Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

₁-Antitrypsin (₁-AT), or ₁-proteinase inhibitor, is the main inhibitor of serine proteases and especially of neutrophil elastase in humans. Current concepts suggest that it plays a key role in lung homeostasis. In particular, ₁-AT deficiency has been associated with the development of pulmonary emphysema, a disease caused by an imbalance between proteases and protease inhibitors (1).

It is well documented that the main source of serum ₁-AT is the liver (2, 3). However, this inhibitor has also been shown to be expressed in extrahepatic cell types, including human blood monocytes (4), pulmonary alveolar and breast-milk macrophages (5), neutrophils (6), activated lymphocytes (7), and human intestinal epithelial cells (8, 9). Recent studies indicate that alveolar epithelium is not only a structural barrier but is also actively involved in the modulation of airways inflammatory reactions through the synthesis of a variety of mediators (10). In this context, we have recently demonstrated that primary cultures of alveolar epithelial type II cells isolated from rat lung, as well as the human epithelial cell line A549, produce biologically active ₁-AT (11), and may therefore contribute to the local lung antiprotease screen.

₁-AT is also considered a positive-acute phase protein (APP), since its plasma level increases during acute and chronic inflammatory processes. The increase in APP plasma concentrations, derived mainly from the liver, and the induction of this hepatic response, are mediated by proinflammatory cytokines (interleukin-1 [IL-1], tumor necrosis factor- [TNF-], and interleukin-6 [IL-6]). IL-6 is the main regulator of hepatic synthesis of ₁-AT (12). More recently, different members of the IL-6 family, which includes oncostatin-M (OSM), leukemia-inhibitory factor (LIF), IL-11, and ciliary neurotrophic factor (CNTF), have been shown to display the same qualitative regulatory pattern of APP gene expression as IL-6 (12). Two additional cytokines, interferon- (IFN-) and transforming growth factor- (TGF-), are now recognized as being able to modulate the production of APP in modifying the hepatocyte response to IL-6 (13, 14). In addition to the stimulatory effect of cytokines, glucocorticoids are needed for maximum stimulation of hepatic APP expression, and act in a synergistic manner with IL-6 or IL-1 (15).

As opposed to the numerous studies of hepatic APP regulation, little is known about the regulation of ₁-AT expression in extrahepatic epithelial cells. However, IL-6 appears to be the main inducer of ₁-AT secretion in other cell types, including CACO2 epithelial cells (9) and macrophages (16). OSM also seems to be involved in the induction of antiproteases in cells of extrahepatic origin. Indeed, Cichy and coworkers have shown that OSM induces ₁-antichymotrypsin synthesis in different epithelial cell lines and in human bronchial epithelial cells (17, 18).

The presence of IL-6, as well as IL-1 and TNF- within the lung during the inflammatory process, has been widely reported (19), but little is known about LIF or OSM. LIF and OSM are known to be produced by monocytes/macrophages (20, 21). We have found elevated levels of LIF and OSM in bronchoalveolar lavage fluid (BALF) from patients with bacterial pneumonias, suggesting a local synthesis of these two cytokines within the alveolar space (unpublished data). Taken together, these results led us to investigate the effect of IL-1, TNF-, IL-6, and the other members of the IL-6 family, including LIF and OSM, on ₁-AT secretion in alveolar epithelial cells, and its modulation by glucocorticoids. We also considered the modulating effect of TGF- and IFN-, since both of these cytokines are produced within the alveolar space during the inflammatory process (10).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents

Cell-culture reagents were purchased from Eurobio (Les Ulis, France). Fetal calf serum (FCS) and amino acid and vitamin supplements were from Gibco (Cergy-Pontoise, France). Hybond N nylon filters were from Amersham (Les Ulis). Molecular biology reagents and RNAzol were from Bioprobe (Montreuil/bois, France). Rabbit antihuman ₁-AT, ₁-acid glycoprotein, and albumin antisera were from Behring (Rueil-malmaison, France); goat peroxidase- conjugated antihuman ₁-AT, ₁-acid glycoprotein, and albumin antibodies were from Cappel-Flobio (Courbevoie, France); and an antirabbit/antimouse streptavidin-peroxidase-immunostaining kit (LSAB kit peroxidase) and diaminobenzidine (DAB) substrates were from Dako (Trappes, France). Rabbit antirat ₁-AT antibodies were a gift from Doctor C. Gauthier (Tours, France). Recombinant human IL-1, TNF-, IL-6, LIF, OSM, TGF-, and IFN- (all from Immugenex, Los Angeles, CA) were diluted to 1 µg/ml in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA), and stored at 70°C until used. Dexamethasone (DEX) and human purified ₁-AT were from Sigma (Saint-Quentin Fallavier, France). An oligolabeling kit (Rediprime) was from Amersham. Tissue-culture plasticware was from Costar (Cambridge, MA).

