Increased ERK signalling promotes inflammatory signalling in primary airway epithelial cells expressing Z α1-antitrypsin

Overexpression of Z α1-antitrypsin is known to induce polymer formation, prime the cells for endoplasmic reticulum stress and initiate nuclear factor kappa B (NF-κB) signalling. However, whether endogenous expression in primary bronchial epithelial cells has similar consequences remains unclear. Moreover, the mechanism of NF-κB activation has not yet been elucidated. Here, we report excessive NF-κB signalling in resting primary bronchial epithelial cells from ZZ patients compared with wild-type (MM) controls, and this appears to be mediated by mitogen-activated protein/extracellular signal-regulated kinase, EGF receptor and ADAM17 activity. Moreover, we show that rather than being a response to protein polymers, NF-κB signalling in airway-derived cells represents a loss of anti-inflammatory signalling by M α1-antitrypsin. Treatment of ZZ primary bronchial epithelial cells with purified plasma M α1-antitrypsin attenuates this inflammatory response, opening up new therapeutic options to modulate airway inflammation in the lung.


INTRODUCTION
Alpha 1 -antitrypsin is a 52-kDa serine protease inhibitor (or serpin), primarily produced by hepatocytes, but also secreted locally by lung epithelial cells and alveolar macrophages (1,2). Its known function is to inhibit a number of serine proteases, including neutrophil elastase and proteinase 3, thereby preventing excessive degradation of the extracellular matrix. It has also been reported to exhibit anti-inflammatory properties, including the inhibition of tumor necrosis factora (TNFa) gene expression (3), inhibition of a disintegrin and metalloprotease (ADAM)17 activity in neutrophils and endothelial cells (4,5) and the regulation of CD14 expression and cytokine release in monocytes (6,7).
The Z mutation (E342K) of a 1 -antitrypsin causes subtle misfolding of the protein that permits polymer formation and accumulation within the endoplasmic reticulum (ER) of hepatocytes or degradation by the proteasome leading to deficiency of the secreted protein (8,9). This causes hepatic cirrhosis through toxic gain-of-function within the liver, most likely due to the retention of polymers, and early-onset lung emphysema, due in large part to the loss of protease inhibition (10). The discovery of polymers in broncho-alveolar lavage fluid and pulmonary tissue (11,12), the pro-inflammatory nature of such extracellular polymers (11,13) and their identification many years after liver transplantation (14) led to the proposal that pulmonary pathology could be induced by polymer-induced toxic gain-of-function with inflammation as an additional mechanism (15).
Secreted proteins are first folded within the ER where quality control systems ensure that only properly folded proteins exit the organelle (16). Accumulation of unfolded or misfolded proteins within the ER induces 'ER stress', thereby activating intracellular signal transduction pathways, collectively called the unfolded protein response (UPR) (16). This complex cellular response evolved to restore ER homeostasis by reducing the load of newly synthesized protein while increasing the complement of molecular chaperones, which enhance ER protein-folding capacity, and increasing the efficiency of misfolded protein degradation (Endoplasmic reticulum-associated degradation, ERAD) (17,18). We have shown previously that mutant Z a 1--antitrypsin is degraded predominantly by ERAD (19). Remarkably, the accumulation of polymers of Z a 1 -antitrypsin within the ER of hepatocytes does not activate the UPR but instead increases the cell's sensitivity to ER-stress upon a 'second hit' owing to impaired ER luminal protein mobility (20)(21)(22).
The transcription factor nuclear factor kappa B (NF-kB) regulates many genes involved in inflammation and cell death, including numerous cytokines and chemokines, e.g. interleukin (IL)-8 (23). Phosphorylation of NF-kB is classically mediated through the phosphorylation of inhibitor kappa-B alpha (IkBa); however, NF-kB can also be activated via mitogenactivated protein kinase (MAPK) signalling cascades (24,25).
