Background
Chronic obstructive pulmonary disease (COPD) is a chronic lung disease commonly associated with inhalation of cigarette smoke (CS). While the pathology of COPD is generally considered to be destructive in nature, epithelial remodeling and sub-epithelial fibrosis of the small airways is now recognized as a key histopathological feature of the disease [
1,
2]. While CS directly activates and damages the epithelium, viral infections, which occur frequently in smokers with and without COPD also influences epithelial phenotype and function [
3]. Moreover, viral-induced exacerbations contribute significantly to disease progression, accelerated decline of lung function and disease morbidity and mortality [
3]. Unfortunately, the mechanisms that drive changes to the epithelium in COPD following exposure to CSE and respiratory viruses remains largely unexplored.
What is known, suggests that damage to the epithelium triggers a temporal cascade of inflammatory and cell signaling events that under normal circumstances leads to inflammation, resolution of inflammation and repair. However, under conditions of chronic inflammation, the epithelium fails to repair effectively with structural and functional changes including goblet cell hyperplasia, squamous cell metaplasia, induction of pro-inflammatory chemokines [
4], matrix metalloproteinases [
5] and epithelial cell apoptosis and proliferation [
6]. The mechanisms by which these changes occur are unclear, although endoplasmic reticulum (ER) stress and oxidative damage have been proposed as initiators of epithelial cell apoptosis in COPD [
7‐
11].
One aberrant epithelial response that has been associated with abnormal repair and fibrosis is epithelial to mesenchymal transition (EMT). This phenomenon, commonly observed in cancer and a component of normal organ development, occurs when epithelial cells transform into highly motile mesenchymal cells. EMT has been shown to contribute to airway disease and fibrosis in several organs, including the lung [
12,
13]. However, the underlying mechanisms and the functional consequences of EMT in the airways from COPD patients remain unclear [
14].
In this study we investigated the epithelial phenotype in COPD and responses to CS extract (CSE) on normal and COPD bronchial epithelial cell function in vitro. We assessed the impact of CSE on apoptosis, proliferation and chemokine production. We show that in the short term, cigarette smoke does not modulate EMT, but rather induces epithelial cell proliferation, apoptosis and CXCL8/IL8 production. Exposing mice to CSE combined with poly I:C challenge revealed that CSE induces apoptosis and inflammation in the lung, but a secondary stimulus of poly I:C amplifies fibrotic changes. These data suggest that CS alone is insufficient to induce the chronic airway remodeling that is observed in COPD and that an extra insult, such as viral infection is required.
Methods
Human COPD cohort
This study was approved by the Research Ethics Board of the University of British Columbia, Canada. Human lung tissue from patients with COPD (
n = 37) and controls (
n = 9) was obtained from patients undergoing surgery and who gave informed consent. All COPD patients were GOLD II/III, the majority were male and ex-smokers undergoing surgery for lung cancer. Control lung tissue was from otherwise healthy, non-smoking subjects whose lungs were not used for transplant from the IIAM as previously described (Hackett et al., [
12]). We attempted to age match as best as possible, but the normal donors were significantly younger than the COPD cohort.
Epithelial cell culture
Normal human bronchial epithelial (NHBE, donor 105116, donor 7 F4334 or 84749) cells and COPD human bronchial epithelial cells (COPD-AEC donor OF3207, Lonza, UK) were maintained in BEGM™ (bronchial epithelial cell growth medium) plus BulletKit™ (Lonza). Both cell types were cultured at 37 °C in the presence of 5% CO
2. Primary bronchial epithelial cells were also isolated from patients and cultured as previously described [
12]. In all in vitro epithelial cell experiments, control and COPD epithelial cells were used before passage 4. Cells were seeded for each experiment so they would be between 70 and 80% confluent at the start of each experiment. In addition, cells were deprived of bovine pituitary extract (BPE), hydrocortisone, human epidermal growth factor (EGF), epinephrine, transferrin and triiodothyronine for 16 h prior to stimulations. Insulin, retinoic acid and GA-1000 remained throughout the experiment.
Induction of EMT
Monolayer cultures were grown to 60–70% confluence on 6-well tissue culture plates (BD Falcon, NJ), before being quiesced and exposed to TGFβ1 (10 ng/ml) for 48 h. Samples were then assayed for mesenchymal (EDA-Fn) and epithelial (E-Cadherin) markers using quantitative polymerase chain reaction and immunoblot.