Cell Cultures

Rat type II pneumocytes were isolated from male Sprague- Dawley rat lungs as previously described (22). For immunocytochemistry, type II pneumocytes were grown to confluence on glass cover slides, incubated for 24 h with or without supernatants from lipopolysaccharide (LPS)-stimulated rat alveolar macrophages (AM), and then rinsed 3 times with Tris-buffered saline (TBS) before immunostaining.

Rat AM recovered from BALF were cultured for 24 h in the presence of Escherichia coli LPS (O55: B5; Sigma), 10 µg/ml, as previously described (23). The collected supernatants are called AM-derived conditioned medium (CM).

The A549 cell line, derived from a patient with alveolar-cell carcinoma of the lung, has been used as a model of human alveolar epithelial type II cells in numerous studies. We used these cells in this study because human alveolar epithelial cells are not easily available. Futhermore, rat primary cell cultures are often contaminated with fibroblasts and macrophages, and to our knowledge, murine or rat recombinant OSM are not yet available, although the cloning of murine OSM has recently been described (24). The A549 human lung epithelial cell line (CCL 185) was purchased from the American Type Culture Collection (ATCC, Rockville, MD). A549 cells were grown to confluence in flasks in Hams' F12 culture medium containing 4 mM L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, 1× minimum essential medium (MEM), nonessential amino acids, and 1× MEM vitamins, supplemented with 15% FCS and kept at 37°C in a humidified incubator with 5% CO₂ in air. To study the inducing effect of cytokines on ₁-AT production, confluent A549 cells (9 × 10⁶ cells/55 cm² flask) were rinsed three times with sterile saline and cultured in 5-ml FCS-free complete medium for 24 h, with or without mediators added (cytokines, dexamethasone [DEX] 10⁶ M), unless otherwise stated. The supernatants were collected, centrifuged (400 × g for 10 min), and stored at 20°C after addition of protease inhibitors (4 mM phenylmethylsulfonyl fluoride [PMSF], 40 mM leupeptin) until ₁-AT was measured. The cells were then either frozen in liquid nitrogen before RNA study, or were submitted to a trypsin treatment and washed to allow their numeration and the DNA assay. DNA was then quantified as described elsewhere (25).

HepG2 cells (ATCC) were grown in MEM supplemented with 10% decomplemented FCS, 25 mM 4-(2-hydroxyethyl)- 1-piperazine-N'-2-ethanesulfonic acid (HEPES), 100 U/liter penicillin, 100 mg/ml streptomycin, and 4 mM glutamine. Albumin, ₁-acid glycoprotein, and ₁-AT secretion in HepG2 cell supernatants was studied over a 24 h period in FCS-free medium containing 10⁶ M DEX, in the presence or absence of cytokines. Supernatants and cells were collected as described for A549 cells.

Elastase Complexation by ₁-AT from Rat Alveolar Type II Cells

The complexation of human neutrophil elastase by ₁-AT from supernatants of rat alveolar type II cells was performed as previously described (11).

Detection of ₁-AT in Rat Alveolar Type II Cells by Immunocytochemistry

After fixation for 10 min in acetone, type II pneumocytes grown on glass cover slides were incubated with rabbit polyclonal antirat ₁-AT antibodies for 1 h at 37°C. Subsequently, the streptavidin-peroxidase antirabbit/antimouse system was applied according to the manufacturer's instructions. ₁-AT was visualized with DAB substrate in TBS containing 0.03% H₂O₂. When the primary antibody was omitted or replaced with a nonimmune normal rabbit serum, no staining was observed.

Enzyme-linked Immunosorbent Assay for Specific Proteins

₁-AT supernatant concentrations were measured with an enzyme-linked immunosorbent assay (ELISA) as elsewhere described (26), as were ₁-acid glycoprotein and albumin concentrations. All measurements were performed in duplicate with different dilutions of the same sample. Secretory leukocyte protease inhibitor (SLPI) was measured with a commercially available kit (R&D, Abingdon, UK). All data are expressed as mean ± SD of secreted protein/µg DNA/24 h.

Northern Blot Analysis

Total cellular RNA was extracted from cells with the RNAzol procedure (27). Ten micrograms of total RNA were subjected to 1% agarose-formaldehyde gel electrophoresis and transferred to a nylon membrane.