Epidermal growth factor (EGF) and related mitogens such as heparin-binding EGF (HB-EGF), amphiregulin (AREG) and transforming growth factor (TGF)-a are synthesized as membrane-bound proteins that upon cleavage by metalloproteases (MPs), including ADAMs, bind to and activate the EGF receptor (EGFR) [reviewed in (26)]. Transactivation of the EGFR can also occur via activation of ADAMs by G-protein-coupled receptor signalling [reviewed in (27)]. Within the lung, EGFR activation can induce epithelial cell proliferation by activating extracellular signal-regulated kinases 1 and 2 (ERK1/2). This is mediated by Ras activation of c-Raf, causing phosphorylation of the mitogen-activated protein/extracellular signal-regulated kinase (MEK), which in turn phosphorylates ERK1/2 (28).
Mutants of members of the serpin superfamily, including a 1--antitrypsin, have been shown to activate NF-kB signalling, postulated to be a response to the formation of protein polymers within the ER (20,21,29). However, this appears to be independent of their ability to prime cells for ER stress (29). Whether the local expression of Z a 1 -antitrypsin by airway epithelial cells in vivo leads to the formation of protein polymers and to the activation of the NF-kB pathway remains unclear. We report here the detection of NF-kB activation in primary bronchial epithelial cells isolated from patients homozygous for the Z mutation (ZZ) and demonstrate this to be mediated by increased ADAM17-dependent EGFR-MEK-ERK signalling in the absence of either detectable polymer formation or ER stress response. Instead, the activation of the EGFR in this setting represents a loss of M a 1 -antitrypsin phenotype.

Z a 1 -antitrypsin activates NF-kB in lung epithelial cells
It is well-recognized that overexpression of Z a 1 -antitrypsin activates the NF-kB response leading to pro-inflammatory cytokine release (20,21,30). We therefore asked whether expression of Z a 1 -antitrypsin regulated by its endogenous promoter in airway epithelial cells could also activate this pathway. Primary bronchial epithelial cells were differentiated into mucin-producing, ciliated epithelial cell layers (Fig. 1A), and a multiplex ELISA (Meso Scale Discovery w , Rockville, MD, USA) of apical washings (air exposed) and basal (liquid exposed) conditioned medium for IL-8, IL-6, TNFa, IL-1b, monocyte chemoattractant protein-1 (MCP-1) and interferon gamma-induced protein-10 (IP-10) was performed (Fig. 1B). This revealed that resting ZZ-differentiated primary bronchial epithelial cells secreted more IL-8 basally when compared with MM cells (P , 0.01). After combined stimulation with oncostatin M (OSM), TNFa and IL-1b (OSM-mix), the ZZ differentiated primary bronchial epithelial cells showed significantly higher release of MCP-1 (P , 0.01), IP-10 (P , 0.05) and IL-1b (P , 0.01) compared with MM controls. The reduced secretion of IL-8 most likely reflects the known inhibitory effect of OSM on IL-1b-induced IL-8 release (31). To determine whether this enhanced release of cytokines was mediated by increased NF-kB signalling, submerged cultures of patientderived ZZ primary bronchial epithelial cells were induced to express increased levels of a 1 -antitrypsin by treatment with OSM and NF-kB activity was assayed using a luciferase reporter system ( Fig. 1C) (29). Low levels of basal NF-kB luciferase activity were detected in MM primary bronchial epithelial cells, whereas ZZ cells showed significantly higher activity at baseline (P , 0.05 compared with MM; Fig. 1C). When a 1 -antitrypsin production was increased by treatment with OSM, the NF-kB activity in ZZ primary bronchial epithelial cells increased significantly (P , 0.001) and the difference between MM and ZZ cells increased still further (P , 0.01). Stimulation with TNFa, a cytokine known to induce NF-kB activation, showed the same difference between MM and ZZ cells (P , 0.05). The same effect was seen in Tet-On A549 cells overexpressing either M or Z a 1 -antitrypsin (Supplementary Material, Fig. S1). To test whether the baseline difference in NF-kB activity was related to the transfection of the reporter constructs, we measured transcription of the NF-kB-dependent chemokine IL-8, which confirmed basal levels of inflammatory signalling were higher in ZZ primary bronchial epithelial cells compared with controls (P , 0.05; Fig. 1D).