Cigarette smoke extract (CSE) 100% stock was generated from mainstream cigarette smoke by the combustion of 1 1R3F reference cigarette (Tobacco Health Research, University of Kentucky, Kentucky, USA) with the filter removed, using a peristaltic pump, through 25 mls of culture medium without FBS. The CSE was then sterile filtered through a 0.22 μM filter, and added to cell cultures at the required concentration.
Immunohistochemical staining of human airway sections
Tissue sections were deparaffinized and rehydrated, and put through antigen retrieval by autoclaving (15 min, 120 C, 30 psi) for 20 min in citrate target retrieval solution (Dako, Mississauga, ON, Canada). Subsequently, endogenous peroxidase was quenched with 3% H2O2 and non-specific interactions blocked for 30 min with 10% goat serum. Antibodies directed against α-smooth muscle actin (ab5694, Abcam, Cambridge, MA), and vimentin (AF2105, R&D Systems) were added overnight at 4 °C in 25% goat serum. Sections were then incubated with either biotinylated goat anti-mouse or goat anti-rabbit secondary antibody (1:100, Vector Labs, Burlingame, CA) for 60 min followed by a 10-min treatment with Streptavidin-HRP (Dako). The antigen of interest was visualized by 3,3-diaminobenzidine (Dako) and counterstained with Harris Hematoxylin Solution (Sigma, Oakville, ON, Canada). Sections were then dehydrated and mounted with Cytoseal 60 (Richard-Allan Scientific, Kalamazoo, MI). Antibody dilutions and all washes were in TRIS-buffered saline solution.
Gene expression analyses
For gene expression analysis from lung tissue, total RNA was obtained using TRIzol reagents (Invitrogen, Paisley, UK) according to manufacturer’s instructions. RNA was reverse transcribed into cDNA using the high capacity cDNA kit (Applied Biosystems) and gene expression levels were determined using qRT-PCR using an ABI Prism 7900HT sequence detector (Applied Biosystems). Transcript levels of genes of interest were normalised to 18s and fold change was to matched animal controls or non-COPD lung using 2-ΔΔCt.
Nu-PAGE and Western blotting
NHBE cells were seeded at 12 × 104 cells/ml and cultured for 24 h in complete medium. Prior to stimulation cells were cultured overnight in basal medium supplemented with only retinoic acid, insulin and antibiotics. TGFβ1 (10 ng/ml, R&D Systems, Abingdon, UK), CSE (5% or 10%), or both combined were added to the cells (media alone was used as a control) and left to incubate for a further 48 h. Cells were scraped and lysed in RIPA buffer (Sigma-Aldrich, Poole, UK) containing protease inhibitors (Roche Diagnostics, East Sussex, UK). Cell suspensions were cleared by centrifugation at 8000 x g for 10 min at 4 °C. 15 μg of protein was loaded onto a Bis-Tris (4–12%) Nu-PAGE gel (Invitrogen, Paisley UK) followed by electrophoresis and then transferred to PVDF membranes using the iBlot dry blotting system (Invitrogen, Paisley, UK). Membranes were blocked in Odyssey blocking buffer (LI-COR) for 2 h and then probed with E-cadherin (clone 67A4 mIgG1, Santa Cruz, Heidelberg, Germany), fibronectin (clone IST-9 mIgG1, Santa Cruz, Heidelberg, Germany) or beta tubulin loading control (LI-COR) diluted in Odyssey blocking buffer (LI-COR) supplemented with 0.02% Tween-20 overnight at 4 °C. Membranes were washed 4 times in PBS-T for 5 min each and were then probed with anti-mouse or anti-rabbit IRDye 800cw conjugated secondary antibody (LI-COR) in Odyssey blocking buffer supplemented with 0.02% Tween-20. Membranes were washed as before and left to dry before scanning on the Odyssey Clx Infrared Imaging System (LI-COR). Western blots were analysed by densitometry using loading control for normalisation and the fold change in expression was determined compared to untreated controls. Experiments were performed on 3 NHBE donors.
CXCL8 ELISA
CXCL8 was measured from the supernatants of epithelial cells using a duoset ELISA (R&D Systems) according to manufacturer’s guidelines. Experiments were done three times using triplicate wells each time.
Apoptosis assay
NHBE and COPD-AEC were plated at 1x105cells/ml and cultured for 24 h to reach around 80–90% confluency. BEGM media with 5, 10 or 20% CSE was added to cells for 6 and 24 h. Caspase-3/-7 activity was determined from treated epithelial cells using the Caspase-Glo® -3/-7 Assay (Promega) according to manufacturer’s guidelines. Experiments were done three times using triplicate wells each time.
