis in vitro.

Although the current study provides compelling new evidence in support of a direct role for IL-23 as a regulator of CS-induced emphysema by promoting inflammatory cell infiltration, oxidative stress injury, and apoptosis, the exact molecular mechanisms by which IL-23 mediates these events at the cellular level are not fully elucidated and warrant further investigation. There are several limitations in our study. First, the emphysema disease model used in our study is not a traditional, prolonged 6-month-long CS-exposed emphysema model. Rather than using the “standard” C57BL6J mice, we selected A/J mice with a genetic background that has increased susceptibility to CS exposure and emphysema development ^15,16 and combined CS exposure and LPS intratracheal instillation to amplify CS effects and accelerate COPD development. The use of this model may potentially limit the interpretation and direct comparison with a traditional disease model consisting solely of CS exposure. However, others have reported the use of the LPS and CS combined exposure to induce murine COPD ⁴⁸, which has potential value as a preclinical disease model for COPD that incorporates different stimuli and exposure routes ⁴⁹. An additional potential limitation is using intraperitoneal (IP) injection to deliver IL-23 antibody instead of a more direct administration route such as intravenous infusion or subcutaneous injection. While technically more feasible than intravenous or subcutaneous delivery for mice, IP injection makes the evaluation of the effective dosages, duration, and local lung tissue antibody concentration estimation challenging. Nonetheless, results generated from our study provide preclinical evidence supporting a therapeutic potential for targeting IL-23 as an upstream regulator of lung inflammation in COPD.

Conclusions

The inhibition of IL-23 in a murine model of emphysema attenuates lung oxidative stress, inflammation, cellular apoptosis, and emphysema severity, suggesting a role for IL-23 as an essential immune mediator of the inflammatory processes in the pathogenesis of CS-induced emphysema.

Data Availability

The data will be available from the corresponding author upon reasonable request.