After prehybridization for up to 8 h at 42°C in 5× standard saline citrate (SSC) (1× SSC: 150 mM NaCl, 15 mM sodium citrate, pH 7.0), 50% formamide, 0.1% sodium dodecyl sulfate (SDS), 0.5% blotto (fat-free milk), 0.5 mg/ml salmon-sperm DNA, and 10% dextran sulfate, membranes were hybridized at 42°C in the same solution containing 10⁶ cpm/ml of labeled full-length complementary DNA (cDNA) probe. Membranes were first hybridized with an ₁-AT cDNA probe specific for human ₁-AT (28), after which autoradiography was performed. The membranes were then dehybridized in the following way: Two washes were performed in 2× SSC, after which membranes were soaked in boiling 0.1% SDS solution and allowed to cool to ambient temperature. Complete dehybridization was checked on an Instant Imager (Hewlett- Packard, Boise, ID). Rehybridization was performed as previously described for ₁-AT, using a glyceraldehyde phosphate dehydrogenase (GAPDH) cDNA probe (Clontech ref.9805.1; Palo Alto, CA). The labeling of the probes with [-³²P] deoxycytosine triphosphate ([-³²P]dCTP) was done by random priming, using an oligolabeling kit. After hybridization, filters were washed and given a final lavage at 55°C for 30 min in 1× SSC containing 0.1% SDS. Autoradiography was performed (time exposures were respectively 1 to 3 wk for ₁-AT and 1 d for GAPDH messages). The relative intensities of the messages were quantified with the Instant Imager, through comparison with the GAPDH message for each sample. Results were expressed as the coefficient (fold) increase over control.

Statistical Analysis

Statistical analysis was done with the Statview SE program on an Apple Macintosh computer (Apple, Inc., Cupertino, CA). Statistical significance was defined at P < 0.05 with analysis of variance (ANOVA) followed by Wilcoxon's paired test.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

AM-derived Conditioned Medium Increases ₁-AT Expression

Among the cells present in alveolar fluid, macrophages are the most important producers of mediators of the acute-phase response. In order to evaluate a possible role of these factors as regulators of ₁-AT synthesis by alveolar epithelial type II cells, we incubated primary rat type II cells with 20% AM-derived CM. We showed the ability of a crude AM-derived CM to upregulate ₁-AT mRNA levels in primary cultures of rat alveolar type II cells (Figure 1A). Moreover, as shown in Figure 1B, the ₁-AT secreted by these cells was biologically active, binding neutrophil elastase. ₁-AT secreted by rat alveolar type II cells was identified as a unique 59-kD band. The difference in electrophoretic mobility of this band as compared with human serum is similar to that observed with A549 cells (11).

View larger version (26K): [in this window] [in a new window]   View larger version (117K): [in this window] [in a new window]   View larger version (87K): [in this window] [in a new window]   View larger version (111K): [in this window] [in a new window]   Figure 1.   1-AT (AAT) synthesis in primary cultures of rat alveolar epithelial type II cells. (A) Stimulating effect of rat alveolar-macrophage-derived conditioned medium. Cells were incubated for 24 h in serum-free medium supplemented with 20% rat alveolar-macrophage-derived conditioned medium. Total cellular RNA was extracted and subjected to Northern blot analysis. Blots were hybridized to the human 1-AT probe (X-ray film was exposed for 3 wk). After stripping, the same blot was reprobed for GAPDH mRNA, the internal control (bottom panel, A). (B) Western blot analysis of human serum 1-AT (lanes 1 and 2) and rat type II pneumocyte supernatant (lanes 3, 4 and 5) before (lanes 1 and 3) and after (lanes 2, 4 and 5) incubation with human neutrophil elastase. In lane 4, rat type II pneumocyte supernatant was incubated with a 2-fold greater amount of elastase than in lane 5. (C) Immunocytochemical detection of 1-AT. Cells were incubated for 24 h in serum-free medium. Nonstimulated cells (a) or cells stimulated with 20% rat alveolar-macrophage-derived conditioned medium (b) were incubated with rabbit antirat 1-AT. Control cells (c) were incubated with normal rabbit serum. Original magnification: ×100.

However, in primary culture of rat type II cells, the presence of contaminating cells cannot be avoided. We therefore performed immunocytochemical staining of isolated rat alveolar type II cells. Our results showed diffuse staining corresponding to ₁-AT in the cytoplasm of the unstimulated cells and in cells cultured in the presence of LPS alone (10 µg/ml). Cells stimulated over a period of 24 h with AM-derived CM exhibited a strong perinuclear ₁-AT specific staining (Figure 1C). Similar results were obtained with A549 cells (data not shown).