Z a 1 -antitrypsin does not form detectable polymers nor causes ER stress in lung epithelial cells NF-kB activation by mutant serpins has previously been associated with the accumulation of protein polymers within the ER (20 -22,29). This accumulation has also been shown to exaggerate ER stress upon a second hit (20,22). To verify whether this mechanism was responsible for the enhanced NF-kB signalling in fully differentiated ZZ primary bronchial epithelial cells, we first measured total secreted and intracellular a 1 -antitrypsin. Resting primary bronchial epithelial cells produced unquantifiable amounts of a 1 -antitrypsin, but after stimulation with OSM-mix for 48 h, a 1 -antitrypsin was detectable in both the apical washes and basal supernatant ( Fig. 2A). Similar results were obtained after exclusion of the current smokers, indicating that the differences in smoking status of the patients between MM and ZZ patients from which cells were obtained did not explain the differences in production of a 1 -antitrypsin (data not shown). Using the 2C1 monoclonal antibody that specifically detects naturally occurring polymers of Z a 1 -antitrypsin (32), we found no evidence of polymer formation ( Fig. 2A). Accordingly, we could detect no differences in ER protein mobility, which we have previously shown occurs in cells containing ER luminal polymers of a 1 -antitrypsin [(22) Supplementary Material, Fig. S2]. To determine whether the absence of polymer formation was a feature of lung epithelial cells, we generated stable transfected A549 lung carcinoma cell lines that conditionally expressed either M or Z a 1 -antitrypsin under the control of a Tet-On responsive promoter. As expected, M a 1 -antitrypsin-expressing A549 cells secreted five-times more a 1 -antitrypsin than did Z a 1--antitrypsin-expressing A549 cells (Fig. 2B). Again, we were unable to detect protein polymers in either the supernatant or cellular lysates (Fig. 2B). As polymer formation is dependent upon a 1 -antitrypsin concentration (8), we compared the relative levels of a 1 -antitrypsin in tissue from an explanted cirrhotic ZZ liver ( Fig. 2C) with those in cultured airway epithelial cells. This revealed a 100-fold higher level of a 1 -antitrypsin in hepatic tissue and significant polymer accumulation (Fig. 2C). While polymerization of a 1 -antitrypsin in vitro is highly dependent upon protein concentration, the concentration dependence of polymerization within the crowded environment of the ER in vivo is not known. Therefore, to determine the critical concentration for the polymerization of Z a 1 -antitrypsin in living cells, we induced the expression of Z a 1 -antitrypsin in Tet-On stable CHO stable cells (22) and measured both total a 1 -antitrypsin and polymer levels (Supplementary Material, Fig. S3A). This revealed that levels of 300 ng a 1 -antitrypsin per 1 mg of total lysate protein are necessary before intracellular polymers can be detected in these cells (Supplementary Material, Fig. S3A). To test this finding in lung epithelial-derived cells, we induced expression of Z a 1 -antitrypsin in Tet-On A549 cells with doxycycline and augmented the protein level by inhibiting ERAD with lactacystin, a selective proteasome inhibitor. This increased the concentration of intracellular Z a 1 -antitrypsin of .300 ng a 1 -antitrypsin per 1 mg of total lysate, whereupon polymers were detected (Supplementary Material, Fig. S3B). It therefore seems likely that the low levels of a 1 -antitrypsin produced by airway epithelia are insufficient to generate detectable polymers.
To define whether Z a 1 -antitrypsin expressed in epithelial cells alters the ER stress response, we induced expression of a 1--antitrypsin in fully differentiated primary bronchial epithelial cells with OSM-mix in the presence or absence of the ER stress-inducing toxin tunicamycin. We detected no differences in the basal or OSM-mix-stimulated levels of spliced X-box binding protein-1 (XBP1s), CCAAT-enhancer-binding protein homologous protein (CHOP) and growth arrest and DNA damage-inducible gene 34 (GADD34) mRNA between MM and ZZ primary bronchial epithelial cells (Fig. 2D). As expected, tunicamycin increased the level of these transcripts; however, there was not an exaggerated ER stress response in ZZ epithelial cells (Fig. 2D). Similarly, we were unable to detect differences in ER stress by western blot for the KDEL-positive chaperones, glucose-regulated protein94 (GRP94) and GRP78, in ZZ and MM cells (Fig. 2E).