Lactate Dehydrogenase (LDH) assay
To determine the levels of released LDH in cell-free supernatants, the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega) was used to measure levels in the supernatants of epithelial cells according to manufacturer’s guidelines. Measurements were performed on three individual experiments using triplicate wells each time.
Proliferation assay
NHBE and COPD-AEC cells were plated at 1x105cells/ml and cultured to 80–90% confluency. At this point, cells were exposed to CSE media at a concentration of 5, 10 and 20% in low serum conditions. Cells were stimulated with media alone, CSE media, full serum complete media or media with 30 ng/ml of PDGF (R&D systems, UK) as a positive control for proliferation for 72 h. At this point cells were pulsed with 30 μl/ml of [3H] thymidine for 6 h and proliferation measured in a topcount liquid scintillation counter (Perkin-Elmer). Results were expressed as counts per minute. Experiments were done three times using triplicate wells each time.
Mice
Six-week-old C57Bl/6 mice were purchased from Taconic Europe (Denmark). Experiments were conducted in accordance with standards established by the Council of Europe, AstraZeneca Global R&D Standards for animal care, European Union Directive 86/609, and Swedish legislation. The studies were approved by the regional Ethics Board, Malmö/Lund, Sweden (reference number M230/09).
Animal exposure to cigarette smoke and poly I:C
The current protocol was adapted from a previously described model by Kang et al. [
16]. Briefly, mice were exposed to cigarette smoke using a whole-body smoke exposure system (SIU-48, Promech Lab AB, Vintrie, Sweden). Mice were exposed to 10 1R3F reference cigarettes with filters removed for a period of approximately 50 min, twice daily, for 18 consecutive days, control animals (i.e., sham-exposed animals) were exposed to room air only. Prior to exposure mice were acclimatized over a 3 day period. On days 8, 11, 15 and 18 of cigarette smoke exposure, mice were lightly anesthetized and 50 μl aliquots of poly (I:C) (50 μg per animal, Sigma Aldrich, Sweden) or its vehicle control were administered via nasal aspiration. Animals were euthanized at day 19 of the protocol and outcome measures assessed.
Evaluation of airway inflammation and remodelling in mice
Tissue samples and bronchoalveolar lavage (BAL) fluid were collected for assessment. For histologic evaluations, the left lung was inflated to 25 cm with 4% buffered formaldehyde (Histolab Products AB, Sweden). Images of lung sections were captured at x20 final magnification on a microscope using a Scanscope (Aperio, CA, USA). In addition, RNA was extracted from tissue for gene expression analysis (see previous section). Total cellularity was determined from BAL fluid and specific blood cell counts were determined for macrophages, neutrophils, lymphocytes and eosinophils.
Immunohistochemical staining
Immunohistochemical staining for Hematoxylin and eosin (H&E), proliferating cell nuclear antigen (PCNA); Novocastra Laboratories Ltd., Newcastle upon Tyne, UK,) and caspase-3 (rabbit anti-human/mouse caspase-3 active, AF835; R&D Systems, UK) were performed. Antibodies were detected with EnVision™ kit (DAKO UK Ltd, Cambridgeshire) and DAB as chromogen, according to the manufacturer’s instructions.
Statistical analysis
Normal distribution was assumed. Data were expressed as means ± SEM, and assessed for significance by Student's t test or ANOVA as appropriate.
Discussion
Given its prime location, it is perhaps not unexpected that the epithelium of people who smoke, and particularly those with COPD would be abnormal. Indeed, chronic inflammation and remodeling of the mucosa are typical features of COPD. We were interested in looking at the effects of CSE on the epithelium and found that using primary epithelial cells from patients with COPD and healthy controls: (1) COPD-AEC have increased basal levels of apoptosis and that exposure to CSE increases caspase activity in healthy and COPD-AEC; (2) CSE increases proliferation and IL-8 release in a concentration-dependent manner although intriguingly, IL-8 release by COPD-AEC was significantly lower than normal-AEC. (3) We then used a mouse model of CSE and poly I:C to mimic viral exacerbations and show that while poly I:C clearly induces lung inflammation in its own right and enhances CSE-induced inflammation, it has no effect on caspase-3/-7 activity either alone or in combination with CSE; (4) finally we show that short-term CSE challenge has no effect on markers of EMT either in vitro or in vivo. Our data suggests that cigarette smoke induces a marked apoptotic burden on epithelial cells, but at the concentrations and times used does not induce EMT either in vitro or in vivo.