These preliminary results indicate that ₁-AT synthesis by rat alveolar type II cells as well as by A549 cells can be induced by an AM-derived CM. The experiments described in the following sections were performed on the A549 cell line.

Stimulation of ₁-AT Production by Cytokines and Modulator Effect of DEX

Confluent A549 cell monolayers were stimulated for 24 h with IL-1, TNF-, OSM, LIF, and IL-6 at a concentration of 20 ng/ml, in the absence and in the presence of DEX (10⁶ M) (Figure 2).

View larger version (23K): [in this window] [in a new window]   Figure 2.   Stimulating effects of IL-1, TNF-, IL-6, LIF, and OSM on 1-AT (AAT) synthesis in A549 cell line. A549 cells were incubated for 24 h in serum-free medium supplemented with 20 ng/ ml IL-1, TNF-, IL-6, LIF, or OSM in the absence or presence of DEX 106 M. (A) Secretion of immunoreactive 1-AT measured in the cell-culture supernatants. Results are mean ± SD of 10 separate experiments. *P < 0.05 versus control (ANOVA followed by Wilcoxon's paired test). (B) Cells were cultured in the presence of DEX 106 M. Total cellular RNA was extracted and subjected to Northern blot analysis. Blots were hybridized to the human 1-AT probe (X-ray film was exposed for 2 wk). After stripping, the same blot was reprobed for GAPDH mRNA, the internal control (bottom panel, B).

In the absence of DEX, IL-1, TNF-, LIF, and IL-6 were poor inducers of ₁-AT secretion. Indeed, IL-6 and LIF had little or no effect. IL-1 or TNF- induced only a 3-fold increase in ₁-AT secretion (P < 0.05). By contrast, OSM was a potent inducer since it resulted in a 30 ± 3-fold increase (mean ± SD) in ₁-AT secretion (Figure 2A).

In the presence of DEX, a 2-fold increase in ₁-AT secretion in unstimulated cells was observed. Stimulation of A549 cells with TNF-, IL-1, and IL-6 resulted in a 2- to 5-fold increase in ₁-AT secretion (P < 0.05), LIF only poorly increased whereas OSM dramatically increased ₁-AT secretion (up to a 55-fold increase, P < 0.05), as shown in Figure 2A. These increases were not due to contaminating LPS, since LPS alone did not induce ₁-AT secretion.

These results were observed at the protein as well as at mRNA levels (Figure 2B). Indeed, in the presence of DEX, except LIF, which failed to significantly increase ₁-AT mRNA, IL-1, IL-6, and TNF-, respectively, induced 1.8-, 2.5-, and 4-fold increases, and OSM a 25-fold increase in specific ₁-AT mRNA. Accordingly, the following experiments were performed in the presence of DEX.

Dose- and Time-dependent Effects of OSM on ₁-AT Production

To determine the optimal concentration of OSM needed to stimulate ₁-AT production, A549 cells were cultured with increasing concentrations of OSM ranging from 2 to 100 ng/ml. As shown in Figure 3A, the maximal cell response was observed at 50 ng/ml, with a plateau between 50 and 100 ng/ml, although a response was already detectable at a concentration as low as 2 ng/ml. ₁-AT mRNA levels paralleled the changes in levels of secreted protein (Figure 3B), which suggests that OSM acts, at least partly, at the transcriptional level.

View larger version (19K): [in this window] [in a new window]   Figure 3.   Dose dependence of 1-AT (AAT) induction by OSM in A549 cells. Cells were incubated for 24 h in serum-free medium supplemented with OSM at the indicated concentrations (ng/ml) in the presence of DEX 106 M. (A) Secretion of immunoreactive 1-AT measured in the cell culture supernatants. Results are mean ± SD of five separate experiments. (B) Total cellular RNA was extracted and subjected to Northern blot analysis. Blots were hybridized to human 1-AT probe (X-ray film was exposed for 1 wk). After stripping, the same blot was reprobed for GAPDH mRNA, the internal control (bottom panel, B).