Loss of M a 1 -antitrypsin leads to increased activation of ERK, which is dependent on MEK and EGFR In order to understand the mechanism of inflammatory signalling in ZZ epithelial cells, we next evaluated activation of the NF-kB pathway components, inhibitor of nuclear factor kappa-B kinase subunit beta (IKKb), IkBa and p65. To evaluate MAPK signalling, we also measured JNK, p38 MAPK and ERK1/2. This revealed a significant difference only in the activation of ERK (P , 0.05; Fig. 3A and Supplementary Material, Fig. S4). Interestingly, depletion of a 1 -antitrypsin by siRNA caused phosphorylation of ERK in MM primary bronchial epithelial cells, but did not alter the phosphorylation of ERK in ZZ cells (Fig. 3A). This effect was specific for ablation of a 1 -antitrypsin, because silencing a non-specific serpin, neuroserpin (NS), did not increase phosphorylation of ERK in MM cells. This suggested that it was the lack of (M) a 1 -antitrypsin, rather than the presence of Z a 1 -antitrypsin, that might be responsible for the phosphorylation of ERK in ZZ primary bronchial epithelial cells. To test this, we treated ZZ primary bronchial epithelial cells with plasma-purified M a 1 -antitrypsin and observed a suppression of ERK phosphorylation (P , 0.05; Fig. 3B). We wished to determine whether this loss of function phenotype reflected a lack of anti-inflammatory activity in Z a 1 -antitrypsin or if cells secreted insufficient Z a 1 -antitrypsin. As concentration of plasma-purified Z a 1 -antitrypsin to a degree required for this experiment would result in its polymerization, we instead transiently transfected HeLa cells with either M or Z a 1 -antitrypsin or empty vector as control. After transfection, the cells produced high levels of a 1 -antitrypsin, with ZZ cells producing 20% of the amount that MM cells produced (444 ng/mg a 1 -antitrypsin in the total lysate versus 1995 ng/mg a 1 -antitrypsin in the total lysate respectively; Supplementary Material, Fig. S5A). Although 14% (33 ng/mg total lysate) of the extracellular Z a 1 -antitrypsin formed polymers (data not shown), the protein was able to inhibit phosphorylation of ERK1/2 to a similar degree as M a 1--antitrypsin (Supplementary Material, Fig. S5B). This result is consistent with a model in which ZZ primary bronchial epithelial cells secrete insufficient a 1 -antitrypsin to inhibit ERK1/2 phosphorylation, rather than Z a 1 -antitrypsin lacking the antiinflammatory activity per se.
To determine the mechanism of ERK phosphorylation in ZZ primary bronchial epithelial cells, we treated cells with U0126, a specific inhibitor of MEK, or an anti-EGFR blocking monoclonal antibody. Both reagents abrogated the phosphorylation of ERK, indicating that the EGFR -MEK-signalling pathway was involved and was activated by the ligand-binding site of EGFR (Fig. 3C). Although under some circumstances, the activation of ERK1/2 can lead to epithelial cell proliferation, we did not observe differences in the rate of MM and ZZ cell proliferation (Supplementary Material, Fig. S6A). Attempts to assess the effect of ERK1/2 inhibition in each cell type were hampered by toxicity, but it did not appear that modulation of ERK1/2 activation affected proliferation of MM or ZZ cells differentially. Treatment of ZZ cells with purified M a 1 -antitrypsin did not appear to affect the rate of proliferation (Supplementary Material, Fig. S6B).

ZZ primary bronchial epithelial cells generate higher levels of ADAM17-dependent EGFR ligands
We reasoned that this increased EGFR signalling might reflect increased cleavage of membrane-tethered ligands by MMPs and/or ADAMs. Therefore, we incubated MM and ZZ primary bronchial epithelial cells with GM6001, a broad-spectrum metalloproteinase (MP) inhibitor, or TAPI-2, a broad-spectrum MP inhibitor with enhanced ADAM17 inhibitory activity. TAPI-2 completely blocked the phosphorylation of ERK1/2 in ZZ cells (P , 0.001), whereas GM6001 failed to affect phosphorylation (Fig. 4A). Furthermore, only treatment with TAPI-2 reduced secretion of IL-8 from ZZ cells (P , 0.01) (Fig. 4B). As this suggested the involvement of ADAM17, we next tested the effect of a specific ADAM17 blocking antibody, D1(A12) (33). At 200 nM, a concentration known to block the activity of ADAM17 (34), we observed a substantial decrease in phosphorylation of ERK1/2 (P , 0.05; Fig. 4C).