One of the strengths of our study was the use of primary AEC from COPD patients where we show for the first time that basal apoptotic load is significantly higher in cultured AEC obtained from these patients than AEC obtained from healthy donors. While caspase activity increased in AEC from normal donors after CSE exposure, we saw no statistically significant effect in COPD-AEC. However, basal caspase activity in COPD-AEC was the same as the maximal CSE-induced activity in normal-AEC, suggesting activity of caspase -3/7 is constitutively high in COPD-AEC. In contrast, Chiappara and colleagues [
17] showed that epithelial caspase-3 expression was similar in patients with COPD and normal controls and that CSE did not influence apoptosis in a transformed human airway epithelial cell line (16HBE). We hypothesize that since Chiappara et al., examined expression but not activity of caspase-3 and used a transformed cell line which are generally resistant to apoptosis are possible reasons for these disparate findings. We also investigated caspase-3 expression as well as caspase-3/7 activity in a mouse model of CSE and poly I:C to mimic viral exacerbations. We show that poly I:C clearly induces lung inflammation and enhances CSE-induced inflammation, that it increases caspase-3 expression in areas of alveolar inflammation but has no effect on lung tissue caspase-3/7 activity either alone or in combination with CSE. A similar model developed by Kang
et al [
16] demonstrated that the combination of CSE and poly I:C were associated with the induction of type I interferon (IFN) and IL-18, followed by the induction of IL-12/IL-23 p40 and IFN-γ, and the activation of PKR (double-stranded RNA-dependent protein kinase) and eIF-2α (eukaryotic initiation factor-2α). These authors also demonstrated that CSE enhanced the effects of influenza, but no other agonists of innate immunity, suggesting that CSE selectively augments the airway and alveolar inflammatory and remodelling responses induced in the murine lung by poly I:C and viruses [
16]. Furthermore, Chiappara et al also showed that expression of two different proliferation markers, PCNA and Ki67 were differentially expressed in epithelium in COPD sections; expression of PCNA was virtually absent in COPD and smoking groups, whereas Ki67 expression was significantly higher in the COPD/smoking groups compared to control. Proliferation was not quantified in vitro. Our data suggest that CSE induces a concentration-dependent proliferative response in vitro, which was maximal at 10% and then sharply declining, presumably due to toxicity. CSE also promoted lung epithelial cell proliferation in vivo as observed by PCNA staining. These results are supported by previous reports that showed smoking generates a dose-dependent pro-proliferative response in epithelial cells of smokers [
18].
Release of CXCL8 from normal-AEC dose dependently increased with CSE exposure. Kode
et al showed similar effects in primary cultures of small airway epithelial cells [
19]. In agreement with previous studies [
20], basal CXCL8 release was similar in normal and COPD-AEC. These findings differ from those of Schneider et al.
, who observed increased basal CXCL8 production in epithelial cells from COPD patients [
21]. These disparate findings may be due the different culture procedures, as Schneider et al generated differentiated air-liquid interface cultures, whereas both Heijink and colleagues and this study used submerged monolayer cultures. Likewise, disease severity may play a role. Although GOLD status was not reported, 8 of the COPD patients included in the study by Schneider had severe emphysema. Nevertheless, Schulz et al [
22] were also not able to detect differences in baseline CXCL8 production in submerged cultured PBEC from currently smoking COPD patients and smoking controls.
Epithelial-mesenchymal transition (EMT) has been categorized into three different types: Type I representing a natural process such as occurs during embryogenesis; Type II – leading to organ fibrosis and Type III, pro-angiogenesis and associated with cancer. In the context of COPD, Sohal and colleagues documented Type III EMT in the epithelium of large airways from patients with COPD [
23] and Nishioka et al., showed that epithelial cells from COPD patients have display a partial EMT phenotype under basal conditions [
24‐
30]. Recent findings indicate that similar changes can also be observed in the small (<1 mm) airways of smokers. However, despite CSE-induced increases in expression of TGFβ1, GSK3β and col1α1, we saw no evidence of EMT, since expression of αSMA and E-cadherin was not changed. Given this gene expression signature was derived from whole lung which may have masked epithelial specific responses, we performed similar experiments using primary cultures of human airway epithelial cells. We show that following exposure to TGFβ1, epithelial cells increase expression of mesenchymal marker EDA-Fn, while down regulating expression of the epithelial marker E-cadherin as expected. Exposure to CSE had no effect on these markers when used alone and did not modify the response to TGFβ1. Recent studies have shown that cigarette smoke condensate (CSC) induces an EMT-like process in BEAS-2B cells [
31]. The reasons for these different findings may relate to the type of cell cultures or duration of exposure to CSE, since Veljkovic and colleagues exposed their cells for 30 days, whereas we used 48 h.
Acknowledgments
The authors would like to thank Tom Kisby for technical assistance.