The kinetics of induction were studied by culturing A549 cells with 20 ng/ml OSM for up to 72 h. The increase in ₁-AT mRNA expression appeared 4 h after OSM addition. The increase in ₁-AT mRNA was linear from 4 h until 24 h after stimulation (3.9-, 6.7-, and 9.8-fold increase over the control value for 4, 10, and 24 h after OSM stimulation, respectively), with maximal stimulation occurring at 24 h. Thereafter, RNA levels remained constant until 72 h (Figure 4B). In parallel, stimulation of A549 cells with OSM resulted in a time-dependent increase in ₁-AT secretion; secretion was detectable as early as 4 h, with maximal secretion being observed at 48 h and lasting until 72 h (Figure 4A).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Time-dependence of 1-AT (AAT) induction by OSM in A549 cells. Cells were incubated in serum-free medium supplemented with 20 ng/ml OSM for indicated times in the presence of DEX 106 M. (A) Secretion of immunoreactive 1-AT measured in the cell-culture supernatants. Representative experiment out of five. (B) Total cellular RNA was extracted and subjected to Northern blot analysis. Blots were hybridized to human 1-AT probe (X-ray film was exposed for 2 wk). After stripping, the same blot was reprobed for GAPDH mRNA, the internal control (bottom panel, B).

TGF--induced Synergistic Effect with OSM on ₁-AT Production

To evaluate the possible modulation of ₁-AT production induced by OSM in A549 cells, different cytokines were tested in combination with OSM. IL-1, TNF-, and IL-6 gave a simple additive effect with OSM, as opposed to TGF-.

TGF- alone (20 ng/ml) induced a 4-fold increase in ₁-AT secretion, which was dose-dependent (Figure 5A). Coincubation of A549 cells with OSM (10 ng/ml) and TGF- (0.1, 1, and 10 ng/ml) resulted in an upregulation of ₁-AT secretion, which was synergistic with respect to that obtained with each cytokine alone (Figure 5A). This effect was dose-dependent, with an increase of up to 4-fold in OSM-induced ₁-AT secretion when cells were incubated with 10 ng/ml of TGF- and OSM. This latter combination resulted in a total 110-fold increase in ₁-AT secretion as compared with that for unstimulated cells. The minimal concentration of TGF- needed to observe the synergistic effect of TGF- on OSM-induced ₁-AT secretion was 1 ng/ ml. In the absence of DEX, although secretion of ₁-AT was reduced, the synergistic effect was maintained (data not shown).

View larger version (13K): [in this window] [in a new window]   Figure 5.   Synergistic effects of TGF- and OSM on 1-AT (AAT) synthesis in A549 cell line. A549 cells were incubated for 24 h in serum-free medium supplemented with 10 ng/ml OSM and TGF- at the concentrations indicated (ng/ml), in the presence of DEX 106 M. TGF- 2 ng/ml was used in the Northern blot analysis. (A) Secretion of immunoreactive 1-AT measured in the cell-culture supernatants. Results are mean ± SD of five separate experiments. (B) Total cellular RNA was extracted and subjected to Northern blot analysis. Blots were hybridized to human 1-AT probe (X-ray film was exposed for 1 wk). After stripping, the same blot was reprobed for GAPDH mRNA, the internal control (bottom panel, B).

Northern blot analysis demonstrated that specific ₁-AT mRNA levels paralleled the changes in the levels of secreted ₁-AT in the presence of 10 ng/ml OSM and 2 ng/ml TGF- (i.e., 1.8-, 5-, and 15-fold increases with TGF-, OSM, and both, respectively [Figure 5B]).

In contrast to the synergistic effect observed with OSM associated with TGF-, TGF- in combination with IL-1, TNF-, LIF, or IL-6 resulted in a simple additive effect of each cytokine (data not shown).

IFN--induced Inhibition of the OSM Effect on ₁-AT Production

IFN- alone, whatever its concentration, did not significantly modify ₁-AT production. By contrast, coincubation of A549 cells with OSM (20 ng/ml) and IFN- (5, 20, 50, and 100 ng/ml) produced a dose-dependent downregulation of ₁-AT secretion, with a decrease of up to 85% in OSM-induced ₁-AT secretion when cells were incubated with 100 ng/ml of IFN- and OSM (Figure 6A). Northern blot analysis demonstrated that specific ₁-AT mRNA levels were decreased when cells were coincubated with IFN- and OSM at the concentrations of 20 ng/ml each (i.e., as shown in Figure 6B, OSM alone induced a 22-fold increase, as compared with an 8.9-fold increase in the presence of IFN-, thus resulting in an inhibition of 60%). However, IFN- alone did not modify ₁-AT mRNA levels.