A previous report suggested that M a 1 -antitrypsin can directly inhibit ADAM17 derived from neutrophils (4), and so we attempted to reproduce this observation. We first performed an in vitro ADAM17 activity assay by incubating 1 nM ADAM17 with 0 -30 mM purified plasma M a 1 -antitrypsin but were unable to detect any inhibition of ADAM17 (Fig. 4D). When using 10 mM b-mercaptoethanol in PBS or PBS alone, two potent ADAM17 inhibitors, we were able to inhibit its activity, confirming the functionality of our assay (Fig. 4D). Moreover, we were unable to detect the formation of a complex between ADAM17 and either form of a 1 -antitrypsin (Fig. 4E). Although it remains controversial as to whether cytoplasmic phosphorylation of ADAM17 plays an important regulatory role (35), we also tested for phosphorylation of ADAM17 in MM and ZZ cells grown with or without supplementary M a 1 -antitrypsin (Supplementary Material, Fig. S7). We detected no differences between these conditions.
To explore this pathway further, we next measured whether ZZ primary bronchial epithelial cells generated more EGFR ligands than controls. We detected significantly higher levels of mRNA encoding HB-EGF and TGFa in ZZ cells compared with MM controls (P , 0.05 and P ¼ 0.05; Fig. 5A). When  the EGFR was blocked using a monoclonal antibody to prevent ligand binding to its receptor, significantly more released TGFa and AREG was detected in ZZ cells compared with MM cells (P , 0.05; Fig. 5B). Consistent with these observations, when we treated MM primary bronchial epithelial cells with conditioned medium derived from ZZ cultures, we observed a significant increase in phosphorylation of ERK1/2 after 1 h (P , 0.01), which returned to baseline after 24 h (P , 0.001; Fig. 5C).
Taken together, these data indicate that the increased NF-kB signalling in ZZ primary bronchial epithelial cells is caused by phosphorylation of ERK1/2. This is due to increased availability of ADAM17-dependent EGFR ligands leading to activation of the EGFR and signalling via MEK (Fig. 6). Surprisingly, we were unable to detect the formation of polymers of a 1 -antitrypsin in ZZ primary bronchial epithelial cells or A549 lung adenocarcinoma cells overexpressing the protein, which may reflect the low levels of a 1 -antitrypsin expression of which these cells are capable.

DISCUSSION
For many years, it was thought that an imbalance between protease and antiprotease activity was solely responsible for the accelerated onset of emphysema in patients homozygous for the Z allele of a 1 -antitrypsin [reviewed in (36)]. However, a 1--antitrypsin has been found to possess additional roles to its antiprotease activity, including a range of anti-inflammatory properties (3)(4)(5), and the Z alleles can increase inflammatory NF-kB signalling when overexpressed pulmonary epithelial cells (20,21,30).
The results of our current study using primary bronchial epithelial cells in which a 1 -antitrypsin is expressed under the control of its endogenous promoter confirm and extend these findings. We found Figure 6. ZZ primary bronchial epithelial cells show an enhanced inflammatory response dependent of the ERK/EGFR/ADAM17 pathway. Impaired expression of a 1 -antitrypsin in primary bronchial epithelial cells leads to increased phosphorylation of ERK1/2, which is dependent on MEK, EGFR and ADAM17. M a 1 -antitrypsin modulates this inflammatory response via a yet undefined mechanism. that at the low levels of a 1 -antitrypsin expression that occur in these cells, clinically relevant polymer formation is unlikely to occur. Sufficient a 1 -antitrypsin is generated by wild-type cells to suppress ERK1/2 and NF-kB signalling, but this antiinflammatory effect appears to be lost in ZZ epithelia, suggesting a novel mechanism for airway pathology in a 1 -antitrypsin deficiency. Surprisingly, while ADAM17 is required to generate secreted EGFR ligands mediating inflammatory signals in these ZZ cells, we were unable to detect a direct inhibition of ADAM17 by M a 1 -antitrypsin.
Many heterologous overexpression systems have reported the presence of Z a 1 -antitrypsin polymers both intracellular and in conditioned medium (20,21). Unexpectedly, but importantly, we were unable to detect 2C1-positive polymers in primary bronchial epithelial cultures. Although the 2C1 monoclonal antibody is highly specific for polymers, it is formally possible that it is less sensitive than other anti-polymer antibodies, for example, ATZ11. However, in our hands both have similar avidity towards Z polymers, but the polyclonal antibody ATZ11 is less specific for polymers, detecting both Z monomers as well as Z polymers (32). Even if very low levels of Z a 1 -antitrypsin polymer are made within bronchial epithelial cells, it is unlikely to affect cellular function as we were unable to detect impaired ER protein mobility nor altered ER stress responsiveness as we have done previously for polymer-expressing cells (22).
We were unable to identify all components required for ERK1/2 activation in ZZ primary bronchial epithelial cells. It is likely there exists a significant level of redundancy in this system as we were able to detect multiple EGFR ligands to be increased in the conditioned medium of ZZ cultures. In our system, a 1 -antitrypsin did not directly inhibit ADAM17 as has been reported for neutrophils (4). Interestingly, it has recently been speculated that endocytosed a 1 -antitrypsin may modulate ADAM17 in endothelial cells (5). If intracellular a 1 -antitrypsin can indeed modulate these pathways, multiple potential mechanisms may explain this effect. ADAM-dependent transactivation of the EGFR by activation of G-protein-coupled receptors can also occur [reviewed in (27)]. For example, IL-8 can induce EGFR phosphorylation in epithelial cells via its receptors C-X-C chemokine receptor1 (CXCR1) and CXCR2 (28,37), while binding of a 1 -antitrypsin to IL-8 has been reported to prevent activation of CXCR1 (4). Although there are contradictory reports concerning the expression of CXCR1 and CXCR2 on bronchial epithelial cells (38,39), we found that IL-8 release is increased in ZZ primary bronchial epithelial cells, which potentially might provide an autocrine inflammatory signal in the absence of sufficient a 1 -antitrypsin. Alternatively, C-C chemokine ligand20 (also known as MIP-3a) and its receptor C-C chemokine receptor6 have been shown to activate ADAM17 causing transactivation of EGFR (40,41).
Taken together, our experiments have demonstrated that airway epithelial cells have anti-inflammatory activity owing to the local synthesis of M a 1 -antitrypsin from its endogenous promoter and that this effect is lacking in ZZ homozygous epithelial cells because of a lack of a 1 -antitrypsin secretion. This raises the possibility that a 1 -antitrypsin augmentation may have unanticipated effects within the airway. It also raises the potential that targeted anti-EGFR therapy might have antiinflammatory effects within the lung.
We carried out siRNA-mediated knockdown of a 1 -antitrypsin using a SMARTpool of ONTARGETplus a 1 -antitrypsin siRNA (Dharmacon, Lafayette, CO, USA) or SERPIN1 siRNA (Dharmacon) as a mock control. In general, 10 nM siRNA and 1 ml RNAiMAX (Invitrogen) were used according to manufacturer's descriptions, and a 1 -antitrypsin expression, measured with qPCR and ELISA, was knocked down .90% after 72 h by this method (Supplementary Material, Fig. S8).
Primary bronchial epithelial cells were differentiated as described previously (44). Briefly, cells were cultured submerged until confluence on semi-permeable Transwell membranes (Corning Costar, Cambridge, MA, USA) in B/D medium with addition of retinoic acid(end-concentration15 ng/ml;Sigma-Aldrich)andsubsequently cultured air-exposed for 14 days to allow mucociliary differentiation.
A549 cells were obtained from American Type Culture Collection (ATCC) and stably transfected with the pTetON vector (Clontech) to obtain Tet-On A549 cells. ELISA confirmed that production of endogenous a 1 -antitrypsin protein was below the level of detection in these cells (data not shown). These were stably transfected with pTRE2-hyg plasmid encoding M or Z a 1 -antitrypsin as described previously (42). Cells were maintained in DMEM with 10% (v/v) tetracycline-free foetal bovine serum, 100 U/ml penicillin, 100 mg/ml streptomycin, 400 mg/ml geneticin and 400 mg/ml hygromycinB (selective antibiotics from Invitrogen) at 378C, 5% CO 2 . Expression of a 1 -antitrypsin was usually induced using 2 mg/ml doxycycline (Sigma-Aldrich) for 48 h. HeLa cells were obtained from ATCC and transient transfected with pcDNA3.1 encoding M or Z a 1 -antitrypsin (or empty vector) as shown previously (19).