View larger version (19K): [in this window] [in a new window]   Figure 6.   Inhibitory effect of IFN- on OSM-induced 1-AT (AAT) synthesis in A549 cell line. A549 cells were incubated for 24 h in serum-free medium supplemented with 20 ng/ml OSM and IFN- at the indicated concentrations (ng/ml) in the presence of DEX 106 M. IFN- 20 ng/ml was used in the Northern blot analysis. (A) Secretion of immunoreactive 1-AT measured in the cell culture supernatants. Results are mean ± SD of five separate experiments. (B) Total cellular RNA was extracted and subjected to Northern blot analysis. Blots were hybridized to human 1-AT probe (X-ray film was exposed for 1 wk). After stripping, the same blot was reprobed for GAPDH mRNA, the internal control (bottom panel, B).

To evaluate the specificity of the modulation of ₁-AT production induced by OSM in A549 cells, we tested other cytokines, namely IL-1, TNF-, and IL-6, in combination with IFN-, and found a similar inhibition of ₁-AT secretion after cytokine induction as that found with OSM (data not shown).

Differential OSM Regulation of SLPI Secretion in the A549 Cell Line

In the same sets of experiments as for ₁-AT secretion, we studied the secretion of SLPI, an antiprotease that is known to be produced by A549 cells as well as by bronchial epithelial cells (Figure 7). In agreement with those of Sallenave and colleagues (29), our results showed that IL-1 induced a 6-fold increase in SLPI secretion (P < 0.05). In our hands, OSM induced a 3-fold increase in basal SLPI secretion (P < 0.05). When TGF- or IFN- were added to OSM, the OSM-induced SLPI secretion was not modified. These latter results contrast with those observed with ₁-AT.

View larger version (12K): [in this window] [in a new window]   Figure 7.   SLPI secretion and regulation in A549 cells. A549 cells were incubated for 24 h in serum-free medium supplemented with 10 ng/ml IL-1, OSM, TGF-, and IFN-, either alone or in association, and in the presence of DEX 106 M. SLPI concentrations in cell-culture supernatants were measured with an ELISA kit. Results are mean ± SD of five separate experiments. *P < 0.05 versus control (ANOVA followed by the Wilcoxon's paired test).

Differential OSM Regulation of ₁-AT Secretion in Hepatocytes

When we studied the effect of cytokines on specific protein secretion in the HepG2 cell line, we found that OSM induced a dose-dependent upregulation of ₁-acid glycoprotein and a dose-dependent downregulation of albumin, which is in agreement with the findings in other studies (30). By contrast, ₁-AT production was slightly increased (1.25-fold) after stimulation with either OSM or OSM in combination with TGF-, as shown in Figure 8.

View larger version (20K): [in this window] [in a new window]   Figure 8.   1-AT (AAT), 1-acid glycoprotein, and albumin production in HepG2 cells incubated with OSM and/or TGF-. Cells were cultured in serum-free DMEM for 24 h and stimulated with 20 ng/ml cytokines in the presence of 106 M DEX. 1-AT (gray bars), 1-acid glycoprotein (open bars), and albumin (solid bars) concentrations in cell culture supernatants were measured with ELISA assays. Results are mean ± SD of five separate experiments.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The major finding of our study was that ₁-AT secretion by A549 cells was greatly induced by OSM, and that TGF- and IFN- were able to modulate this OSM-stimulated secretion. This regulation was specific to ₁-AT, and different from the liver response.

Damage to the alveolar epithelium of patients with acute or chronic inflammatory disease is believed to be caused by the release of proinflammatory mediators from effector cells recruited to the site of the inflammatory reaction, and particularly by proteases released by neutrophils. Therefore, the presence of protease inhibitors, such as ₁-AT, may contribute to repair of the damaged lung. We have recently shown that primary cultures of rat type II pneumocytes, as well as A549 cells, secrete ₁-AT (11). Furthermore, this ₁-AT is biologically active, binding neutrophil elastase. This local synthesis of ₁-AT is potentially regulated by cytokines produced locally whithin the alveolar spaces. Indeed, we showed the ability of a crude AM-derived conditioned medium to upregulate ₁-AT mRNA levels in primary cultures of rat alveolar type II cells. Although contaminating cells (e.g., fibroblasts and macrophages) may complicate interpretation of the results when primary cell cultures are studied, we have shown that the upregulation of ₁-AT by the AM-derived conditioned medium was effectively localized to rat alveolar type II cells.

Similar upregulation of ₁-AT by an AM-derived conditioned medium was obtained with A549 cells. Therefore, we further investigated with the A549 cell line the effects of cytokines which, among those released by activated macrophages, might be involved in the upregulation of ₁-AT gene expression and secretion.