Patient groups
PiZZ a 1 -antitrypsin-deficient patients were recruited in the Leiden University Medical Center (LUMC; Leiden, the Netherlands). ZZ primary bronchial epithelial cells were acquired by bronchial biopsy, with approval from the Medical Ethical Committee of the LUMC. Briefly, bronchial biopsies were washed with PBS, divided in 2-mm pieces and placed into a fibronectin/ collagen-coated 24-well plate. Twice daily, the explants were fed with 20 ml B/D medium until they became adherent (maximum of three days). Then, primary bronchial epithelial cells were expanded submerged with 500 ml B/D medium, replaced triweekly. When cultures reached confulence cells were frozen down in liquid nitrogen until further use. MM primary bronchial epithelial cells were obtained from tumour-free resected lung tissue as described previously (43) and matched to ZZ primary bronchial epithelial cells according to sex, GOLD-stage (0-III) and smoking status (non-, ex-or current smoker) (Supplementary Material, Table S1). In each case, care was taken to ensure that cells were sourced only from the 2nd-3rd branches of the bronchial tree in order that results from each were directly comparable.
Measurement of total a 1 -antitrypsin, polymerized a 1 -antitrypsin and cytokine release Total and polymerized a 1 -antitrypsin were measured in whole cell lysate (intracellular) and supernatant (secreted) by ELISA as described previously (32). Cytokine release of fully differentiated cells was measured using a 4-plex Meso Scale Discovery kit (IL-6, IL-8, TNFa and IL-1b) and singleplex kits (MCP-1 and IP-10; Meso Scale Discovery w ). EGFR ligands were measured using commercially available ELISAs following manufacturer's protocol (R&D systems). IL-8 release in supernatant of submerged primary bronchial epithelial cultures was quantified using an IL-8 ELISA kit (Sanquin, Amsterdam, the Netherlands).

Protein mobility assay
Submerged cultured primary bronchial epithelial cells were grown overnight on coated 35-mm glass-bottom petri dishes (MatTek Corporation, Ashland, MA, USA) and transiently transfected with an ER-GFP plasmid (45). Live cells were imaged on an LSM510 confocal microscope (DuoScan; Carl Zeiss Inc., Thornwood, NY, USA) at 378C as defined in Ordonez et al. (22). Briefly, ER-GFP was visualized with an x63/1.4NA oil objective at 488-nm laser and fluorescence recovery after photobleaching experiments were performed. Fluorescence recovery curves were obtained by transforming fluorescence intensities into a percentage scale in which the prebleach time point represents 100% of fluorescence intensity.

Luciferase activity assays
Transfection with luciferase reporter plasmids was typically performed in 6-well plates with 1 mg of either p(5x)ATF6-luc (Firefly) or pELAM1-luc (Firefly) and 50 ng of pRL-TK (Renilla) as a transfection efficacy control (29). Cells were transfected for 6 h with 2 ml Lipofectamine LTX (Invitrogen) in serum-and antibiotic-free OptiMEM according to manufacturer's instructions and lysed the next day using the recommended protocol of the Dual-Luciferase Reporter Assay (Promega).

Quantitative RT-PCR
Total RNA was isolated and normalized mRNA levels were calculated using RPL13A and ATP5B as housekeeping genes (48). Primers used are described in Supplementary Material, Table S2 (49,50). IQ SYBRGreen supermix (Bio-rad, Herculus, CA, USA) was used for amplification of the cDNA.

Statistical analysis
Results from primary bronchial epithelial cells are expressed as single patients (each dot is the average of one patient in duplicate). Results from Tet-On A549 are shown as mean + SEM. Data were analysed using two-way repeated-measures analysis of variance (ANOVA) or Student t-test as appropriate. Differences with P-values , 0.05 were considered to be statistically significant.

SUPPLEMENTARY MATERIAL
Supplementary Material is available at HMG online.