DEX enhanced ₁-AT secretion by A549 cells whatever the stimulation status. Therefore, in our further discussion, we consider only the results of experiments done with the presence of DEX. In agreement with our results, other investigators have reported a similar potentiating effect of DEX on APP synthesis by hepatocytes and on antiproteases (i.e., ₁-antichymotrypsin and SLPI) by airway epithelial cells (18, 31). However, this potentiating effect is not a general mechanism, since the secretion of ₁-AT by human monocytes has been shown to be downregulated by DEX (32).

In the present study, we showed a modest upregulation of ₁-AT secretion and gene expression by IL-1 and TNF-. A similar upregulation of other antiproteases, such as elafin and SLPI secreted by A549 cells (29), and of ₁-antichymotrypsin secreted by the lung adenocarcinoma cell line HTB 55 (18) has been reported. IL-6 also increased ₁-AT secretion. Although IL-1 and TNF- are potent inducers of IL-6 secretion in A549 cells (23), a direct regulatory role of IL-1 and TNF- on ₁-AT gene expression seems more likely. Indeed, under optimal conditions of ₁-AT secretion, the presence of DEX inhibited almost all IL-6 secretion by A549 cells, thus excluding a pathway involving IL-6 (data not shown).

Although LIF was a poor inducer of ₁-AT synthesis in A549 cells, OSM dramatically increased ₁-AT secretion and gene expression in a dose- and time-dependent manner. A discrepancy in the effect produced by the different members of the IL-6 family has already been reported; indeed, a weaker inducing effect on hepatic synthesis of haptoglobin and fibrinogen with LIF as compared with IL-6 and OSM has been reported (33). Cichy and coworkers showed a more potent stimulating effect of OSM than of LIF or IL-6 on ₁-antichymotrypsin synthesis by the lung adenocarcinoma cell line HTB 55 (18).

OSM, LIF, and IL-6 share characteristics in terms of both structure and mechanism of action, since they cause signal transduction through the common gp 130 component of their receptors. Whereas other cytokines in the IL-6-family use specific binding subunits complexed with gp 130 to form a multimeric signaling receptor, OSM binds directly to gp 130 (34). OSM is also unique in using two different receptor complexes formed by the association of the OSM/gp 130 complex and either the LIF receptor or the recently cloned OSM receptor  (35, 36). This specificity may explain the biologic responses that seem to be exclusively induced by OSM. Since neither IL-6 nor LIF have potent stimulatory effects on ₁-AT synthesis, and since the gp 130 and the LIF receptor are known to be present on A549 cells (33), our results suggest that A549 cells may express the specific OSM receptor  unit. In accord with this hypothesis, Piquet-Pellorce and colleagues have demonstrated that OSM, but neither IL-6 nor LIF, dose-dependently inhibited A549 cell proliferation (33). It should be noted, however, that this inhibition was observed only 4 days after OSM stimulation. The OSM concentrations that we used to induce ₁-AT secretion modified neither cell morphology nor cell proliferation during a 24 h culture period. Therefore, the inducing effect of OSM on ₁-AT synthesis cannot be related to an effect on cell proliferation, differentiation, or morphology.

We failed to reproduce the potent ₁-AT-inducing effect of OSM on primary cultures of rat type II pneumocytes. Richards and colleagues have shown that recombinant human OSM induces APP synthesis in rat hepatocytes (30). However, they have shown that although human OSM induced an equivalent or greater response than IL-6 in human HepG2 cells, it was a poor inducer of APP synthesis in rat hepatocytes. Moreover, murine OSM cDNA showed only 48% homology with human OSM (24). These converging data argue for a species specificity of the effect of OSM, which could explain why OSM failed to induce ₁-AT secretion in rat type II cells.

We have also shown in this study that OSM-induced ₁-AT secretion by A549 cells could be modulated by TGF- or IFN-.

TGF- is considered to be a dominant immunomodulator, and has been shown to enhance the expression of ₁-AT in human selected hepatoma cell lines (13), as well as to modulate the action of cytokines by enhancing or reducing their effects on APP gene expression in rat hepatoma cells (37). Our results showed that TGF- on its own induced a modest increase in the expression of the ₁-AT gene in A549 cells, and had only an additive effect on IL-1-, TNF--, IL-6-, and LIF-induced ₁-AT expression. By contrast, TGF- in combination with OSM produced a potent and synergistic upregulation of ₁-AT expression. To our knowledge, such a potent inducing effect for ₁-AT has never been reported whatever the cell type.

The synergistic activity of TGF- and OSM may be relevant in vivo, since TGF- and OSM are both present in the airways during inflammatory processes. Indeed, host-defense effector cells present in the lung, such as monocytes, AM, and activated lymphocytes (21, 38), as well as alveolar epithelial type II cells during lung fibrosis (39), are potential local sources of OSM and/or TGF- within the lung. Furthermore, we found increased OSM concentrations in the BALF of patients with acute bacterial pneumonia (unpublished data). Moreover, the high TGF- concentrations found in epithelial lining fluid of the normal human respiratory tract are further increased during an inflammatory process (39, 40).

The mechanisms underlying the effects of TGF- are poorly elucidated even for hepatocytes. To our knowledge, alveolar epithelial cells do not synthesize OSM, and we failed to induce OSM secretion even by TGF- (data not shown). Therefore, the upregulation of OSM effect induced by TGF- cannot be explained by such a mechanism. In the rat hepatoma cell line H₃₅, enhancement of the IL-6-induced response by TGF- can be partly explained by overexpression of the 80-kD ligand-binding subunit of the IL-6 receptor (37). The upregulation of the OSM receptor by TGF- in A549 cells represents an attractive hypothesis to explain our results, but remains to be tested.

IFN- has been shown to decrease the acute-phase response initiated by IL-6 and to a lesser extent by TNF- in human hepatoma cells (14). However, IFN- is also able to upregulate C2 complement and C1-inhibitor synthesis in human blood monocytes and HepG2 cells (41). In our study, we showed that IFN- downregulated OSM-induced ₁-AT secretion and mRNA levels. This downregulation was not specific to the OSM effect since it was also observed with IL-1, IL-6, LIF, and TNF--induced ₁-AT secretion. However, the mechanisms underlying the effects of IFN- on ₁-AT synthesis by A549 cells need further investigation.

It has recently been shown that SLPI, another antiprotease, is synthesized by A549 cells (29). In our hands, OSM slightly upregulated SLPI secretion, and this OSM-induced secretion was modified neither by TGF- nor by IFN-. Taken together, these results show that ₁-AT and SLPI secretion by A549 cells are regulated in different ways, and that the synergism of the OSM/TGF- combination is specific to ₁-AT.

₁-AT regulation by OSM also appears to be tissue specific. ₁-AT secretion by HepG2 cells was poorly enhanced by OSM or by OSM in combination with TGF-, contrasting with the regulation of other APPs such as ₁-acid glycoprotein and albumin, and corresponding to the findings of Richards and colleagues (30). It is widely admitted that ₁-AT secretion by hepatocytes is poorly upregulated by recombinant human cytokines, even by IL-6 and by LIF (2- and 1.5-fold increase, respectively), as compared with other APPs (42). The results of all these experiments, including ours, provide convincing evidence that expression of the ₁-AT gene in different cell types is under the control of different modulators and mechanisms of regulation. This special effect of OSM on pulmonary epithelial cells may be important within the lung. Indeed, by inducing local synthesis of protease inhibitors, OSM may participate in lung homeostasis during an inflammatory process. Interestingly, paralleling our work, another study, done by Cichy and colleagues, found that bronchial epithelial cells produce ₁-AT and that this production is upregulated by OSM and dexamethasone, whereas IL-6 and LIF have no effect (43). However, basal ₁-AT secretion appears to be greater in bronchial cells than in alveolar cells. On the other hand, the inducing effect of OSM is weaker in this cell type, leading to a 2- to 6-fold increase in ₁-AT secretion. These differences are certainly due to the different origins of these lung epithelial cells (i.e., bronchial versus alveolar). Cichy and colleagues' study (43) and ours suggest that the upper as well as the distal epithelial airways may contribute to local ₁-AT production, forming a continuous antiprotease screen during an inflammatory process.

In conclusion, this study provides further evidence that extrahepatic cell types and tissues may contribute locally to the protease-antiprotease balance during the inflammatory response. Taken together, our data suggest that synthesis of ₁-AT in lung epithelial cells may be triggered at the early stage of the inflammatory or infectious process by the alarm cytokines IL-1 and TNF-, and at a later stage by the potent stimulator OSM, secreted by host- defense effector cells. OSM, which is present in epithelial lining fluid during an infectious process, may play an important role in triggering local ₁-AT synthesis, leading to increased levels of ₁-AT in the alveolar fluids close to the sites of elastase release by neutrophils. Specific control of OSM-induced ₁-AT synthesis by TGF- and/or IFN- may be critical in contributing to the maintenance of a proper protease-antiprotease balance in the alveolar space, which is especially sensitive to damage by proteolytic enzymes.